Early Opioid Prescriptions for Work-Related Musculoskeletal ...

Early Opioid Prescriptions for Work-Related Musculoskeletal Disorders of the Low Back: Understanding Dispensing Patterns,

Their Determinants, and Their Impact on Work Disability

by

Nancy Gaspar Carnide

A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy in Epidemiology

Dalla Lana School of Public Health University of Toronto

©Copyright by Nancy Gaspar Carnide 2017

ii

Early Opioid Prescriptions for Work-Related Musculoskeletal Disorders

of the Low Back: Understanding Dispensing Patterns, Their

Determinants, and Their Impact on Work Disability

Nancy Gaspar Carnide

Doctor of Philosophy in Epidemiology

Dalla Lana School of Public Health University of Toronto

2017

Abstract

Opioids are prescribed early for low back pain (LBP), despite little evidence supporting this

practice. The dissertation objective was to understand factors associated with early opioid

provision after a LBP workers’ compensation claim and their relationship with work disability.

A systematic review (Manuscript 1) found evidence of a positive association between early

opioids and prolonged work disability for musculoskeletal disorders, but internal validity was

jeopardized by exposure measurement and confounding biases.

Informed by review findings, an historical cohort study was conducted of workers’

compensation claimants with LBP using dispensing data on LBP-indicated drug classes: opioids,

nonsteroidal anti-inflammatory drugs (NSAIDs), and skeletal muscle relaxants (SMRs).

Opioids, NSAIDs, and SMRs were dispensed to 19.7%, 21.2%, and 6.3% of claimants,

respectively, one year pre-injury, and 39.0%, 50.2%, and 28.4% one year post-injury

(Manuscript 2). Dispensing was stable pre-injury, increased within eight weeks post-injury, and

declined thereafter, although not to pre-injury levels. A small subgroup demonstrated

iii

prolonged dispensing, suggesting delayed recovery.

Claimant factors associated with early post-injury dispensing included pre-injury days supplied

with the three drug classes, prior adjuvant medication dispenses, diagnosis, prior health care,

age, sex, and occupation (Manuscript 3). Prescriber sex, birth year, and specialty were

associated with drug class(es) at first prescription and the prescriber accounted for 25-36% of

outcome variability.

Examining the relationship between these drug classes dispensed within eight weeks post-

injury and wage replacement benefits after eight weeks (Manuscript 4) revealed claimants

receiving opioids, particularly with NSAIDs and/or SMRs, were more likely to receive benefits

compared to NSAIDs/SMRs only. Strong and weak opioid combinations and days supply for all

classes were positively associated with receiving benefits.

There is evidence of increased opioid, NSAID, and SMR dispensing within eight weeks of a LBP

claim. Early dispensing is multifactorial and prescribers play a role in dispensing variation. Early

opioids compared to NSAIDs/SMRs appear to confer a greater risk of work disability.

Methodological strengths include capturing pre-injury factors and all prescriptions and partially

addressing confounding by indication/severity. However, residual confounding remains

concerning. Consideration of severity, prescriber, and system/workplace measures is needed to

advance understanding of opioid use for LBP and work disability.

iv

Acknowledgements

I am truly thankful to have had the support of some amazing people and organizations along

this journey to my PhD.

First, I would like to express my sincerest gratitude to my supervisor, Dr. Sheilah Hogg-Johnson,

for being a constant source of support, encouragement, patience, and knowledge. Your advice

throughout this process has been invaluable to me. I would also like to thank my amazing

committee members, Dr. Pierre Côté, Dr. Mieke Koehoorn, and Dr. Andrea Furlan. Your critical

appraisal of my work and the various perspectives that you brought to the table have been

instrumental in shaping this body of work. It has been a pleasure to work with all of you.

My sincere appreciation goes to Hyunmi Lee for her work in setting up the cohort and

assistance with data analysis. You consistently demonstrated a willingness to help and were an

incredible sounding board for all things SAS-related. This was certainly appreciated in what felt

like a daunting task of making sense of all this administrative data.

Thank you to my Institute for Work & Health colleagues for your support and interest in my

research. I especially wish to thank Emma Irvin and Dwayne Van Eerd for providing their

expertise and guidance on my systematic review and injecting some fun into the process. I also

want to thank Shanti Raktoe and my “stooges”, Colette Severin and Albana Canga, for always

providing me with an ear and a good laugh to relieve stress.

I would like to thank the staff at Population Data BC and the Partnership for Work, Health, and

Safety for their assistance with all things data-related, especially Tim Choi and Lillian Tamburic. I

also wish to thank the data stewards for granting me access to an incredibly rich source of data

for my dissertation.

The analyses in this thesis would not have been possible without the generous funding provided

by the Canadian Institutes of Health Research (CIHR) and WorkSafeBC.

A big thank you goes to the Institute for Work & Health for providing me with personal funding

and institutional support. The big window and quiet space to work was also a bonus. I would

v

also like to express my appreciation for the personal funding I received from the CIHR Vanier

Scholarship Program, CIHR Work Disability Prevention Program, and the QEII C.P. Shah

Graduate Scholarship Program in Science and Technology.

Last, but absolutely not least, I could not have done this without the love and support of my

family and friends.

Mae and Pai, you sacrificed so much in your own lives to ensure that we had a better future. I

hope I have made you proud. Thank you also to my Linden parents for their encouragement

and interest in my work. All of you helped me enormously throughout and I am truly grateful.

To Lucy, Delia, Kristian, Katie, Michael, Julia, and Maggie: thank you for being an amazing

support system for me over the years. Marta, you never got to see me officially finish this

degree, but I hope you know that I could not have done this without you.

To my wonderfully patient and supportive husband Robert. Where to even begin? Two days

after we said “I do”, I dove headfirst into my PhD…and you still stuck around! Your steadfast

support and encouragement were instrumental in getting me to the finish line. You pulled me

through some pretty rough periods and took on additional responsibility at home so that I could

focus on my work. I know I owe you…but can we at least put a time limit on how long?

And finally Oliver, my beautiful snuggle muffin. Although your arrival came at a semi-awkward

time (a month before my thesis data finally arrived after a two year long wait) and though

getting a PhD is pretty darn great, I will always consider you to be my biggest and best

accomplishment in life. You gave me the motivation to keep going. Mama has finally finished

her “big project” and hearing the words “good job mommy!” is the best reward this tired

student could ever ask for.

vi

Disclaimer

All inferences, opinions, and conclusions drawn in this dissertation are those of the authors and

do not reflect the opinions or policies of the Data Stewards or the Province of Ontario.

vii

Table of Contents

Abstract ........................................................................................................................................... ii

Acknowledgements ......................................................................................................................... iv

Disclaimer ........................................................................................................................................ vi

List of Tables .................................................................................................................................... x

List of Figures ................................................................................................................................. xii

List of Appendix Tables ................................................................................................................. xiii

List of Appendix Figures ................................................................................................................. xv

Abbreviations ................................................................................................................................ xvi

CHAPTER 1 INTRODUCTION ............................................................................................................ 1

1.1 Introduction ....................................................................................................................... 1

1.2 Overall Thesis Objective and Format of the Dissertation ................................................. 4

1.3 Candidate’s Roles and Responsibilities ............................................................................. 5

1.4 References ......................................................................................................................... 7

CHAPTER 2 ADDITIONAL BACKGROUND ....................................................................................... 13

2.1 Basic Spine Anatomy ....................................................................................................... 13

2.2 Low Back Pain .................................................................................................................. 14

2.3 Burden of Low Back Pain ................................................................................................. 16

2.4 Prognosis of Low Back Pain ............................................................................................. 17

2.5 Pharmacologic Treatment of Low Back Pain ................................................................... 18

2.6 References ....................................................................................................................... 27

CHAPTER 3 EARLY PRESCRIPTION OPIOID USE FOR MUSCULOSKELETAL DISORDERS AND WORK

OUTCOMES: A SYSTEMATIC REVIEW OF THE LITERATURE ...................................................... 41

3.1 Introduction ..................................................................................................................... 43

3.2 Methods ........................................................................................................................... 44

3.3 Results .............................................................................................................................. 48

3.4 Discussion ........................................................................................................................ 58

3.5 Conclusion ....................................................................................................................... 63

3.6 References ....................................................................................................................... 65

CHAPTER 4 SPECIFIC CHAPTER OBJECTIVES AND DATA SOURCE INFORMATION ........................ 79

viii

4.1 Summary and Chapter Objectives ................................................................................... 79

4.2 Chapter Objectives .......................................................................................................... 81

4.3 Administrative Datasets .................................................................................................. 82

4.4 References ....................................................................................................................... 89

CHAPTER 5 PRESCRIPTION DISPENSING PATTERNS BEFORE AND AFTER A WORKERS’

COMPENSATION CLAIM: AN HISTORICAL COHORT STUDY OF WORKERS WITH LOW BACK

PAIN INJURIES IN BRITISH COLUMBIA ...................................................................................... 94

5.1 Introduction ..................................................................................................................... 96

5.2 Methods ........................................................................................................................... 98

5.3 Results ............................................................................................................................ 105

5.4 Discussion ...................................................................................................................... 110

5.5 Conclusion ..................................................................................................................... 117

5.6 References ..................................................................................................................... 118

CHAPTER 6 FACTORS ASSOCIATED WITH EARLY OPIOID, NSAID, AND MUSCLE RELAXANT

DISPENSING AFTER A WORK-RELATED LOW BACK INJURY .................................................... 137

6.1 Introduction ................................................................................................................... 139

6.2 Methods ......................................................................................................................... 141

6.3 Results ............................................................................................................................ 150

6.4 Discussion ...................................................................................................................... 158

6.5 Conclusion ..................................................................................................................... 165

6.6 References ..................................................................................................................... 167

CHAPTER 7 THE RELATIONSHIP BETWEEN EARLY PRESCRIPTION DISPENSING PATTERNS AND

WORK DISABILITY IN A COHORT OF LOW BACK PAIN WORKERS’ COMPENSATION CLAIMANTS

192

7.1 Introduction ................................................................................................................... 194

7.2 Methods ......................................................................................................................... 196

7.3 Results ............................................................................................................................ 207

7.4 Discussion ...................................................................................................................... 212

7.5 Conclusion ..................................................................................................................... 219

7.6 References ..................................................................................................................... 220

CHAPTER 8 DISCUSSION .............................................................................................................. 233

ix

8.1 Major Findings and Contribution to the Literature ....................................................... 233

8.2 External and Internal Validity ........................................................................................ 238

8.3 Implications for Future Research .................................................................................. 246

8.4 Conclusions and Implications for Policy and Practice ................................................... 254

8.5 References ..................................................................................................................... 256

APPENDIX A SUPPLEMENTARY TABLES FOR MANUSCRIPT 1 (CHAPTER 3) ................................ 261

APPENDIX B METHODS APPENDICES FOR MANUSCRIPTS 2 TO 4 (CHAPTERS 5 TO 7) ............... 270

APPENDIX C SUPPLEMENTARY FIGURES FOR MANUSCRIPT 2 (CHAPTER 5) .............................. 303

APPENDIX D SUPPLEMENTARY TABLES FOR MANUSCRIPT 3 (CHAPTER 6) ................................ 311

APPENDIX E SUPPLEMENTARY TABLES FOR MANUSCRIPT 4 (CHAPTER 7) ................................ 314

x

List of Tables

Table 2-1 Overview of clinical guideline recommendations for the pharmacologic treatment of

acute and subacute episodes of low back pain ..................................................................... 36

Table 3-1 Characteristics of the cohort studies included in the systematic review of the

relationship between early opioid use for musculoskeletal disorders and work outcomes 72

Table 3-2 Description of the exposure, outcome, and confounding measures in each study

included in the systematic review ......................................................................................... 73

Table 3-3 Description of the main findings of each study included in the systematic review ..... 75

Table 3-4 Results of the risk of bias assessment of each study included in the systematic review

............................................................................................................................................... 78

Table 4-1 A description of the datasets and elements used in the analyses for Chapters 5 to 7 92

Table 5-1 Characteristics of the cohort of workers’ compensation claimants with an allowed

short-term disability claim for low back pain between 1998 and 2009 (n=97,124) ........... 127

Table 5-2 Prescription dispensing patterns one year before injury date compared to one year

after injury date among claimants with low back pain short-term disability claims

(n=97,124) ............................................................................................................................ 128

Table 6-1 Baseline characteristics of the overall cohort of workers’ compensation claimants with

short-term disability claims for injuries to the low back between 1998 and 2009 (n=95,413)

............................................................................................................................................. 174

Table 6-2 Multinomial logistic regression analysis of factors associated with drug class(es)

dispensed within eight weeks of injury (n=95,413) (no drugs as the reference, n=39,842)

............................................................................................................................................. 177

Table 6-3 Multilevel multinomial logistic regression analysis of factors associated with drug

class(es) at first prescription within 8 weeks of injury (n=54,197) (NSAIDs and/or SMRs only

as the reference, n=32,846) ................................................................................................ 184

Table 7-1 Baseline characteristics of the cohort of workers’ compensation claimants with short-

term disability claims for injuries to the low back from 1998-2009 and with at least one

prescription opioid, NSAID, or SMR dispense in the first eight weeks after injury (n=55,571)

............................................................................................................................................. 226

Table 7-2 Results of a zero-inflated negative binomial regression analysis of the association

between early drug class(es) dispensed, drug class(es) at first prescription, and strength of

xi

opioid received in the first eight weeks after injury and time on benefits and benefit status

after the eight week exposure window up to 1 year post-injury ........................................ 230

Table 7-3 Results of a zero-inflated negative binomial regression analysis of the association

between cumulative days supplied and average daily MED in the first eight weeks after

injury and time on benefits and benefit status after the eight week exposure window up to

1 year post-injury ................................................................................................................. 232

xii

List of Figures

Figure 3-1 Flowchart describing the selection of studies for the systematic review of the

relationship between early opioid prescriptions for musculoskeletal disorders and work

outcomes ............................................................................................................................... 71

Figure 5-1 Claim-level and claimant-level exclusions to derive the final cohort sample of

workers’ compensation claimants with allowed low back pain short-term disability claims

occurring between 1998 and 2009. ..................................................................................... 126

Figure 5-2 Proportion of claimants dispensed their first prescription by week after injury among

claimants with at least one dispensed opioid, NSAID, or SMR prescription. Each bar

represents a 1-week period post-injury. ............................................................................. 132

Figure 5-3 Proportion of claimants with at least one dispensed prescription in 8-week blocks

from 48 weeks pre- to 48 weeks post-injury (n=97,124). ................................................... 133

Figure 5-4 Mean cumulative days supply / claimant in 8-week blocks from 48 weeks pre- to 48

weeks post-injury for claimants with at least one day supplied of opioids and/or SMRs with

codeine. ............................................................................................................................... 134

Figure 5-5 Proportion of claimants receiving at least one dispensed weak, strong, short-acting or

long-acting opioid prescription in 8-week blocks from 48 weeks pre- to 48 weeks post-

injury for claimants with at least one dispensed opioid in the period. ............................... 135

Figure 5-6 Average daily MED / claimant in 8-week blocks from 48 weeks pre- to 48 weeks post-

injury for claimants with at least one day supplied of opioids and/or SMRs with codeine.

............................................................................................................................................. 136

file:///F:/Comments%20on%20Full%20Draft/Carnide-PhD%20Thesis%20Revisions%20Formatted%20(June%2030,%202017).docx%23_Toc486589497



xiii

List of Appendix Tables

Appendix Table A-1 Examples of work outcome measures eligible for inclusion in the systematic

review .................................................................................................................................. 262

Appendix Table A-2 Sample search strategy for the systematic review .................................... 263

Appendix Table A-3 Items considered in the risk of bias assessment ........................................ 268

Appendix Table B-1 ICD-9 diagnoses used to identify LBP-related claims in the WorkSafeBC

claim file ............................................................................................................................... 271

Appendix Table B-2 ICD-9 and ICD-10-CA diagnoses used to exclude claimants with LBP-related

hospitalizations or serious LBP-related outpatient health care visits and/or hospitalizations

within 5 days of injury date ................................................................................................. 274

Appendix Table B-3 Specific prescription drugs included in our study according to drug class 278

Appendix Table B-4 Ratios used to convert opioid daily dose to a standardized morphine-

equivalent dose (MED) ........................................................................................................ 280

Appendix Table B-5 ICD-9 and ICD-10-CA diagnoses used to identify comorbid health conditions

in the two years before injury date ..................................................................................... 281

Appendix Table B-6 CCI and CCP codes used to identify spinal x-rays in the DAD file and fee item

codes used to identify spinal x-rays in the MSP file ............................................................ 300

Appendix Table B-7 CCI and CCP codes used to identify spinal surgeries in the DAD file ......... 301

Appendix Table D-1 Multinomial logistic regression analysis of the association between pre-

injury supplementary health care and drug class(es) dispensed within 8 weeks of injury

among claimants with injury years 1998-2001 (n=34,657) (no drugs as the reference,

n=15,405) ............................................................................................................................. 312

Appendix Table D-2 Multilevel multinomial logistic regression analysis of factors associated

with drug class(es) at first prescription within 8 weeks of injury among claimants with injury

years 1998-2001 (n=18,777) (NSAIDs and/or SMRs only as the reference, n=11,449) ...... 313

Appendix Table E-1 Results of a zero-inflated negative binomial regression analysis of the

association between early drug class(es) dispensed, drug class(es) at first prescription, and

strength of opioid received in the first 4 weeks and time on benefits and benefit status

after the four week exposure window up to 1 year post-injury ......................................... 315

xiv


association between cumulative days supplied and average daily MED in the first 4 weeks

and time on benefits and benefit status after four week exposure window up to 1 year

post-injury ............................................................................................................................ 317


association between early drug class(es) dispensed, drug class at first prescription, and

strength of opioid received and time on benefits and benefit status after exposure window

up to 1 year post-injury among claimants with injury years 1998-2001 ............................ 318


association between cumulative days supplied and average daily MED and time on benefits

and benefit status after exposure window up to 1 year post-injury among claimants with

injury years 1998-2001 ........................................................................................................ 322

xv

List of Appendix Figures

Appendix Figure C-1 Proportion of claimants dispensed their first opioid prescription by week

after injury among claimants with at least one dispensed opioid prescription. Each bar


Appendix Figure C-2 Proportion of claimants dispensed their first NSAID prescription by week

after injury among claimants with at least one dispensed NSAID prescription. Each bar


Appendix Figure C-3 Proportion of claimants dispensed their first SMR prescription by week

after injury among claimants with at least one dispensed SMR prescription. Each bar


Appendix Figure C-4 Count of dispensed opioid prescriptions in 8-week blocks from 48 weeks

pre- to 48 weeks post-injury among claimants with at least one opioid prescription. ...... 307

Appendix Figure C-5 Count of dispensed NSAID prescriptions in 8-week blocks from 48 weeks

pre- to 48 weeks post-injury among claimants with at least one NSAID prescription. ...... 308

Appendix Figure C-6 Count of dispensed SMR prescriptions in 8-week blocks from 48 weeks pre-

to 48 weeks post-injury among claimants with at least one SMR prescription. ................. 309

Appendix Figure C-7 Median cumulative days supply / claimant in 8-week blocks from 48 weeks

pre- to 48 weeks post-injury for claimants with at least one day supplied of opioids and/or

SMR with codeine. ............................................................................................................... 310

xvi

Abbreviations

BC = British Columbia

CCI = Canadian Classification of Health Interventions

CCP = Canadian Classification of Diagnostic, Therapeutic, and Surgical Procedures

CI = Confidence interval

DAD = Discharge Abstract Database

DALY = Disability-adjusted life year

DEA = Drug Enforcement Agency

GP = General practitioner

HR = Hazard ratio

ICD-9 = International Statistical Classification of Diseases and Related Health

Problems, Ninth Revision

ICD-10-CA = International Statistical Classification of Diseases and Related Health

Problems, Tenth Revision, Canada

IQR = Interquartile range

IRR = Incidence rate ratio

LBP = Low back pain

MEA = Morphine-equivalent amount

MED = Morphine-equivalent dose

MG = Milligram

MSD = Musculoskeletal disorder

MSP = Medical Services Plan

MVA = Motor vehicle accident

N/A = Not applicable

NSAID = Nonsteroidal anti-inflammatory drug

OR = Odds ratio

RCT = Randomized controlled trial

RR = Relative risk

RX = Prescription

xvii

SD = Standard deviation

SMR = Skeletal muscle relaxant

US = United States

WC = Workers’ compensation

1

CHAPTER 1 INTRODUCTION

1.1 Introduction

Opioids have been used for the relief of pain for centuries, though not without controversy.(1)

After almost a century of resistance to their use, clinical consensus slowly began to shift in the

eighties and nineties. The genesis of this change appears to trace back to the publication of

several studies and opinion papers, advocating the use of opioids for non-cancer pain and

downplaying the potential for addiction with long-term exposure.(2-4) In 1995, The American

Pain Society introduced the concept of pain as the ‘fifth vital sign’, creating further dialogue

around the necessity to prioritize pain treatment and prevent the needless suffering.(5) By

then, the seeds of a transformation around pain treatment had taken root.

Perhaps the most decisive factor in the move away from opiophobia occurred in 1996, when

Purdue Pharma introduced OxyContin to the market. The company launched an aggressive and

extensive marketing campaign, promoting this long-acting formulation of oxycodone for the

initial treatment of non-cancer pain.(6) Over the next several years, sales would skyrocket, but

by early 2000, reports of widespread abuse began to emerge.(6, 7)

This confluence of events, in hindsight, created a perfect storm that has culminated in what has

become a major public health crisis, perhaps one of the largest we will see in our generation.

Until recently, there has been a substantial and steady increase in the use of prescription

opioids in North America and skyrocketing rates of opioid abuse and overdose have paralleled

these trends.(8-14) From 1999 to 2011, the age-adjusted rate of opioid-related poisoning

deaths in the United States (US) increased 264% from 1.4 per 100,000 to 5.1.(14) Similar

2

increases were noted in Ontario, Canada, where the overall rate increased 242% from 1991 to

2010.(11) While there is some evidence to suggest ongoing efforts to stem the epidemic may be

having a positive impact on reducing prescribing and associated harms,(8, 10, 11, 14-16) there

has been a corresponding surge in the rates of use of heroin and illicitly-manufactured fentanyl

and associated overdose deaths, as abusers of prescription opioids turn to the more freely

available illicit opioid supply.(15-20) And despite some of the progress made, Canada and the

US continue to represent the largest global consumers of prescription opioids per capita.(21)

Back in 2005, in the midst of this rapidly developing crisis, researchers from Washington State

sounded the initial alarm on opioid use among workers’ compensation claimants.(22) From

1996 to 2002, researchers found the proportion of opioid prescriptions reimbursed by the state

workers’ compensation system that were for the most potent opioids increased from 19.3% to

37.2%. During this same time period, accidental poisoning deaths also increased substantially,

with thirty-two deaths classified as definitely or probably related to accidental overdose of

opioids. In 62.5% of these deaths, claims were for low back pain.

Since that seminal paper was published, numerous studies and industry reports have

documented increasing use of opioids among injured workers in the workers’ compensation

system.(23-27) In the province of Ontario, Canada, opioid prescription costs incurred by the

Workplace Safety and Insurance Board increased almost 700% between 2001 to 2009, from

CAD $4.4 million to $35 million (23) and in a National Council of Compensation Insurance

report, 34% of claimants in the United States with prescriptions received opioids in 2008, up

from 27% in 2001.(27) Extensive studies have also demonstrated evidence of adverse outcomes

associated with the use of opioids among claimants, including increased compensation costs

3

and prolonged work disability.(24, 28-38) This accumulation of findings has led to considerable

concern among workers’ compensation organizations across North America. In an effort to

tackle this growing problem, several have developed policies and guidelines to limit

reimbursement of opioids, curb long-term use, and ensure the appropriate provision of opioids

to injured workers.(39-42) The American College of Occupational and Environmental Medicine

also issued two practice guidelines around the use of opioids for the treatment of pain and for

workers with safety-sensitive occupations.(43, 44)

There is recent evidence to suggest the tide may be turning, with opioid prescribing among

workers’ compensation claimants on the decline in some jurisdictions.(25, 26, 45-47) However,

opioids continue to be commonly prescribed. In 41 states of the United States, for example,

opioids represented four of the top ten prescription drugs paid by workers’ compensation in

2014.(48) In California, opioids were the most commonly prescribed drug class, as well as most

costly drug class from 2006 to 2014.(25)

Low back pain (LBP) is among one of the leading diagnoses of disabling work-related injuries in

Canada and the United States (49, 50) and a common reason for which injured workers receive

opioids.(51, 52) Recent studies of patient populations suggest opioids are increasingly provided

early in the course of a LBP episode,(53, 54) a trend that has also been shown among workers’

compensation claimants with LBP claims.(28-30, 33, 37, 55-58) Yet there is little high quality,

randomized controlled trial evidence to support this practice (59, 60) and clinical guidelines

caution against the use of opioids as an initial treatment measure for LBP.(61-63)

Individuals with acute episodes of LBP will typically experience rapid improvements in pain and

4

function soon after onset, even in the absence of active intervention.(64-66) However, research

has shown that the longer a worker is off work, the less likely they are to return.(67, 68) Work

plays a vital role, not only economically for the individual and society, but also as a determinant

of health, quality of life, self-worth, and social standing and identity.(69) Whether the provision

of opioids during the early stages of a new LBP episode has beneficial or harmful effects on

work outcomes is an important issue to consider.

1.2 Overall Thesis Objective and Format of the Dissertation

The overall purpose of this dissertation is to determine the relationship between the early

provision of opioids and work disability. This dissertation is organized using a paper-based

format and is comprised of a series of four manuscripts. In this introduction (Chapter 1), I have

provided some brief information on the prescription opioid crisis in general and how the issue

has impacted injured workers and the workers’ compensation system, including workers with

LBP-related injuries. Chapter 2 then describes relevant background information on LBP and the

prescription medications often used in its management. This information is relevant as pain is

often the main symptom of injuries to the low back, which is a focus of analyses in Chapters 5

through 7. The first manuscript, a systematic review examining the relationship between the

early provision of opioids for musculoskeletal disorders and work outcomes, is found in Chapter

3. In this review, I more broadly consider other types of musculoskeletal disorders for which

pain is also a common feature in addition to LBP, in order to increase the relevance of the

review findings. This review also takes on the form of a critical appraisal of the literature. As will

become apparent later, most studies identified were of LBP populations and it was discovered

through this review that the existing literature is fraught with bias, which calls into question the

5

validity of their findings.

The results of this systematic review inform the data analyses described in three additional

manuscripts. The specific objectives for these analyses and a brief rationale are provided at the

beginning of Chapter 4, along with a brief description of the administrative data sources used.

The analyses for each of these manuscripts are based on the same cohort of workers’

compensation claimants with LBP claims in British Columbia from 1998 to 2009. Chapter 5

presents the results of a descriptive analysis examining and contrasting one year pre- to post-

injury dispensing patterns of opioids, as well as nonsteroidal anti-inflammatory drugs (NSAIDs)

and skeletal muscle relaxants (SMRs). In Chapter 6, the focus is on the first eight weeks after

injury and identifying the factors associated with dispensing of these three drug classes in this

early period. Finally, Chapter 7 describes a series of analyses in which the relationship between

dispensing patterns of opioids, NSAIDs, and SMRs in the first eight weeks after injury and work

disability after eight weeks and up to one year are examined. The final chapter (Chapter 8)

concludes by highlighting the main findings from each of the four manuscripts and their

contribution to the literature, the limitations of these analyses and thoughts on future research

in this area, and finally, the implications of these findings for policy and practice.

1.3 Candidate’s Roles and Responsibilities

The overarching topic and each of the individual objectives were conceived by myself, in

consultation with my committee. I designed and led the systematic review (Manuscript 1) in

Chapter 3. I was responsible for developing the search (in consultation with a librarian), as well

as the relevancy, quality appraisal, and data extraction instruments. In order to secure funding

6

for the analyses conducted for Manuscripts 2 to 4 (Chapters 5 to 7), I developed and led the

submission of three grant proposals, with feedback from my committee, to the Canadian

Institutes of Health Research, WorkSafeBC, and the Workplace Safety and Insurance Board of

Ontario. All three were successful. I also led the submission and coordination of the data access

application process and data retention applications. I was responsible for securing the initial

ethics approval and all renewals. In conjunction with a data analyst, I participated in

establishing the cohort and data cleaning activities. I was responsible for all data analyses for

Manuscripts 2 to 4 (Chapters 5 to 7), with input from my committee. Finally, I was responsible

for all initial drafts of each of the four individual manuscripts and for completing revisions based

on the suggestions of my co-authors.

7

1.4 References

1. Rosenblum A, Marsch LA, Joseph H, Portenoy RK. Opioids and the treatment of chronic pain: controversies, current status, and future directions. Exp Clin Psychopharmacol. 2008;16(5):405-16.

2. Melzack R. The tragedy of needless pain. Sci Am. 1990;262(2):27-33.

3. Portenoy RK, Foley KM. Chronic use of opioid analgesics in non-malignant pain: report of 38 cases. Pain. 1986;25(2):171-86.

4. Porter J, Jick H. Addiction rare in patients treated with narcotics. N Engl J Med. 1980;302(2):123.

5. Quality improvement guidelines for the treatment of acute pain and cancer pain. American Pain Society Quality of Care Committee. JAMA. 1995;274(23):1874-80.

6. United States General Accounting Office (GAO). Prescription Drugs: OxyContin Abuse and Diversion and Efforts to Address the Problem. 2003. Report No.: GAO-04-110. Available from: http://www.gao.gov/new.items/d04110.pdf. Last Accessed: March 14, 2017.

7. Cicero TJ, Inciardi JA, Munoz A. Trends in abuse of Oxycontin and other opioid analgesics in the United States: 2002-2004. J Pain. 2005;6(10):662-72.

8. Dart RC, Surratt HL, Cicero TJ, Parrino MW, Severtson SG, Bucher-Bartelson B, et al. Trends in opioid analgesic abuse and mortality in the United States. N Engl J Med. 2015;372(3):241-8.

9. Fischer B, Jones W, Rehm J. Trends and changes in prescription opioid analgesic dispensing in Canada 2005-2012: an update with a focus on recent interventions. BMC Health Serv Res. 2014;14:90.

10. Gladstone EJ, Smolina K, Morgan SG. Trends and sex differences in prescription opioid deaths in British Columbia, Canada. Inj Prev. 2016;22(4):288-90.

11. Gomes T, Mamdani MM, Dhalla IA, Cornish S, Paterson JM, Juurlink DN. The burden of premature opioid-related mortality. Addiction. 2014;109(9):1482-8.

12. Murphy Y, Goldner EM, Fischer B. Prescription Opioid Use, Harms and Interventions in Canada: A Review Update of New Developments and Findings since 2010. Pain Physician. 2015;18(4):E605-14.

13. Pezalla EJ, Rosen D, Erensen JG, Haddox JD, Mayne TJ. Secular trends in opioid prescribing in the USA. J Pain Res. 2017;10:383-7.

http://www.gao.gov/new.items/d04110.pdf

8

14. Warner M, Hedegaard H, Chen LH. Trends in Drug-Poisoning Deaths Involving Opioid Analgesics and Heroin: United States, 1999-2012. Hyattsville, MD: National Center for Health Statistics; 2014. Available from: https://www.cdc.gov/nchs/data/hestat/drug_poisoning/drug_poisoning_deaths_1999-2012.pdf. Last Accessed: March 14, 2017.

15. Rudd RA, Aleshire N, Zibbell JE, Gladden RM. Increases in Drug and Opioid Overdose Deaths--United States, 2000-2014. MMWR Morb Mortal Wkly Rep. 2016;64:1378-82.

16. Rudd RA, Seth P, David F, Scholl L. Increases in Drug and Opioid-Involved Overdose Deaths - United States, 2010-2015. MMWR Morb Mortal Wkly Rep. 2016;65:1445-52.

17. Compton WM, Jones CM, Baldwin GT. Relationship between Nonmedical Prescription-Opioid Use and Heroin Use. N Engl J Med. 2016;374(2):154-63.

18. Ministry of Public Safety & Solicitor General. Illicit Drug Overdose Deaths in BC: January 1, 2007 to January 31, 2017. Burnaby, BC: Office of the Chief Coroner; 2017. Available from: http://www2.gov.bc.ca/assets/gov/public-safety-and-emergency-services/death-investigation/statistical/illicit-drug.pdf. Last Accessed: March 14, 2017.

19. Frank RG, Pollack HA. Addressing the Fentanyl Threat to Public Health. N Engl J Med. 2017;376(7):605-7.

20. Howlett K, Giovannetti J, Vanderklippe N, Perreaux L. A Killer High: How Canada Got Addicted to Fentanyl. Globe & Mail (online). 2017; Available from: http://www.theglobeandmail.com/news/investigations/a-killer-high-how-canada-got-addicted-tofentanyl/article29570025/. Last Accessed: March 14, 2017.

21. International Narcotics Control Board. Opioid Consumption Motion Chart. Madison, Wisconsin: The Board of Regents of the University of Wisconsin System; 2016; Available from: https://ppsg.medicine.wisc.edu/chart. Last Accessed: December 20, 2016.

22. Franklin GM, Mai J, Wickizer T, Turner JA, Fulton-Kehoe D, Grant L. Opioid dosing trends and mortality in Washington State workers' compensation, 1996-2002. Am J Ind Med. 2005;48(2):91-9.

23. 2012-2016 Strategic Plan: Measuring Results. Q1 2013 Report. Toronto, Ontario: Workplace Safety & Insurance Board; 2013.

24. Bernacki EJ, Yuspeh L, Lavin R, Tao XG. Increases in the use and cost of opioids to treat acute and chronic pain in injured workers, 1999 to 2009. J Occup Environ Med. 2012;54(2):216-23.

25. Hayes S, Swedlow A. Trends in the Use of Opioids in California's Workers' Compensation System. California: California Workers' Compensation Institute; 2016.

https://www.cdc.gov/nchs/data/hestat/drug_poisoning/drug_poisoning_deaths_1999-2012.pdf

https://www.cdc.gov/nchs/data/hestat/drug_poisoning/drug_poisoning_deaths_1999-2012.pdf

http://www2.gov.bc.ca/assets/gov/public-safety-and-emergency-services/death-investigation/statistical/illicit-drug.pdf

http://www2.gov.bc.ca/assets/gov/public-safety-and-emergency-services/death-investigation/statistical/illicit-drug.pdf

http://www.theglobeandmail.com/news/investigations/a-killer-high-how-canada-got-addicted-tofentanyl/article29570025/

http://www.theglobeandmail.com/news/investigations/a-killer-high-how-canada-got-addicted-tofentanyl/article29570025/

https://ppsg.medicine.wisc.edu/chart

9

26. Ireland J, Young B, Swedlow A. Changes in Schedule II & Schedule III Opioid Prescriptions and Payments in California Workers' Compensation. Oakland, California: California Workers' Compensation Institute; 2012.

27. Laws C. Narcotics in Workers Compensation. Florida: National Council on Compensation Insurance (NCCI); 2012.

28. Busse JW, Ebrahim S, Heels-Ansdell D, Wang L, Couban R, Walter SD. Association of worker characteristics and early reimbursement for physical therapy, chiropractic and opioid prescriptions with workers' compensation claim duration, for cases of acute low back pain: an observational cohort study. BMJ Open. 2015;5(8):e007836.

29. Franklin GM, Stover BD, Turner JA, Fulton-Kehoe D, Wickizer TM, Disability Risk Identification Study C. Early opioid prescription and subsequent disability among workers with back injuries: the Disability Risk Identification Study Cohort. Spine. 2008;33(2):199-204.

30. Gross DP, Stephens B, Bhambhani Y, Haykowsky M, Bostick GP, Rashiq S. Opioid prescriptions in canadian workers' compensation claimants: prescription trends and associations between early prescription and future recovery. Spine. 2009;34(5):525-31.

31. Lavin RA, Tao XG, Yuspeh L, Bernacki EJ. Impact of the combined use of benzodiazepines and opioids on workers' compensation claim cost. J Occup Environ Med. 2014;56(9):973-8.

32. Lavin RA, Tao XG, Yuspeh L, Kalia N, Bernacki EJ. Relationship Between Opioid Prescribing Patterns and Claim Duration and Cost. J Occup Environ Med. 2016;58(3):e90-3.

33. Lee SS, Choi Y, Pransky GS. Extent and Impact of Opioid Prescribing for Acute Occupational Low Back Pain in the Emergency Department. J Emerg Med. 2016;50(3):376-84 e1-2.

34. Tao XG, Lavin RA, Yuspeh L, Weaver VM, Bernacki EJ. Is Early Prescribing of Opioid and Psychotropic Medications Associated With Delayed Return to Work and Increased Final Workers' Compensation Cost? J Occup Environ Med. 2015;57(12):1315-8.

35. Tao XG, Lavin RA, Yuspeh L, Weaver VM, Bernacki EJ. The association of the use of opioid and psychotropic medications with workers' compensation claim costs and lost work time. J Occup Environ Med. 2015;57(2):196-201.

36. Volinn E, Fargo JD, Fine PG. Opioid therapy for nonspecific low back pain and the outcome of chronic work loss. Pain. 2009;142(3):194-201.

37. Webster BS, Verma SK, Gatchel RJ. Relationship between early opioid prescribing for acute occupational low back pain and disability duration, medical costs, subsequent surgery and late opioid use. Spine. 2007;32(19):2127-32.

10

38. White JA, Tao X, Talreja M, Tower J, Bernacki E. The effect of opioid use on workers' compensation claim cost in the State of Michigan. J Occup Environ Med. 2012;54(8):948-53.

39. California Division of Workers' Compensation. Medical Treatment Utilization Schedule (MTUS) Opioids Treatment Guidelines. Part 1: Executive Summary, Introduction, and Recommendations. California, 2016.

40. Department WCPQ. Practice Directive #C10-1. Claims With Opioids Prescribed. Richmond, BC: WorkSafeBC; 2009.

41. Office of the Medical Director. Guideline for Prescribing Opioids to Treat Pain in Injured Workers. Olympia, Washington: Washington State Department of Labor & Industries; 2013.

42. Workplace Safety and Insurance Board (WSIB) of Ontario. Enhanced narcotics management for injured workers. 2010; Available from: http://www.wsib.on.ca/WSIBPortal/faces/WSIBDetailPage?cGUID=WSIB014326&rDef=WSIB_RD_ARTICLE&_afrLoop=2075015899742000&_afrWindowMode=0&_afrWindowId=null#%40%3FcGUID%3DWSIB014326%26_afrWindowId%3Dnull%26_afrLoop%3D2075015899742000%26rDef%3DWSIB_RD_ARTICLE%26_afrWindowMode%3D0%26_adf.ctrl-state%3D18vkgcbv2w_4. Last Accessed: December 20, 2016.

43. Hegmann KT, Weiss MS, Bowden K, Branco F, DuBrueler K, Els C, et al. ACOEM practice guidelines: opioids and safety-sensitive work. J Occup Environ Med. 2014;56(7):e46-53.

44. Hegmann KT, Weiss MS, Bowden K, Branco F, DuBrueler K, Els C, et al. ACOEM practice guidelines: opioids for treatment of acute, subacute, chronic, and postoperative pain. J Occup Environ Med. 2014;56(12):e143-59.

45. Franklin GM, Mai J, Turner J, Sullivan M, Wickizer T, Fulton-Kehoe D. Bending the prescription opioid dosing and mortality curves: impact of the Washington State opioid dosing guideline. Am J Ind Med. 2012;55(4):325-31.

46. Garg RK, Fulton-Kehoe D, Turner JA, Bauer AM, Wickizer T, Sullivan MD, et al. Changes in opioid prescribing for Washington workers' compensation claimants after implementation of an opioid dosing guideline for chronic noncancer pain: 2004 to 2010. J Pain. 2013;14(12):1620-8.

47. Thumula V, Wang D, Liu T-C. Interstate Variations in Use of Opioids, 3rd Edition. Cambridge, Massachusetts: Workers Compensation Research Institute; 2016. Report No.: WC-16-22.

48. Lipton B, Colon D. Workers Compensation and Prescription Drugs: 2016 Update. Florida: National Council on Compensation Insurance; 2016.

http://www.wsib.on.ca/WSIBPortal/faces/WSIBDetailPage?cGUID=WSIB014326&rDef=WSIB_RD_ARTICLE&_afrLoop=2075015899742000&_afrWindowMode=0&_afrWindowId=null#%40%3FcGUID%3DWSIB014326%26_afrWindowId%3Dnull%26_afrLoop%3D2075015899742000%26rDef%3DWSIB_RD_ARTICLE%26_afrWindowMode%3D0%26_adf.ctrl-state%3D18vkgcbv2w_4





11

49. Association of Workers' Compensation Boards of Canada. National Work Injury, Disease and Fatality Statistics 2010-2012. Mississauga, ON: Association of Workers' Compensation Boards of Canada; 2014.

50. Bureau of Labor Statistics. Survey of occupational injuries and illnesses. Nonfatal injuries and illnesses. Case and demographic characteristics for work-related injuries and illnesses involving days away from work. Washington, D.C.: U.S. Department of Labor; 2014; Available from: http://www.bls.gov/iif/oshcdnew.htm. Last Accessed: August 4, 2015.

51. Lipton B, Laws C, Li L. Workers Compensation Prescription Drug Study. 2010 Update. National Council on Compensation Insurance Holdings Inc.; 2011.

52. Swedlow A, Ireland J, Johnson G. Prescribing Patterns of Schedule II Opioids in California Workers' Compensation. Oakland, California: California Workers' Compensation Institute; 2011.

53. Ivanova JI, Birnbaum HG, Schiller M, Kantor E, Johnstone BM, Swindle RW. Real-world practice patterns, health-care utilization, and costs in patients with low back pain: the long road to guideline-concordant care. The spine journal : official journal of the North American Spine Society. 2011;11(7):622-32.

54. Ndlovu M, Bedson J, Jones PW, Jordan KP. Pain medication management of musculoskeletal conditions at first presentation in primary care: analysis of routinely collected medical record data. BMC musculoskeletal disorders. 2014;15:418.

55. Cifuentes M, Webster B, Genevay S, Pransky G. The course of opioid prescribing for a new episode of disabling low back pain: opioid features and dose escalation. Pain. 2010;151(1):22-9.

56. Mahmud MA, Webster BS, Courtney TK, Matz S, Tacci JA, Christiani DC. Clinical management and the duration of disability for work-related low back pain. J Occup Environ Med. 2000;42(12):1178-87.

57. Stover BD, Turner JA, Franklin G, Gluck JV, Fulton-Kehoe D, Sheppard L, et al. Factors associated with early opioid prescription among workers with low back injuries. Journal of Pain. 2006;7(10):718-25.

58. Tacci JA, Webster BS, Hashemi L, Christiani DC. Clinical practices in the management of new-onset, uncomplicated, low back workers' compensation disability claims. J Occup Environ Med. 1999;41(5):397-404.

59. Abdel Shaheed C, Maher CG, Williams KA, Day R, McLachlan AJ. Efficacy, Tolerability, and Dose-Dependent Effects of Opioid Analgesics for Low Back Pain: A Systematic Review and Meta-analysis. JAMA Intern Med. 2016;176(7):958-68.

http://www.bls.gov/iif/oshcdnew.htm

12

60. Chou R, Huffman LH. Medications for acute and chronic low back pain: a review of the evidence for an American Pain Society/American College of Physicians clinical practice guideline. Ann Intern Med. 2007;147(7):505-14.

61. Dagenais S, Tricco AC, Haldeman S. Synthesis of recommendations for the assessment and management of low back pain from recent clinical practice guidelines. Spine J. 2010;10(6):514-29.

62. Koes BW, van Tulder M, Lin CW, Macedo LG, McAuley J, Maher C. An updated overview of clinical guidelines for the management of non-specific low back pain in primary care. Eur Spine J. 2010;19(12):2075-94.

63. Wong JJ, Côté P, Sutton DA, Randhawa K, Yu H, Varatharajan S, et al. Clinical practice guidelines for the noninvasive management of low back pain: A systematic review by the Ontario Protocol for Traffic Injury Management (OPTIMa) Collaboration. Eur J Pain. 2017;21(2):201-16.

64. Costa CML, Maher CG, Hancock MJ, McAuley JH, Herbert RD, Costa LO. The prognosis of acute and persistent low-back pain: a meta-analysis. CMAJ. 2012;184(11):E613-24.

65. Pengel LH, Herbert RD, Maher CG, Refshauge KM. Acute low back pain: systematic review of its prognosis. BMJ. 2003;327(7410):323.

66. Hayden JA, Dunn KM, van der Windt DA, Shaw WS. What is the prognosis of back pain? Best Pract Res Clin Rheumatol. 2010;24(2):167-79.

67. Abenhaim L, Rossignol M, Valat JP, Nordin M, Avouac B, Blotman F, et al. The role of activity in the therapeutic management of back pain. Report of the International Paris Task Force on Back Pain. Spine (Phila Pa 1976). 2000;25(4 Suppl):1S-33S.

68. Wynne-Jones G, Cowen J, Jordan JL, Uthman O, Main CJ, Glozier N, et al. Absence from work and return to work in people with back pain: a systematic review and meta-analysis. Occup Environ Med. 2014;71(6):448-56.

69. Black C. Working for a Healthier Tomorrow. Dame Carol Black's Review of the Health of Britain's Working Age Population. London, UK: Parliamentary Press, TSO; 2008. Available from: https://www.gov.uk/government/publications/working-for-a-healthier-tomorrow-work-and-health-in-britain. Last Accessed: March 17, 2017.

https://www.gov.uk/government/publications/working-for-a-healthier-tomorrow-work-and-health-in-britain

https://www.gov.uk/government/publications/working-for-a-healthier-tomorrow-work-and-health-in-britain

13

CHAPTER 2 ADDITIONAL BACKGROUND

2.1 Basic Spine Anatomy

The spine is composed of 33 stacked vertebrae divided into five segments (from top to bottom):

the cervical spine (C1 to C7 vertebrae), the thoracic spine (T1 to T12), the lumbar spine (L1 to

L5), the sacral region (5 vertebrae fused into a single bony element known as the sacrum), and

the coccygeal region (three to four coccygeal vertebrae fused into a single coccyx or

tailbone).(1) Each vertebra consists of two main parts. The vertebral body is the weight-bearing

component of the vertebra. The vertebral arch is anchored to the vertebral body by two

pedicles and the roof of the arch is formed by two laminae that extend from each pedicle and

fuse at the midline. The projections from a vertebral arch (known as processes) act as

attachments for muscles and ligaments, levers for the movement of muscles, and sites of

articulation with adjacent vertebrae.(1)

Adjacent vertebrae are mainly connected by intervertebral discs and zygapophysial joints. The

vertebral arches are connected by pairs of synovial joints that stabilize the spine by limiting

motion,(2) while the intervertebral discs link the vertebral bodies. Each disc consists of a strong

outer ring of collagen and fibrocartilage, called the annulus fibrosus, that surrounds the nucleus

pulposus, a gelatinous centre that acts as a shock absorber between vertebrae. Numerous

ligaments also help to support the joints in connecting the vertebrae.(1)

The space between each adjacent vertebral body and arch align to form the spinal canal, which

encloses and protects the spinal cord and nerves. There are 31 pairs of spinal nerves that

emanate from the vertebral canal between the pedicles of adjacent vertebrae: eight cervical, 12

14

thoracic, five lumbar, five sacral, and one coccygeal. Spinal nerves are connected to the spinal

cord by nerve roots.(3) Intervertebral foramina are formed on each side between adjacent

vertebra, allowing spinal nerves and blood vessels to pass in and out of the spinal canal.(1)

There are also many vertebral muscles that help to support the spine that can be grouped into

three categories: flexors, needed for flexing, bending, lifting, and arching; obliques, needed for

rotation and posture; and extensors, needed for standing and lifting.(4)

The vertebral bodies of the lumbar spine are the largest of the spine and bear the majority of

the body’s weight. This is particularly true for lumbar spine’s lowest two segments, L4-L5 and

L5-S1, which are most susceptible to degradation and injury.(5, 6) The intervertebral discs act as

shock absorbers and allow for flexibility, while the facet joints in the lumbar spine articulate to

enable bending and twisting of the lower back.(5) The sacrum supports the lumbar vertebrae

and transmits load from the trunk of the body to the pelvis and lower extremities.(3)

2.2 Low Back Pain

Low back pain (LBP) has been defined as “pain and discomfort, localised below the costal

margin and above the inferior gluteal folds, with or without leg pain“.(7) It can arise from the

many structures in the lumbar spine, including the intervertebral discs, facet joints, and

ligaments connecting the vertebrae, sprains and strains to the musculature of the back, and

irritation of the nerve roots exiting the spine.(2, 3)

LBP is not a disease, but a symptom and is often classified according to its underlying cause.

Diagnostic classification of LBP is generally consistent across clinical guidelines, classifying LBP

15

into one of three broad categories: specific LBP, nonspecific LBP, and LBP associated with

radiculopathy or radicular syndrome.(8) Specific LBP is that which is due to serious or specific

pathology, such as cancer, infections, fractures, cauda equina syndrome, and inflammatory

disorders. Nonspecific LBP is considered to be that which cannot be attributed to any

recognizable spinal disease or abnormality.(6, 8) Nonspecific LBP, which includes lumbar sprains

and strains, myofascial pain, and spondylosis, is the most common, accounting for

approximately 85% of low back pain cases seen in health care.(9) Finally, LBP associated with

radiculopathy or radicular syndrome is low back pain that is accompanied by radiating leg

pain.(8) Lumbar radiculopathy is dysfunction of a nerve root, namely compression,

inflammation, and/or injury to a nerve root in the low back, and is associated with pain, sensory

impairment, weakness, or diminished deep tendon reflexes in a nerve root distribution. Sciatica

is the most common symptom of radicular pain and refers to pain radiating down the back of

the leg into the foot in the distribution of the sciatic nerve. Radiculopathy is most commonly

due to lumbar disc herniations and spinal stenosis.(10-12)

A herniated disc is a herniation of the inner nucleus pulposus of an intervertebral disc through

the outer annulus fibrosus, which can lead to irritation or compression of adjacent nerve

roots.(12) Spinal stenosis is the narrowing of the spinal canal and may result in compression of

the nerve roots.(12) Spinal stenosis and symptomatic disc herniation account for approximately

3% and 4% of low back pain cases in primary care, respectively.(11) Both nonspecific LBP and

LBP with radiculopathy are generally amenable to conservative therapeutic management.(11,

12)

16

LBP is also frequently classified based on the duration of the symptoms and disability

experienced. Acute LBP has been variably defined as pain present for up to 4 weeks or 6 weeks,

while subacute pain is considered to be pain lasting longer than the acute period, but no longer

than 12 weeks after symptom onset. Chronic low back pain is consistently defined as any pain

present for more than three months.(6, 8, 11, 13)

2.3 Burden of Low Back Pain

LBP is an important public health problem. The age-standardized point prevalence of LBP

globally has been estimated to be 9.4%.(14) In developed countries, most individuals will

experience at least one episode of LBP in their lives, with lifetime prevalence among adults

estimated to range from 58% to 84%.(15-20)

In occupational settings, LBP is one of the most prevalent physical health conditions to affect

workers.(21-23) Estimates from the 2012 National Health Interview Survey suggest

approximately 24% of employed individuals in the United States (US) experienced LBP in the

previous three months.(21) In the 2010 Canadian Community Health Survey, 17.5% of

employed Canadians reported currently experiencing chronic back problems diagnosed by a

health professional.(23)

Back pain is also one of the most common work-related disorders in North America. In the US in

2015, LBP-related injuries accounted for approximately 21% of all occupational musculoskeletal

disorders resulting in time off work.(24) In Canada, work-related back pain, namely of the low

back, is the single most common reason for filing a lost-time workers’ compensation claim,

accounting for approximately 25% of accepted lost-time claims from 2013 to 2015.(25)

17

LBP is also a leading cause of disability. Among 310 conditions studied in the Global Burden of

Disease 2015 study, LBP ranked highest globally in years lived with disability and fourth overall

in disability-adjusted life years (DALYs), accounting for approximately 60 million DALYs

globally.(26, 27) LBP due to occupational exposure to ergonomic risk factors was also estimated

to cause 21.8 million DALYs in 2010, with 35% of all DALYs due to occupational risk factors.(28)

DALYs were highest between 35 and 55 for LBP due to occupational ergonomic factors,

representing prime working years for most workers.(14, 28)

LBP causes substantial work disability.(23, 29-35) In 2012, approximately 291 million workdays

were lost due to back pain in the US alone.(34) It also poses a substantial economic burden in

the workplace. In the US, the indirect costs associated with LBP have been estimated to range

from US $4.6 billion to US$28 billion. While these studies did not examine total costs, in studies

of other countries, indirect costs, primarily driven by absenteeism, early retirement, and

presenteeism, accounted for a median of 85% of all costs associated with back pain.(36) In

Canada, it has been estimated that back problems account for the largest incremental

productivity loss (approximately $621 million CAD) of all major chronic health conditions among

workers.(23)

2.4 Prognosis of Low Back Pain

The short-term prognosis of LBP is typically favourable. Among individuals that seek care for

acute episodes of back pain, marked improvements in pain and function are often rapidly seen

soon after onset.(35, 37-39) In one systematic review,(39) both pain and functional disability

decreased by an average of 58% in the first month and 82% of individuals initially off work

18

returned within one month. Another review found that 68% of workers with back pain

(primarily of the low back) returned to work after one month, 86% between one and six

months, and 93% after six months.(35)

Indeed, most individuals with LBP do not go on to develop chronic disabling symptoms.

However, it is now recognized that LBP is an episodic condition where recurrences of pain and

disability are common.(37, 40, 41) On average, an estimated 60% of individuals with LBP (range

44% to 78%) experience relapses of pain and 33% (range 26% to 37%) experience relapses of

work absence.(41)

Prognosis also appears to vary according to the outcome assessed.(42, 43) A recent systematic

review examined studies of individuals with less than eight weeks of LBP.(42) Researchers

found the median proportion of primary care patients with a poor work-related outcome was

11% at three to six months (range 2% to 20%), as well as at one year (range 9% to 13%), while a

poor outcome related to pain and functional status was present for a median of 26% of patients

at three to six months (range 2% to 48%) and 21% at one year (range 7% to 42%). This

discrepancy in estimates suggests that workers may return to work, but continue to experience

residual pain and functional disability and estimates of prognosis based solely on return-to-

work rates may underestimate the continued burden of LBP experienced by individuals.

2.5 Pharmacologic Treatment of Low Back Pain

LBP has been shown to be one of the most common reasons for seeking healthcare.(44-46)

Pharmacotherapy is typically the basis of LBP management among physicians (45, 47-49) and

accounts for approximately 13% of direct medical costs for LBP, equivalent to the costs

19

associated with primary care and more than twice the costs of diagnostic imaging.(36) Opioids,

nonsteroidal anti-inflammatory drugs (NSAIDs), and skeletal muscle relaxants (SMRs) are the

drugs most commonly prescribed for LBP, including for acute LBP episodes.(45, 47-52)

2.5.1 Guidelines for the Pharmacologic Treatment of Acute and Subacute Low Back Pain

Several guidelines have been published that recommend best practices in the pharmacologic

treatment of acute and/or subacute LBP (11, 53-62) and have been reviewed extensively.(8, 63,

64) An overview of the pharmacologic recommendations in these guidelines is presented in

Table 2-1. Note that this list does not necessarily represent an exhaustive compilation of all

guidelines available. The guidelines were selected if they provided recommendations specific to

acute or subacute LBP and were published in English. Guidelines dating back to 2000 were

included as they are of relevance to the study period (1998 to 2009) for subsequent data

analyses in this dissertation (Chapters 5 to 7). Note that the quality of the evidence upon which

recommendations are made varies considerably across the guidelines and not all guidelines

grade the strength of evidence.

Overall, most guidelines (published from 2000 to 2015) are consistent in their recommendation

that acetaminophen or other simple analgesics be considered as first-line therapeutic options in

the treatment of acute or subacute nonspecific LBP, with NSAIDs generally recommended as a

second-line option.(53, 57-59, 61, 62) Some guidelines also consider NSAIDs as appropriate

first-line treatments or do not distinguish between simple analgesics and NSAIDs as the optimal

initial treatment.(11, 54) Several also cautiously recommend muscle relaxants (11, 53, 55, 57-

59, 61, 62) or opioids (11, 53, 56-58, 61, 62) for severe refractory pain, often stipulating that

20

only short courses be provided (11, 53, 55-58, 61, 62) and at lowest possible doses.(56, 58) A

few guidelines make general recommendations for drug classes indicated for LBP, namely,

simple analgesics, NSAIDs, and SMRs, without mention of a preferential order of treatment.(55,

56, 60) Several guidelines also recommend that medication in general be taken on a time-

contingent, rather than as-needed basis.(57, 58, 61, 62)

A number of guidelines make specific recommendations to avoid use of certain drug classes for

treatment of acute and/or subacute LBP, including opioids,(54) steroids,(53, 55, 58)

anticonvulsants,(57) antidepressants,(11, 53, 57) benzodiazepines,(54) and SMRs for mild to

moderate pain.(53)

Few guidelines provide recommendations specifically for acute and/or subacute LBP with

radiculopathy. In those that do, two recommend simple analgesics and NSAIDs as initial

treatment.(53, 59) For refractory pain, others recommend acetaminophen with an opioid,(59,

61) SMRs,(59, 61) steroids,(53, 59) gabapentin (an anticonvulsant),(11) and norepinephrine

reuptake inhibitors (antidepressants).(53) The Norwegian guideline also recommends against

NSAIDs.(61)

2.5.2 Description of the Most Commonly Recommended Drug Classes for Acute and Subacute Low Back Pain

2.5.2.1 Opioids

Prescription opioids can be categorized as natural alkaloids from the opium poppy (known as

opiates; e.g., morphine, codeine), synthetic derivatives (e.g., fentanyl, meperidine), or semi-

synthetic derivatives (e.g., hydromorphone, oxycodone).(65) Opioids exert their effects by

21

binding to one of three main opioid receptors located in the brain, spinal cord, peripheral

nerves, and gastrointestinal tract: 1) mu receptors (mu1 and mu2), responsible for supraspinal

analgesia, respiratory depression, euphoria, sedation, decreased gastrointestinal motility, and

physical dependence; 2) kappa receptors, responsible for spinal analgesia, sedation, dyspnea,

dependence, dysphoria, and respiratory depression; and 3) delta receptors, responsible for

psychomimetic and dysphoric effects . Most prescription opioids exert their primary analgesic

activity at the mu receptors.(65, 66)

The strength of an opioid binding to its receptor is measured by its affinity. The strength of

activity from this binding or ability to activate an opioid receptor and produce a biological

response is measured by its efficacy. The degree of affinity and efficacy of an opioid can be used

to categorize opioids as agonists, partial agonists, antagonists, and agonists-antagonists.(65-67)

Agonists have both affinity and efficacy. Pure opioid agonists (e.g., morphine, hydromorphone,

fentanyl) stimulate mu receptors and have no ceiling to their analgesic effect, such that the

degree of analgesia induced is limited only by intolerable dose-related adverse effects. Partial

agonists (e.g., buprenorphine) have affinity, but only partial efficacy and exhibit a ceiling effect

on the degree of analgesia that they can produce. That is, increasing the dose beyond a certain

level will only yield greater opioid side effects.(65-67)

Antagonists (e.g., naloxone) have affinity for mu receptors, but no efficacy and can block the

action of an agonist. These opioids are not used for analgesia, but rather to reverse acute

adverse effects of agonists, as well as for abuse deterrence, detoxification, and opioid

maintenance treatment. Opioid agonists-antagonists (e.g., pentazocine, butorphanol) have

22

both agonist and antagonist properties. They have high mu affinity, but poor mu receptor

efficacy and therefore, may act functionally as mu opioid receptor antagonists, as well as having

kappa agonist properties. As a result, they can induce dysphoria, unlike pure agonists, limiting

their abuse potential. They can be used as analgesics, but have a ceiling to their analgesic

effect.(65-67)

Opioids are provided by a number of routes of administration, primarily oral, transdermal, and

parenteral, but some also by rectal, intranasal, and transmucosal routes. Opioids are often

classified as either weak or strong based on their potency relative to morphine, the prototypical

opioid.(68-75) Weak opioids (e.g., codeine, tramadol, propoxyphene) are used to treat mild to

moderate pain, while strong opioids (e.g., morphine, oxycodone, fentanyl) are typically used for

more severe pain. Similarly, opioids can be described based on their duration of action. Short-

acting formulations are characterized by rapid increases in serum blood levels and a shorter

duration of effect, while long-acting formulations release the drug more gradually into the

bloodstream and have a longer half-life.(76) Several opioids are provided in both short- and

long-acting formulations, including hydromorphone, morphine, and oxycodone. Transdermal

fentanyl and buprenorphine are designed to release over a period of time and are inherently

long-acting. Short-acting opioids are typically recommended for short-term management of

acute pain. There are some advocates of using long-acting opioids in chronic pain in order to

maintain more consistent blood serum levels and pain relief, as well as reduce euphoric effects.

However, the evidence to support superiority of long- versus short-acting formulations is

lacking and long-acting formulations can lead to tolerance of effects and potential dose

escalations.(77)

23

The most common adverse effects of opioids include sedation, constipation, nausea, vomiting,

and pruritis.(77) However, opioids are also prone to misuse, abuse, and addiction, as well as

overdose.(78-82) In the US, 165,000 deaths from 1999-2014 were related to prescription

opioids and at least half of all drug overdose deaths involved an opioid.(78) In Ontario, the

annual rate of opioid-related deaths increased 242% from 1991 to 2010, and by 2010, nearly 1

out of every 8 deaths in individuals aged 25-34 were opioid-related.(80)

As a result, all prescription opioids are considered controlled substances in North America,(83,

84) with the exception of tramadol. Tramadol is an atypical opioid with partial mu agonist

activity that is considered by some to be a non-opioid analgesic.(65) When it was first

introduced, it was marketed as being of lower abuse potential and was not scheduled as a

controlled substance in either the US or Canada. However, there is some evidence to suggest

that tramadol abuse is on the rise.(85) In 2014, the US added tramadol to the list of controlled

substances,(86) but in Canada, tramadol is still not scheduled in the Controlled Drugs and

Substances Act, despite calls to do so.(87)

2.5.2.2 Nonsteroidal Anti-Inflammatory Drugs

Nonsteroidal anti-inflammatory drugs (NSAIDs) are non-opioid analgesics used to relieve the

pain and inflammation of musculoskeletal and joint conditions, injuries, and headaches. For

musculoskeletal pain, the main routes of administration are oral and topical.

NSAIDs act by blocking cyclooxygenase (COX) enzymes. These enzymes act as mediators in the

synthesis of prostanoids, which are involved in various processes, including inflammatory

reactions, protection of the gastrointestinal tract and mucosa, platelet function, and

24

endothelial function. There are two main forms of cyclooxygenase. COX-1 is present in most

tissues and performs protective functions throughout the body, including protecting the lining

of the stomach from acid. COX-2 enzymes are induced in various cells during tissue damage and

inflammation, including the joints and muscles. The analgesic and anti-inflammatory effects of

NSAIDs are achieved by the inhibition of COX-2.(77, 88-90)

NSAIDs can be categorized based on their COX inhibition activity. Traditional NSAIDs are

nonselective COX inhibitors (e.g., ibuprofen, naproxen) and generally demonstrate little

preference for either COX enzyme. COX-2 selective inhibitors can inhibit both, but preferentially

inhibit COX-2, with different drugs demonstrating varying degrees of selectivity for COX-2.

Coxibs (e.g., rofecoxib, celecoxib) were developed specifically to be selective for COX-2 and are

typically the drugs that are referred to as COX-2 inhibitors. However, a number of previously

developed and structurally dissimilar NSAIDs have also been found to be COX-2 selective,

including meloxicam, etodalac, and diclofenac.(89-91)

The main adverse events seen with NSAIDs are typically due to the ratio of COX-1 to COX-2

inhibition. Drugs demonstrating greater COX-1 inhibition, namely the traditional NSAIDs, are

associated with increased risks for gastrointestinal adverse effects, including serious harms such

as perforation and bleeding. One of the benefits of COX-2 selective inhibitors is the lower risk of

gastrointestinal effects. However, COX-2 is also responsible for production of Prostaglandin I2,

which has cardioprotective properties, including promoting vasodilation and decreasing platelet

aggregation. Inhibition of COX-2 by selective inhibitors has been associated with an increased

risk of cerebrovascular accidents and myocardial infarction.(77, 89, 92) In particular, rofecoxib

25

was pulled from the market in 2004, five years after its introduction, due to concerns regarding

excess risk of cardiovascular outcomes.(93)

2.5.2.3 Skeletal Muscle Relaxants

Skeletal muscle relaxants (SMRs) are a diverse group of drugs that are chemically dissimilar, but

are grouped based on having a similar indication. They are generally categorized into two broad

groups based on their main approved indication. Anti-spasticity SMRs, including baclofen and

tizandine, are used for the treatment of spasticity associated with central nervous system

disorders, such as spinal cord injury, stroke, and multiple sclerosis. The mechanism of action

differs for each drug.(77, 94) Anti-spasmodic SMRs are indicated for the treatment of muscle

spasms due to musculoskeletal disorders, including low back pain, fibromyalgia, and tension

headaches. Their mechanism of action is unknown, but is presumed to act through their central

nervous system effects. Anti-spasticity drugs are sometimes used for the management of

musculoskeletal conditions despite little evidence to support this practice.(77, 94-96)

SMRs are associated with a high prevalence of central nervous system adverse effects, namely

dizziness and sedation. Carisoprodol, in particular, is prone to abuse and addiction, as it

metabolizes to meprobamate, a controlled substance.(77, 94, 96)

2.5.3 Effectiveness of Opioids, NSAIDs, and SMRs for Acute and Subacute LBP Episodes

In two systematic reviews, only two randomized controlled trials were found that specifically

examined the effectiveness of opioids for acute and subacute episodes of LBP, one finding

opioids to be effective in reducing pain of acute LBP, but not improving return to work, for

26

acute LBP and another study of subacute LBP patients finding no significant difference in pain

intensity when comparing tramadol to tramadol with acetaminophen.(97, 98) Indeed, claims of

effectiveness and recommendations in some LBP guidelines have been based on evidence from

other pain conditions.(98, 99) Authors of a recent systematic review of reviews found no

difference in outcomes when comparing oral NSAIDs to placebo, a muscle relaxant, or other

NSAIDs for non-specific LBP. When looking at non-specific LBP with radiculopathy, outcomes

were also similar when comparing various NSAIDs and the evidence comparing NSAIDs to

placebo was inconclusive.(100) In contrast, two systematic reviews found SMRs were effective

in providing short-term symptom relief and global improvement when compared to placebo.

No differences were seen between SMRs. However, these findings were based on a minimal

number of small, high-quality trials.(98, 101)

Despite guideline recommendations, it appears there is limited high quality evidence to support

the effectiveness of these three drug classes for improving pain and function in recent onset

episodes of LBP. There is also a dearth of evidence on the comparative effectiveness of these

drug classes in managing LBP.

27

2.6 References

1. Drake RL, Vogl W, Mitchell AWM. Gray’s Anatomy for Students. Philadelphia, PA: Churchill Livingstone/Elsevier; 2015.

2. Vora AJ, Doerr KD, Wolfer LR. Functional anatomy and pathophysiology of axial low back pain: disc, posterior elements, sacroiliac joint, and associated pain generators. Phys Med Rehabil Clin N Am. 2010;21(4):679-709.

3. Bogduk N. Clinical Anatomy of the Lumbar Spine and Sacrum. 4th ed. London, UK: Elsevier Churchill Livingstone; 2005.

4. Ullrich Jr PF. Back Muscles and Low Back Pain. 2009; Available from: http://www.spine-health.com/conditions/spine-anatomy/back-muscles-and-low-back-pain. Last Accessed: January 4, 2017.

5. Davis E. Lumbar Spine Anatomy and Pain. 2013; Available from: http://www.spine-health.com/conditions/spine-anatomy/lumbar-spine-anatomy-and-pain. Last Accessed: January 4, 2017.

6. Violante FS, Mattioli S, Bonfiglioli R. Low-back pain. Handb Clin Neurol. 2015;131:397-410.

7. Burton AK, Balague F, Cardon G, Eriksen HR, Henrotin Y, Lahad A, et al. Chapter 2. European guidelines for prevention in low back pain : November 2004. Eur Spine J. 2006;15 Suppl 2:S136-68.


9. Deyo RA, Rainville J, Kent DL. What can the history and physical examination tell us about low back pain? JAMA. 1992;268(6):760-5.

10. Ben-Yishay A. Lumbar Radiculopathy. 2012; Available from: http://www.spine-health.com/conditions/lower-back-pain/lumbar-radiculopathy. Last Accessed: January 4, 2017.

11. Chou R, Qaseem A, Snow V, Casey D, Cross JT, Jr., Shekelle P, et al. Diagnosis and treatment of low back pain: a joint clinical practice guideline from the American College of Physicians and the American Pain Society. Ann Intern Med. 2007;147(7):478-91.

12. Jenkins H. Classification of low back pain. Australas Chiropr Osteopathy. 2002;10(2):91-7.

13. Frank JW, Brooker AS, DeMaio SE, Kerr MS, Maetzel A, Shannon HS, et al. Disability resulting from occupational low back pain. Part II: What do we know about secondary

http://www.spine-health.com/conditions/spine-anatomy/back-muscles-and-low-back-pain

http://www.spine-health.com/conditions/spine-anatomy/back-muscles-and-low-back-pain

http://www.spine-health.com/conditions/spine-anatomy/lumbar-spine-anatomy-and-pain

http://www.spine-health.com/conditions/spine-anatomy/lumbar-spine-anatomy-and-pain

http://www.spine-health.com/conditions/lower-back-pain/lumbar-radiculopathy

http://www.spine-health.com/conditions/lower-back-pain/lumbar-radiculopathy

28

prevention? A review of the scientific evidence on prevention after disability begins Spine. 1996;21(24):2918-29.

14. Hoy D, March L, Brooks P, Blyth F, Woolf A, Bain C, et al. The global burden of low back pain: estimates from the Global Burden of Disease 2010 study. Ann Rheum Dis. 2014;73(6):968-74.

15. Cassidy JD, Carroll LJ, Côté P. The Saskatchewan health and back pain survey. The prevalence of low back pain and related disability in Saskatchewan adults. Spine (Phila Pa 1976). 1998;23(17):1860-6; discussion 7.

16. Fujii T, Matsudaira K. Prevalence of low back pain and factors associated with chronic disabling back pain in Japan. Eur Spine J. 2013;22(2):432-8.

17. Gross DP, Ferrari R, Russell AS, Battie MC, Schopflocher D, Hu RW, et al. A population-based survey of back pain beliefs in Canada. Spine (Phila Pa 1976). 2006;31(18):2142-5.

18. Leboeuf-Yde C, Klougart N, Lauritzen T. How common is low back pain in the Nordic population? Data from a recent study on a middle-aged general Danish population and four surveys previously conducted in the Nordic countries. Spine (Phila Pa 1976). 1996;21(13):1518-25; discussion 25-6.

19. Walker BF, Muller R, Grant WD. Low back pain in Australian adults. health provider utilization and care seeking. J Manipulative Physiol Ther. 2004;27(5):327-35.

20. Walsh K, Cruddas M, Coggon D. Low back pain in eight areas of Britain. J Epidemiol Community Health. 1992;46(3):227-30.

21. Blackwell DL, Lucas JW, Clarke TC. Summary health statistics for U.S. adults: national health interview survey, 2012. Vital Health Stat 10. 2014(260):1-161.

22. Collins JJ, Baase CM, Sharda CE, Ozminkowski RJ, Nicholson S, Billotti GM, et al. The assessment of chronic health conditions on work performance, absence, and total economic impact for employers. Journal of Occupational & Environmental Medicine47(6):547-57. 2005.

23. Zhang W, McLeod C, Koehoorn M. The relationship between chronic conditions and absenteeism and associated costs in Canada. Scand J Work Environ Health. 2016;42(5):413-22.

24. Nonfatal cases involving days away from work: selected characteristics (2011 forward) [data file] [Internet]. n.d. Available from: https://www.bls.gov/data/#injuries. Last Accessed: December 14, 2016.

25. Association of Workers’ Compensation Boards of Canada (AWCBC). National Work Injury/Disease Statistics Program (NWISP). Table 5. Number of Accepted Lost Time Claims, by Part of Body and Jurisdiction, 2013-2015. Mississauga, ON: 2017; Available from:

https://www.bls.gov/data/#injuries

29

http://awcbc.org/wp-content/uploads/2017/02/Lost-Time-Claims-by-PartofBody-and-Jurisdiction-2013-2015.pdf. Last Accessed: May 3, 2017.

26. G. B. D. 2015 DALYs, Hale Collaborators. Global, regional, and national disability-adjusted life-years (DALYs) for 315 diseases and injuries and healthy life expectancy (HALE), 1990-2015: a systematic analysis for the Global Burden of Disease Study 2015. Lancet. 2016;388(10053):1603-58.

27. G. B. D. 2015 Disease and Injury Incidence and Prevalence Collaborators. Global, regional, and national incidence, prevalence, and years lived with disability for 310 diseases and injuries, 1990-2015: a systematic analysis for the Global Burden of Disease Study 2015. Lancet. 2016;388(10053):1545-602.

28. Driscoll T, Jacklyn G, Orchard J, Passmore E, Vos T, Freedman G, et al. The global burden of occupationally related low back pain: estimates from the Global Burden of Disease 2010 study. Ann Rheum Dis. 2014;73(6):975-81.

29. Bielecky A, Chen C, Ibrahim S, Beaton DE, Mustard CA, Smith PM. The impact of co-morbid mental and physical disorders on presenteeism. Scand J Work Environ Health. 2015;41(6):554-64.

30. Burton WN, Pransky G, Conti DJ, Chen CY, Edington DW. The association of medical conditions and presenteeism. Journal of Occupational & Environmental Medicine46(6 Suppl):S38-45. 2004.

31. Goetzel RZ, Hawkins K, Ozminkowski RJ, Wang S. The health and productivity cost burden of the "top 10" physical and mental health conditions affecting six large U.S. employers in 1999. J Occup Environ Med. 2003;45(1):5-14.

32. Ricci JA, Stewart WF, Chee E, Leotta C, Foley K, Hochberg MC. Back pain exacerbations and lost productive time costs in United States workers. Spine. 2006;31(26):3052-60.

33. Stewart WF, Ricci JA, Chee E, Morganstein D, Lipton R. Lost productive time and cost due to common pain conditions in the US workforce. JAMA. 2003;290(18):2443-54.

34. United States Bone and Joint Initiative. The Burden of Musculoskeletal Diseases in the United States (BMUS). Rosemont, IL: 2014; Available from: http://www.boneandjointburden.org. Last Accessed: May 3, 2017.


36. Dagenais S, Caro J, Haldeman S. A systematic review of low back pain cost of illness studies in the United States and internationally. Spine J. 2008;8(1):8-20.

http://awcbc.org/wp-content/uploads/2017/02/Lost-Time-Claims-by-PartofBody-and-Jurisdiction-2013-2015.pdf

http://awcbc.org/wp-content/uploads/2017/02/Lost-Time-Claims-by-PartofBody-and-Jurisdiction-2013-2015.pdf

http://www.boneandjointburden.org/

30


38. Costa CML, Maher CG, Hancock MJ, McAuley JH, Herbert RD, Costa LO. The prognosis of acute and persistent low-back pain: a meta-analysis. CMAJ. 2012;184(11):E613-24.


40. Côté P, Baldwin ML, Johnson WG, Frank JW, Butler RJ. Patterns of sick-leave and health outcomes in injured workers with back pain. Eur Spine J. 2008;17(4):484-93.

41. Hestbaek L, Leboeuf-Yde C, Manniche C. Low back pain: what is the long-term course? A review of studies of general patient populations. European spine journal : official publication of the European Spine Society, the European Spinal Deformity Society, and the European Section of the Cervical Spine Research Society. 2003;12(2):149-65.

42. Chou R, Shekelle P. Will this patient develop persistent disabling low back pain? JAMA. 2010;303(13):1295-302.

43. Mehling WE, Gopisetty V, Bartmess E, Acree M, Pressman A, Goldberg H, et al. The prognosis of acute low back pain in primary care in the United States: a 2-year prospective cohort study. Spine (Phila Pa 1976). 2012;37(8):678-84.

44. Deyo RA, Mirza SK, Martin BI. Back pain prevalence and visit rates: estimates from U.S. national surveys, 2002. Spine (Phila Pa 1976). 2006;31(23):2724-7.

45. Hart LG, Deyo RA, Cherkin DC. Physician office visits for low back pain. Frequency, clinical evaluation, and treatment patterns from a U.S. national survey. Spine (Phila Pa 1976). 1995;20(1):11-9.

46. Hing E, Rui P, Palso K. National Ambulatory Medical Care Survey: 2013 State and National Summary Tables. Available from: https://www.cdc.gov/nchs/data/ahcd/namcs_summary/2013_namcs_web_tables.pdf. Last Accessed: December 16, 2016.

47. Gore M, Tai KS, Sadosky A, Leslie D, Stacey BR. Use and costs of prescription medications and alternative treatments in patients with osteoarthritis and chronic low back pain in community-based settings. Pain Pract. 2012;12(7):550-60.


https://www.cdc.gov/nchs/data/ahcd/namcs_summary/2013_namcs_web_tables.pdf

31

49. Vogt MT, Kwoh CK, Cope DK, Osial TA, Culyba M, Starz TW. Analgesic usage for low back pain: impact on health care costs and service use. Spine (Phila Pa 1976). 2005;30(9):1075-81.

50. Gore M, Sadosky A, Stacey BR, Tai KS, Leslie D. The burden of chronic low back pain: clinical comorbidities, treatment patterns, and health care costs in usual care settings. Spine (Phila Pa 1976). 2012;37(11):E668-77.


52. Ritzwoller DP, Crounse L, Shetterly S, Rublee D. The association of comorbidities, utilization and costs for patients identified with low back pain. BMC Musculoskelet Disord. 2006;7:72.

53. Low back disorders. In: Hegmann KT, editor. Occupational medicine practice guidelines Evaluation and management of common health problems and functional recovery in workers. 3rd ed. ed. Elk Grove Village, IL: American College of Occupational and Environmental Medicine (ACOEM); 2011.

54. Accident Compensation Corporation. New Zealand Acute Low Back Pain Guide. Wellingon, New Zealand: Accident Compensation Corporation; 2004.

55. Agence Nationale d'Accréditation et d'Évaluation en Santé. Guidelines Department. Diagnosis and Management of Acute Low Back Pain (<3 Months) With or Without Sciatica. Paris, France: Agence Nationale d'Accréditation et d'Évaluation en Santé; 2000.

56. Goertz M, Thorson D, Bonsell J, Bonte B, Campbell R, Haake B, et al. Adult Acute and Subacute Low Back Pain. Bloomington, MN: Institute for Clinical Systems Improvement; 2012.

57. Australian Acute Musculoskeletal Pain Guidelines Group. Evidence-Based Management of Acute Musculoskeletal Pain. Brisbane, Australia: The National Health and Medical Research Council; 2003.

58. Toward Optimized Practice (TOP) Low Back Pain Working Group. Evidence-informed primary care management of low back pain: Clinical practice guideline. Edmonton, AB: Toward Optimized Practice; 2015.

59. Negrini S, Giovannoni S, Minozzi S, Barneschi G, Bonaiuti D, Bussotti A, et al. Diagnostic therapeutic flow-charts for low back pain patients: the Italian clinical guidelines. Eura Medicophys. 2006;42(2):151-70.

60. Rossignol M, Arsenault B, Dionne C, Poitras S, Tousignant M, Truchon M, et al. Clinic on Low-Back Pain in Interdisciplinary Practice (CLIP) guidelines. Montreal, QC: Direction de santé publique, Agence de la santé et des services sociaux de Montréal; 2007.

32

61. The Norwegian Back Pain Network Tcu. Acute low back pain. Interdisciplinary clinical guidelines. Oslo, Norway: The Norwegian Back Pain Network; 2002.

62. van Tulder M, Becker A, Bekkering T, Breen A, del Real MT, Hutchinson A, et al. Chapter 3. European guidelines for the management of acute nonspecific low back pain in primary care. Eur Spine J. 2006;15 Suppl 2:S169-91.



65. Trescot AM, Datta S, Lee M, Hansen H. Opioid pharmacology. [Review] [65 refs]. Pain Physician11(2 Suppl):S133-53. 2008.

66. Salinas OJ, Merritt CK. Chapter 7. Analgesics: Opiate Agonists, Mixed Agonists/Antagonists, and Antagonists for Acute Pain Management. In: Kaye AD, Kaye AM, Urman RD, editors. Essentials of Pharmacoogy for Anesthesia, Pain Medicine, and Critical Care. New York, NY: Springer-Verlag; 2015. p. 113-23.

67. Vallejo R, Barkin RL, Wang VC. Pharmacology of opioids in the treatment of chronic pain syndromes. Pain Physician. 2011;14(4):E343-60.

68. Equianalgesic Dosing of Opioids for Pain Management. PL Detail Document. Stockton, CA: Therapeutic Research Center; July 2015.

69. Washington State Agency Medical Directors’ Group (AMDG). Interagency Guideline on Prescribing Opioids for Pain. 2015 [updated; cited.; Available from: http://www.agencymeddirectors.wa.gov/Files/2015AMDGOpioidGuideline.pdf. Last Accessed: September 2, 2015.

70. American Pain Society. Principles of Analgesic Use in the Treatment of Acute Pain and Cancer Pain. 6th Edition ed. Glenview, IL; 2008.

71. Compendium of pharmaceuticals and specialties, online version (e-CPS) [Internet]. Available from: https://www.pharmacists.ca/products-services/compendium-of-pharmaceuticals-and-specialties/ Last Accessed: March 12, 2015.

72. National Opioid Use Guideline Group (NOUGG). Canadian Guideline for Safe and Effective Use of Opioids for Chronic Non-Cancer Pain. Part B. Recommendations for Practice. Appendix B-8. Hamilton, ON: National Pain Centre, McMaster University; 2010.

http://www.agencymeddirectors.wa.gov/Files/2015AMDGOpioidGuideline.pdf

https://www.pharmacists.ca/products-services/compendium-of-pharmaceuticals-and-specialties/


33

73. Svendsen K, Borchgrevink P, Fredheim O, Hamunen K, Mellbye A, Dale O. Choosing the unit of measurement counts: the use of oral morphine equivalents in studies of opioid consumption is a useful addition to defined daily doses. Palliat Med. 2011;25(7):725-32.

74. Turner BJ, Liang Y. Drug Overdose in a Retrospective Cohort with Non-Cancer Pain Treated with Opioids, Antidepressants, and/or Sedative-Hypnotics: Interactions with Mental Health Disorders. J Gen Intern Med. 2015;30(8):1081-96.

75. Von Korff M, Saunders K, Thomas Ray G, Boudreau D, Campbell C, Merrill J, et al. De facto long-term opioid therapy for noncancer pain. Clin J Pain. 2008;24(6):521-7.

76. Argoff CE, Silvershein DI. A comparison of long- and short-acting opioids for the treatment of chronic noncancer pain: tailoring therapy to meet patient needs. Mayo Clin Proc. 2009;84(7):602-12.

77. Chou R. Pharmacological management of low back pain. Drugs. 2010;70(4):387-402.

78. Centers for Disease Control and Prevention. Wide-ranging online data for epidemiologic research (WONDER). Atlanta, GA: CDC, National Center for Health Statistics; 2016; Available from: http://wonder.cdc.gov. Last Accessed: June 20, 2016.

79. Fischer B, Gittens J, Rehm J. Characterizing the 'awakening elephant' of prescription opioid misuse in North America: epidemiology, harms, interventions. Contemporary Drug Problems. 2008;35:397-426.



82. Substance Abuse and Mental Health Services Administration. Drug Abuse Warning Network 2011: National Estimates of Drug-Related Emergency Department Visits. Rockville, MD; 2013.

83. United States Drug Enforcement Administration. Drug Scheduling. Available from: https://www.dea.gov/druginfo/ds.shtml. Last Accessed.

84. Controlled Drugs and Substances Act, Stat. S.C. 1996, c. 19 (1996).

85. Bush DM. Emergency Department Visits for Drug Misuse or Abuse Involving the Pain Medication Tramadol. The CBHSQ Report. Rockville, MD; 2013.

86. Schedules of Controlled Substances: Placement of Tramadol Into Schedule IV, FR Doc No: 2014-15548 (2014).

http://wonder.cdc.gov/

https://www.dea.gov/druginfo/ds.shtml

34

87. Juurlink DN. Prescription Opioids Are Behind Canada's Greatest Health Crisis. Huffington Post Canada (online). 2015 December 2, 2015; Available from: http://www.huffingtonpost.ca/david-juurlink/opioid-addiction-epidemic_b_8691624.html. Last Accessed: January 7, 2017.

88. Bacchi S, Palumbo P, Sponta A, Coppolino MF. Clinical pharmacology of non-steroidal anti-inflammatory drugs: a review. Antiinflamm Antiallergy Agents Med Chem. 2012;11(1):52-64.

89. Hunter TS, Robison C, Gerbino PP. Emerging evidence in NSAID pharmacology: important considerations for product selection. Am J Manag Care. 2015;21(7 Suppl):S139-47.

90. Peterson K, McDonagh M, Thakurta S, Dana T, Roberts C, Chou R, et al. Drug Class Review: Nonsteroidal Antiinflammatory Drugs (NSAIDs). Final Update 4 Report. Portland, OR: Oregon Health & Science University; 2010. Available from: https://www.ncbi.nlm.nih.gov/pubmedhealth/PMH0009765/. Last Accessed: February 11, 2017.

91. Hawkey CJ. COX-1 and COX-2 inhibitors. Best Pract Res Clin Gastroenterol. 2001;15(5):801-20.

92. Bekhit M, Navab K, Ghobrial A, Aust T. Chapter 9. Nonopioid Analgesic and Adjunct Drugs. In: Kaye AD, Kaye AM, Urman RD, editors. Essentials of Pharmacology for Anesthesia, Pain Medicine, and Critical Care. New York, NY: Springer-Verlag; 2015. p. 147-65.

93. Sibbald B. Rofecoxib (Vioxx) voluntarily withdrawn from market. CMAJ. 2004;171(9):1027-8.

94. Lo JC, Kaye AD. Chapter 10. Benzodiazepines and Muscle Relaxants. In: Kaye AD, Kaye AM, Urman RD, editors. Essentials of Pharmacoogy for Anesthesia, Pain Medicine, and Critical Care. New York, NY: Springer-Verlag; 2015. p. 167-78.

95. Chou R, Peterson K. Drug Class Review: Skeletal Muscle Relaxants. Final Report. Portland, OR: Oregon Health & Science University; 2005. Available from: https://www.ncbi.nlm.nih.gov/pubmedhealth/PMH0005346/. Last Accessed: February 11, 2017.

96. See S, Ginzburg R. Choosing a skeletal muscle relaxant. Am Fam Physician. 2008;78(3):365-70.


http://www.huffingtonpost.ca/david-juurlink/opioid-addiction-epidemic_b_8691624.html

https://www.ncbi.nlm.nih.gov/pubmedhealth/PMH0009765/

https://www.ncbi.nlm.nih.gov/pubmedhealth/PMH0005346/

35


99. Deyo RA, Von Korff M, Duhrkoop D. Opioids for low back pain. BMJ. 2015;350:g6380.

100. Wong JJ, Côté P, Ameis A, Varatharajan S, Varatharajan T, Shearer HM, et al. Are non-steroidal anti-inflammatory drugs effective for the management of neck pain and associated disorders, whiplash-associated disorders, or non-specific low back pain? A systematic review of systematic reviews by the Ontario Protocol for Traffic Injury Management (OPTIMa) Collaboration. Eur Spine J. 2016;25(1):34-61.

101. van Tulder MW, Touray T, Furlan AD, Solway S, Bouter LM. Muscle relaxants for non-specific low back pain. Cochrane Database Syst Rev. 2003(2):CD004252.

36

Table 2-1 Overview of clinical guideline recommendations for the pharmacologic treatment of acute and subacute episodes of low back pain

Organization Country Year Target Population Recommended Therapies

Agence Nationale d’Accréditation et d’Évaluation en Santé (55)

France 2000 Acute low back pain (<3 months) with or without sciatica

Recommendations based on a moderate level of evidence: Simple analgesics, NSAIDs, and muscle relaxants are indicated for use No evidence available for combining these drug classes Recommendations based on a weak level of evidence: Systemic corticosteroids have not been shown to be effective

Norwegian Back Pain Network (61)

Norway 2002 Acute non-specific low back pain (<3 months)

Recommendations based on a strong level of evidence: Acetaminophen as a first-line option, then NSAIDs, then acetaminophen

with opioid If necessary, additional muscle relaxant may be considered for a short

time If pain is excessive, analgesics should be taken at regular intervals

Acute nerve root pain/radiculopathy (<3 months)

Recommendations based on a strong level of evidence: Acetaminophen as a first-line option, then acetaminophen with opioid Analgesics should be taken at regular intervals If necessary, additional muscle relaxant may be considered for a short

time NSAIDs are not recommended

Australian Acute Musculoskeletal Pain Guidelines Group (57)

Australia 2003 Acute nonspecific low back pain (<3 months)

Consensus recommendations: Acetaminophen or other simple analgesics as a first-line option for mild

to moderate pain NSAIDs may be used as a second-line therapy when acetaminophen is

insufficient for pain relief Oral opioids may be necessary for severe pain, preferably short-acting

agents at regular intervals, not a pain-contingent basis Adjuvant agents, such as anticonvulsants and antidepressants, are not

recommended

37

Organization Country Year Target Population Recommended Therapies Muscle relaxants cannot be routinely recommended due to adverse

events

Accident Compensation Corporation (54)

New Zealand

2004 Acute low back pain (<3 months)

Recommendations based on a strong level of evidence: Acetaminophen and NSAIDs may be effective Use of opioids or diazepam (benzodiazepine) has been shown to be

harmful

Working Group of the Care and Research Institute (IRCCS) Fondazione Don Carlo Gnocchi ONLUS (59)

Italy 2006 Acute low back pain (≤30 days)

Recommendations based on a strong level of evidence: Acetaminophen should be considered first-line therapy NSAIDs are effective SMRs should not be considered first-line therapy

Acute sciatica (≤30 days)

Recommendations based on a strong level of evidence: Acetaminophen, NSAIDs, SMRs, and tramadol are useful Acetaminophen with weak opioid may be an effective alternative when

NSAIDs or acetaminophen alone do not control pain Recommendations based on a moderate level of evidence: Epidural steroids can reduce radicular pain for a short period if pain not

controlled by other pharmacologic treatment options Recommendations based on a weak level of evidence: Systemic steroids may be useful for a short period

Working Group of the European Commission, Research Directorate-General, department of Policy, Coordination and Strategy (62)

Europe 2006 Acute and subacute nonspecific low back pain (≤12 weeks)

Acetaminophen as a first-line option, then NSAIDs as second-line option A combination of NSAIDs and weak opioids, acetaminophen and weak

opioids, or NSAIDs and muscle relaxants may be considered a third-line option

Only consider adding a short course of muscle relaxants on its own or to NSAIDs if pain uncontrolled by acetaminophen or NSAIDs

Medication is preferably taken at regular intervals

38

Organization Country Year Target Population Recommended Therapies American College of Physicians and American Pain Society (11)

United States

2007 Acute low back pain (<4 weeks)

Recommendations with at least fair evidence: Acetaminophen and NSAIDs are first-line options for most patients NSAIDs should be used at lowest effective doses for shortest periods

necessary A time-limited course of opioids are an option when used judiciously in

those with severe, disabling pain not controlled (or unlikely to be controlled) with acetaminophen and NSAIDs

Short-term use of anti-spasmodic SMRs are an option; no evidence for recommending anti-spasticity SMRs

Short-term use of gabapentin (anticonvulsant) may have small benefits in those with radiculopathy; no evidence for recommending other anticonvulsants

A time-limited course of benzodiazepines may provide short-term relief of pain

Antidepressants are not recommended

Clinic on Low-Back Pain in Interdisciplinary Practice (60)

Canada 2007 Acute low back pain (≤4 weeks)

Recommendations based on a strong level of evidence: Acetaminophen, NSAIDs, muscle relaxants, and combination muscle

relaxants with NSAIDs or acetaminophen can be used Recommendations based on a weak level of evidence: Opioids may be useful compared to non-opioids No evidence: Insufficient information on antidepressants

Subacute low back pain (4-12 weeks)

Recommendations based on a weak level of evidence: Acetaminophen, NSAIDs, and opioids may be useful

American College of Occupational and Environmental Medicine (adapted by state workers’ compensation boards, including NY,

United States

2011 Acute work-related low back pain (not defined)

Recommendations based on a strong level of evidence: Acetaminophen or aspirin as first-line therapy NSAIDs are second-line options Recommendations based on a moderate level of evidence: Muscle relaxants as a second-line treatment for moderate to severe

pain not controlled by NSAIDs Glucosteroids not recommended

39

Organization Country Year Target Population Recommended Therapies

Colorado, and Louisiana) (53)

Recommendations based on a low level of evidence: Limited use of opioids for severe pain without radiculopathy Routine use of opioids not recommended Recommendations based on insufficient level of evidence: No recommendation for use of botulinum toxin injections Antidepressants not recommended Muscle relaxants for mild to moderate pain not recommended

Subacute work-related low back pain (not defined)

Recommendations based on a strong level of evidence: Acetaminophen or aspirin as first-line therapy Recommendations based on a moderate level of evidence: NSAIDs are second-line options Recommendations based on a low level of evidence: Routine use of opioids not recommended Recommendations based on insufficient level of evidence: No recommendation for use of botulinum toxin injections Antidepressants not recommended Muscle relaxants for chronic use in subacute pain not recommended

Work-related

radicular pain syndromes (including sciatica)

Recommendations based on a strong level of evidence: Acetaminophen or aspirin as first-line therapy Recommendations based on a low level of evidence: NSAIDs may be used Norepinephrine reuptake inhibitor antidepressants may be used Glucocorticosteroids for acute severe pain Recommendations based on insufficient level of evidence: Muscle relaxants as second- or third-line treatment for acute pain No recommendation for use of botulinum toxin injections Glucocorticosteroids for mild to moderate pain not recommended

40

Organization Country Year Target Population Recommended Therapies Institute for Clinical Systems Improvement (56)

United States

2012 Acute (≤6 weeks) and subacute (6-12 weeks) low back pain

Recommendations based on a moderate level of evidence: NSAIDs may be used for short-term acute and subacute pain relief Muscle relaxants may be used as an option for acute pain, but possible

side effects should be considered Recommendations based on a low level of evidence: Cautious and responsible use of opioids may be considered for those

carefully selected with severe acute pain not controlled with acetaminophen and NSAIDs, at a minimum effective dose for a limited period of time (usually 1-2 weeks)

Toward Optimized Practice (58)

Canada 2015 Acute and subacute nonspecific low back pain (<3 months)

Acetaminophen as first-line option, NSAIDs are second-line option Analgesics preferably taken at regular intervals Only consider adding a short course of muscle relaxants or

benzodiazepines on their own or to NSAIDs, if acetaminophen or NSAIDs have failed to reduce pain

Opioids should only be considered for carefully selected patients with severe acute pain not controlled with acetaminophen and NSAIDs, at a minimum effective dose only for a limited period of time (1-2 weeks)

Oral and systemic steroids are not recommended Inconclusive evidence to recommend for or against antidepressants,

anticonvulsants, or marijuana

41

CHAPTER 3 EARLY PRESCRIPTION OPIOID USE FOR MUSCULOSKELETAL

DISORDERS AND WORK OUTCOMES: A SYSTEMATIC REVIEW OF THE LITERATURE

Nancy Carnide, Sheilah Hogg-Johnson, Pierre Côté, Emma Irvin, Dwayne Van Eerd, Mieke

Koehoorn, Andrea Furlan

This is a non-final version of an article published in final form in: Carnide N, Hogg-Johnson S,

Côté P, Irvin E, Van Eerd D, Koehoorn M, Furlan AD. Early prescription opioid use for

musculoskeletal disorders and work outcomes: A systematic review of the literature. Clinical

Journal of Pain. 2017 Jul;33(7):647-658.

http://journals.lww.com/clinicalpain/pages/default.aspx

Abstract

Objective Musculoskeletal disorders (MSDs) are a common source of work disability. Opioid

prescribing for MSDs has been on the rise, despite a lack of data on effectiveness. The objective

was to conduct a systematic review to determine whether early receipt of opioids is associated

with future work outcomes among workers with MSDs compared to other analgesics, no

analgesics, or placebo.

Methods MEDLINE, EMBASE, CINAHL, and CENTRAL were searched from inception to 2014 and

reference lists were scanned. Studies were included if opioids were prescribed within 12 weeks

of MSD onset. Eligible outcomes included absenteeism, work status, receiving disability

payments, and functional status. Two reviewers independently reviewed articles for relevance,

risk of bias, and data extraction using standardized forms. Data synthesis using best evidence

http://journals.lww.com/clinicalpain/pages/default.aspx

42

synthesis methods was planned.

Results Five cohort studies met the inclusion criteria, all including workers filing wage

compensation claims. Four studies demonstrated a significant association between early

opioids and prolonged work disability. One study found a shorter time between prescriptions to

be associated with shorter work disability. However, all studies were found to be at a high risk

of bias and a best evidence synthesis could not be conducted. The main limitations identified

were with exposure measurement and control of confounding.

Conclusion Current literature suggests opioids provided within the first 12 weeks of onset of

MSD are associated with prolonged work disability. However, the conclusions of these studies

need testing in a high-quality study that addresses the methodological shortcomings identified

in the current review.

43

3.1 Introduction

Musculoskeletal disorders (MSDs) are highly prevalent in both general and working

populations.(1-4) Among workers, they are among the most common cause of work-related

injuries in North America (5, 6) and are a major source of work disability.(4, 7, 8)

Clinical treatment guidelines emphasize managing acute MSDs with reassurance, avoiding over-

investigation and over-medicalization, and simple symptomatic treatments using

acetaminophen and nonsteroidal anti-inflammatories.(9-12) A short-course of opioid analgesics

may be considered for severe, disabling pain unresponsive (or unlikely to be responsive) to non-

opioid analgesics.(9-12) Recent research findings, however, suggest current management

practices deviate from guideline recommendations. Opioids are commonly prescribed for

MSDs, even at first medical presentation, which has led to significant increases in prescribing

over time.(13-17) Evidence from studies of workers’ compensation claimants with back

disorders points to widespread opioid use throughout the course of a claim, as well as shortly

after claim filing.(18-24) These trends have persisted despite the associated risks of misuse,

addiction, and overdose (25, 26) and a lack of available and convincing data on effectiveness,

particularly with long-term use and for functional outcomes.(25, 27, 28)

For workers with MSDs, improvement in function and return to productive employment is

crucial. An early time-limited course of opioids may enable a worker to take an active role in

recovery, while delaying analgesia may result in delays in return to regular role functioning. On

the other hand, side effects of opioid use, such as somnolence and cognitive impairment (29)

may adversely affect work functioning and put the worker at risk of injury. Furthermore,

aggressive early treatment may increase sick role behaviours that, in turn, inadvertently

44

promote prolonged disability.(30)

The question of whether opioids provided early in the course of an MSD influence work

functioning is important, particularly given the trend toward increasing use. The objective of

this systematic review was to identify and synthesize the best evidence on whether receipt of

opioids within 12 weeks of onset of an MSD is associated with future work outcomes compared

to other analgesics, no analgesics, or placebo. A period of 12 weeks was selected as researchers

and clinicians traditionally consider continued pain and disability beyond this time to indicate a

transition to chronicity.(30, 31)

3.2 Methods

This review follows the Preferred Reporting Items for Systematic Reviews and Meta-Analysis

guidelines,(32) where applicable. There is no published protocol for this review.

3.2.1 Eligibility Criteria

Study characteristics

Studies were eligible if they were original research manuscripts published in peer-reviewed

journals and the study design was a randomized controlled trial (RCT), cohort study, or case-

control study.

Exclusion criteria included letters, editorials, commentaries, unpublished manuscripts,

dissertations, government reports, conference proceedings, meeting abstracts, consensus and

guideline statements, and books. Cross-sectional studies, case reports, case series, and

qualitative studies were not eligible. Only English language studies were included.

45

Participants

Studies examining working aged adults (i.e., 18-65 years old) with musculoskeletal disorders,

pain, injuries, or conditions (herein referred to as MSDs) of any body site were included. Both

non-specific MSDs (e.g., sprains, strains, tendonitis, herniated discs, sciatica) and specific MSDs

(e.g., fractures, dislocations, osteoarthritis) disorders were eligible. Studies based on samples of

surgical patients, post-operative pain, or MSDs due to neoplasms, infections, autoimmune

disorders, amputations, neurologic disease, inflammatory arthritic conditions, burns,

lacerations, or pregnancy were excluded.

Interventions

Opioid prescriptions had to be provided within 12 weeks following MSD onset (defined as date

of first symptoms, injury date, first healthcare visit, or date of filing an insurance claim). Eligible

routes of administration included oral, transdermal, rectal, intranasal, or transmucosal routes.

No restrictions were placed on specific opioid drug, dosage, or duration of use.

Comparison groups

Studies were eligible if they compared opioids to other analgesics, no analgesics, a placebo or

compared different opioids.

Outcomes

Studies with work outcome measures that fell into one of the following categories were eligible

(see Appendix A, Table A-1, which provides examples for each category): 1) general work

measures (e.g., work disability); 2) absenteeism-related and work status measures (e.g., return-

to-work status, duration of work absence); 3) disability payment measures (e.g., time on wage

replacement benefits); and 4) work functioning measures (e.g., work limitations, productivity).

46

3.2.2 Information Sources

The electronic search was developed in consultation with a health sciences librarian. The

following databases were searched from inception to July 2014: MEDLINE, EMBASE, CINAHL,

and the Cochrane Central Register of Controlled Trials. The electronic search strategy was

restricted to English-language studies and guided by the PICO (Population, Intervention,

Comparison, Outcome) framework,(33) incorporating search terms from three broad

categories: 1) Population - musculoskeletal terms; 2) Intervention - opioid analgesics; and 3)

Outcome - work disability outcomes. The search strategy was developed in MEDLINE (see

Appendix A, Table A-2, which provides the detailed MEDLINE strategy) and adapted to other

databases. Database-specific controlled vocabulary terms and text words were included in the

search. Reference lists of included studies and systematic reviews identified through the search

were examined to identify additional studies. Citations were maintained in Reference Manager

(34) and duplicates were removed manually.

3.2.3 Study Selection

Retrieved articles were screened for relevancy using a two-stage process. At each stage, two

researchers in rotating pairs independently screened studies for inclusion using stage-specific

criteria.

In Stage 1, titles and abstracts were assessed for eligibility using a subset of the

inclusion/exclusion criteria (publication language, timing of opioid use, and route of

administration were not considered at this stage). Articles proceeded to the next stage if all

47

criteria were met, there was insufficient information to assess relevancy, or there were

discrepancies between reviewers.

Full-text articles were obtained for studies proceeding from Stage 1 and were independently

screened for relevance using the full list of inclusion/exclusion criteria. Articles identified

through reference lists were also screened at this stage. Only articles meeting all criteria were

retained. Original study authors were contacted by email (up to three attempts) to request

clarification if there was insufficient information to determine relevancy. Discrepancies

between reviewers were resolved by consensus discussion, and where necessary, consultation

with an independent third reviewer.

3.2.4 Assessment of Risk of Bias and Data Synthesis

The risk of bias in each relevant study was independently evaluated by rotating pairs of

reviewers using standardized forms specifically created for this review and adapted from the

Cochrane Risk of Bias Tool (33) and the Quality in Prognosis Studies tool.(35) This assessment

focused mainly on issues relevant to internal validity: selection bias, attrition bias, exposure and

outcome measurement bias, and confounding (see Appendix A, Table A-3, which describes

each risk of bias item assessed). The quality of the analysis was also assessed and, for RCTs

specifically, performance bias (e.g., contamination, compliance) was considered. For each type

of bias, reviewers were asked to consider a number of items that may indicate the potential for

that specific bias to affect study findings. Reviewers were also asked to provide a summary

assessment of whether they believed, based on these items, the study was at risk of the

48

particular bias in question. For each bias, studies were classified as being at high, moderate, or

low risk.

Researchers met to reach consensus on each study’s risk of bias, for each domain and overall.

Following best evidence synthesis methods,(36) only scientifically admissible studies would

form the basis for a qualitative evidence synthesis in this review. Studies were considered

scientifically admissible if they were free of significant methodological flaws and biases (i.e., at

a low risk of bias).

3.2.5 Data Extraction

Using standardized forms developed for this review, the following information was extracted:

general information (e.g., first author, publication date, and country where the study was

conducted), study design, source population, inclusion/exclusion criteria,

exposure/intervention and work outcomes, confounders, sample size, participant

characteristics, participation and follow-up rates, and main study findings. Two researchers

independently extracted data and discrepancies between reviewers were resolved by

consensus. Study authors were contacted by email to obtain missing information or obtain

clarification.

3.3 Results

3.3.1 Study Selection

The electronic searches yielded 3,028 articles (Figure 3-1). A total of 2,675 studies were

excluded at Stage 1 screening, leaving 353 articles. An additional 44 potentially relevant articles

49

were identified from reference lists, for a total of 397 full-text articles to screen for relevance at

Stage 2. Five studies (21, 22, 24, 37, 38) met all inclusion criteria and were retained. The

remaining 392 articles were excluded primarily because of incorrect publication type or study

design and not examining opioid use as a predictor of a work outcome. For six articles, authors

were contacted to obtain further information to determine relevance, but did not respond.

These articles were excluded due to insufficient information.

All five relevant studies were considered to be at a high risk of bias and, therefore, scientifically

inadmissible, preventing a best evidence synthesis. However, to advance this important area of

research and to inform future research, a detailed critical appraisal of their methods is

presented below.

3.3.2 Study Characteristics

No RCTs were identified. The five studies were observational cohort studies conducted in the

United States,(21, 24, 38) Canada,(22) and Australia (37) (Table 3-1). The source population in

four studies (21, 22, 24, 38) was workers’ compensation claimants with compensable work-

related back injuries, while the fifth study (37) was based on motor vehicle accident (MVA)

insurance claimants with musculoskeletal or orthopaedic injuries. The proportion in each study

receiving an early opioid prescription (i.e., within 12 weeks of onset) varied from 5.1% to 34.1%

in three studies.(21, 22, 24) One study did not provide this information,(37) while another (38)

examined only workers who received an early prescription.

All studies used administrative medical billing data to identify opioid prescriptions paid by the

respective insurance organization (Table 3-2). The time window to define prescriptions as early

50

ranged from 10 days to one month following claim onset (22, 24, 37, 38) and six weeks

following the first medical visit.(21) Opioid exposures included binary indicators of prescription

receipt,(22, 37) number of days between prescriptions,(38) cumulative morphine-equivalent

amount,(21, 24) and number of prescriptions and prescription days.(21)

Work outcomes in all studies were constructed from insurance data on wage replacement

benefits. Information on potential confounders (detailed in Table 3-2) was obtained from

administrative data in four studies (22, 24, 37, 38) and telephone interviews in one study.(21)

3.3.3 Summary of Evidence

3.3.3.1 Comparing Workers With Early Opioid Prescriptions to Workers Without Any Early Opioid Prescriptions

Four studies (21, 22, 24, 37) examining the association between early opioid prescriptions and

future work outcomes (when compared to receiving no prescriptions) found a significant

association between early prescriptions and adverse outcomes (Table 3-3). Gross et al.(22)

found that an early opioid prescription was significantly associated with increased wage

replacement benefit duration (hazard ratio [HR] 0.49, 95% confidence interval [CI] 0.48-0.50),

while Berecki-Gisolf et al.(37) found early prescriptions to be associated with receipt of benefits

after 6 months (odds ratio [OR] 2.1, 95% CI 1-5-2.9) and 18 months (OR 2.2, 95% CI 1.5-3.1).

Two other studies (21, 24) found a relationship between increasing early total cumulative

morphine equivalent amount and longer wage replacement benefit duration.

Franklin et al.(21) examined exposure to opioids and wage replacement benefit status at one

year following claim initiation. Compared to not receiving any early opioid prescription,

51

increasing number of prescriptions and quartiles of total morphine equivalent amount (all

quartiles, but one) were associated with receiving benefits at one year. Receiving opioids for 1-

7 days (OR 1.4, 95% CI 0.9-2.2) was not associated with receiving benefits at one year.

However, those who received opioids for more than 7 days were more likely to receive benefits

at one year (OR 2.2, 95% CI 1.5-3.1). Early receipt of more than 7 days of US Drug Enforcement

Agency (DEA) Schedule III opioids (i.e., lower abuse potential than Schedule II opioids,

moderate to low potential for physical and psychological dependence (39)) was also found to

be significantly associated with receiving benefits at one year (OR 2.7, 95% CI 1.8-4.1), but this

association was not seen for Schedule II opioids (i.e., drugs with high potential for abuse and

potential for severe psychological or physical dependence(39)) (OR 1.3, 95% CI 0.8-2.3).

3.3.3.2 Among Only Workers With at Least One Early Opioid Prescription

Cifuentes et al.(38) found that a longer time between prescriptions in the first month after

claim onset was associated with a longer time on benefits within two years of follow-up

(relative risk [RR] 1.14, 95% CI 1.06-1.22).

3.3.4 Risk of Bias Within Studies

The risk of bias assessment results are presented in Table 3-4. Risk of attrition bias was

considered to be low in all studies given outcomes were extracted from administrative data and

were not dependent on study participation. Likewise, it is generally believed that the outcome

of wage replacement receipt is validly and reliably recorded in insurers’ databases, as it

represents a measure for which an insurer is responsible for paying. Therefore, risk of outcome

52

measurement bias was considered low. Statistical analyses were also considered appropriate

given the data structure and study design.

The main limitations in all studies were with exposure measurement and control of

confounding. One study was also threatened by selection bias.(21)

3.3.4.1 Exposure Measurement Bias

Validity of prescription data

All studies relied on data for prescriptions paid by an insurance system. The validity of these

data is presumed high since prescriptions represent an item reimbursed by the system and

there is no reliance on worker recall. However, it is unknown whether opioid prescriptions may

have been paid for outside of these systems (e.g., private insurance), either for the condition

being compensated or for an unrelated health condition. This may have led to underestimation

of exposure to opioid prescriptions in all studies and an overestimation of the association

between early opioids and work disability. A recent study of workers’ compensation claims in

Canada found that, while 80% of claims had all opioid prescriptions captured by workers’

compensation data, the proportion decreased with increasing total morphine equivalent

amount per claim.(40) One study in this review found that 17% of strong opioid prescriptions

overall were not captured in the workers’ compensation billing database and 35% of strong

prescriptions were underestimated among workers with shorter benefit durations.(22) This

differential misclassification may have led to an overestimate of the association between early

opioid prescriptions and disability duration in these data.

53

There is also potential measurement bias given the inability to determine whether prescriptions

dispensed were actually consumed. At least two previous studies suggest underuse of

analgesics is common,(41, 42) finding approximately three quarters of study participants

prescribed an opioid in the preceding 12 months having leftover medication.(41, 42) Actual use

in all studies may have been overestimated, particularly among those with less severe injuries,

potentially leading to an underestimation of the association.

Further, in Berecki-Gisolf et al.,(37) analyses were restricted to workers hospitalized for 1-7

days following an MVA. Drugs prescribed in hospital were not captured. As a result, opioid

prescriptions, captured only in the first 10 days, may have been underestimated, particularly for

those with longer hospital stays (and potentially more severe injuries), leading to a potential

overestimate of the association. However, this may be a minimal threat given the binary nature

of the exposure variable.

Validity of ‘early’ opioid definition

In Berecki-Gisolf et al.,(37) onset was defined as the date of an MVA, creating a valid inception

point for defining early opioid prescriptions. In three of the workers’ compensation studies,

early opioid exposure was defined using administrative dates of claim registration and accident

date.(22, 24, 38) Using these dates assumes onset of the MSD was sudden. However, many

MSDs exhibit a gradual, progressive onset and workers may, in fact, experience symptoms

before filing a work loss claim.(43) MSDs, including back pain, are also episodic in nature.(44,

45) As such, workers who file new claims may be at different clinical stages of their conditions

and may seek treatment prior to filing a claim. Findings of previous studies examining potential

54

pre-claim health care utilization have been conflicting.(46-48) We are aware of one recent

study to examine pre-claim analgesic use, finding that 10.5% of individuals claiming

compensation for an MVA had at least one opioid prescription in the year prior.(49) None of the

three studies (22, 24, 38) accounted for potential pre-claim symptoms and treatment when

establishing the early opioid window and it is unclear whether the opioids provided were truly

measures of early use or whether the observed associations may have been overestimates that,

in part, reflect prior symptomatology and/or opioid use.

In the remaining study,(21) opioid prescriptions received within six weeks of the first medical

visit defined early opioid exposure. In Washington state, “the first medical visit is usually just

after the injury, but prior to date of claim filing” (personal correspondence). Therefore,

misclassification of early prescription was less likely.

Characterization of prescription patterns

In most studies, opioid exposure was defined as binary indicators of early exposure (22, 37) and

cumulative dose.(21, 24) Binary indicators, while important, are broad measures that treat all

opioid prescriptions as equivalent and may obscure potential differential effects of other

important opioid features, such as daily dose and opioid strength. The result may be a bias

toward the null and potential underestimation of the association that mirrors non-differential

misclassification.(50) However, collapsing categories has also been shown to result in

differential misclassification when the categories have different underlying risks of outcome

and probabilities of exposure misclassification.(51)

55

In two studies,(21, 24) early total morphine equivalent amount was calculated, with higher

amounts representing greater opioid exposure. Dose is an important feature to assess, but

defining exposure in this manner is problematic, as this measure assumes that duration and

dose have equal effects on the outcome. The effect of a brief high-dose exposure is, therefore,

presumed similar to that of a prolonged low-dose exposure. While numerically equivalent, their

effects may not be.

Immortal time bias

Both early opioid exposure and the outcome of cumulative time on benefits were estimated

using the same inception point (cohort entry) in three studies.(21, 22, 24) The time between

cohort entry and date of first prescription is ‘immortal’ for the exposed group: (52) to have

received the prescription implies the claimant ‘survived’ on benefits up to the first prescription.

Time on benefits prior to receipt of an opioid prescription is actually ‘unexposed’ time and

measuring time on benefits as of cohort entry attributes days accumulated prior to the first

prescription to opioid exposure, potentially leading to a biased association between early

opioids and longer duration of work disability in these studies. This is likely most problematic in

the study examining a larger six-week exposure window.(21)

3.3.4.2 Confounding Bias

Confounding by indication and severity

Confounding by indication and severity is likely a major source of residual confounding in these

studies. Those who exhibit more disabling MSDs (e.g., greater pain intensity, longer duration of

pain) are potentially more likely to be prescribed opioids (53-55) and may have a more difficult

56

time returning to work.(56, 57) Two studies (24, 38) used ICD-9 codes to measure severity, but

the validity of this method for MSDs is questionable.(58) In the Franklin et al. study,(21)

function and pain intensity were controlled. However, for many, interviews occurred after the

first opioid prescription. These modifiable variables could have then been mediators rather

than confounders and adjustment may have led to an underestimation of the associations

demonstrated.

In four studies,(21, 22, 24, 37) comparisons were made between claimants who received opioid

prescriptions and those who did not. This latter group may include: 1) workers receiving at least

one non-opioid analgesic and having an indication for an analgesic; 2) workers receiving no

analgesics and potentially a higher probability of recovery. This type of comparison can magnify

confounding by indication and may have overestimated the associations between early opioids

and adverse work outcomes seen in these studies. Notably, Gross et al.(22) found similar

associations between early NSAID and early muscle relaxant prescriptions and longer time on

benefits, further suggesting residual confounding by indication may at least partially explain the

findings of these studies.

Cifuentes et al.(38) examined workers with early prescriptions and found, after controlling for

severity, that shorter time between prescriptions (considered an indicator of close clinical

monitoring) was associated with shorter work disability. This is in contrast to the expected

result of longer disability if close monitoring had only represented greater severity. The authors

concluded that this suggests time between prescriptions represents an effect distinct from

clinical severity. It should be noted, however, that the opioid prescription features considered

57

as potential markers of severity (e.g., daily dose) were likely misclassified given days supply was

assumed, not measured.

Confounding by other key variables

In four studies, analysis was restricted to a limited set of administrative variables used for

insurance claims processing.(22, 24, 37, 38) Many variables of theoretical relevance to the

association were not available, including comorbidities (substance abuse, mental health

disorders, other pain-related conditions), co-occurring medications (namely psychotropic

medications), co-occurring supplementary health care, prescribing physician characteristics

(including physician tendency to prescribe and comfort with prescribing opioids), occupational

physical demands, and psychosocial factors, such as self-efficacy. Pre-claim healthcare

utilization is also relevant, as current clinical management of claimants and future work

disability is likely influenced by a history of musculoskeletal symptoms and previous medication

experience. While one study excluded claimants with a prior workers’ compensation claim for

low back pain,(24) none considered prior health care or analgesic prescriptions. Likewise,

opioids are commonly prescribed following orthopedic surgery (59, 60) and time off work is

often required for post-operative recovery. Only one study considered surgery as a potential

confounder.(38) Socioeconomic factors may also contribute to both prescribing and disability.

Adjustment for these factors was limited to pre-injury earnings,(22) occupation,(37) and

education (21) in three studies. Therefore, the threat of residual confounding in these studies is

considered high.

58

Franklin et al.(21) gathered data on a number of relevant variables using participant interviews,

including recovery expectations and fear avoidance beliefs. However, as mentioned previously,

interview delays may have led to adjustment of mediating rather than confounding variables.

3.3.4.3 Selection Bias

Risk of selection bias was minimal in four studies.(22, 24, 37, 38) In the fifth study,(21) the

interview participation rate was low (50.3%) and participants were more likely than non-

participants to be on wage replacement benefits at one year. Participants may, therefore, have

represented more severe cases and may have been more likely to receive an opioid

prescription. Selection bias, if present, could have led to an overestimate of the association in

this study.

3.4 Discussion

We systematically identified and critically reviewed the available peer-reviewed literature

examining the relationship between receipt of opioids early in the course of an MSD and future

work outcomes. We found only five studies, all based on workers claiming compensation for

musculoskeletal injuries. There was a consistent finding across four studies demonstrating a

relationship between receipt of early opioid prescriptions and adverse work outcomes. Greater

dose and duration of use were also associated with longer disability. An additional study

examined only workers with early prescriptions and found that a shorter time between

prescriptions (considered to be an indicator of close clinical monitoring) was associated with a

59

shorter time on disability. However, the methodological deficiencies of the literature limit the

validity of study conclusions.

There was an increased risk of selection bias in one study,(21) which may have led to an

overestimate of the association between early opioid prescriptions and work disability. Across

all studies, the most serious threats to internal validity were: 1) exposure measurement bias,

specifically the incompleteness of prescription data due to a reliance on insurer billing data;

defining opioid receipt as ‘early’ without account of pre-claim health care and analgesic

prescriptions; methods of characterizing early prescription patterns that do not adequately

account for prescribing complexities; and immortal time bias, where time on benefits

accumulated prior to the first prescription may have been attributed to opioid exposure; and 2)

confounding bias, where important factors, such as pre-claim health care and analgesic

prescriptions and comorbidities have not been measured and included in prior analyses. In

particular, residual confounding by indication and severity is likely substantial and one of the

main sources of bias in these studies. However, the magnitude of these biases and their

combined impact on the direction of the findings in each individual study is difficult to estimate,

limiting inference on the true nature of the relationship between early opioids and work.

The question of whether early opioid receipt influences future work outcomes among workers

with MSDs is important – for workers in pain, for physicians tasked with treating workers, and

for employers who require a healthy and productive worker. For work-related disorders,

insurers are also key stakeholders who require definitive answers on how to best treat a worker

in order to expeditiously return him/her to sustainable work.

60

We believe there is a need for a more cautionary and critical approach to the interpretation of

this literature. There is a lack of convincing and valid data on the impact of early opioid use on

work outcomes. We cannot conclude with any degree of certainty that early opioid use

adversely impacts work outcomes, despite the findings of the studies in this review. On the

other hand, there is also no available data to suggest early opioid use improves work outcomes.

This poses a dilemma for all relevant parties, namely clinicians, on how best to treat workers

presenting with musculoskeletal complaints. Clinicians should remain vigilant and cautious in

their approach to treatment of musculoskeletal pain among workers, initiating therapy only if

the anticipated benefits outweigh the risks, ensuring open and transparent communication

with the patient throughout therapy to ensure that changes in pain and function are positive,

clinically relevant and meaningful for the patient, and, as always, monitor for any aberrant

behaviours.

Further investigation is required to make valid inferences about the relationship between early

receipt of opioid prescriptions and work outcomes among workers with MSDs. We recommend

that future research address the limitations of previous studies highlighted in this review.

Namely, improved measurement of opioid exposure is needed. Careful consideration of the

timing of opioid prescriptions relative to disorder onset and in relation to previous treatment is

important, as is maintaining the temporality of the relationship in analysis. Likewise, the data

source selected to define opioid exposure should be comprehensive and complete. Opioid

exposure variables should also adequately address the complexities of drug prescribing. We

might expect workers are provided with drugs from various analgesic drug classes, may switch

drugs, and change doses in response to inadequate analgesia or side effects. Studies should

61

consider the impact of daily dose and how opioid prescriptions provided to workers fit within

the overall context of prescriptions from other drug classes.

Improved control of confounding is another crucial methodological issue. Minimizing the

presence of confounding by indication and severity is absolutely key to determining whether

opioid prescriptions are independently associated with adverse work outcomes rather than

simply representing a marker for severity. Priority must be given to addressing sources of

confounding related to factors that may influence the decision to prescribe opioids, including

measures of disorder severity, duration of pain, functional status, distress, and prior analgesic

experience, as well as other factors previously described, including comorbidities, other

healthcare utilization, and socioeconomic status.

The findings of this review are limited by the small quantity and narrow scope of the available

literature in this area. We were only able to identify five studies, all of which were

observational cohort studies focused on workers claiming wage compensation through

workers’ compensation or MVA insurance schemes. There is evidence of underreporting of

work-related injuries to workers’ compensation in the literature and workers with a shorter

work absence have been shown to be less likely to file a claim.(61, 62) Workers receiving

opioids, but not entering the compensation system (and perhaps with little or no time off work)

would not have been captured in the studies of this review. Four of the five studies also only

included workers with back-related conditions.(21, 22, 24, 38) Future studies should look at this

question using other study designs, including RCTs, and in workers with other MSDs and among

those outside of compensation systems.

62

The literature has also been limited to examining administrative work outcomes based on wage

replacement benefits. However, these outcomes fail to consider the recurrent nature of the

MSD and work disability trajectory and have been shown to underestimate the true extent of

work disability.(63) They are also of limited relevance to workers with MSDs who do not enter

the compensation system. An expansion to other relevant work outcomes is needed, including

recurrences of work disability and at-work functioning.

It may be argued that the findings of the studies in this review remain pertinent, as use of early

opioids may represent a potential “red flag” for insurance schemes to identify claimants at risk

of prolonged disability. However, these findings have been used to guide change in policies and

practices around opioid reimbursement in workers’ compensation settings, suggesting the

reported associations are being interpreted as causally related.

If opioids adversely impact work outcomes, irrespective of severity, what are the potential

underlying mechanisms? It is possible that significant opioid-related side effects, particularly

when first provided and with increasing doses,(29) may make it more difficult for workers to

return to and be productive at work, but this has not yet been studied. Future studies could

examine the side effect profile of workers on opioids to determine whether adverse effects act

as mediating factors.

It is also possible that early use of opioids may lead to prolonged use, potentially placing

workers at risk for opioid misuse and addiction.(25) The literature on opioid misuse among

workers is scant and limited to descriptive analyses.(64, 65) Examining the relationship between

63

problematic opioid use patterns and work outcomes and how it contrasts with that associated

with appropriate use is another potential avenue for further research.

Our review has several strengths. The quality of studies in this review was rigorously evaluated

and a thorough critical appraisal of the evidence was provided, rather than simply pooling

results of studies limited by methodological bias. The search strategy included scanning

reference lists for relevant studies and use of a broad and inclusive set of search terms for the

electronic database searches. Reviewer guides and forms were developed a priori to

standardize the review process and minimize the potential for subjectivity or misinterpretation.

Every effort was also made to contact the original study investigators when clarification was

required. This review also has limitations. We were unable to review the non-English literature

and may have missed relevant studies published in other languages. We also limited studies to

those published in the peer-reviewed literature. Therefore, the potential for publication bias

cannot be excluded, particularly given our search yielded only five studies, all with significant

findings.

3.5 Conclusion

Appropriately informed clinical and policy decisions regarding the provision of opioids for the

treatment of musculoskeletal pain among workers can only be made with valid and robust

evidence. However, our understanding of the relationship between early opioid prescriptions

and future work outcomes among individuals with MSDs is restricted by the paucity of studies

and methodological limitations of the published literature. Future research is needed with a

64

focus on addressing issues of opioid measurement and confounding and an expansion to other

populations and outcomes.

65

3.6 References

1. Côté P, van der Velde G, Cassidy JD, Carroll LJ, Hogg-Johnson S, Holm LW, et al. The burden and determinants of neck pain in workers: results of the Bone and Joint Decade 2000-2010 Task Force on Neck Pain and Its Associated Disorders. Spine. 2008;33(4 Suppl):S60-74.

2. Hoy D, Bain C, Williams G, March L, Brooks P, Blyth F, et al. A systematic review of the global prevalence of low back pain. Arthritis Rheum. 2012;64(6):2028-37.

3. Luime JJ, Koes BW, Hendriksen IJ, Burdorf A, Verhagen AP, Miedema HS, et al. Prevalence and incidence of shoulder pain in the general population; a systematic review. Scand J Rheumatol. 2004;33(2):73-81.

4. McDonald M, DiBonaventura M, Ullman S. Musculoskeletal pain in the workforce: the effects of back, arthritis, and fibromyalgia pain on quality of life and work productivity. J Occup Environ Med. 2011;53(7):765-70.

5. Association of Workers' Compensation Boards of Canada. National Work Injury, Disease and Fatality Statistics 2010-2012. Mississauga, ON: Association of Workers' Compensation Boards of Canada; 2014.

6. Bureau of Labor Statistics. Survey of occupational injuries and illnesses. Nonfatal injuries and illnesses. Case and demographic characteristics for work-related injuries and illnesses involving days away from work. Washington, D.C.: U.S. Department of Labor; 2014; Available from: http://www.bls.gov/iif/oshcdnew.htm. Last Accessed: August 4, 2015.



9. Bigos S, Bowyer O, Braen G, et al. Acute Low Back Problems in Adults. Clinical Practice Guideline No. 14. AHCPR Publication No. 95-0642. Rockville,MD: Agency for Health Care Policy and Research, Public Health Service, U.S. Department of Health and Human Services; 1994.


11. van Tulder M, Becker A, Bekkering T, Breen A, del Real MT, Hutchinson A, et al. Chapter 3. European guidelines for the management of acute nonspecific low back pain in primary care. Eur Spine J. 2006;15 Suppl 2:S169-91.

http://www.bls.gov/iif/oshcdnew.htm

66

12. Waddell G. Preventing incapacity in people with musculoskeletal disorders. Br Med Bull. 2006;77-78:55-69.

13. Caudill-Slosberg MA, Schwartz LM, Woloshin S. Office visits and analgesic prescriptions for musculoskeletal pain in US: 1980 vs. 2000. Pain. 2004;109(3):514-9.

14. Deyo RA, Smith DH, Johnson ES, Donovan M, Tillotson CJ, Yang X, et al. Opioids for back pain patients: primary care prescribing patterns and use of services. J Am Board Fam Med. 2011;24(6):717-27.

15. Larochelle MR, Zhang F, Ross-Degnan D, Wharam JF. Trends in opioid prescribing and co-prescribing of sedative hypnotics for acute and chronic musculoskeletal pain: 2001-2010. Pharmacoepidemiology and Drug Safety. 2015;24(8):885-92.

16. Mafi JN, McCarthy EP, Davis RB, Landon BE. Worsening trends in the management and treatment of back pain. JAMA Internal Medicine. 2013;173(17):1573-81.


18. Berecki-Gisolf J, Collie A, McClure RJ. Prescription opioids for occupational injury: results from workers' compensation claims records. Pain medicine. 2014;15(9):1549-57.


20. Franklin GM, Rahman EA, Turner JA, Daniell WE, Fulton-Kehoe D. Opioid use for chronic low back pain: A prospective, population-based study among injured workers in Washington state, 2002-2005. Clin J Pain. 2009;25(9):743-51.





67

25. Chou R, Turner JA, Devine EB, Hansen RN, Sullivan SD, Blazina I, et al. The effectiveness and risks of long-term opioid therapy for chronic pain: a systematic review for a National Institutes of Health Pathways to Prevention Workshop. Ann Intern Med. 2015;162(4):276-86.

26. Vowles KE, McEntee ML, Julnes PS, Frohe T, Ney JP, van der Goes DN. Rates of opioid misuse, abuse, and addiction in chronic pain: a systematic review and data synthesis. Pain. 2015;156(4):569-76.

27. Chaparro LE, Furlan AD, Deshpande A, Mailis-Gagnon A, Atlas S, Turk DC. Opioids compared with placebo or other treatments for chronic low back pain: an update of the Cochrane Review. Spine. 2014;39(7):556-63.


29. Benyamin R, Trescot AM, Datta S, Buenaventura R, Adlaka R, Sehgal N, et al. Opioid complications and side effects. Pain Physician. 2008;11(2 Suppl):S105-20.

30. Frank J, Sinclair S, Hogg-Johnson S, Shannon H, Bombardier C, Beaton D, et al. Preventing disability from work-related low-back pain. New evidence gives new hope--if we can just get all the players onside. CMAJ. 1998;158(12):1625-31.

31. Frank JW, Brooker AS, DeMaio SE, Kerr MS, Maetzel A, Shannon HS, et al. Disability resulting from occupational low back pain. Part II: What do we know about secondary prevention? A review of the scientific evidence on prevention after disability begins Spine. 1996;21(24):2918-29.

32. Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gotzsche PC, Ioannidis JP, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate healthcare interventions: explanation and elaboration. BMJ. 2009;339:b2700.

33. Cochrane Handbook for Systematic Reviews of Interventions. The Cochrane Collaboration, 2011; Available from: www.cochrane-handbook.org.

34. Reference Manager Professional Network Edition. 9.5N ed. Berkeley, CA: ISI ResearchSoft; 2000.

35. Hayden JA, van der Windt DA, Cartwright JL, Côté P, Bombardier C. Assessing bias in studies of prognostic factors. Ann Intern Med. 2013;158(4):280-6.

36. Slavin RE. Best evidence synthesis: an intelligent alternative to meta-analysis. J Clin Epidemiol. 1995;48(1):9-18.

37. Berecki-Gisolf J, Collie A, McClure R. Work disability after road traffic injury in a mixed population with and without hospitalisation. Accid Anal Prev. 2013;51:129-34.

http://www.cochrane-handbook.org/

68

38. Cifuentes M, Powell R, Webster B. Shorter time between opioid prescriptions associated with reduced work disability among acute low back pain opioid users. J Occup Environ Med. 2012;54(4):491-6.

39. U.S. Drug Enforcement Agency Drug Schedules. Available from: http://www.dea.gov/druginfo/ds.shtml. Last Accessed: January 11, 2016.

40. Shafer LA, Raymond C, Ekuma O, Kraut A. The impact of opioid prescription dose and duration during a workers compensation claim, on post-claim continued opioid use: A retrospective population-based study. Am J Ind Med. 2015;58(6):650-7.

41. Adult use of prescription opioid pain medications - Utah, 2008. MMWR Morb Mortal Wkly Rep. 2010;59(6):153-7.

42. Lewis ET, Cucciare MA, Trafton JA. What do patients do with unused opioid medications? Clin J Pain. 2014;30(8):654-62.

43. Yassi A, Sprout J, Tate R. Upper limb repetitive strain injuries in Manitoba. Am J Ind Med. 1996;30(4):461-72.

44. Enthoven P, Skargren E, Oberg B. Clinical course in patients seeking primary care for back or neck pain: a prospective 5-year follow-up of outcome and health care consumption with subgroup analysis. Spine. 2004;29(21):2458-65.

45. Von Korff M, Saunders K. The course of back pain in primary care. Spine. 1996;21(24):2833-7; discussion 8-9.

46. Atlas SJ, Wasiak R, van den Ancker M, Webster B, Pransky G. Primary care involvement and outcomes of care in patients with a workers' compensation claim for back pain. Spine. 2004;29(9):1041-8.

47. Côté P, Yang X, Kristman V, Hogg-Johnson S, Van Eerd D, Rezai M, et al. The association between workers' compensation claims involving neck pain and future health care utilization: a population-based cohort study. Journal of Occupational Rehabilitation. 2013;23(4):547-56.

48. Koehoorn M, Cole DC, Hertzman C, Lee H. Health care use associated with work-related musculoskeletal disorders among hospital workers. Journal of Occupational Rehabilitation. 2006;16(3):411-24.

49. Berecki-Gisolf J, Hassani-Mahmooei B, Collie A, McClure R. Prescription opioid and benzodiazepine use after road traffic injury. Pain medicine. 2016;17(2):304-13.

50. Rothman KJ, Greenland S, Lash TL. Modern Epidemiology. Philadelphia, PA: Lippincott Williams & Wilkins; 2008.

http://www.dea.gov/druginfo/ds.shtml

69

51. Wacholder S, Dosemeci M, Lubin JH. Blind assignment of exposure does not always prevent differential misclassification. Am J Epidemiol. 1991;134(4):433-7.

52. Suissa S. Immortal time bias in pharmaco-epidemiology. Am J Epidemiol. 2008;167(4):492-9.

53. Fillingim RB, Doleys DM, Edwards RR, Lowery D. Clinical characteristics of chronic back pain as a function of gender and oral opioid use. Spine. 2003;28(2):143-50.


55. Turk DC, Okifuji A. What factors affect physicians' decisions to prescribe opioids for chronic noncancer pain patients? Clin J Pain. 1997;13(4):330-6.

56. Crook J, Milner R, Schultz IZ, Stringer B. Determinants of occupational disability following a low back injury: a critical review of the literature. Journal of Occupational Rehabilitation. 2002;12(4):277-95.

57. Steenstra IA, Verbeek JH, Heymans MW, Bongers PM. Prognostic factors for duration of sick leave in patients sick listed with acute low back pain: a systematic review of the literature. Occup Environ Med. 2005;62(12):851-60.

58. Buchbinder R, Goel V, Bombardier C. Lack of concordance between the ICD-9 classification of soft tissue disorders of the neck and upper limb and chart review diagnosis: one steel mill's experience. Am J Ind Med. 1996;29(2):171-82.

59. Chan EY, Blyth FM, Nairn L, Fransen M. Acute postoperative pain following hospital discharge after total knee arthroplasty. Osteoarthritis and cartilage / OARS, Osteoarthritis Research Society. 2013;21(9):1257-63.

60. Waljee JF, Zhong L, Hou H, Sears E, Brummett C, Chung KC. The Use of Opioid Analgesics following Common Upper Extremity Surgical Procedures: A National, Population-Based Study. Plast Reconstr Surg. 2016;137(2):355e-64e.

61. Rosenman KD, Gardiner JC, Wang J, Biddle J, Hogan A, Reilly MJ, et al. Why most workers with occupational repetitive trauma do not file for workers' compensation. Journal of occupational and environmental medicine / American College of Occupational and Environmental Medicine. 2000;42(1):25-34.

62. Stock S, Nicolakakis N, Raiq H, Messing K, Lippel K, Turcot A. Underreporting work absences for nontraumatic work-related musculoskeletal disorders to workers' compensation: results of a 2007-2008 survey of the Quebec working population. Am J Public Health. 2014;104(3):e94-e101.

70

63. Pole JD, Franche RL, Hogg-Johnson S, Vidmar M, Krause N. Duration of work disability: a comparison of self-report and administrative data. Am J Ind Med. 2006;49(5):394-401.


65. Parhami I, Hyman M, Siani A, Lin S, Collard M, Garcia J, et al. Screening for addictive disorders within a workers' compensation clinic: an exploratory study. Subst Use Misuse. 2012;47(1):99-107.

71

Figure 3-1 Flowchart describing the selection of studies for the systematic review of the relationship between early opioid prescriptions for musculoskeletal disorders and work outcomes

MEDLINE

n=988

EMBASE

n=2,909

CINAHL

n=308

CENTRAL

n=220

Duplicates excluded

(n=1,397) Databases with potentially relevant publications are merged with duplicates

removed (n=3,028)

Stage 1 Screen: titles and abstracts screened for relevance (n=3,028)

Stage 1 Screen: articles excluded (n=2,675)

Stage 2 Screen: full-text articles screened for relevance (n=397)

Final papers retained for quality appraisal (n=5)

Stage 2 Screen: articles excluded (n=392) Primary reasons:

Publication type: n=183

Study design: n=78

Did not examine opioids as a predictor of work outcome: n=86

Not MSD population: n=23

Timing of opioids not within 12 weeks: n=11

Not working age: n=5

Insufficient information (no response from authors): n=6

Additional full-text articles identified from reference lists (n=44)

72

Table 3-1 Characteristics of the cohort studies included in the systematic review of the relationship between early opioid use for musculoskeletal disorders and work outcomes

Studya Source Population Select Sample Characteristics

Studies comparing workers with early opioids prescriptions to those with no early prescriptions

Berecki-Gisolf et al., 2013

Victoria, Australia (18)

Persons injured in motor vehicle accidents with musculoskeletal or orthopaedic injuries from 2005-2007 and claiming income compensation from the Transport Accident Commission (state-run insurance organization for traffic accident injuries)

n = 5,970 Mean age: NR Sex: 63% male Prescribed an early opioid: NR

Gross et al., 2009

Alberta, Canada (22)

Injured workers filing lost-time claims with the Workers’ Compensation Board of Alberta for back sprain/strain from 2000-2005

n = 47,784 Mean age: 37.4 years (SD 11.4) Sex: 66% male Prescribed an early opioid: 5.1%

Franklin et al., 2008 Washington State, United States (21)

Injured workers filing a back injury claim with the Washington State Workers’ Compensation Program (the state fund) from 2002-2004

n = 1,843 Mean age: 39.4 years (SD 11.2) Sex: 68% male Prescribed an early opioid: 34.1%

Webster et al., 2007 United States (46 states, District of Columbia) (24)

Injured workers filing lost-time low back pain claims from 2002-2003 to Liberty Mutual, a large private workers’ compensation insurer representing approximately 10% of the private US workers’ compensation market

n = 8,443 Mean age: 40.3 years (SD 10.4) Sex: 71.8% male Prescribed an early opioid: 21.2%

Studies comparing opioid prescription characteristics among workers with early opioids prescriptions

Cifuentes et al., 2012

United States (46 states and District of Columbia) (19)

Injured workers filing lost-time low back pain claims from 2004-2007 to Liberty Mutual, a large private workers’ compensation insurer representing approximately 10% of the private US workers’ compensation market

n = 1,422 Mean age: 41.9 years (SD NR) Sex: 70.4% male Prescribed an early opioid: 100%

Abbreviations: NR=not reported; SD=standard deviation; US=United States

73

Table 3-2 Description of the exposure, outcome, and confounding measures in each study included in the systematic review Citation Opioid Exposure(s) Work Outcome(s) Confounders

Studies comparing workers with early opioids prescriptions to those with no early prescriptions


(18)

Data source: MVA insurance medical billing data Information available on: Drug ID codes, prescription date, strength, route of administration, form Exposure Definition: Opioid prescription within 10 days of the MVA (yes/no)

Data source: MVA insurance administrative data Outcome Definition: 1. Compensated time off work after 6 months (yes/no) (i.e., having any loss of earnings payments after 6 months from date of accident) 2. Loss of earning capacity (yes/no) (i.e., payments beyond 18 months)

Data source: MVA insurance administrative data Potential Confounders: Age, gender, occupation, accident location, type of injury

Gross et al.,

2009 (22)

Data source: WC medical billing data Information available on: Drug ID codes, prescription date Exposure Definition: Opioid prescription within 2 weeks of claim registration (yes/no)

Data source: WC administrative data Outcome Definition: Cumulative days receiving temporary total wage replacement benefits up to one year after initial claim receipt

Data source: WC administrative data Potential Confounders: Age, gender, gross pre-injury salary, # of previous claims, year of claim

Franklin

et al., 2008 (21)

Data source: WC medical billing data Information available on: Drug ID codes, prescription date, strength, # of pills, days supply, US DEA drug schedule Exposure Definition: Within 6 weeks of the first medical visit: 1. # of opioid prescription days; 2. # of opioid prescription days by drug schedule; 3. Total cumulative MEA1; 4. # of opioid prescriptions

Data source: WC administrative data Outcome Definition: 1. Receipt of temporary total wage replacement benefits (yes/no) at one year after initial claim receipt 2. Cumulative days receiving temporary total wage replacement benefits up to one year after initial claim receipt

Data source: Telephone interviews Potential Confounders: Age, gender, race/ethnicity, education, injury severity, pain intensity, radiating pain, functional status, tobacco and alcohol use, recovery expectations, fear-avoidance beliefs, SF-36 mental health score

74

Citation Opioid Exposure(s) Work Outcome(s) Confounders Webster

et al., 2007 (24)

Data source: WC medical billing data Information available on: Drug ID codes, prescription date, strength, # of pills Exposure Definition: Total cumulative MEA2 within the first 15 days of the accident date listed on the claim

Data source: WC administrative data Outcome Definition: Cumulative days receiving temporary total and partial wage replacement benefits up to two years after initial claim receipt

Data source: WC administrative data Potential Confounders: Age, gender, job tenure, injury severity


Cifuentes et al.,

2012 (19)

Data source: WC medical billing data Information available on: Drug ID codes, prescription date, strength, # of pills Exposure Definition: Average number of days between opioid prescriptions in the first month after claim onset

Data source: WC administrative data Outcome Definition: Cumulative days receiving temporary total wage replacement benefits after the first month and up to two years after initial claim onset

Data source: WC administrative data Potential Confounders: Age, gender, job tenure, injury year, injury severity, lumbar surgery, comorbidity, medical visits, daily dose, % of opioid days on strong/pure/long-acting opioids

Abbreviations: ID=identification; MEA=morphine-equivalent amount; MVA=motor vehicle accident; SF-36=Short-Form 36; US DEA=United States Drug Enforcement Agency; WC=workers’ compensation

75

Table 3-3 Description of the main findings of each study included in the systematic review Outcome Study Exposure Adjusted Findings

Studies comparing workers with early opioids prescriptions to those with no early prescriptionsa

Cumulative duration of wage

replacement benefits

Gross et al., 2009

(22)

Prescription (yes/no) within 2 weeks claim registration

The unadjusted mean/median days on benefits by exposure status was not reported. Receiving at least one prescription associated with a longer time on benefits (HR 0.49, 95% CI 0.48-0.50).

Webster et al.,

2007 (24)

Total cumulative MEA within 15 days of accident

date

The unadjusted mean (SD) days on benefits were 121.1 (204.3) days for those receiving 0mg, 124.1 (202.3) days for those receiving 1-140mg, 149.6 (217.4) days for those receiving 141-225mg, and 175.5 (232.7) days for those receiving 226-450mg. Total number of days on benefits were, on average, 21.9 days (95% CI 3.2-40.6) higher among those receiving 141-225mg, 43.8 days (95% CI 23.7-63.9) higher among those receiving 226-450mg, and 69.1 days (95% CI 49.3-89.0) higher among those receiving >450mg. Receiving 1-140mg was not associated with total number of days on benefits (5.2 days, 95% CI -14.6-25.0).

Franklin et al., 2008

(21)

Total cumulative MEA within 6 weeks of first

medical visit

The unadjusted mean/median number of days on benefits by MEA was not reported. For each 100mg MEA increase, total number of days on benefits increased, on average, by 1.5 days (95% CI 0.2-2.7, p=0.025) over the first year of the claim.

76

Outcome Study Exposure Adjusted Findings

Receipt of wage replacement

benefits at one year (yes/no)


(21)

Total cumulative MEA within 6 weeks of first

medical visit

The unadjusted proportion of claimants receiving benefits at 1 year was 9% for those with no opioids, 16% for those receiving up to 150mg, 22% for those receiving 151-300mg, 22% for those receiving 301-650mg, and 35% for those receiving >650mg.

Receiving up to 150mg (OR 1.9, 95% CI 1.2-3.1), 151-300mg (OR 2.0, 95% CI 1.2-3.3), and >650mg (OR 1.9, 95% CI 1.2-2.9) were associated with receiving benefits at 1 year.

Receiving 301-650mg was not associated with receiving benefits at 1 year (OR 1.6, 95% CI 0.9-2.6).


(21)

# of days within 6 weeks of first medical visit

The unadjusted proportion of claimants receiving benefits at 1 year was 9% for those with no opioids, 14% for those receiving opioids for 1-7 days, and 31% for those receiving opioids for > 7 days. Receiving opioids for > 7 days was associated with being on benefits at 1 year (OR 2.2, 95% CI 1.5-3.1). Receiving opioids for 1-7 days was not associated with being on benefits at 1 year (OR 1.4, 95% CI 0.9-2.2).


(21)

# of days within 6 weeks of first medical visit by DEA

schedule

The unadjusted proportion of claimants receiving benefits at 1 year was 9% for those with no opioids, 15% for those receiving opioids 1-7 days of Schedule II or III opioids, 27% for those receiving > 7 days of Schedule III opioids, and 33% for those receiving >7 days of Schedule II opioids. Receiving > 7 days of Schedule III opioids (OR 2.7, 95% CI 1.8-4.1) and 1-7 days of Schedule II or III opioids (OR 1.5, 95% CI 1.0-2.3) were associated with receiving benefits at 1 year. Receiving >7 days of Schedule II opioids was not associated with receiving benefits at 1 year (OR 1.3, 95% CI 0.8-2.3)

77

Outcome Study Exposure Adjusted Findings


(21)

# of prescriptions within 6 weeks of first medical visit

The unadjusted proportion of claimants receiving benefits at 1 year was 9% for those with no opioids, 15% for those receiving 1 prescription, 23% for those receiving 2 prescriptions, 36% for those receiving 3 prescriptions, and 38% for those receiving >3 prescriptions. Receiving 1 (OR 1.5, 95% CI 1.0-2.3), 2 (OR 1.8, 95% CI 1.1-3.0), 3 (OR 2.5, 95% CI 1.4-4.3), and >3 prescriptions (OR 2.2, 95% CI 1.3-3.6) were all associated with receiving benefits at 1 year.

Receipt of wage

replacement benefits between >6 months and 17 months (yes/no)


(18)

Prescription (yes/no) within 10 days of traffic accident

The unadjusted proportion of claimants receiving benefits after 6 months by exposure status was not reported. Receiving at least one prescription was associated with receiving benefits after 6 months (OR 2.1, 95% CI 1.5-2.9).

Receipt of loss of earning capacity

payments after 18 months (yes/no)


(18)

Prescription (yes/no) within 10 days of traffic accident

The unadjusted proportion of claimants receiving loss of earning capacity payments by exposure status was not reported. Receiving at least one prescription was associated with receiving loss of earning capacity payments (OR 2.2, 95% CI 1.5-3.1).


Cumulative duration of wage

replacement benefits

Cifuentes et al.,

2012 (19)

Time between prescriptions in the first month after

claim onset

Unadjusted mean/median days on benefits according to days between opioid prescriptions not reported. Longer time between opioid prescriptions was associated with longer time on benefits (RR 1.14, 95% CI 1.06-1.22)

Abbreviations: HR=hazard ratio; MEA=morphine-equivalent amount; OR=odds ratio; RR=relative risk a Reference group: Did not receive an early opioid prescription (as per study-specific definition of early)

78

Table 3-4 Results of the risk of bias assessment of each study included in the systematic review

Domain

Berecki-Gisolf et al.,

2013 (18)

Cifuentes et al., 2012 (19)

Gross et al., 2009 (22)

Franklin et al., 2008 (21)

Webster et al., 2007 (24)

Selection Bias

Attrition Bias

Exposure Measurement Bias

Outcome Measurement Bias

Confounding Bias

Analysis

Low risk

Moderate risk

High risk

79

CHAPTER 4 SPECIFIC CHAPTER OBJECTIVES AND DATA SOURCE INFORMATION

4.1 Summary and Chapter Objectives

The systematic review in Chapter 3 has highlighted the limitations of the previous literature

examining the effect of early opioid provision on work outcomes among workers with

musculoskeletal disorders, namely workers’ compensation claimants with low back pain (LBP)

injuries. Some of the key limitations of this literature include incomplete prescription data, a

lack of account of pre-injury health care and drug prescriptions, immortal time bias, and

residual confounding, namely confounding by indication and severity.

Informed by the findings of this systematic review, the following three manuscripts (Chapters 5

to 7) involve an analysis of administrative data from a population-based historical cohort of

workers’ compensation claimants with LBP injuries requiring short-term disability benefits in

British Columbia (BC), Canada. There are two main data providers: Population Data BC, one of

the world’s largest collections of health data, and PharmaNet, a province-wide network

containing prescription dispensing data from all community and hospital outpatient pharmacies

in BC. These data are a rich source of linkable, individual-level information on workers’

compensation claims, outpatient billing records, hospital discharges, pharmacy dispensing, and

sociodemographics. Unlike previous studies, these data contain all medical services and

prescriptions provided to claimants, without restriction to items funded by workers’

compensation.

In addition to opioids, these analyses also consider two drug classes commonly used for the

treatment of LBP: nonsteroidal anti-inflammatory drugs (NSAIDs) and skeletal muscle relaxants

80

(SMRs). The detailed dispensing data provide an opportunity to characterize and contrast pre-

and post-injury prescription patterns to identify whether there is evidence of increasing pre-

injury dispensing leading up to the claim and whether post-injury dispensing levels return to

pre-injury levels by one year after injury. These data are also leveraged to better understand

how pre-injury health care and prescription patterns may influence early post-injury dispensing.

Finally, these data allow a re-examination of the relationship between early opioid prescriptions

and work disability, while addressing some of the limitations of the previous literature. To avoid

immortal time bias, analyses are conducted using an event-based cohort approach (1), whereby

the relationship between opioid prescription patterns in the first eight weeks of claim are

examined with work outcomes following this exposure period. To minimize confounding by

indication and severity, analyses are limited to individuals with at least one prescription for

opioids, NSAIDs, or SMRs, resulting in a sample of workers who all have, to some degree, an

indication for receiving these drugs and perhaps a more homogeneous sample of workers with

respect to severity. This also allows an analysis of the comparative effect of opioids versus

NSAIDs and SMRs on work disability outcomes. Pre-injury health care and prescription

dispensing, which could influence post-injury care and recovery, are also accounted for in these

analyses. This is a significant improvement over previous studies, where pre-injury factors were

not available in workers’ compensation billing data.

In order to identify true early exposure to opioids, a ‘new user’ design was initially considered,

where claimants with a pre-injury history of dispensing (i.e., prevalent users) would be excluded

from the analysis. Claimants with prior low back pain health care utilization and workers’

compensation claims were also considered for removal from the cohort. Ultimately, however,

81

the decision was made to retain these claimants, primarily because restricting the sample in

this manner would likely not be reflective of the claimant population typically seen by

WorkSafeBC. This would compromise the external validity of the study, making the findings

potentially less generalizable and relevant for clinicians and policy makers. However, in order to

ensure internal validity was not compromised by this decision, potential confounding due to

pre-injury dispensing, health care utilization, and workers’ compensation claims was considered

in all analyses.

BC is a province in Canada in which the opioid crisis has had a particularly devastating

impact.(2) In the context of workers’ compensation, WorkSafeBC, the provincial workers’

compensation organization, has led the way in Canada in the development of policies and

guidance around opioid use and its reimbursement among workers’ compensation claimants,

with the initial policy implemented in 1989.(3) Still, in 2008, opioids represented 13 of the top

50 drugs prescribed to claimants and paid by WorkSafeBC and 25.2% of prescription costs

incurred by the provincial workers’ compensation organization, WorkSafeBC.(4) LBP claims also

account for a quarter of all lost-time claims made to WorkSafeBC.(5) As such, British Columbia

is an ideal province in which to conduct this research.

4.2 Chapter Objectives

The overall objective of this dissertation is to determine whether the early provision of opioid

prescriptions to workers after a new, compensated work-related LBP injury is associated with

future work disability.

82

The specific objectives are to:

1. To describe and compare pre- and post-injury prescription opioid, NSAID, and SMR

dispensing patterns

2. To examine the factors associated with prescription opioid, NSAID, and SMR dispensing

in the first eight weeks after injury

3. To examine the relationship between early prescription opioid, NSAID, and SMR

dispensing patterns in the first eight weeks after injury and their relationship with work

disability after eight weeks and up to one year

Eight weeks was chosen to coincide with WorkSafeBC policy at the time of the study that

limited reimbursement for prescription opioids to the first eight weeks after injury or surgery

for most claims.(6) In 2015, this policy was revised to limit payment to the first four weeks.(7)

4.3 Administrative Datasets

Data for the analyses in the following three chapters were obtained from administrative data

delivered by Population Data BC.(8, 9) Population Data BC is a multi-university provincial

population health data service containing longitudinal, person-specific, de-identified health

data for British Columbia’s four and a half million residents. Routine data collected by various

public bodies, including provincial health service and hospitalization records, workers’

compensation claims, and census data are among Population Data BC’s data holdings.

PharmaNet dispensing data are not currently held in-house at Population Data BC, but data are

linkable to Population Data BC data holdings.

83

Data from six databases from 1996 through to 2010 were linked by Population Data BC staff

using both deterministic and probabilistic matching techniques (Table 4-1): WorkSafeBC Claims

and Firm Level Files,(10) PharmaNet,(11) Medical Services Plan (MSP) Payment Information

File,(12) Discharge Abstract Database (DAD),(13) Consolidation File (MSP Registration and

Premium Billing),(14) and MSP Practitioner File.(15) Data were linkable through claimant’s

Personal Health Number, WorkSafeBC claim numbers, and practitioner’s MSP registration

numbers. Data were provided anonymously at the individual level using anonymous study

identifiers. WorkSafeBC, PharmaNet, British Columbia Ministry of Health, and the College of

Physicians and Surgeons of British Columbia approved access to and use of the data facilitated

by Population Data BC in this study.

4.3.1 WorkSafeBC Claims and Firm Level Files

WorkSafeBC is the provincial workers’ compensation system in BC established by provincial

legislation in 1917.(16) The agency oversees a publicly administered, no-fault single payer

insurance system in the workplace and is funded by employers who pay insurance premiums.

Workers with work-related injuries and diseases, regardless of the cause, may be compensated

for time missed from work, permanent disability, health care costs, rehabilitative services, and

fatalities. Employers in BC are legally mandated to have WorkSafeBC coverage, with a few

exceptions, such as the self-employed, though optional coverage can be obtained.(17)

Approximately 92.5% to 94% of the workforce in BC from 2000 to 2009 was eligible for

compensation coverage from WorkSafeBC.(18)

84

For each claim in the WorkSafeBC data files, data include anonymous claimant and claim

identifiers, injury and claim registration dates, claim eligibility status (e.g., allowed, pending,

rejected), and an indicator to identify whether the claim is for short-term disability (temporary

wage loss due to missed work), long-term disability (for workers deemed as having a

permanent disability and unlikely to ever return to work), health care only (health care services

considered necessary to treat injuries), vocational rehabilitation, or fatality. Payment

information for each claim includes the date of the first payment made and the amount paid by

year and month of payment.

Each claim also has an associated International Statistical Classification of Diseases and Related

Health Problems – Ninth Revision (ICD-9) diagnostic code (19) as reported by the treating

physician and a Statistics Canada Standard Occupational Classification code (20) to identify

occupation at the time of injury.

4.3.2 PharmaNet

PharmaNet is a province-wide, real-time system that captures all prescription medications

dispensed from community and hospital outpatient pharmacies in BC, regardless of the

payment source. Therefore, prescriptions paid out-of-pocket, through public or private

insurance, through workers’ compensation benefits, or other means are all captured.

Information for each prescription dispensed includes anonymous claimant, prescriber, and

pharmacy identifiers, date of dispensing, Drug Identification Number (DIN), American Hospital

Formulary Service (AHFS) classification system code,(21) quantity dispensed, days supply, and

dosage form. A DIN is a unique eight-digit number assigned by Health Canada to uniquely

85

identify all drug products sold in Canada.(22) Using the DIN, data were merged with Health

Canada Drug Product Database (DPD) data (23) to obtain information on product name, active

ingredient(s), drug strength, and route of administration.

4.3.3 Medical Services Plan (MSP)

Canada has a universal health care system that is publicly funded through federal, provincial,

and territorial taxes. BC also requires residents to pay health premiums based on family size

and income to further cover health care services. Health care is largely administered by the

provincial and territorial governments and all plans must insure all medically necessary services

provided by hospitals, medical practitioners, and dentists working in a hospital setting, though

what is deemed necessary may differ across plans.(24) Health coverage also does not include

prescription drugs, except for certain groups, such as lower income populations.

Medical Services Plan (MSP) is the universal health insurance program in BC and covers

approximately 96% of residents, excluding those covered by federal programs, such as Armed

Forces and First Nations. MSP data consist of billing records for all medically required

outpatient services provided by fee-for-service practitioners to individuals covered by MSP,

including examinations and consultations and laboratory and diagnostic procedures. Each

billing record in MSP includes anonymous patient and practitioner identifiers, one fee item

code that denotes the insured medical service provided, service date, and up to five ICD-9

diagnoses to indicate the reason for that service, with the first diagnosis considered the primary

diagnosis assigned to the claim. Diagnoses are considered valid only to the first three digits.(25)

Several additional diagnostic codes created by MSP are also used.(26)

86

Up until 2001, a limited number of visits to supplementary health care practitioners (e.g.,

chiropractic, massage therapy, physiotherapy) were provided through MSP to all insured

individuals. Individuals less than 65 years of age were entitled to up to 12 visits per year and

those over the age of 65 received a maximum of 15 visits, all at a charge of $10 per visit. Lower

income residents received the same number of visits for free. From 2002 onward, these

services were only insured by MSP in medically necessary cases for lower income individuals.

4.3.4 Discharge Abstract Database (DAD)

Hospitals in Canada (with the exception of those in the province of Quebec) submit their

separation (discharge) records to the Canadian Institute of Health Information (CIHI) for

inclusion in the DAD.(27) The DAD file contains data for hospital discharges (including deaths,

sign-outs, and transfers) of acute care inpatients and day surgery patients from BC facilities.

Separation records for residents admitted to facilities in other provinces or territories are also

included. Emergency room visits are not recorded in this dataset, though admissions to acute

care from emergency are noted.

Each separation record contains an anonymous patient identifier and an anonymous

practitioner identifier for the physician most responsible for the patient’s care during

hospitalization, as well as the admission and separation dates. Separation records up until

2001/2002 contain up to 16 ICD-9 diagnostic codes. From 2001/2002 onward, the number of

diagnostic codes expanded to 25 and diagnoses were coded using the International Statistical

Classification of Diseases and Related Health Problems - Tenth Revision, Canada (ICD-10-

CA).(28) From 2001/2002 to 2006/2007, CIHI also provided the corresponding ICD-9 codes.

87

Each diagnosis is assigned a diagnostic type to refer to the relationship of the diagnosis to the

patient’s hospitalization. The first diagnosis listed is always the most responsible diagnosis,

referring to the condition most accountable for an individual’s stay and/or longest length of

stay in hospital and does not necessarily reflect the condition for which the individual sought

care. Other diagnostic types include pre-admit comorbidity (a condition that existed prior to

admission and meets criteria for comorbidity), post-admit comorbidity (a condition that arises

after admission and meets criteria for comorbidity), and secondary diagnosis (a condition for

which a patient may or may not have received treatment during their stay, but does not meet

criteria for comorbidity). Criteria for comorbidity include a condition that requires treatment

beyond maintenance of the pre-existing condition, increases length of stay by at least 24 hours,

and significantly affects the treatment received.(29)

Until 2001/2002, up to 10 procedures coded using the Canadian Classification of Diagnostic,

Therapeutic, and Surgical Procedures system (CCP) were captured.(30) Thereafter, the number

of procedure codes increased to 20 and were coded using the Canadian Classification of Health

Interventions (CCI).(31) As with diagnostic codes, CIHI provided back translations of CCI codes

to CCP from 2001/2002 until 2006/2007.

4.3.5 MSP Registration and Premium Billing (Consolidation File)

This file is the central demographics dataset and includes basic demographic data on all

individuals registered as eligible to receive health services in BC. This includes sex, year and

month of birth, the first three digits of the Canadian postal code, and health authority of the

individual’s residence. In BC, there are five regional health authorities (Fraser Health,

88

Vancouver Coastal Health, Vancouver Island Health, Northern Health, and Interior Health),

whose purpose is to govern, plan, and deliver health care services within their respective

geographic areas.(32) For each year of the dataset, the start date of an individual’s registration

in MSP and the total number of days registered for MSP in that year are also recorded.

4.3.6 MSP Practitioner File

This file contains basic data on practitioners enrolled in MSP and who provide eligible health

care services using the fee-for-service funding model. Data in this file include practitioner sex,

year of birth, and year of graduation. Up to three specialties are also provided by year of

registration in MSP.

89

4.4 References


2. British Columbia Node of the Canadian Research Initiative on Substance Misuse. Together, We Can Do This. Strategies to Address British Columbia’s Prescription Opioid Crisis. 2015. Available from: http://www.cfenet.ubc.ca/sites/default/files/uploads/news/releases/opioid-safety-news_release_nov-19-2015.pdf. Last Accessed: March 6, 2017.

3. WorkSafeBC. The Prescription of Narcotics and Other Drugs of Addiction. Richmond, BC: WorkSafeBC; 1989.

4. Dunn C, Rothfels P, Nguyen T, Martin C, Sambandam R, Noertjojo K, et al. Developing Practice Directive on Claims With Opioids Prescribed at WorkSafeBC. 2009.

5. WorkSafeBC. WorkSafeBC Statistics 2009. Richmond, BC: Workers Compensation Board Province of British Columbia; 2009.


7. Department WCPQ. Practice Directive #C10-1. Claims With Opioids, Sedative/Hypnotics or Other Prescribed Potentially Addictive Drugs. Richmond, BC: WorkSafeBC; 2015.

8. Population Data BC. 2017; Available from: https://www.popdata.bc.ca/. Last Accessed: March 3, 2017.

9. Pencarrick Hertzman C, Meagher N, McGrail KM. Privacy by Design at Population Data BC: a case study describing the technical, administrative, and physical controls for privacy-sensitive secondary use of personal information for research in the public interest. J Am Med Inform Assoc. 2013;20(1):25-8.

10. WorkSafeBC [creator] (2012): WorkSafeBC Claims and Firm Level Files. V2. Population Data BC [publisher]. Linked Data Set. WorkSafeBC (2013).

11. British Columbia Ministry of Health [creator] (2012): PharmaNet. V2. British Columbia Ministry of Health [publisher]. Data Extract. Data Stewardship Committee (2013).

12. British Columbia Ministry of Health [creator] (2013): Medical Services Plan (MSP) Payment Information File. V2. Population Data BC [publisher]. Data Extract. MOH (2013).

13. Canadian Institute for Health Information [creator] (2013): Discharge Abstract Database (Hospital Separations). V2. Population Data BC [publisher]. Data Extract. CIHI (2013).

http://www.cfenet.ubc.ca/sites/default/files/uploads/news/releases/opioid-safety-news_release_nov-19-2015.pdf

http://www.cfenet.ubc.ca/sites/default/files/uploads/news/releases/opioid-safety-news_release_nov-19-2015.pdf

https://www.popdata.bc.ca/

90

14. British Columbia Ministry of Health [creator] (2013): Consolidation File (MSP Registration & Premium Billing). V2. Population Data BC [publisher]. Data Extract. MOH (2013).

15. British Columbia Ministry of Health [creator] (2013): Medical Services Plan (MSP) Practitioner File. V2. Population Data BC [publisher]. Data Extract. College of Physicians and Surgeons of British Columbia (CPSBC) (2013).

16. WorkSafeBC. Who we are. 2017; Available from: https://www.worksafebc.com/en/about-us/who-we-are. Last Accessed: March 11, 2017.

17. WorkSafeBC. Who does and doesn't need coverage?. 2017; Available from: https://www.worksafebc.com/en/insurance/need-coverage/who-needs-coverage. Last Accessed: November 23, 2016.

18. Association of Workers’ Compensation Boards of Canada (AWCBC) Customized KSM Report. Percentage of Workforce Covered [%]. British Columbia. 2000-2009 [data file]. Association of Workers’ Compensation Boards of Canada (AWCBC). (2013). http://awcbc.org/?page_id=9755

19. World Health Organization. The International Statistical Classification of Diseases and Related Health Problems, Ninth Revision (ICD-9). Geneva, CH: World Health Organization; 1977.

20. Statistics Canada. Standard Occupational Classification (SOC) 1991. Ottawa, ON: Statistics Canada; 2016; Available from: http://www23.statcan.gc.ca/imdb/p3VD.pl?Function=getVD&TVD=143375. Last Accessed: September 9, 2016.

21. American Hospital of Hospital Pharmacists. AHFS Drug Information. Bethesda, MD: Published by authority of the Board of Directors of the American Society of Hospital Pharmacists; 2011.

22. Health Canada. Drug Identification Number (DIN). 2009; Available from: http://www.hc-sc.gc.ca/dhp-mps/prodpharma/activit/fs-fi/dinfs_fd-eng.php. Last Accessed: January 28, 2017.

23. Health Canada. Drug Product Database. 2015; Available from: http://www.hc-sc.gc.ca/dhp-mps/prodpharma/databasdon/index-eng.php. Last Accessed: January 28, 2017.

24. Health Canada. Canada’s Health Care System. Ottawa, ON: Health Canada; Available from: http://www.hc-sc.gc.ca/hcs-sss/pubs/system-regime/2011-hcs-sss/index-eng.php. Last Accessed: September 22, 2016.

25. Hu W. Diagnostic Codes in MSP Claim Data, Summary Report. Victoria: Medical Services Plan; 1996.

https://www.worksafebc.com/en/about-us/who-we-are

https://www.worksafebc.com/en/about-us/who-we-are

https://www.worksafebc.com/en/insurance/need-coverage/who-needs-coverage

http://awcbc.org/?page_id=9755

http://www23.statcan.gc.ca/imdb/p3VD.pl?Function=getVD&TVD=143375

http://www.hc-sc.gc.ca/dhp-mps/prodpharma/activit/fs-fi/dinfs_fd-eng.php

http://www.hc-sc.gc.ca/dhp-mps/prodpharma/activit/fs-fi/dinfs_fd-eng.php

http://www.hc-sc.gc.ca/dhp-mps/prodpharma/databasdon/index-eng.php

http://www.hc-sc.gc.ca/dhp-mps/prodpharma/databasdon/index-eng.php

http://www.hc-sc.gc.ca/hcs-sss/pubs/system-regime/2011-hcs-sss/index-eng.php

91

26. Ministry of Health. Medical Services Plan Additional Diagnostic Codes. Available from: http://www2.gov.bc.ca/assets/gov/health/practitioner-pro/medical-services-plan/additional-diag-codes.pdf. Last Accessed: October 29, 2016.

27. Canadian Institute for Health Information. Discharge Abstract Database Metadata. Ottawa, ON: Canadian Institute for Health Information. Available from: https://www.cihi.ca/en/discharge-abstract-database-metadata. Last Accessed: September 23, 2017.

28. Canadian Institute for Health Information. Volume 1 - International Statistical Classification of Diseases and Related Health Problems, Tenth Revision, Canada (ICD-10-CA) - Tabular List. Ottawa, ON: Canadian Institute for Health Information; 2006.

29. Canadian Institute for Health Information. Canadian Coding Standards for Version 2012 ICD-10-CA and CCI. Ottawa, ON: Canadian Institute for Health Information; 2012.

30. Statistics Canada. Canadian Classification of Diagnostic, Therapeutic and Surgical Procedures. Ottawa, ON: Minister of Industry, Science and Technology; 1993.

31. Canadian Institute for Health Information. Volume 3 - Canadian Classification of Health Interventions (CCI) - Tabular List. Ottawa, ON: Canadian Institute for Health Information; 2006.

32. Ministry of Health. Health Authorities. Government of British Columbia; Available from: http://www2.gov.bc.ca/gov/content/health/about-bc-s-health-care-system/partners/health-authorities. Last Accessed: September 20, 2017.

http://www2.gov.bc.ca/assets/gov/health/practitioner-pro/medical-services-plan/additional-diag-codes.pdf

http://www2.gov.bc.ca/assets/gov/health/practitioner-pro/medical-services-plan/additional-diag-codes.pdf

https://www.cihi.ca/en/discharge-abstract-database-metadata

http://www2.gov.bc.ca/gov/content/health/about-bc-s-health-care-system/partners/health-authorities


92

Table 4-1 A description of the datasets and elements used in the analyses for Chapters 5 to 7

Data Custodian Data File Relevant Data Elements

WorkSafeBC Claims and Firm Level Files Anonymous claimant identifier Anonymous claim identifier Anonymous consolidated claim identifier Consolidation indicator Claim injury date Claim registration date Claim eligibility status Claim active status ICD-9 medical diagnosis codes for claim Statistics Canada occupation code First STD payment date First final STD payment date STD claim indicator Summary year Summary month Summary STD and income continuity days paid Summary amount

PharmaNet Data Stewardship Committee

PharmaNet data (prescription dispensing records from BC community and hospital outpatient pharmacies)

Anonymous claimant identifier Anonymous prescriber identifier Anonymous pharmacy identifier Prescriber specialty Service date DIN AHFS code AHFS category Dispensed quantity Dispensed days supply Dosage form Drug brand or generic name1 Active ingredient(s)1 Drug generic name1 Drug strength1 Route of administration1

British Columbia Ministry of Health

Medical Services Plan (MSP) Payment Information File (billing records for outpatient health care services)

Anonymous claimant identifier Anonymous practitioner identifier Service date Service code Paid service units Fee item Paid amount Service location code ICD-9 diagnostic code

93

Data Custodian Data File Relevant Data Elements

British Columbia Ministry of Health (continued)

Discharge Abstract Database (Hospital Separations) (data for inpatient and day surgery separations)

Anonymous claimant identifier Anonymous practitioner identifier (most

responsible physician) Admission date Discharge (separation) date Level of care ICD-9 diagnostic codes ICD-10-CA diagnostic codes Diagnosis type Transferred to hospital code Transferred from hospital code CCP procedure/intervention codes CCI procedure/intervention codes Procedure/intervention dates

Consolidation File (MSP Registration & Premium Billing)

Anonymous claimant identifier Year of birth Month of birth Registration start date Total days registered during the year Three digit postal code Health authority Neighbourhood income quintile

College of Physicians and Surgeons of British Columbia

Medical Services Plan (MSP) Practitioner File

Anonymous claimant identifier Anonymous practitioner identifier Sex Year of birth Specialty codes Specialty date (year)

1Obtained by linking dispensing records to Health Canada Drug Product Database data with the DIN Abbreviations: AHFS=American Hospital Formulary System; CCP=Canadian Classification of Diagnostic, Therapeutic and Surgical Procedures; CCI=Canadian Classification of Health Interventions; DIN=Drug Identification Number; ICD-9=International Classification of Diseases 9th Revision; ICD-10-CA=International Classification of Diseases 10th Revision, Canada; STD=short-term disability

94

CHAPTER 5 PRESCRIPTION DISPENSING PATTERNS BEFORE AND AFTER A

WORKERS’ COMPENSATION CLAIM: AN HISTORICAL COHORT STUDY OF WORKERS WITH LOW BACK PAIN INJURIES IN BRITISH COLUMBIA

Nancy Carnide, Sheilah Hogg-Johnson, Andrea Furlan, Pierre Côté, Mieke Koehoorn

Abstract

Objectives To describe and compare prescription opioid, nonsteroidal anti-inflammatory drug

(NSAID), and skeletal muscle relaxant (SMR) dispensing patterns before and after a

compensated work-related low back injury and to determine whether post-injury dispensing

returns to pre-injury levels.

Study Design Historical cohort study.

Population Workers from British Columbia, Canada filing a new, short-term disability workers’

compensation claim for a low back injury occurring between 1998 and 2009 with at least one

compensated lost workday.

Methods Administrative data on opioid, NSAID, and SMR prescriptions dispensed in community

and hospital outpatient pharmacies one year before and after injury data were collected.

Descriptive statistical methods were used to describe pre- to post-injury dispensing overall and

stratified into six eight-week periods before and after injury.

Results The cumulative proportion of claimants dispensed opioids, NSAIDs, and SMRs in the

year before injury was 19.7%, 21.2%, and 6.3%, respectively, increasing to 39.0%, 50.2%, and

28.4% one year after injury. Median time to first post-injury prescription was 6, 4, and 3 days,

respectively, for opioids, NSAIDs, and SMRs. Among claimants with at least one prescription

95

dispense, pre- and post-injury median cumulative days supply were 8 and 10 for opioids, 15 and

18 for NSAIDs, and 10 and 10 SMRs. Among claimants dispensed opioids, approximately 95%

received weak opioids and 99% received short-acting opioids, with an average daily morphine-

equivalent dose of 31mg/day both pre- and post-injury. The proportion of claimants receiving a

dispense for any of the three drug classes in any given eight week period was generally stable

prior to injury, followed by a sharp increase within eight weeks post-injury. Dispensing dropped

thereafter and by 41-48 weeks post-injury, dispensing was nearing pre-injury levels, but

remained elevated. This post-injury increase in opioid, NSAID, and SMR dispensing was

attributable to 1.8%, 0.8%, and 0.4% of claimants, respectively. At 41-48 weeks, cumulative

days supply for each drug class also remained 20 to 30% higher compared to pre-injury. Among

claimants with opioid prescriptions, the proportion receiving strong and long-acting opioids

increased over time.

Conclusion All three drug classes are commonly dispensed, particularly in the immediate weeks

after a compensated LBP-related injury. For most claimants, based on dispensing patterns,

injuries appear to have been abrupt in onset and post-injury prescription dispensing is of short

duration. However, a small subgroup of claimants receive prolonged courses of drugs,

suggesting they are experiencing a delay in their recovery.

96

5.1 Introduction

Low back pain (LBP) is one of the most prevalent physical health conditions to affect workers (1-

3) and is a leading cause of work disability,(3-7) accounting for 149 million lost work-days each

year in the United States (US).(4) In Canada, back problems are responsible for the largest

incremental productivity loss of all major chronic health conditions among workers.(3)

The provision of pharmaceuticals is the mainstay of LBP medical management.(8-10) Opioids,

nonsteroidal anti-inflammatory drugs (NSAIDs), and skeletal muscle relaxants (SMRs) are the

most common drugs prescribed,(8-11) despite a lack of high quality evidence demonstrating

their effectiveness for LBP.(12-15) Opioids, specifically, are also associated with significant

harms, including abuse, addiction, and overdose.(16-20)

Over the last decade, the use of opioids in workers’ compensation populations has been a

major focus of research, with several studies documenting the high prevalence of opioid

prescriptions following a work-related LBP injury.(21-29) The majority of these studies,

however, have been conducted in the US and have provided a limited characterization of

prescription patterns. Comparatively, little is also known about prescription patterns of NSAIDs

and SMRs after a work-related LBP injury.

The extent of exposure to opioids after injury may also have been underestimated. Previous

studies used data of prescriptions reimbursed by the workers’ compensation system, thus not

capturing prescriptions paid out-of-pocket or through private insurance. The results of one

study found work-related LBP claimants use additional out-of-pocket sources to pay treatment

costs.(30) In unpublished data from our research of LBP claimants in British Columbia (BC),

97

69.3%, 68.0%, and 67.3% of opioid, NSAID, and SMR dispensing records, respectively, identified

in a prescription monitoring program database were not captured by workers’ compensation

data within eight weeks after injury.(31)

Pre-injury exposure to these medications is also unknown. Much concern has been raised

regarding the early provision of opioids after injury.(21, 22, 24-26, 29) However, it is not clear

whether claimants may have been receiving prescriptions prior to their injuries or before filing

their claims. A report by the National Council on Compensation Insurance found approximately

80% of lost-time claims in the US had a lag between injury and claim reporting.(32) It would not

be surprising that some workers delay reporting, yet seek treatment before reaching a

threshold where time off work is now required. This was the conclusion of one group of

researchers who found a steady increase in the number of pre-claim health care contacts

leading up to claim filing for a musculoskeletal disorder among health care workers.(33) In a

similar study of neck pain claimants, this pattern was not evident.(34) These studies, however,

both demonstrated that health care utilization after a claim remained elevated compared to

before claim. To date, we are not aware of any study to compare pre- and post-injury injury

prescription drug patterns among workers’ compensation LBP claimants in a similar manner.

The objectives of this study were, therefore, to describe and contrast prescription opioid,

NSAID, and SMR dispensing patterns one year prior to a work-related low back injury through

one year post-injury and to determine whether post-injury dispensing returns to pre-injury

levels by one year after injury in a Canadian sample of workers’ compensation claimants.

98

5.2 Methods

5.2.1 Study Design and Setting

We conducted an historical cohort study of workers’ compensation claimants in British

Columbia (BC), Canada using linked administrative data. The source population included all

employees working for employers insured through the provincial workers’ compensation board

in BC and who filed a new short-term disability claim for a LBP injury occurring between 1998

and 2009. Data from five databases capturing information on workers’ compensation claims

(WorkSafeBC Claims and Firm Level Files), pharmacy dispensing (PharmaNet), outpatient health

care (Medical Services Plan [MSP] Payment Information File), hospitalizations (Discharge

Abstract Database [DAD]), and demographics (MSP Consolidation File) (35-39) from 1996

through to 2010 were linked and delivered by Population Data BC.(40, 41)

WorkSafeBC is the provincial workers’ compensation system in BC. The agency oversees a

publicly administered, no-fault single payer insurance system in the workplace to compensate

workers with work-related injuries and diseases for time missed from work, permanent

disability, health care costs, rehabilitative services, and fatalities. Employers in BC are legally

mandated to have WorkSafeBC coverage, with a few exceptions, such as the self-employed.(42)

During the study period, the employed labour force in BC ranged from approximately 1.9 to 2.2

million employees,(43) with 92.5% to 94% of the workforce eligible for compensation coverage

from 2000 to 2009.(44)

PharmaNet is a province-wide, real-time system that captures detailed information for all

prescription medications dispensed from community and hospital outpatient pharmacies in BC,

regardless of payment source. Medical Services Plan (MSP) is the universal health insurance

99

program in BC and covers approximately 96% of residents, excluding those covered by federal

programs. MSP data consist of billing records for all medically required outpatient services

provided by fee-for-service practitioners, including examinations and consultations and

laboratory and diagnostic procedures. The Discharge Abstract Database (DAD) file contains

diagnostic and intervention data for all hospital discharges (deaths, sign-outs, transfers) of

acute care inpatients and day surgery patients from BC facilities. The MSP and DAD datasets

were used in the creation of the cohort. Finally, the MSP Registration and Premium Billing file

includes basic demographic data on all individuals registered as eligible to receive health

services in BC. WorkSafeBC, PharmaNet, and the British Columbia Ministry of Health approved

access to and use of the data facilitated by Population Data BC in this analysis. The study was

approved by the University of Toronto Health Sciences Research Ethics Board (protocol

reference number 26885).

5.2.2 Study Participants

Linked data for workers who filed at least one new workers’ compensation short-term disability

claim with WorkSafeBC for LBP between 1998 and 2009 were extracted and provided to the

research team by Population Data BC staff. All claims for these workers dating back to 1991

were provided and a number of claim-level inclusion criteria were applied. Specifically, claims

were included if they met the following criteria: 1) claim was not a consolidated claim or linked

to a consolidated claim. Claims may be consolidated for administrative reasons, such as when

duplicate claims have been inadvertently created for the same worker and injury event.

Consolidation may also occur if it is determined that a new claim is a compensable consequence

100

of a previous injury for which a claim has been initiated.; 2) claimant was 18 years or older at

injury; 3) claim was an allowed short-term disability claim with at least one day of wage

replacement benefits received in the first eight weeks following injury date; 4) claim was for a

nonspecific LBP disorder identified using four digit International Classification of Diseases, 9th

Revision (ICD-9) codes (45) (Appendix B, Table B-1) occurring between January 1, 1998 and

December 31, 2009; 5) claim had no LBP-related hospitalization or no serious LBP-related

outpatient encounter (Appendix B, Table B-2) within five days after injury; 6) claimant was a BC

resident and continuously registered as eligible to receive health services for at least two years

before and one year after injury date; 7) injury date was equal to or preceded claim registration

date. Injury date was chosen to establish the inception point for entry into the cohort rather

than registration date to best approximate clinical onset of the injury. In our data, 98% of

claimants had injury dates that preceded claim registration date, with the mean (standard

deviation) number of days between injury date and registration date being 6 (5.4) days (median

of 5).

After applying these criteria, some claimants had more than one eligible claim. From this pool

of eligible claims, we selected one index claim per claimant as the earliest claim: 1) that had no

other allowed claims (all types of claims, irrespective of diagnosis) in the year after injury; and

2) where the total benefits paid in the year after injury was greater than zero. If these

conditions were not met by at least one claim, the claimant was excluded. Following selection

of the index claim, one additional claimant-level exclusion was applied: claimants were

excluded if they had at least one hospitalization and/or two outpatient health care billings for

cancer (ICD-9 codes 140-209 and 235-239; International Classification of Diseases, 10th Revision,

101

Canada [ICD-10-CA] codes C00-C97 and D37-D48) two years before through one year after

injury.

5.2.3 Study Variables

5.2.3.1 Outcomes

Detailed information for each prescription dispensed was obtained, including dispense date,

active ingredient(s), quantity dispensed, days supply, dosage form, drug strength, and route of

administration. Dispensing records from the following drug classes were retained for this study:

opioids (American Hospital Formulary Service (AHFS) classification system codes,(46) 28:08.08,

28:08.12), NSAIDs (28:08.04.08, 28:08.04.24, 28:08.04.92), and SMRs (12.20.04, 12.20.08,

12.20.12) (see Appendix B, Table B-3 for a complete list). All routes of administration were

eligible with the exception of topical formulations. In the two years around injury (one year pre-

and post-injury), 99.4% of all opioid, 99.1% of all NSAID, and 99.7% of all SMR dispensing

records were oral formulations. Over-the-counter medications were excluded as they are only

infrequently entered into the system.

A number of dispensing patterns for the three drug classes were constructed for the one year

periods before and after injury date and in eight-week periods pre- and post-injury. We chose

eight-week increments as WorkSafeBC policy at the time of the study limited reimbursement of

prescription opioids to the first eight weeks after injury or surgery for most claims.(47)

Receipt of each drug class (yes/no) was estimated as the presence of at least one dispensing

record for the specific drug class. A categorical variable describing the specific prescription

combinations dispensed among claimants with at least one drug dispense was also created (i.e.,

102

NSAIDs only, SMRs only, NSAIDS and SMRs, opioids only, opioids and NSAIDs, opioids and SMRs,

all three drug classes). Dispensing of specific drugs within each drug class was estimated as

individual dichotomous yes/no variables according to the active ingredient (e.g., codeine,

naproxen) among claimants with at least one dispensed prescription of a particular drug class.

Number of dispensed prescriptions (1, 2, 3-4, ≥5) among claimants with at least one dispense

was estimated for each drug class by counting the number of unique dispensing records based

on claimant and prescriber identification numbers, date of service, and active ingredient.

Cumulative days supply dispensed for each drug class was calculated by summing the days

supply across all prescriptions for a given class. When dispensing records for the same drug

class overlapped in time, days were counted only once.

Time since last dispensed prescription before injury was calculated as the number of days from

date of injury to the date of the last dispensed prescription for a particular drug class within the

year prior to injury date. Likewise, time to the first dispensed prescription after injury was

calculated as the number of days between date of injury and date of the first dispensed

prescription for a particular drug class within the year following injury. These two variables

were derived only among those claimants who had at least one prescription dispensed for the

particular drug class in the year before or the year after injury, respectively.

A number of other variables were also derived among claimants receiving at least one opioid

dispensed prescription. Two dichotomous variables describing receipt of at least one weak

opioid (yes/no) and receipt of at least one strong opioid (yes/no) were constructed. For this

study, anileridine, codeine, meperidine, pentazocine, propoxyphene, and tramadol were

103

considered to be weak opioids, while buprenorphine, butorphanol, fentanyl, hydromorphone,

methadone, morphine, and oxycodone were considered strong.(48-55)

Receipt of at least one short-acting opioid (yes/no) and receipt of at least one long-acting opioid

(yes/no) were estimated. In this study, codeine, hydromorphone, morphine, oxycodone, and

tramadol were provided as both short- and long-acting formulations. Fentanyl and

buprenorphine were available in the data as patches, which are designed to release over a

period of time, and were therefore considered long-acting. All other opioids were considered

short-acting only (e.g., meperidine, propoxyphene).

To obtain the average daily morphine-equivalent dose (MED) for oral formulations, the daily

dose for each day supplied with an opioid was first calculated as (drug strength*quantity

dispensed)/days supply. A similar calculation for daily dose was conducted for transdermal

formulations, incorporating the strength per hour delivered by each patch: (drug strength in

microgram/hour * hours supplied)/days supply. A standardized daily dose was then calculated

by multiplying the daily dose by previously published morphine-equivalent conversion ratios

(Appendix B, Table B-4).(48-55) For transdermal formulations, the result was also divided by

1,000 to convert from micrograms to milligrams. The average daily MED for each claimant was

then calculated as the sum of the daily MED over the sum days supplied with opioids in a given

time period. Opioids dispensed for parenteral, buccal, rectal, or nasal routes were not included

in the calculation because the unit of measure in the dispensed quantity field was not clear.

Opioids by these routes represented only 0.6% of all opioid dispensing records in the year

before and after injury. In the calculation, we also included the daily dose of codeine in SMR

104

combinations containing codeine. These dispensing records accounted for approximately 5.1%

of all records used in the MED calculation in the two years around injury.

5.2.3.2 Other Variables

A number of other variables were constructed to describe the cohort, including sex, year of

injury (in two year groups from 1998 to 2009), four-digit ICD-9 diagnosis code associated with

the claim, number of workers’ compensation claims in the two years before injury (0, 1, ≥2) and

regional health authority (Fraser Health, Vancouver Coastal Health, Vancouver Island Health,

Northern Health, Interior Health). Year and month of birth were used to approximate age at

injury date, grouped as 18-24, 25-34, 35-44, 45-54, and ≥55. ICD-9 diagnosis was re-categorized

based on the first three digits: 722 (Intervertebral Disc Disorders), 724 (Other and Unspecified

Disorders of Back), 846 (Sprains and Strains of Sacroiliac Region), and 847 (Sprains and Strains

of Other and Unspecified Parts of Back). Neighbourhood income quintile was assigned to each

claimant by Population Data BC analysts based on Census income data and claimant’s postal

code using Postal Code Conversion File+ (PCCF+) software developed by Statistics Canada.(56)

5.2.4 Statistical Analyses

Descriptive statistical methods were used to describe pre- to post-injury dispensing patterns,

namely proportions for categorical variables and mean, standard deviation (SD), median, and

interquartile range (IQR) for continuous variables. Ninety-five per cent confidence intervals

(95%CI) were also estimated around proportions and means. This was done overall, comparing

one year pre-injury to one year period post-injury, as well as stratified into six eight-week

105

periods before and after injury date. The data analysis for this paper was generated using SAS

software Version 9.3. Copyright © 2000-2010 SAS Institute Inc. SAS and all other SAS Institute

Inc. product or service names are registered trademarks or trademarks of SAS Institute Inc.,

Cary, NC, USA.

5.3 Results

A total of 142,993 eligible claims linked to 113,434 claimants resulted after applying the claim-

level inclusion criteria (Figure 5-1). Choosing one index claim per claimant resulted in a total of

99,233 claimants each with one eligible index claim. A total of 2,109 claimants with a history of

cancer-related health care utilization were then excluded, resulting in a final cohort of 97,124

claimants.

5.3.1 Sample Characteristics

The cohort was comprised of 62.3% men (Table 5-1). Mean (SD) age at injury was 40.5 (11.2),

with 30.8% between 35 and 44 years of age. A large proportion of claimants were located in the

Fraser regional health authority (42.5%), the most densely population health authority. Only

5.5% were in the Northern health authority region, with claimants fairly evenly distributed

among the remaining three regions, representing the population distribution in the province.

The vast majority of claimants had a diagnosis of sprain or strain (87%) and only 3% were

diagnosed with an intervertebral disc disorder. Around 33% had at least one previous workers’

compensation claim accepted by WorkSafeBC in the two years before injury.

106

5.3.2 Comparing Prescription Dispensing Patterns One Year Pre-Injury to One Year Post-Injury

The one-year period prevalence of receiving at least one of the three drug classes doubled from

34.0% (95% CI 33.7%-34.3%) pre-injury to 68.0% (95%CI 67.7%-68.3%) post-injury (Table 5-2).

Approximately one in five claimants received at least one opioid or NSAID in the year before

injury. In the year after injury, 38.8% received opioids (95%CI 38.5%-39.1%) and 50.2% received

NSAIDs (95%CI 49.8%-50.5%), representing a 98% and 138% relative increase from pre-injury,

respectively. Although less commonly dispensed, the largest increase in prevalence was for

SMRs: from 6.3% (95%CI 6.1%-6.4%) pre-injury to 28.4% (95%CI 28.1%-28.7%) post-injury,

representing a 351% relative increase over time.

Among claimants receiving at least one opioid prescription, the vast majority received weak and

short-acting opioids. There was a small increase from pre- to post-injury in the proportion of

claimants receiving strong opioids, from 10.8% (95%CI 10.4%-11.3%) to 14.8% (95%CI 14.4%-

15.2%). The most commonly dispensed drugs for each drug class were similar both pre- and

post-injury. Among claimants dispensed opioids, the vast majority of claimants received

codeine, while oxycodone was the second most common. Both were primarily dispensed as

combinations with acetaminophen or acetylsalicylic acid. Only 0.8% received codeine not

combined with other non-opioid drugs and 1.2% of claimants received oxycodone not in

combination with other drugs pre-injury, increasing to 2% post-injury. Less than 1% of claimants

received long-acting oxycodone. Naproxen and diclofenac were the most commonly dispensed

NSAIDs among claimants receiving at least one NSAID prescription. Most claimants receiving

107

SMRs were dispensed cyclobenzaprine, while approximately one fifth received methocarbamol

(all combined with low dose codeine).

The incidence proportion of claimants who received each of the three drug classes in the year

after injury was 33.5% (95%CI 33.2%-33.9%) for opioids, 45.6% (95%CI 45.3%-46.0%) for

NSAIDs, and 26.8% (95%CI 26.6%-27.1%) for SMRs. (Table 5-2). Among claimants with at least

one prescription in the prior year, the proportion dispensed a post-injury prescription was even

higher, particularly for opioids (60.2%, 95%CI 59.5%-60.9%) and NSAIDs (67.1%, 95%CI 66.5%-

67.8%). A greater proportion of claimants received combinations of drug classes in the post-

injury period compared to only one drug class.

Most claimants received only one to two dispensed prescriptions both pre- and post-injury

across all drug classes: 72.7% pre-injury/77.7% post-injury for claimants receiving opioids,

82.8% pre-injury/78.7% post-injury for claimants receiving NSAIDs, and 89.5% pre-injury/88.1%

post-injury for claimants receiving SMRs. Cumulative days supplied ranged from a median of 8

(IQR 5-20) and 10 days (IQR 5-24) pre- and post-injury, respectively, for opioids, to 15 (IQR 10-

31) and 18 days (IQR 10-37) for NSAIDs. The median cumulative number of days supplied of

SMRs remained steady at 10 from pre- to post-injury.

On average, the last prescriptions received by claimants in the year before injury were

dispensed just over five months prior to injury across all drug classes, while the median number

of days to first prescription after injury was 6 (IQR 1-79), 4 (IQR 1-26), and 3 (IQR 1-13) days for

opioids, NSAIDs, and SMRs, respectively (Table 5-2). The majority of claimants received their

first prescription for at least one of the drug classes in the first two weeks (Figure 5-2), with a

108

similar pattern seen for each drug class separately (Appendix C, Figures C1-C3). Among those

receiving strong opioids, specifically, the median time to first prescription was 25 days (IQR 3-

154).

Among claimants with at least one day supply of opioids and/or SMRs with codeine, daily

morphine equivalent dose (MED) averaged approximately 31mg/day both pre- and post-injury.

Less than one per cent had an average daily MED of at least 120mg/day in either period.

5.3.3 Prescription Dispensing Patterns From 48 Weeks Pre- to 48 Weeks Post-Injury in 8-Week Increments

For all drug classes, the proportion of claimants dispensed a prescription in any given eight-

week period was fairly stable leading up to the injury, with a slight dip in the immediate eight

week period prior to injury (Figure 5-3). The average proportion of claimants in any given eight-

week period receiving at least one opioid, NSAID, or SMR was 5.2%, 5.0%, and 1.3%,

respectively (background prevalence). In the eight-week period immediately after injury, the

prevalence sharply peaked, with 27.8%, 40.7%, and 24.8% of claimants receiving at least one

opioid, NSAID, and SMR, respectively. The prevalence dropped immediately in the following

eight-week period and continued to drop, stabilizing over time. Among claimants receiving

prescriptions in the first eight weeks, most did not receive any prescriptions after that period

(16,824/27,023, 62.3% opioids; 27,512/39,563, 69.5% NSAIDs; 20,298/24,111, 84.2% SMRs).

Comparing pre- and post-injury, 2.0%, 1.1%, and 0.2% received opioids, NSAIDs, and SMRs,

respectively in at least four out of the six eight-week periods pre-injury, compared to 4.0%,

2.6%, and 0.7%, respectively, post-injury. Between 41 and 48 weeks post-injury, the prevalence

109

of receiving at least one prescription was nearly at pre-injury levels, but still slightly higher at

7.0%, 5.8%, and 1.7%, for opioids, NSAIDs, and SMRs, respectively. Comparing this to the

background prevalence in the pre-injury period, this post-injury increase in opioid, NSAID, and

SMR dispensing was attributable to 1.8%, 0.8%, and 0.4% of claimants, respectively. A similar,

but less dramatic pattern was seen for number of prescriptions (Appendix C, Figures C4-C6).

Mean cumulative days supplied showed a small increasing trend leading up to the claim for all

drug classes, particularly in the 16 weeks pre-injury, decreasing in the immediate post eight-

week period, due to an influx of claimants receiving prescriptions in this period for a short days

supply (Figure 5-4). Mean cumulative days supply increased thereafter, remaining higher post-

injury compared to pre-injury, though stable through 48 weeks post-injury. By 41-48 weeks

post-injury, the mean cumulative days supplied per claimant among claimants with at least one

dispense was 22.0 for NSAIDs, 19.2 for SMRs, and 17.8 for opioids, compared to 18.4, 14.8, and

14.0 days, respectively, 41-48 weeks pre-injury. A similar trend was seen using median

cumulative days (Appendix C, Figure C-7).

Among claimants dispensed opioid prescriptions, the proportion receiving weak and short-

acting opioids declined from pre-injury to post-injury, while the proportion of claimants

receiving strong and long-acting opioids increased (Figure 5-5). By 41-48 weeks post-injury,

19.0% of claimants dispensed opioids received strong opioids and 5.9% were receiving long-

acting opioids. This is compared to 41-48 weeks pre-injury where only 11.6% and 2.6% were

receiving strong and long-acting opioids, respectively. The eight-week period immediately after

injury was the exception to this trend, with the proportion of claimants receiving weak and

short-acting opioids increasing slightly, and the proportion dispensed strong and long-acting

110

opioids decreasing. Average daily MED generally remained stable from pre- to post-injury

(Figure 5-6).

5.4 Discussion

In this Canadian sample of workers’ compensation claimants with LBP-related short-term

disability claims, we found that short courses of prescription drug classes often used in the

management of LBP are commonly dispensed, both before and after injury. Most claimants

dispensed opioids received weak and short-acting opioids at low average daily doses. The

prevalence of dispensing pre-injury was shown to be fairly stable, followed by a rapid increase

in the proportion of claimants with dispensed prescriptions in the first eight weeks after injury.

Thereafter, the prevalence decreased quickly, but remained slightly elevated between 41 and

48 weeks compared to pre-injury. The small proportion of claimants continuing to receive

dispensed prescriptions almost one year after injury tended to also receive them for longer and

the proportion of claimants receiving strong and long-acting opioid formulations increased over

time.

Before the injury, 34% of claimants in our sample received at least one dispensed prescription

for opioids, NSAIDs, and/or SMRs, with opioids (19.6%) and NSAIDs (21.1%) most commonly

dispensed. This estimate for opioids is higher than that seen in a recently published population-

based study in BC that found that approximately 12% of residents filled at least one opioid

prescription in any given year from 2005 to 2013.(57) Some of this discrepancy may be due to

differences in sample composition, as that study included individuals of all ages and was not

specific to a working population. Another Canadian survey in 2008 found that 21.6% of

111

individuals over the age of 15 reported use of opioid pain relievers in the prior year.(58) We

were unable to locate published sources detailing past-year prevalence of NSAIDs and SMRs in

population-based samples. However, past-month prevalence of use among US adults from 1999

to 2010 has been estimated to be 7.0% for NSAIDs and 2.0% for SMRs,(59) which is slightly

higher than our prevalence estimates in any given eight-week period before injury.

Not surprisingly, the dispensing prevalence for all three drug classes increased substantially in

the one year post-injury period. The largest increase was seen for SMRs (from 6.3% pre-injury

to 28.4% post-injury), though NSAIDs and opioids remained most commonly dispensed after

injury (50.2% and 38.8% of claimants, respectively). In previous studies of US workers’

compensation claimants with LBP claims, one-year prevalence of opioid prescriptions ranged

from 38% to 42% in the year after injury.(23, 27) A small study of US LBP claimants also found

83% and 66% received NSAIDs and SMRs, respectively, over one year after injury.(27) Estimates

from these studies were derived from prescriptions reimbursed by workers’ compensation for

the compensated injury, while our study included data on all dispensed prescriptions

irrespective of payment source or indication. The fact that our estimates were still similar or

even lower could reflect system-level differences in reimbursement, where American workers’

compensation organizations reimburse injury-related prescriptions to a greater extent than

suggested in our analysis of BC claimants.(31) However, given that our data also may have

included prescriptions for other indications, it is likely that true differences in prescribing

patterns exist, with greater prescribing occurring south of the border. These differences in

prescribing may also be reflective of WorkSafeBC’s policy efforts at the time of the study period

to limit the use of opioids after a workplace injury.(60)

112

The only previous Canadian study found between 7% and 10% of claimants with LBP claims in

Alberta had at least one opioid prescription within a year after injury, a much lower estimate

than in our study.(25) Authors of this study found some evidence to suggest underestimation of

strong opioid prescriptions in their data, though they were unable to verify this for weak

opioids. Nonetheless, prevalence was likely underestimated.

The time to first prescription was short for all drug classes, with SMRs demonstrating the

shortest median time (three days), followed by NSAIDs (four days) and opioids (six days). When

stratified by strength of opioid received, it took a median of 25 days for claimants to fill their

first strong opioid prescription compared to seven days for weak opioids. These findings are

suggestive of non-adherence to treatment guidelines that recommend against opioids and

SMRs as first-line treatment for LBP.(61, 62) However, some consideration of the context may

be important. Our sample included only workers with at least one day of lost work time and the

findings of some research suggest such workers may represent a more severe population than

those who do not require time off.(63, 64) Nonetheless, research is needed to understand the

determinants of early exposure to these drug classes among workers’ compensation claimants.

Only a handful of studies have examined this, but have focused solely on opioids.(25, 65, 66)

While overall dispensing was prevalent, the majority of claimants in this sample received one to

two dispensed prescriptions both pre- and post-injury and with a short days supply. A very

small minority of claimants demonstrated a pattern of dispensing that was suggestive of long-

term use. The proportion of claimants receiving various combinations of the three drug classes

was also higher post-injury than pre-injury, where many claimants received only one of the

113

three drug classes. This likely reflects the complex management of the injury and attempts at

prescribing various drugs to ameliorate pain.

The vast majority of claimants with at least one opioid dispense received weak and short-acting

formulations, namely codeine combinations with acetaminophen or acetylsalicylic acid.

Historically, codeine has typically been perceived as relatively innocuous and in Canada, low

dose formulations are available over-the-counter. However, codeine poses a risk of toxicity for

some individuals (67) and serious morbidity can result from its abuse.(68-71) Oxycodone was

the second most commonly dispensed opioid in this sample. Surprisingly, less than one per cent

of claimants with an opioid prescription received long-acting oxycodone, despite the fact that

this analysis was conducted using data from a decade that saw rapid growth in use and abuse of

OxyContin across North America.(16, 72, 73)

Average daily MED among claimants receiving opioid dispenses was approximately 31mg/day

pre- and post-injury. Among claimants receiving opioids between 41 and 48 weeks post-injury,

the average daily MED was 37.8mg/day. These values are well below WorkSafeBC’s high dose

threshold of 120mg/day at the time of the study.(47) Less than one per cent had an average

daily MED equivalent to or exceeding 120mg/day, a finding that suggests the vast majority of

claimants were not exposed to high opioid doses that have been found to be associated with a

number of serious harms, including mortality.(74-76) While this finding is encouraging, it is

important to note that harms have been shown to occur at lower doses, such as in a study of

workers’ compensation claimants in Washington State, where most opioid poisonings and

adverse effects from 2004 to 2010 occurred in cases without prescribed chronic supplies and

with prescribed daily doses of less than 120mg/day.(77)

114

The pre-injury period was relatively stable with respect to the prevalence of dispensing of each

of the three drug classes. However, there was a peculiar, albeit minor, dip in prevalence in the

eight weeks before injury, particularly for opioids, that is difficult to explain. One potential

explanation may be due to our selection of injury date as the inception point for our cohort,

which was done in an attempt to better reflect the clinical picture of the LBP injury. The median

time between injury date and claim registration date was five days. Presumably claim

registration date is valid, as it is documented upon claim submission. However, injury date may

be applied retroactively and there may be some misclassification in this measure if there is any

ambiguity in the timing of injury, particularly if the injury was more gradual in onset. It is

possible that a subgroup of claimants may have begun to experience symptoms prior to the

official injury date on claim and increased their use of over-the-counter drugs or sought non-

pharmacologic measures to address their symptoms. In unpublished analyses, we found some

evidence of an increase in the proportion of claimants from our cohort (with injury years 1998

to 2001) who received MSP-insured chiropractic visits leading up to the injury.(31) Further, we

lacked data on over-the-counter medications and, in Canada, a number of drugs from all three

drug classes are available without prescription. Workers experiencing symptoms or already

using these medications may also delay obtaining a prescription in anticipation of receiving

workers’ compensation benefits, as the cost of the drugs would be covered if the claim is

accepted. Another possible explanation may be related to duration of use prior to claim. Among

individuals receiving these drug classes before injury date, there was a trend towards an

increase in mean cumulative days supply and it could be that claimants receiving these

prescriptions were receiving them for longer periods of time, thus leading to fewer dispensed

115

prescriptions. However, these are all hypothetical explanations and this observation requires

verification in other samples.

Dispensing then peaked in the first eight weeks after injury, decreasing quickly thereafter. This

pattern suggests that, for most claimants, prescriptions were new and exposure was acute in

nature. However, dispensing remained slightly elevated through 48 weeks compared to pre-

injury. This increase in opioid, NSAID, and SMR dispensing was attributable to approximately

1.8%, 0.8%, and 0.4% of claimants, respectively. Claimants who continued to be dispensed

these prescriptions in the later periods after injury were also receiving them for slightly longer

periods of time than seen pre-injury and, for claimants receiving opioids, increasingly more

claimants were receiving stronger and long-acting formulations as time passed after injury,

similar to observations from a US study of LBP claimants.(22) Therefore, while most claimants in

our cohort appeared to be exposed to these prescription drugs for a short window of time,

there appears to be a very small proportion of claimants who may be struggling with recovery.

This is consistent with other studies that found most claimants do not go on to long-term use,

but those that do have greater work loss and little improvement in pain and function.(23, 28)

Importantly, this group may also be particularly vulnerable to drug-related harms from

prolonged use of these drug classes.

To our knowledge, there are no other published studies examining pre- to post-injury

prescription patterns of workers’ compensation claimants to which we can compare our results.

Our findings are, however, generally consistent with those of two previous studies of claimants

with musculoskeletal injuries examining health care utilization patterns.(33, 34) In both, health

care utilization peaked immediately after claim, decreased, but remained elevated compared to

116

pre-injury. Researchers in Australia also recently compared opioid dispensing one year before a

motor vehicle accident to dispensing 18 months after and found a sustained increase in total

morphine equivalent amounts at 18 months compared to pre-injury.(78) Unlike Koehoorn et

al.,(33) who found an increase in health care utilization leading up to claim, we saw very little

evidence to suggest an increase in prescription dispensing preceding the injury date on the

claim.

There are a number of strengths to our study. We have provided a comprehensive picture of

dispensing of prescription drugs commonly used for LBP in a large sample of workers’

compensation claimants with LBP claims from a broad range of injury years. Unlike previous

studies in this area, we have been able to contrast pre- and post-injury prescription dispensing

patterns. Our study also provides estimates from a Canadian sample as a comparator to the

many studies from the US. Our prescription data comes from a comprehensive prescription

monitoring program that provides detailed information on all drugs dispensed in community

pharmacies, allowing us to capture drugs paid by workers’ compensation, as well as those paid

through all other means.

Our study also has limitations not yet described. The generalizability of findings to other

jurisdictions, to non-workers’ compensation claimants, or to claimants without lost work time is

unknown. This study was also conducted using data from 1998 to 2009 and current dispensing

patterns may differ. Prescription data were based on dispensing records and we do not know

whether the prescriptions were actually consumed. Therefore, exposure may have been

overestimated. However, the findings still likely reflect clinical practice patterns in this

population. We also did not have data on drugs received in hospital and this likely led to

117

underestimation of exposure, though likely this was minor since we excluded individuals with

LBP-related hospitalizations in the first five days after injury, a period of extensive dispensing

activity in our sample. The prescription-related measures derived in our study were also based

on prescription dispensing records irrespective of indication and cannot necessarily be

attributed to the LBP injury. However, we believe it is still important to identify the drugs to

which claimants are exposed regardless of indication and the pattern of dispensing suggests for

most claimants, prescriptions were new.

5.5 Conclusion

In this Canadian sample of workers’ compensation LBP claimants, dispensing of prescription

opioids, NSAIDs, and SMRs was common both pre- and post-injury and occurred quickly after

injury. Most claimants receiving opioids received low doses of weak and short-acting

formulations. Dispensing patterns also suggest that for most, injuries were acute and treatment

was time limited. However, the small, sustained increase in post-injury dispensing compared to

pre-injury and the increase in dispensing of strong and long-acting opioids and days supply

across all drug classes suggests a small proportion of workers struggle with recovery and will

continue to use these medications almost a year after injury. A more empirical examination of

the factors associated with post-injury prescription drug patterns, particularly for the active

early injury period and for sustained exposure, is needed. Confirmation of our findings in other

samples using comprehensive and more recent prescription data is also warranted.

118

5.6 References

1. Blackwell DL, Lucas JW, Clarke TC. Summary health statistics for U.S. adults: national health interview survey, 2012. Vital Health Stat 10. 2014(260):1-161.

2. Collins JJ, Baase CM, Sharda CE, Ozminkowski RJ, Nicholson S, Billotti GM, et al. The assessment of chronic health conditions on work performance, absence, and total economic impact for employers. Journal of Occupational & Environmental Medicine47(6):547-57. 2005.

3. Zhang W, McLeod C, Koehoorn M. The relationship between chronic conditions and absenteeism and associated costs in Canada. Scand J Work Environ Health. 2016;42(5):413-22.

4. Guo HR, Tanaka S, Halperin WE, Cameron LL. Back pain prevalence in US industry and estimates of lost workdays. Am J Public Health. 1999;89(7):1029-35.




8. Hart LG, Deyo RA, Cherkin DC. Physician office visits for low back pain. Frequency, clinical evaluation, and treatment patterns from a U.S. national survey. Spine (Phila Pa 1976). 1995;20(1):11-9.





119


14. van Tulder MW, Touray T, Furlan AD, Solway S, Bouter LM. Muscle relaxants for non-specific low back pain. Cochrane Database Syst Rev. 2003(2):CD004252.

15. Wong JJ, Côté P, Ameis A, Varatharajan S, Varatharajan T, Shearer HM, et al. Are non-steroidal anti-inflammatory drugs effective for the management of neck pain and associated disorders, whiplash-associated disorders, or non-specific low back pain? A systematic review of systematic reviews by the Ontario Protocol for Traffic Injury Management (OPTIMa) Collaboration. Eur Spine J. 2016;25(1):34-61.

16. Cicero TJ, Inciardi JA, Munoz A. Trends in abuse of Oxycontin and other opioid analgesics in the United States: 2002-2004. J Pain. 2005;6(10):662-72.








24. Franklin GM, Stover BD, Turner JA, Fulton-Kehoe D, Wickizer TM, Disability Risk Identification Study C. Early opioid prescription and subsequent disability among workers

120

with back injuries: the Disability Risk Identification Study Cohort. Spine. 2008;33(2):199-204.






30. Pransky G, Benjamin K, Hill-Fotouhi C, Himmelstein J, Fletcher KE, Katz JN, et al. Outcomes in work-related upper extremity and low back injuries: results of a retrospective study. Am J Ind Med. 2000;37(4):400-9.

31. Hogg-Johnson S, Carnide N, Côté P, Furlan AD, Koehoorn M. Early Opioid Prescriptions for Work-Related Musculoskeletal Disorders of the Back: Understanding Utilization Patterns, Determinants and Impact on Work Disability. Final Report. WorkSafeBC Grant # RS2011-OG12. Toronto, ON: Institute for Work & Health; 2016.

32. Sheppard T, Lipton B, Robertson J. The Relationship Between Accident Report Lag and Claim Cost in Workers Compensation Insurance. Florida: National Council on Compensation Insurance; 2015. Available from: https://www.ncci.com/Articles/Documents/II_Relationship-Accident-Report-Lag-Claim-Cost-WC-Insurance.pdf. Last Accessed: January 17, 2017.

33. Koehoorn M, Cole DC, Hertzman C, Lee H. Health care use associated with work-related musculoskeletal disorders among hospital workers. Journal of Occupational Rehabilitation. 2006;16(3):411-24.

34. Côté P, Yang X, Kristman V, Hogg-Johnson S, Van Eerd D, Rezai M, et al. The association between workers' compensation claims involving neck pain and future health care utilization: a population-based cohort study. Journal of Occupational Rehabilitation. 2013;23(4):547-56.

https://www.ncci.com/Articles/Documents/II_Relationship-Accident-Report-Lag-Claim-Cost-WC-Insurance.pdf

https://www.ncci.com/Articles/Documents/II_Relationship-Accident-Report-Lag-Claim-Cost-WC-Insurance.pdf

121






40. Population Data BC. 2017; Available from: https://www.popdata.bc.ca/. Last Accessed: March 3, 2017.

41. Pencarrick Hertzman C, Meagher N, McGrail KM. Privacy by Design at Population Data BC: a case study describing the technical, administrative, and physical controls for privacy-sensitive secondary use of personal information for research in the public interest. J Am Med Inform Assoc. 2013;20(1):25-8.


43. CANSIM Table 282-0002. Labour force survey estimates (LFS), by sex and detailed age group.Customized. 1998-2009, British Columbia. [Internet]. Statistics Canada. 2017. Available from: http://www5.statcan.gc.ca/cansim/a26?lang=eng&id=2820002. Last Accessed: October 24, 2016.





https://www.popdata.bc.ca/


http://www5.statcan.gc.ca/cansim/a26?lang=eng&id=2820002


122


49. Washington State Agency Medical Directors’ Group (AMDG). Interagency Guideline on Prescribing Opioids for Pain. 2015; Available from: http://www.agencymeddirectors.wa.gov/Files/2015AMDGOpioidGuideline.pdf. Last Accessed: September 2, 2015.

50. American Pain Society. Principles of Analgesic Use in the Treatment of Acute Pain and Cancer Pain. 6th Edition ed. Glenview, IL; 2008.






56. Statistics Canada. Postal Code Conversion File Plus (PCCF+) (82F0086X). Available from: http://www5.statcan.gc.ca/olc-cel/olc.action?lang=en&ObjId=82F0086X&ObjType=2. Last Accessed: February 1, 2017.

57. Smolina K, Gladstone E, Morgan SG. Determinants of trends in prescription opioid use in British Columbia, Canada, 2005-2013. Pharmacoepidemiol Drug Saf. 2016;25(5):553-9.

58. Health Canada. Archived - Canadian Alcohol and Drug Use Monitoring Survey. Summary of Results for 2008. Ottawa, ON: Health Canada; 2009; Available from: http://www.hc-sc.gc.ca/hc-ps/drugs-drogues/stat/_2008/summary-sommaire-eng.php. Last Accessed: January 28, 2017.

59. Kantor ED, Rehm CD, Haas JS, Chan AT, Giovannucci EL. Trends in Prescription Drug Use Among Adults in the United States From 1999-2012. JAMA. 2015;314(17):1818-31.

60. WorkSafeBC. The Prescription of Narcotics and Other Drugs of Addiction. Richmond, BC: WorkSafeBC; 1989.




http://www5.statcan.gc.ca/olc-cel/olc.action?lang=en&ObjId=82F0086X&ObjType=2

http://www.hc-sc.gc.ca/hc-ps/drugs-drogues/stat/_2008/summary-sommaire-eng.php

http://www.hc-sc.gc.ca/hc-ps/drugs-drogues/stat/_2008/summary-sommaire-eng.php

123



63. Smith JM, Bombardier C, Frank JW, Hogg-Johnson S, Kerr MS, Shannon HS, et al. Clinical profile of cases in a worksite study of low back pain according to WCB claim and lost-time status at baseline and 26 weeks. Arthritis Rheum. 1997;40:S263.

64. Hogg-Johnson S, Bombardier C, Guzman J, Smith J, Kerr M, Sinclair S, et al. A comparison of symptoms and disability from acute low back pain (LBP) among three study settings. Arthritis Rheum. 1997;40:S263.


66. Webster BS, Cifuentes M, Verma S, Pransky G. Geographic variation in opioid prescribing for acute, work-related, low back pain and associated factors: a multilevel analysis. American Journal of Industrial Medicine52(2):162-71. 2009.

67. Madadi P, Amstutz U, Rieder M, Ito S, Fung V, Hwang S, et al. Clinical practice guideline: CYP2D6 genotyping for safe and efficacious codeine therapy. J Popul Ther Clin Pharmacol. 2013;20(3):e369-96.

68. Frei MY, Nielsen S, Dobbin MD, Tobin CL. Serious morbidity associated with misuse of over-the-counter codeine-ibuprofen analgesics: a series of 27 cases. Med J Aust. 2010;193(5):294-6.

69. McAvoy BR, Dobbin MD, Tobin CL. Over-the-counter codeine analgesic misuse and harm: characteristics of cases in Australia and New Zealand. N Z Med J. 2011;124(1346):29-33.

70. Roxburgh A, Hall WD, Burns L, Pilgrim J, Saar E, Nielsen S, et al. Trends and characteristics of accidental and intentional codeine overdose deaths in Australia. Med J Aust. 2015;203(7):299.

71. Sproule BA, Busto UE, Somer G, Romach MK, Sellers EM. Characteristics of dependent and nondependent regular users of codeine. J Clin Psychopharmacol. 1999;19(4):367-72.

72. Dhalla IA, Mamdani MM, Sivilotti ML, Kopp A, Qureshi O, Juurlink DN. Prescribing of opioid analgesics and related mortality before and after the introduction of long-acting oxycodone. CMAJ Canadian Medical Association Journal181(12):891-6. 2009.

124

73. Van Zee A. The promotion and marketing of oxycontin: commercial triumph, public health tragedy. Am J Public Health. 2009;99(2):221-7.

74. Bohnert AS, Valenstein M, Bair MJ, Ganoczy D, McCarthy JF, Ilgen MA, et al. Association between opioid prescribing patterns and opioid overdose-related deaths. JAMA. 2011;305(13):1315-21.

75. Gomes T, Mamdani MM, Dhalla IA, Paterson JM, Juurlink DN. Opioid dose and drug-related mortality in patients with nonmalignant pain. Arch Intern Med. 2011;171(7):686-91.

76. Gomes T, Redelmeier DA, Juurlink DN, Dhalla IA, Camacho X, Mamdani MM. Opioid dose and risk of road trauma in Canada: a population-based study. JAMA Intern Med. 2013;173(3):196-201.

77. Fulton-Kehoe D, Garg RK, Turner JA, Bauer AM, Sullivan MD, Wickizer TM, et al. Opioid poisonings and opioid adverse effects in workers in Washington state. Am J Ind Med. 2013;56(12):1452-62.

78. Berecki-Gisolf J, Hassani-Mahmooei B, Collie A, McClure R. Prescription Opioid and Benzodiazepine Use After Road Traffic Injury. Pain Med. 2016;17(2):304-13.

125

Claims in provided data extract

n=916,804

Claims with consolidated claims

n=35,637 excluded

<18 years of age at injury

n=3,642 excluded

Not a STD claim and/or no paid STD days within first 8

weeks after injury

n=413,761 excluded

STD claim is non-allowed claim

n=1,666 excluded

Allowed STD claims with at least one day paid in first 8 weeks after injury

n=462,098

Non-LBP diagnosis and/or injury year not between 1998

and 2009

n=289,969 excluded

Claim Level

Exclusions

LBP hospitalization or serious LBP

hospitalization/outpatient visit within 5 days after injury

n=1,088 excluded

Claimant a non-BC resident and/or not registered for

health services continuously 2 years pre- to 1 year post-

injury

n=28,170 excluded

Injury date after registration date

n=5 excluded

Claim Level

Exclusions

(continued next page)

126

Eligible cohort claims: n=142,993

Eligible cohort claimants: n=113,434

Claim with at least one other allowed claim in year after

injury date and/or summed total of benefits paid ≤0

n=43,760 claims excluded

n=14,201 claimants excluded

Eligible cohort claimants with one allowed LBP STD claim from 1998 to 2009

n=99,233

Claimants with at least 2 health care visits and/or 1

hospitalization for cancer within 2 years pre- and 1 year

post-injury

n=2,109 claimants excluded

Final cohort claimants

n=97,124

Figure 5-1 Claim-level and claimant-level exclusions to derive the final cohort sample of workers’ compensation claimants with allowed low back pain short-term disability claims occurring between 1998 and 2009.

Abbreviations: LBP = low back pain; STD = short-term disability

Claimant Level

Exclusion

Claimant Level

Exclusion

127

Table 5-1 Characteristics of the cohort of workers’ compensation claimants with an allowed short-term disability claim for low back pain between 1998 and 2009 (n=97,124)

Variable n ( % )

Age at injury 18-24 10,180 (10.5) 25-34 21,444 (22.1) 35-44 29,929 (30.8) 45-54 25,415 (26.2) ≥55 10,156 (10.5)

Sex1

Women 36,606 (37.7) Men 60,517 (62.3)

Neighbourhood income quintile1

1 (lowest) 21,169 (21.8) 2 22,295 (23.0) 3 20,874 (21.5) 4 18,234 (18.8) 5 (highest) 12,930 (13.3)

Regional health authority1

Interior 14,116 (14.5) Fraser 41,237 (42.5) Vancouver Coastal 18,834 (19.4) Vancouver Island 17,538 (18.1) Northern 5,303 ( 5.5)

Injury year

1998/99 18,413 (19.0) 2000/01 17,105 (17.6) 2002/03 16,002 (16.5) 2004/05 15,625 (16.1) 2006/07 15,756 (16.2) 2008/09 14,223 (14.6)

First three digits of ICD-9 diagnosis on claim

722 – Intervertebral disc disorders 2,954 ( 3.0) 724 – Other and unspecified disorders of back 9,763 (10.1) 846 – Sprains and strains of sacroiliac region 21,031 (21.7) 847 – Sprains and strains lumbar, sacrum, coccyx, unspecified site of back 63,376 (65.3)

Number of workers’ compensation claims within 2 years prior to injury2

0 65,461 (67.4) 1 21,984 (22.6) ≥2 9,679 (10.0)

Abbreviations: ICD-9=International Statistical Classification of Diseases and Related Health Problems, 9th Revision 1Number of claimants with data missing on the following variables: n=1 sex; n=1,622 neighbourhood income quintile; n=96

regional health authority. 2Includes all types of claims (short-term disability, long-term disability, health care only, vocational rehabilitation) for any

diagnosis (not limited to low back pain claims).

128

Table 5-2 Prescription dispensing patterns one year before injury date compared to one year after injury date among claimants with low back pain short-term disability claims (n=97,124)

One Year Pre-Injury One Year Post-Injury

Dispensing Pattern n % (95% CI) n % (95% CI)

Prevalence of at least one dispensed prescription

At least one opioid, NSAID, and/or SMR 32,977 34.0 (33.7-34.3) 66,051 68.0 (67.7-68.3) Opioids1,2,3 18,995 19.6 (19.3-19.8) 37,639 38.8 (38.5-39.1)

Weak opioids 18,095 95.3 (95.0-95.6) 35,772 95.0 (94.8-95.3) Strong opioids 2,053 10.8 (10.4-11.3) 5,566 14.8 (14.4-15.2)

Short-acting opioids 18,914 99.6 (99.5-99.7) 37,499 99.6 (99.6-99.7) Long-acting opioids

349 1.8 (1.7-2.0) 1,085 2.9 (2.7-3.1)

Codeine 17,569 92.5 34,459 91.6 Oxycodone 1,612 8.5 4,345 11.5 Meperidine 430 2.3 1,200 3.2 Tramadol 400 2.1 1,586 4.2 Hydromorphone 313 1.7 969 2.6 Morphine 263 1.4 838 2.2

NSAIDs1,2 20,499 21.1 (20.9-21.4) 48,714 50.2 (49.8-50.5)

Naproxen 10,561 51.5 29,208 60.0 Diclofenac 4,166 20.3 11,426 23.5 Ketorolac 1,767 8.6 3,524 7.2 Celecoxib 1,754 8.6 4,927 10.1 Ibuprofen 1,479 7.2 3,096 6.4 Indomethacin 1,394 6.8 2,053 4.2 Rofecoxib 1,369 6.7 3,687 7.6

Skeletal muscle relaxants1 6,068 6.3 (6.1-6.4) 27,603 28.4 (28.1-28.7)

Cyclobenzaprine 4,881 80.4 22,858 82.8 Methocarbamol 1,112 18.3 4,834 17.5 Baclofen 245 4.0 1,054 3.8

Proportion receiving at least one dispensed prescription stratified by one-year pre-injury prescription history

No pre-injury history (incidence proportion) At least one opioid, NSAID, and/or SMR (n=64,147)

39,767 62.0 (61.6-62.4)

Opioids (n=78,129) 26,201 33.5 (33.2-33.9) NSAIDs (n=76,625) 34,955 45.6 (45.3-46.0) SMRs (n=91,056) 24,439 26.8 (26.6-27.1)

Pre-injury history

At least one opioid, NSAID, and/or SMR (n=32,977)

26,284 79.7 (79.3-80.1)

Opioids (n=18,995) 11,438 60.2 (59.5-60.9) NSAIDs (n=20,499) 13,759 67.1 (66.5-67.8) SMRs (n=6,068) 3,164 52.1 (50.9-53.4)

129

One Year Pre-Injury One Year Post-Injury Dispensing Pattern n % (95% CI) n % (95% CI)

Prescription combinations received among claimants dispensed at least one prescription (n=32,977 pre-injury, n=66,051 post-injury)

NSAIDs only 10,880 33.0 (32.5-33.5) 15,657 23.7 (23.4-24.0) SMRs only 1,286 3.9 (3.7-4.1) 4,093 6.2 (6.0-6.4) NSAIDs and SMRs 1,816 5.5 (5.3-5.8) 8,662 13.1 (12.9-13.4) Opioids only 10,106 30.7 (30.2-31.1) 9,247 14.0 (13.7-14.3) Opioids and NSAIDs only 5,923 18.0 (17.6-18.4) 13,544 20.5 (20.2-20.8) Opioids and SMRs only 1,086 3.3 (3.1-3.5) 3,997 6.1 (5.9-6.2) All 3 drug classes 1,880 5.7 (5.5-6.0) 10,851 16.4 (16.2-16.7)

# of prescriptions among claimants dispensed at least one prescription

Opioids (n=18,995 pre-injury, n=37,639 post-injury)

1 11,561 60.9 (60.2-61.6) 20,502 54.5 (54.0-55.0) 2 3,181 16.8 (16.2-17.3) 6,849 18.2 (17.8-18.6) 3-4 1,973 10.4 (10.0-10.8) 4,746 12.6 (12.3-12.9) ≥5 2,280 12.0 (11.5-12.5) 5,542 14.7 (14.4-15.1)

NSAIDs (n=20,499 pre-injury, n=48,714 post-injury)

1 13,156 64.2 (63.5-64.8) 28,149 57.8 (57.4-58.2) 2 3,804 18.6 (18.0-19.1) 10,194 20.9 (20.6-21.3) 3-4 2,388 11.7 (11.2-12.1) 6,743 13.8 (13.5-14.2) ≥5 1,151 5.6 (5.3-5.9) 3,628 7.5 (7.2-7.7)

SMRs(n=6,068 pre-injury, n=27,603 post-injury)

1 4,553 75.0 (73.9-76.1) 19,770 71.6 (71.1-72.2) 2 879 14.5 (13.6-15.4) 4,544 16.5 (16.0-16.9) 3-4 369 6.1 (5.5-6.7) 2,204 8.0 (7.7-8.3) ≥5 267 4.4 (3.9-4.9) 1,085 3.9 (3.7-4.2)

Mean (SD) 95% CI Mean (SD) 95% CI

Dispensing Pattern Median IQR Median IQR

Cumulative days supply per claimant among claimants with at least one day of supply4

Opioids (n=19,217 pre-injury, n=37,720 post-injury)

28.0 (60.8) 8.0

27.2-28.9 5.0-20.0

29.3 (57.1) 10.0

28.7-29.9 5.0-24.0

Weak opioids (n=18,312 pre, n=35,865 post)

24.0 (51.6) 7.0

23.2-24.7 4.0-17.0

24.7 (47.3) 10.0

24.2-25.2 5.0-20.0

Strong opioids (n=2,080 pre, n=5,569 post)

54.7 (99.0) 10.0

50.4-58.9 5.0-37.0

45.5 (82.0) 10.0

43.3-47.6 5.0-36.0

NSAIDs (n=21,082 pre-injury, n=49,038 post-injury)

35.4 (54.5) 15.0

34.6-36.1 10.0-31.0

35.7 (48.8) 18.0

35.2-36.1 10.0-37.0

SMRs (n=6,237 pre-injury, n=27,710 post-injury)

25.0 (42.6) 10.0

24.0-26.1 7.0-24.0

22.2 (33.1) 10.0

21.8-22.6 8.0-23.0

130

One Year Pre-Injury One Year Post-Injury Mean (SD) 95% CI Mean (SD) 95% CI

Dispensing Pattern Median IQR Median IQR

# of days between last prescription dispensed and injury date among claimants with at least one prescription in year before injury

Opioids (n=18,995) 162.1 (108.2) 154.0

160.6-163.7 64.0-253.0

Weak opioids (n=18,095) 164.9 (107.7) 158.0

163.3-166.5 68.0-256.0

Strong opioids (n=2,053) 152.4 (111.8) 141.0

147.6-157.3 45.0-245.0

NSAIDs (n=20,499) 164.5 (106.2) 156.0

163.0-165.9 70.0-254.0

Skeletal muscle relaxants (n=6,068) 175.4 (105.7) 172.0

172.8-178.1 84.0-265.0

# of days between injury date and first prescription dispensed among claimants with at least one prescription in year after injury

Opioids (n=37,639) 59.7 (96.5) 6.0

58.7-60.6 1.0-79.0

Weak (n=35,772) 61.8 (97.5) 7.0

60.8-62.8 1.0-86.0

Strong (n=5,566) 87.1 (108.3) 25.0

84.2-89.9 3.0-154.0

NSAIDs (n=48,714) 40.3 (80.1) 4.0

39.6-41.0 1.0-26.0

Skeletal muscle relaxants (n=27,603) 28.7 (68.0) 3.0

27.9-29.5 1.0-13.0

Average daily MED among claimants with at least one days supply of opioids or SMRs containing codeine (n=19,901 pre-injury, n=40,573 post-injury)4,5

31.2 (31.8) 27.0

30.8-31.6 18.0-36.7

31.6 (29.0) 27.0

31.3-31.8 18.0-37.5

Dispensing Pattern n % (95% CI) n % (95% CI)

Average daily MED >=120mg among claimants with at least one days supply of opioids or SMRs containing codeine (n=19,901 pre-injury, n=40,573 post-injury)4,5

157

0.8 (0.7-0.9) 308

0.8 (0.7-0.8)

At least one day of high MED (>=120mg) among claimants with at least one days supply of opioids or SMRs containing codeine (n=19,901 pre-injury, n=40,573 post-injury)4,5

479

2.4 (2.2-2.6) 1,336

3.3 (3.1-3.5)

Abbreviations: IQR=interquartile range; MED=morphine-equivalent dose; mg=milligrams; N/A=not applicable; NSAIDs=nonsteroidal anti-inflammatory drugs; Rx=prescription; SD=standard deviation; SMRs=skeletal muscle relaxants.

131

1 Most commonly dispensed drugs are described for each drug class. The denominator for the proportions presented include only individuals with at least one dispense for the relevant drug class.

2 The following drugs were only available during a portion of the study period: tramadol (2004-2009), celecoxib (1999-2009), and rofecoxib (1999-2004).

3 For data on opioid strength and length of action, the denominator for the proportions presented include only individuals with at least one opioid dispense.

4 Sample size is based on individuals with at least one day supply of each drug class in the specific time period. Some claimants may have been dispensed a prescription outside of the period (e.g., >1 year pre-injury), but the days supply spilled over into the relevant time period. Thus, the sample size is slightly larger than the number of claimants who received a prescription in the relevant time period.

5 Morphine equivalent dose was not calculated for methadone or opioids provided via parenteral or intranasal routes (<1% of all dispensing records).

132

Figure 5-2 Proportion of claimants dispensed their first prescription by week after injury among claimants with at least one dispensed opioid, NSAID, or SMR prescription. Each bar represents a 1-week period post-injury.

Note: Blue=first 8 weeks post-injury / yellow=9-52 weeks post-injury

133

Figure 5-3 Proportion of claimants with at least one dispensed prescription in 8-week blocks from 48 weeks pre- to 48 weeks post-injury (n=97,124).

Note: Blue=pre-injury / yellow=post-injury

134

Figure 5-4 Mean cumulative days supply / claimant in 8-week blocks from 48 weeks pre- to 48 weeks post-injury for claimants with at least one day supplied of opioids and/or SMRs with codeine.

135

Figure 5-5 Proportion of claimants receiving at least one dispensed weak, strong, short-acting or long-acting opioid prescription in 8-week blocks from 48 weeks pre- to 48 weeks post-injury for claimants with at least one dispensed opioid in the period.

Note: Blue=pre-injury / yellow=post-injury

136

Figure 5-6 Average daily MED / claimant in 8-week blocks from 48 weeks pre- to 48 weeks post-injury for claimants with at least one day supplied of opioids and/or SMRs with codeine.

137

CHAPTER 6 FACTORS ASSOCIATED WITH EARLY OPIOID, NSAID, AND MUSCLE

RELAXANT DISPENSING AFTER A WORK-RELATED LOW BACK INJURY

Nancy Carnide, Sheilah Hogg-Johnson, Pierre Côté, Mieke Koehoorn, Andrea Furlan

Abstract

Objectives In a sample of workers’ compensation claimants with low back pain (LBP) claims, to

determine the claimant-level factors associated with early drug class(es) dispensed after injury

and to examine the association between claimant and prescriber factors and drug class(es) at

the first post-injury dispense.

Study Design Historical cohort study.

Population Workers from British Columbia, Canada filing a new, short-term disability workers’

compensation claim for a LBP injury occurring between 1998 and 2009 with at least one

compensated lost workday.

Methods Administrative data on workers’ compensation was linked with prescription

dispensing and health care services data. Main drug classes of interest were opioids,

nonsteroidal anti-inflammatory drugs (NSAIDs), and skeletal muscle relaxants (SMRs). The

linked database provided information on claimant-level factors, including sociodemographics,

occupation, injury diagnosis, and pre-existing comorbidities, as well as pre-injury prescription

dispensing, physician visits, hospitalizations, spine x-rays, and spine surgeries. The association

between these claimant-level factors and drug class(es) received in the first eight weeks after

injury were examined using multinomial logistic regression. The association between claimant-

138

and prescriber-level factors (sex, year of birth, specialty) with drug class(es) at first dispense

was also examined in a multilevel multinomial logistic regression analysis, with prescriber as the

clustering group.

Results The claimant-level factors found to be consistently associated with both dispensing

outcomes were prior cumulative days supplied with one of the three drug classes, as well as

prior dispenses for antidepressants, anticonvulsants, and sedative-hypnotic/anxiolytics, ICD-9

diagnosis on the claim, age at injury, a number of pre-existing comorbidities, prior physician

visits and hospitalizations, year of injury, sex, health authority, occupation, and prior workers’

compensation claims. The direction of the relationships varied by outcome category. Results of

the multilevel analysis suggest 25-36% of the variability in the drug class(es) received at first

dispense was accounted for by the prescribers. Prescriber sex, specialty, and birth year were

associated with drug class(es) received at first dispense, but adding them to the model did not

account for this between-prescriber variation.

Conclusion The investigation of claimant-level characteristics suggests early post-injury

dispensing is multifactorial, with a number of factors found to be associated with dispensing,

including pre-injury dispensing and health care, as well as injury diagnosis, age, sex, and

occupation. Prescriber variation is also important in accounting for differences in drug class at

first prescription, but could not be explained by basic prescriber characteristics. Further

research is needed to understand the prescriber factors that influence early prescribing in this

population given their importance in the current study. Future research on compensation

system factors or more detailed work-related characteristics may prove beneficial to

understanding the drivers of early dispensing.

139

6.1 Introduction

In recent years, opioid prescribing for low back pain (LBP) has been on the rise (1-3) and data

suggest opioids may now be the most commonly prescribed drug class for LBP.(4, 5) However,

this practice is controversial. There is a dearth of high quality and consistent evidence from

randomized controlled trials on the effectiveness of opioids for LBP. This is particularly evident

for acute LBP, for outcomes beyond pain relief, and for comparative effectiveness.(6, 7) On the

other hand, there is extensive evidence from observational studies demonstrating the risk of

serious harms associated with opioid use.(8-11) Most individuals with acute episodes of LBP will

also experience rapid improvements in pain and function,(12-14) casting doubt on the necessity

for aggressive early treatment with opioids. Indeed, clinical guidelines generally recommend

opioids be reserved only for severe or refractory pain.(15, 16)

Yet recent clinical studies have shown prescribing of opioids early in the course of a LBP episode

is common,(4, 17) a trend also seen among workers’ compensation claimants. The prevalence

of early prescriptions among injured workers with a lost-time claim for LBP has varied from

5.1% (18) to 21.2% (19) within two weeks of claim onset; 9.4% within four weeks of claim; (20)

12.1% within two days of the first emergency department visit; (21) and 35% (22, 23) within six

weeks of the first medical visit. Early opioid dispensing among LBP claimants has also been

shown to be associated with a number of harms, including prolonged work disability (18-20, 22)

and higher medical and workers’ compensation costs.(19, 21)

Although much attention has been paid to documenting early opioid exposure and its effects,

less is known about its determinants. To our knowledge, only two studies have examined the

factors associated with early opioid prescriptions following a work-related LBP injury.(23, 24)

140

Stover et al. found daily tobacco use, pain radiating below the knee, and greater injury severity

were positively associated with receiving opioids within six weeks of the first post-injury

medical visit,(23) while Webster et al. found that state-level factors, namely greater household

income inequality, lower number of physicians per capita, and greater workers’ compensation

cost containment efforts, explained the majority of geographic variation in early

prescribing.(24) Another study of claimants with various injuries found increasing injury

severity, older age, male sex, greater number of previous claims, and higher income were

positively associated with receiving an early opioid.(18)

In addition to opioids, we previously demonstrated that nonsteroidal anti-inflammatory drugs

(NSAIDs) and skeletal muscle relaxants (SMRs) are commonly dispensed to workers with LBP

injuries.(25) However, no study of injured workers receiving compensation for LBP has

investigated the factors associated with early dispensing of these other drug classes. (4, 26, 27)

Similarly, previous studies of compensation claimants have not examined the influence of pre-

injury and prescriber factors on post-injury dispensing.

Using a large administrative database rich in information on health care utilization, prescription

dispensing, and sociodemographics, our objective was to determine the claimant-level factors

associated with early opioid, NSAID, and SMR dispensing in a sample of workers’ compensation

claimants with LBP claims. We chose eight weeks to define the early period as WorkSafeBC

policy at the time of the study limited reimbursement of opioids to the first eight weeks after

injury or surgery for most claims.(28) We also sought to investigate the claimant- and

prescriber-level factors that may be associated with the drug class(es) received at first

prescription dispense after injury.

141

6.2 Methods


We conducted an historical cohort study of workers’ compensation claimants in British

Columbia (BC), Canada who filed a new short-term disability claim for a LBP-related injury from

1998 to 2009. The methods have been described in detail elsewhere.(25) We obtained

longitudinal, de-identified administrative data for the period 1996 through 2010 from six

linkable databases that included data on: workers’ compensation claims (WorkSafeBC Claims

and Firm Level Files), pharmacy dispensing (PharmaNet), outpatient health care (Medical

Services Plan [MSP] Payment Information File), hospital discharges (Discharge Abstract

Database [DAD]), and practitioner (MSP Practitioner File) and claimant (MSP Consolidation File)

demographics.(29-34)

WorkSafeBC is the provincial workers’ compensation organization in BC, providing no-fault

single payer insurance to workers with work-related injuries and diseases. During the study

period, 92.5% to 94% of the workforce was eligible for coverage by WorkSafeBC.(35) Workers

excluded from coverage include the self-employed and other smaller segments of the labour

force.(36)

PharmaNet is a province-wide prescription monitoring program that captures detailed

information for all prescription medications dispensed from community and hospital outpatient

pharmacies in BC. Medical Services Plan (MSP) is the universal health insurance program in BC,

providing coverage for approximately 96% of residents. MSP data consist of billing records for

all medically required outpatient services provided by fee-for-service practitioners. Until 2001,

limited benefits for supplementary health care practitioners (e.g., chiropractic, massage

142

therapy, physiotherapy) were also provided. The Discharge Abstract Database (DAD) file

contains diagnostic and intervention data for all hospital discharges of inpatients and day

surgery patients from acute care hospitals in BC. The MSP Practitioner File and MSP

Registration and Premium Billing (Consolidation) file contains basic demographic data on

practitioners enrolled in MSP and individuals registered to receive health services in BC,

respectively.

The BC Ministry of Health, WorkSafeBC, PharmaNet, and the College of Physicians and Surgeons

of BC approved access to and use of the data facilitated by Population Data BC in this study.

This study was approved by the University of Toronto Health Sciences Research Ethics Board

(protocol reference number 26885).

6.2.2 Study Population

We obtained data for workers who filed at least one new workers’ compensation short-term

disability claim with WorkSafeBC for LBP between January 1, 1998 and December 31, 2009. All

claims for these workers dating back to 1991 were provided by Population Data BC to the

researchers. We included claimants with short-term disability claims for a nonspecific LBP

disorder identified using four digit International Classification of Diseases, 9th Revision (ICD-9)

codes (37) (Appendix B, Table B-1). Short-term disability was defined as at least one day of

wage replacement benefits in the first eight weeks following injury. The claimant had to be at

least 18 years at injury with no LBP-related hospitalization and/or no serious LBP-related

outpatient service within five days after injury (Appendix B, Table B-2). Finally, claimants had to

be BC residents continuously eligible for health services from two years before to one year

143

post-injury. Injury date had to be equal to or precede claim registration date. Consolidated

claims were excluded. Claims may be consolidated for administrative reasons, such as when

duplicate claims have been inadvertently created for the same worker and injury event.

Consolidation may also occur if it is determined that a new claim is a compensable consequence

of a previous injury for which a claim has been initiated.

After applying these criteria, some claimants had more than one eligible claim. From this pool

of eligible claims, we selected one index claim per claimant. The index claim was defined as the

earliest LBP short-term disability claim where: 1) there were no other allowed claims in the year

after injury; and 2) the total benefits paid in the year after injury was greater than zero.

Claimants were then excluded if they had at least one hospitalization and/or two outpatient

billings for cancer (ICD-9 codes 140-209 and 235-239; International Statistical Classification of

Diseases and Related Health Problems, Tenth Revision, Canada [ICD-10-CA](38) codes C00-C97

and D37-D48) two years before through one year after injury.

6.2.3 Outcomes

Dispensing data from the following drug classes of interest were obtained from the PharmaNet

dataset: opioids (American Hospital Formulary Service [AHFS] (39) codes 28:08.08, 28:08.12),

NSAIDs (28:08.04.08, 28:08.04.24, 28:08.04.92), and SMRs (12.20.04, 12.20.08, 12.20.12). See

Appendix B, Table B-3 for a complete list of active ingredients included by drug class. We

excluded over-the-counter medications as they are not required to be entered into the

PharmaNet database.

144

We examined two outcomes related to the dispensing of these three prescription drug classes

in the first eight weeks post-injury:

i. Drug class(es) dispensed in first eight weeks after injury: This outcome aimed to measure

which drug classes were dispensed during the first eight weeks post-injury. For each drug

class, the presence of at least one dispensing record in the eight week period defined a

dispense. For each claimant, possible categories included no dispensing records for either

opioids, NSAIDs, or SMRs; dispensed at least one NSAID and/or SMR only; dispensed at

least one opioid only; dispensed at least one opioid and at least one NSAID and/or SMR.

ii. Drug class(es) at first dispensed prescription after injury: This outcome was constructed in

a similar manner and aimed to measure the drug classes dispensed at the first

prescription dispense in the first eight weeks after injury. This outcome was restricted to

claimants with at least one drug dispense for one of the three drug classes in the first

eight weeks. Categories for this outcome were dispensed at least one NSAID and/or SMR

only; dispensed at least one opioid only; dispensed at least one opioid and at least one

NSAID and/or SMR.

6.2.4 Claimant-Level Factors

The claimant-level factors for our analysis were limited to those available in the administrative

data, but to the extent possible, informed by factors identified in prior research of various pain

populations as associated with dispensing of one of the three drug classes (e.g.,

sociodemographics, severity, prior compensation claims, comorbidities) or our own hypotheses

145

of potentially important factors (e.g., occupation, pre-injury health care patterns, pre-injury

experience with these drug classes, prescriber factors).

6.2.4.1 Sociodemographic and Work-Related Factors

Sex, regional health authority (Fraser Health, Vancouver Coastal Health, Vancouver Island

Health, Northern Health, and Interior Health) (40) and neighbourhood income quintile were

obtained directly from the MSP Consolidation File. Year and month of birth were used to

approximate age at injury date (18-24, 25-34, 35-44, 45-54, ≥55 years).

Claimant occupation, obtained from the WorkSafeBC claim file and coded using Statistics

Canada’s Standard Occupational Classification,(41) was linked to data from The National

Occupational Classification Career Handbook.(42) This handbook provides global ratings to 923

occupational profiles to define skills, worker characteristics, and other indicators, including

physical activities and environmental conditions. Ratings for each occupational characteristic

were assigned to each occupational code by trained occupational analysts using a modified

Delphi procedure. For this study, two occupational characteristics were considered: physical

demands of the occupation and whether the job involved working near or with equipment,

instruments, machinery, or power/hand tools that may be a potential source of accident or

injury. A dichotomous yes/no variable for each claimant identified whether their occupation

involved heavy physical strength requirements (i.e., handling loads of more than 20kg) and/or

involved use of equipment/machinery/instruments.

146

6.2.4.2 Injury-Related Factors

Year of injury (grouped in two year groups from 1998 to 2009), ICD-9 diagnosis code associated

with the claim, and number of prior workers’ compensation claims (0, 1, ≥2) were obtained from

the WorkSafeBC claims file. ICD-9 diagnosis was re-categorized based on the first three digits:

722 (Intervertebral Disc Disorders), 724 (Other and Unspecified Disorders of Back), 846 (Sprains

and Strains of Sacroiliac Region), and 847 (Sprains and Strains of Other and Unspecified Parts of

Back).

6.2.4.3 Comorbidities

Pre-existing health conditions were identified using MSP outpatient and DAD hospitalization

data, including pain-related conditions, mental health and substance use disorders, and other

chronic conditions. Conditions were selected if they represented an indication for one of the

drug classes, a contraindication or warning, or were common chronic conditions that could be

indicative of overall poor health.

Claimants with one unique hospital record and/or two outpatient billing records on separate

visits in the two years before injury with at least one relevant ICD-9 or ICD-10-CA diagnosis

(Appendix B, Table B-5) were considered to have the condition. Case definitions were

informed by the findings of prior validation studies (references in Appendix B, Table B-5). The

combination of claimant ID and date of service or separation date was used to identify unique

visits.

147

6.2.4.4 Health Care Utilization and Other Prescription Factors

Prior spinal x-ray (yes/no) was defined as at least one MSP outpatient billing record with a

relevant fee item code that denotes the insured medical service provided or one DAD hospital

separation record with a relevant Canadian Classification of Diagnostic, Therapeutic, and

Surgical Procedures system (CCP) (43) or Canadian Classification of Health Interventions (CCI)

(44) procedure code (Appendix B, Table B-6) in the year before injury. Prior spine surgery

(yes/no) was defined as the presence of at least one applicable CCP or CCI procedure code

(Appendix B, Table B-7) in the DAD two years before injury. Prior hospitalizations for any

reason (yes/no) was defined as at least one hospital separation in the year before injury in the

DAD.

Number of outpatient physician health care visits in the year before injury was determined using

MSP outpatient data for three groups of physicians: general practitioners (GPs), medical

specialists commonly seen for pain-related complaints (herein referred to as pain-related

specialists), and other medical specialists. Up to three specialty fields were available. Physicians

were assigned a specialty of GP if no other specialty was listed. Physicians with at least one

specialty field that included physiatry, neurology, neuropsychiatry, neurosurgery, orthopaedic

surgery, internal medicine, anesthesiology, occupational medicine, emergency medicine,

osteopathic medicine, or rheumatology were assigned a designation of ‘pain-related specialist’.

If no pain-related specialty was listed, but at least one other specialty was present, the

physician was designated as ‘other medical specialist’. Outpatient billing records identified as a

consultation or examination were considered visits. The unique combination of physician ID,

claimant ID, and service date defined one visit. Number of GP visits was categorized as 0, 1-2, 3-

148

4, 5-7, and ≥8, while number of pain-related or other medical specialist visits was categorized as

0, 1, 2, and ≥3. For claimants with injury years 1998 to 2001, we also constructed separate

variables for the number of outpatient supplementary care visits to physiotherapists,

chiropractors, and massage therapists in the year before injury (0, 1-5, ≥6) using a similar

procedure. This was done only for claimants with injury years 1998 to 2001 when MSP provided

limited insured benefits for these practitioners.

Using PharmaNet data, the cumulative days supply in receipt of dispensed opioids, NSAIDs, and

SMRs in the year before injury (0, 1-14, ≥15) was calculated by summing the days supply across

all dispensed prescriptions for a given drug class. When dispensing records overlapped in time

for the same drug class, days were counted only once.

We also obtained data from PharmaNet on specific drugs that may be used as adjuvants in LBP

management (45): antidepressants (28:16.04.16, 28:16.04.24, 28:16.04.28), anticonvulsants

(28:12.08, 28:12.12, 28:12.92), and sedative hypnotics/anxiolytics (28:24.08, 28:24.92). Receipt

of dispensed antidepressant, anticonvulsant, and sedative hypnotic/anxiolytic prescriptions in

the year before injury (yes/no) was defined as the presence of at least one dispensing record for

one of these drug classes in the year prior to injury.

6.2.5 Prescriber-Level Factors

Data on prescriber sex, year of birth (<1945, 1945-1954, 1955-1964, ≥1965), and specialty were

obtained from the MSP Practitioner File. Up to three specialties by year of registration in MSP

were provided. Specialty was assigned in a similar manner described above and categorized as

GP, pain-related specialist, other medical specialist, and dental surgeon.

149


Descriptive statistical methods were used to describe baseline characteristics and prescription

dispensing patterns in the first eight weeks following injury date.

The association between the claimant-level factors and the outcome of drug class(es) dispensed

(no drugs as reference) was examined using a multinomial logistic regression model. The

unadjusted association with each variable was first estimated and then all variables were

entered into a final, multivariable model.

We used a multilevel, multinomial logistic regression model to examine the association

between claimant- and prescriber-level factors and the outcome of drug class(es) at first post-

injury dispense (NSAIDs and/or SMRs only as reference) with prescriber as the clustering group.

Following bivariate analyses, three models were fit using the Laplace estimation method: a

random intercept only model (Model 1), a model with only claimant-level factors (Model 2),

and a model with claimant- and prescriber-level factors (Model 3). To estimate between

prescriber variability, we calculated intraclass correlation coefficients (ICCs) for each model

following a previously described method.(46, 47)

In sensitivity analyses, we also included prior supplementary health care as claimant-level

factors that may be associated with the outcomes. These analyses were limited to claimants

with injury years 1998 to 2001 since MSP ceased to insure these services after this time.

The data analysis for this paper was generated using SAS software Version 9.3. Copyright ©

2000-2010 SAS Institute Inc. SAS and all other SAS Institute Inc. product or service names are

registered trademarks or trademarks of SAS Institute Inc., Cary, NC, USA.

150

6.3 Results

6.3.1 Study Sample

We obtained data on 916,804 claims from Population Data BC. A total of 142,993 eligible claims

linked to 113,434 claimants resulted after applying the claim-level inclusion criteria and one

eligible index claim per claimant was chosen for 99,233 claimants. A total of 2,109 claimants

with a history of cancer-related health care utilization were then excluded, resulting in a cohort

of 97,124 claimants.

Data on sex (n=1), neighbourhood income (n=1,622), and health authority (n=96) were missing

for a total of 1,711 claimants, resulting in an overall cohort sample of 95,413 claimants for our

analyses examining the outcome of drug class(es) dispensed.

For the outcome of first prescription drug class dispensed, analysis was further limited to

claimants with at least one dispense in the first eight weeks (n=55,571). From this sample, 490

claimants were excluded as they had more than one prescriber at first dispense and 103 were

excluded as the first prescriber prescribed to more than one claimant and changed specialty

over the study period. Finally, 781 claimants were excluded as their first prescriber had missing

data on sex (n=612), birth year (n=577), and specialty (n=386), leaving a total of 54,197

claimants for this analysis.


In the overall cohort (n=95,413), mean (standard deviation, SD) age at injury was 40.5 (11.2).

(Table 6-1). Men made up 62.2% of the cohort. Claimants from the Fraser regional health

151

authority (42.8%) and those with sprains and strains (87.0%) made up the largest proportion.

Just over half had occupations involving heavy physical demands and/or use of

equipment/machinery. One third had a previous accepted workers’ compensation claim.

Pre-existing pain-related comorbidities, namely back and/or neck pain (33.2%) and arthritis and

other joint pain (17.5%) and mood and anxiety disorders (16.6%) were common. Most

claimants had at least one visit to a GP in the previous year (88.5%). Approximately one fifth of

claimants also had at least one visit to a pain-related specialist (17.1%) or other specialist

(22.5%). Prior hospitalizations (7.7%), spine x-rays (7.0%), and spine surgeries (0.2%) were not

common. A considerable proportion of claimants received opioids (19.7%) and NSAIDs (21.7%)

in the year before injury. Fewer claimants received a dispense for an SMR (6.4%) antidepressant

(5.1%), an anticonvulsant (2.6%), or a sedative hypnotic/anxiolytic (9.0%).

6.3.3 Factors Associated With Drug Class(es) Dispensed in First Eight Weeks After Injury

In the overall cohort (n=95,413), 27.7%, 40.7%, and 24.8% of workers received at least one

opioid, NSAID, or SMR, respectively, in the first eight weeks post injury. When considering the

combinations of drug classes dispensed, 41.8% received no drugs, 30.5% received NSAIDS

and/or SMRs only, 8.1% received opioids only, and 19.6% received opioids with NSAIDs and/or

SMRs.

When compared to receiving no drugs, the claimant-level factors found to be associated with

receiving each of the drug combinations were cumulative days supplied with one of the three

drug classes in the year before injury, ICD-9 diagnosis, age at injury, prior physician visits and

152

hospitalizations, prior prescription dispense for an antidepressant, anticonvulsant, or sedative-

hypnotic/anxiolytic, occupation, year of injury, pre-existing health conditions, sex,

neighbourhood income, health authority, and prior workers’ compensation claims (Table 6-2).

In the multivariable model, claimants with 1-14 days supply of opioids in the prior year had 1.74

times higher odds of receiving opioids only (95%CI 1.62-1.87) or opioids with NSAIDs and/or

SMRs (95%CI 1.65-1.84) than no drug compared to claimants with no prior supply of opioids.

The odds increased further for claimants with ≥15 days of opioids in the prior year, to 6.85

greater odds (95%CI 6.22-7.55) of receiving opioids only and 4.91 times higher odds (95%CI

4.52-5.34) of being dispensed opioids with NSAIDs and/or SMRs. A similar relationship was seen

for prior cumulative days of NSAIDs, with a higher odds of receiving all drug combinations with

increasing number of days supplied, particularly for NSAIDs and/or SMRs only (1-14 days: OR

1.52, 95%CI 1.44-1.61); ≥15 days: 2.02, 95%CI 1.91-2.13) and opioids with NSAIDs and/or SMRs

(1-14 days: OR 1.38, 95%CI 1.29-1.47); ≥15 days: 1.97, 95%CI 1.86-2.09). Similarly, claimants

with prior dispenses for SMRs were at a higher odds of receiving NSAIDs and/or SMRs only (1-

14 days: OR 1.23, 95%CI 1.13-1.35); ≥15 days: 1.47, 95%CI 1.32-1.65) and opioids with NSAIDs

and/or SMRs (1-14 days: OR 1.42, 95%CI 1.29-1.57); ≥15 days: 1.62, 95%CI 1.45-1.82) compared

to claimants with no prior supply, but not opioids only.

When compared to claimants with an ICD-9 diagnosis of 847 (Sprains and Strains of Other and

Unspecified Parts of Back), claimants with an ICD-9 diagnosis of 722 (Intervertebral Disc

Disorders) had significantly higher odds of receiving dispensed NSAIDs and/or SMRs only (OR

1.82, 95%CI 1.62-2.05), opioids only (OR 3.28, 95%CI 2.85-3.78), and opioids with NSAIDs

and/or SMRs (OR 5.04, 95%CI 4.52-5.60) than no drug. Claimants with a diagnosis of 724 (Other

153

and Unspecified Disorders of Back) had 19% lower odds of receiving NSAIDs and/or SMRs only

(OR 0.81, 95%CI 0.77-0.85) or opioids with NSAIDs and/or SMRs (OR 0.83, 95%CI 0.78-0.88).

The odds of receiving any of the drug classes compared to no drugs was significantly higher

with older ages, though the relationship was not linear. In particular, when compared to

claimants aged 18-24, those aged 35-44 had significantly higher odds of receiving NSAIDs

and/or SMRs only (OR 1.40, 95%CI 1.33-1.48), opioids only (OR 1.63, 95%CI 1.48-1.79), and

opioids with NSAIDs and/or SMRs (OR 2.05, 95%CI 1.91-2.21) than no drug.

Claimants with increasing number of prior GP visits in the year before injury were at higher

odds of receiving NSAIDs and/or SMRs only compared to no drugs, ranging from an OR of 1.08

for 1-2 visits (95%CI 1.02-1.14) to an OR of 1.15 for ≥8 visits (95%CI 1.08-1.23). Claimants with

≥8 GP visits also had higher odds of receiving opioids only (OR 1.19, 95%CI 1.07-1.33). Claimants

with prior pain specialist visits were at approximately 10% reduced odds of receiving NSAIDs

and/or SMRs only, but those with three or more visits were at higher odds of receiving opioids

only (OR 1.14, 95%CI 1.00-1.29). Having at least one dispense for an antidepressant,

anticonvulsant, or sedative-hypnotic/anxiolytic in the year before injury was significantly

associated with receiving either of the three drug combinations, particularly opioid

combinations.

Claimants with an occupation involving heavy physical demands and/or use of machinery were

at higher odds of receiving NSAIDs and/or SMRs only (OR 1.12, 95%CI 1.08-1.16), opioids only

(OR 1.16, 95%CI 1.09-1.23), and opioids with NSAIDs and/or SMRs (OR 1.24, 95%CI 1.19-1.29).

Increasing year of injury (i.e., more recent injury) was significantly associated with receiving

154

NSAIDs and/or SMRs only and opioids with NSAIDs and/or SMRs compared to no drugs, but not

opioids only. A number of comorbidities were also found to be significantly associated with a

higher odds of receiving at least one of the three drug class combinations, including

neuropathic pain, headache disorders, prior self-inflicted harm/suicide attempt, diabetes,

upper gastrointestinal disorders, and chronic respiratory disorders.

Women and claimants with higher neighbourhood income quintiles had significantly lower odds

of receiving any of the drug class combinations versus no drugs. There was also some variability

with respect to regional health authority. After adjustment, claimants with increasing prior

workers’ compensation claims, pre-existing back and/or neck pain, and increasing number of

hospitalizations had significantly reduced odds of being dispensed all drug combinations.

6.3.4 Factors Associated With First Drug Class(es) Dispensed

Among claimants receiving at least one prescription in the first eight weeks, the proportion of

claimants receiving NSAIDs and/or SMRs only, opioids only, and opioids with NSAIDs and/or

SMRs as their first dispense was 60.6%, 19.8%, and 19.6%, respectively.

Results of the multi-level analysis of claimant- and prescriber-level factors associated with first

drug class(es) dispensed are found in Table 6-3. The ICCs in the intercept only model (Model 1)

were 0.36 for opioids only and 0.25 for opioids with NSAIDs and/or SMRs, suggesting 36% of the

variability in the odds of receiving opioids only as the first dispense and 25% of the variability in

the odds of receiving opioids with NSAIDs and/or SMRs, respectively, was accounted for by the

prescribers. Including claimant- and prescriber-level factors in the models (Models 2 and 3) did

155

little to change the ICC values and effect estimates for these models were similar. Therefore,

we detail the findings for Model 3 only.

After including both claimant- and prescriber-level factors (Model 3), the claimant-level factors

found be associated with drug class(es) at first dispense were similar: cumulative days supplied

with one of the three drug classes, ICD-9 diagnosis on the claim, age at injury, occupation, sex,

prior physician visits, hospitalizations, and spine surgery, pre-existing health conditions, prior

prescription dispense for an antidepressant, anticonvulsant, or sedative-hypnotic/anxiolytic,

occupation, prior workers’ compensation claims, and health authority.

Increasing number of days supplied with opioids in the previous year was associated with a

higher odds of receiving opioids only (1-14 days: OR 1.62, 95%CI 1.51-1.75; ≥15 days: OR 5.13,

95%CI 4.66-5.65) and opioids with NSAIDs and/or SMRs (1-14 days: OR 1.49, 95%CI 1.39-1.60;

≥15 days: OR 2.82, 95%CI 2.56-3.12). The odds of receiving opioids with NSAIDS and/or SMRs

were also higher for claimants receiving SMRs in the year before injury (1-14 days: OR 1.14,

95%CI 1.00-1.28; ≥15 days: OR 1.22, 95%CI 1.07-1.40). On the other hand, claimants with prior

NSAIDs had lower odds of receiving opioids only (1-14 days: OR 0.81, 95%CI 0.74-0.89; ≥15

days: OR 0.73, 95%CI 0.67-0.79) or opioids with NSAIDs and/or SMRs compared to NSAIDs

and/or SMRs only (1-14 days: OR 0.94, 95%CI 0.86-1.03; ≥15 days: OR 0.91, 95%CI 0.84-0.98).

When compared to claimants with an ICD-9 diagnosis of 847 (Sprains and Strains of Other and

Unspecified Parts of Back), claimants with a diagnosis of 722 (Intervertebral Disc Disorders) had

significantly higher odds of dispensed opioids only (OR 1.68, 95%CI 1.48-1.90) and opioids with

NSAIDs and/or SMRs (OR 1.72, 95%CI 1.53-1.94) than NSAIDs and/or SMRs only. Claimants with

156

a diagnosis of 724 (Other and Unspecified Disorders of Back) also had higher odds of receiving

opioids only (OR 1.16, 95%CI 1.05-1.27) and opioids with NSAIDs and/or SMRs (OR 1.20, 95%CI

1.10-1.31).

The odds of receiving any of the drug classes compared to NSAIDs and/or SMRs was also

significantly higher with older ages. This relationship did not consistently demonstrate a

gradient and was strongest for claimants receiving opioids with NSAIDs and/or SMRs.

Occupation involving heavy physical demands and/or use of machinery was associated with an

approximately 10% higher odds of receiving either drug class combination at first prescription,

while female sex and having one hospitalization in the prior year compared to none was

associated with a reduction in odds of receiving either combination.

Several factors, including some comorbidities (such as prior back/neck pain, headache

disorders, diabetes, heart and cerebrovascular disease), previous spine surgery, prior

antidepressant, anticonvulsant, and sedative-hypnotic/anxiolytic dispenses, and prior number

of pain-related specialist visits were found to be associated with a higher odds of receiving

opioids only compared to NSAIDs and/or SMRs only. A reduction in odds of receiving opioids

with NSAIDs and/or SMRs, but not opioids only, was found with increasing number of GP visits

and thyroid disorders. There was also a trend toward lower odds of receiving opioids only for

those with prior claims.

With some exceptions, claimants from health authorities other than Fraser had greater odds of

receiving opioid combinations compared to NSAIDs and/or SMRs only. The odds of receiving

opioids with NSAIDs and/or SMRs was also greater for claimants with injury years spanning

157

2002-2009 compared to 1998/1999, while the odds of receiving opioids only were reduced for

those with injury years 2000/2001 and 2008/2009.

Claimants with a female prescriber at first prescription were less likely to receive opioids only

(OR 0.57, 95%CI 0.51-0.65) and opioids with NSAIDs and/or SMRs (OR 0.80, 95%CI 0.73-0.89)

than NSAIDs and/or SMRs only when compared to male prescribers. Claimants whose first

prescriber was a pain-related specialist (OR 4.94, 95%CI 3.82-6.39), other specialist (OR 5.92,

95%CI 4.17-8.40), or a dental surgeon (OR 11.3, 95%CI 5.29-24.06) had significantly higher odds

of receiving opioids only than NSAIDs and/or SMRs only when compared to claimants with GPs

as first prescriber. Claimants with a pain-related specialist as first prescriber also had greater

odds of receiving opioids with NSAIDs and/or SMRs (OR 1.77, 95%CI 1.37-2.29). Claimants with

younger prescribers also had significantly reduced odds of receiving opioids only.

6.3.5 Sensitivity Analyses Examining Prior Supplementary Health Care Visits

When looking at claimants with injuries in 1998-2001 (Appendix D, Table D-1), claimants with

any chiropractor visits in the prior year and ≥6 massage therapy visits had lower odds of

receiving all drug combinations compared to no drugs in the first eight weeks. Claimants with

physiotherapy visits had lower odds of receiving opioids only and opioids with NSAIDs and/or

SMRs than no drugs.

For drug class(es) at first dispense (Appendix D, Table D-2), there was no significant association

with the number of chiropractor or massage therapy visits. Claimants with ≥6 billed

physiotherapy visits had significantly lower odds of receiving opioids only and opioids with

NSAIDs and/or SMRs at first prescription than NSAIDs and/or SMRs only.

158

6.4 Discussion

In the first eight weeks after injury, almost 60% of claimants received at least one prescription

dispense for an opioid, NSAID, or SMR. Most claimants received prescription NSAID and/or SMR

dispenses in this early period, as well as at first dispense. The factors found to be consistently

associated with both dispensing outcomes were prior cumulative days supplied with one of the

three drug classes, prior dispenses for antidepressants, anticonvulsants, and sedative-

hypnotic/anxiolytics, ICD-9 diagnosis on the claim, age at injury, pre-existing comorbidities,

prior physician visits and hospitalizations, year of injury, sex, health authority, occupation, and

prior workers’ compensation claims. Prescriber sex, specialty, and birth year were associated

with drug class at first post-injury dispense, but did little to account for between prescriber

variation.

In this sample, 27.7%, 40.7%, and 24.8% of claimants were dispensed at least one opioid,

NSAID, or SMR, respectively, in the first eight weeks after injury. Early opioid dispensing

prevalence in previous studies of LBP claimants has ranged from five to 35% using insurer

reimbursement data.(18-23) Our estimate of early opioid prevalence appears to fall somewhere

in between. Only one study in Alberta, Canada reported on the prevalence of early NSAID/SMR

prescriptions, finding approximately 5.3% received a prescription within two weeks of LBP

claim.(18) Differences in estimates likely reflect the various definitions of various time periods

used to define “early”, geographic and system-level variation in prescribing patterns, and

differences in data source.

We have also considered how claimants take combinations of these drug classes in the early

weeks after injury. Most were dispensed either NSAIDs and/or SMRs only (30.5%) or opioids

159

with NSAIDS and/or SMRs (19.6%). A total of 8.1% of the overall cohort received only opioids.

At first post-injury dispense, we found one fifth of claimants received only opioids and an

additional one fifth received opioids with NSAIDs and/or SMRs at the first post-injury dispense.

This pattern of dispensing is seemingly in contradiction to clinical guidelines that generally

recommend against opioids as first line treatment for acute low back pain,(15, 16) making it

important to understand the determinants of these patterns.

Our study advances understanding of how pre-injury factors can influence the drug classes

received in the early weeks after a work-related injury. Indeed, one of the factors associated

with early dispensing in this sample was a recent prior history of dispensing for drugs from one

of the three drug classes. Given we could not attribute prescriptions to their indication, it is

possible that some of this is a result of individuals receiving ongoing prescriptions for a pre-

existing and treated comorbidity. However, our previous analyses suggest that, for most

claimants, dispensing in the early weeks after injury is relatively new and generally does not

reflect persistent dispensing.(25) Our results are also consistent with those of previous studies

of non-compensation samples demonstrating a relationship between prior medication history

and prescribing at the first LBP visit.(17, 48) Our findings suggest that a claimant’s previous

history with these medications, which may include prior effectiveness and experience with

adverse events, is likely considered by the physician in the decision to prescribe these drugs in

the early post-injury period.

ICD-9 diagnosis, namely intervertebral disc disorder (ICD-9 722), was also associated with early

dispensing. In our cohort, 98.6% of claimants with an ICD-9 722 had the more specific code of

722.1, which refers to a displacement of the intervertebral disc without myelopathy, a diagnosis

160

classified as more severe.(49, 50) While ICD-9 diagnosis is only a surrogate and imperfect

measure of severity, results suggest claimants with potentially more severe injuries (or greater

perceived severity) were more likely to receive prescriptions, including stronger prescriptions.

Stover et al. also found severity of injury to be associated with an increased likelihood of

receiving early opioids after a compensated LBP injury.(23) Interestingly, studies of non-

compensation populations have been conflicting, with some finding characteristics, such as pain

intensity, radiating pain, and functional status to be associated with prescribing (51-56) and

others demonstrating no relationship.(48, 57, 58) Unfortunately, we did not have detailed data

on injury severity or pain symptoms.

Previous studies have also been conflicting about the importance of age in drug prescribing for

LBP or other non-cancer pain.(1, 17, 18, 23, 27, 51, 53, 57-61) The results of our study suggest

older claimants experience a greater degree of overall dispensing and dispensing for stronger

drug combinations than younger claimants. These findings may reflect the fact that pain has

been shown to increase with age (62, 63) and may also be partially due to dispensing for other

pain indications, as increasing age is associated with having a greater number of

comorbidities.(64) However, these results persisted after adjusting for several pre-existing

conditions, as well as prior health care utilization. Further, the relationship with age was not

always linear. In general, the strongest effect sizes were seen among claimants aged 35-54

years. The reasons for this pattern are unclear. This particular age group may face more

competing pressures between home and work life, as well as more financial and career strain,

and therefore may require stronger medications to manage their pain and maintain their

161

responsibilities. This may also represent a healthy worker effect, whereby older workers in

poorer health may no longer be in the labour force.

Other claimant-level factors associated with early dispensing included various pre-existing

comorbidities, increasing frequency of outpatient GP and pain-related specialist visits in the

prior year, and previous prescription dispenses for other medications that can be used as

adjuvants in LBP management. Together, these factors may be a constellation of indicators

representative of poor general health, which is consistent with some studies examining opioids

that have found individuals with various comorbidities to be more likely to receive opioids.(61,

65)

Surprisingly, injuries occurring in more recent years were mainly associated with an increased

likelihood of receiving NSAIDS and/or SMRs only and opioids with NSAIDs and/or SMRs rather

than opioids only. Other research has demonstrated an increase in opioid use among

compensation claimants in this time period (66-70), but did not differentiate between opioids

only versus opioids provided in combination with other non-opioid drugs. Our findings may

reflect WorkSafeBC’s policy initiatives during this time aimed at curbing opioid prescribing

during this time.(28)

Claimants with occupations involving heavy physical demands and/or use of machinery were

also more likely to receive any of the drug combinations compared to no drug, and to receive

opioid combinations at first dispense compared to NSAIDs and/or SMRs. Occupation has not

been previously examined as a potential determinant of dispensing, but these findings are

noteworthy. Claimants who return to a job after injury that is more physically taxing may

162

require stronger medication to manage any residual symptoms and continue to function in the

workplace. Of concern, however, is that the sedating effects of opioids and SMRs may pose a

safety risk for claimants with these occupations. This observation merits greater attention in

future research.

Women were generally less likely to receive any of the drug combinations or to receive

combinations involving opioids at first dispense. Previous findings, mostly in patient

populations, concerning the role of sex in drug dispensing for LBP have been conflicting, with

most demonstrating no relationship.(1, 23, 53, 57-59, 61) However, some studies found women

were more likely to receive opioids or SMRs for LBP,(27, 51, 65) which conflicts with our

findings. There was also variation by health authority, consistent with the geographic variability

seen in other studies, namely for opioids.(1, 24, 52, 54, 59, 71)

After adjustment, claimants with increasing prior workers’ compensation claims, pre-existing

back and/or neck pain, and increasing number of hospitalizations were less likely to be

dispensed all drug combinations compared to no drugs. In examining this further, we found

these surprising relationships were seen after adjusting mainly for prior drug history. One

potential reason may be that workers who have had prior related pain and claims have

identified other methods of managing pain and no longer rely on these medications as a coping

strategy. Interestingly, we also saw that claimants with prior compensation claims were at

reduced odds of receiving opioids only compared to NSAIDS and/or SMRs only at first dispense.

During the study period, WorkSafeBC policy imposed time limits on the reimbursement of

opioids.(28) It is unclear whether workers with prior recent claims may have experienced

163

barriers to further opioid reimbursement for the current injury given a past history of

dispensing. These observations require confirmation in other studies.

In general, there appeared to be a protective effect of prior visits to supplementary care

providers, although these associations were not consistently seen. Previous studies have shown

a protective relationship between chiropractic visits and opioid use, but an inverse relationship

for physiotherapy.(27, 61) Our analysis was limited to a small subset of our cohort based on

earlier injury years when MSP provided limited benefits for these practitioners. Some

misclassification is possible as we were not able to capture visits paid through other means.

Still, our results suggest workers with a history of seeking alternative health care may be less

likely to fill early prescriptions after injury.

Unlike prior studies of compensation samples, we were able to link claimant and prescriber

data for the first prescription dispensed. Approximately 36% and 25% of the variability in the

odds of receiving opioids only and the odds of receiving opioids with NSAIDs and/or SMRs,

respectively, as a first prescription was accounted for by the prescribers. A recent study of

emergency department physicians also found significant variation in the rates of opioid

prescribing among physicians practicing within the same department.(72) Little is known,

however, about the characteristics of prescribers that may influence opioid prescribing in pain

samples. Most studies have focused on prescriber specialty with conflicting findings.(73-75) In

our study, prescriber sex, birth year, and specialty were found to be significantly associated

with the first drug class dispensed. In particular, claimants with a first dispense prescribed by a

female practitioner or a younger practitioner were less likely to receive a first dispense that

included opioids compared to claimants with male and older practitioners. Claimants with pain-

164

related specialists as a first prescriber were more likely to receive a first dispense involving

opioids compared to GPs, while those whose first prescribers were other specialists or dental

surgeons were more likely to receive opioids only.

Despite these findings, adding these factors to the model did little to explain between

prescriber variation. Further research is needed to better understand the prescriber factors that

act as determinants of early post-injury prescribing in this population. Other factors, such as

prescriber comfort, satisfaction, and prior experience with prescribing, as well as prescriber

perceptions about risks have been previously shown to influence prescribing behaviour,(76) but

were not available in our study. In the context of a workers’ compensation claimant, there may

also be unique system-level and workplace factors that play a role in a physician’s decision-

making process, such as the desires of the insurer regarding treatment, an employer’s response

to the injury, and availability of work accommodations. The contribution of these factors has

yet to be examined.

Our study has a number of strengths. Our cohort was comprised of a large sample of

compensation claimants with injuries spanning over a 12-year period. We have considered a

wide variety of factors as potential determinants of early post-injury dispensing and, unlike

previous studies, have accounted for pre-injury health care utilization and drug dispensing,

which turned out to be associated with post-injury dispensing. We have looked beyond opioids

and considered other drugs commonly used for LBP. We have also demonstrated the

importance of the prescriber and basic prescriber factors in the first drug class received after

injury. Finally, our analyses were longitudinal, where only factors occurring prior to the injury

date or at the time of injury were considered.

165

Our study has additional limitations not yet described. Importantly, due to our use of

administrative data, we did not have information on individual-level factors previously

demonstrated as determinants of dispensing of one of the three drug classes, such as race.(23,

59, 60, 71, 73) Importantly, we lacked data on measures of severity that may be considered by

practitioners in the prescribing of these drug classes, including pain intensity, functional

limitations, and distress. Our data, while comprehensive in capturing dispensed prescriptions,

do not provide information on the pattern of consumption. However, the findings still speak to

the potential factors associated with clinical practice (with the caveat that prescriptions not

filled will also not have been captured in our data). It is also unclear whether our findings may

be generalized to injured workers who do not seek compensation or those who do not

experience lost time. Finally, given the prevalence of our outcomes, our odds ratio estimates

are likely an overestimate of the magnitude of risk.

6.5 Conclusion

In this study, we have demonstrated the multifactorial nature of dispensing in the early stages

after a compensated LBP injury. Many factors identified, including prior dispensing and health

care utilization patterns, diagnosis, age, pre-existing comorbidities, and occupation are also

factors that could, in theory, influence recovery trajectories post-injury. This is important

because previous studies demonstrating negative effects of early opioid dispensing among

injured workers have not accounted for many of these factors. Prescriber variation is also

important in accounting for differences in dispensing. Ultimately, the choice of medication

prescribed is likely the result of a complex set of personal and provider characteristics, claimant

166

and prescriber preferences, and system- and work-related factors. There is a need for further

research to understand the contribution of these factors to early prescribing in this population.

167

6.6 References

1. Luo X, Pietrobon R, Hey L. Patterns and trends in opioid use among individuals with back pain in the United States. Spine (Phila Pa 1976). 2004;29(8):884-90; discussion 91.

2. Mafi JN, McCarthy EP, Davis RB, Landon BE. Worsening trends in the management and treatment of back pain. JAMA Internal Medicine. 2013;173(17):1573-81.

3. Martin BI, Turner JA, Mirza SK, Lee MJ, Comstock BA, Deyo RA. Trends in health care expenditures, utilization, and health status among US adults with spine problems, 1997-2006. Spine (Phila Pa 1976). 2009;34(19):2077-84.


5. Salt E, Gokun Y, Rankin Kerr A, Talbert J. A Description and Comparison of Treatments for Low Back Pain in the United States. Orthop Nurs. 2016;35(4):214-21.








168












169

24. Webster BS, Cifuentes M, Verma S, Pransky G. Geographic variation in opioid prescribing for acute, work-related, low back pain and associated factors: a multilevel analysis. American Journal of Industrial Medicine52(2):162-71. 2009.

25. Carnide N, Hogg-Johnson S, Furlan AD, Côté P, Koehoorn M. Prescription dispensing patterns before and after a workers' compensation claim: An historical cohort study of workers with low back pain injuries in British Columbia. In preparation. 2017.











36. WorkSafeBC. Who does and doesn't need coverage? : 2017; Available from: https://www.worksafebc.com/en/insurance/need-coverage/who-needs-coverage. Last Accessed: November 23, 2016.



170




40. Ministry of Health. Health Authorities. Government of British Columbia; Available from: http://www2.gov.bc.ca/gov/content/health/about-bc-s-health-care-system/partners/health-authorities. Last Accessed: September 20, 2017.

41. Statistics Canada. Standard Occupational Classification (SOC) 1991. Ottawa, ON: Statistics Canada; 2016; Available from: http://www23.statcan.gc.ca/imdb/p3VD.pl?Function=getVD&TVD=143375. Last Accessed: September 9, 2016.

42. Canada HRD. National Occupational Classification Career Handbook. Ottawa, ON: Government of Canada; 2011.




46. Snijders TAB, Bosker RJ. Multilevel Analysis: An Introduction to Basic and Advanced Multilevel Modeling. Thousand Oaks, CA: Sage; 1999.

47. Ene M, Leighton EA, Blue GL, Bell BA. Multilevel Models for Categorical Data using SAS PROC GLIMMIX: The Basics. Paper 3430-2015. 2015; Available from: https://support.sas.com/resources/papers/proceedings15/3430-2015.pdf. Last Accessed: April 27, 2016.

48. Muller S, Bedson J, Mallen CD. The association between pain intensity and the prescription of analgesics and non-steroidal anti-inflammatory drugs. European journal of pain. 2012;16(7):1014-20.



http://www23.statcan.gc.ca/imdb/p3VD.pl?Function=getVD&TVD=143375

https://support.sas.com/resources/papers/proceedings15/3430-2015.pdf

171

49. Dasinger LK, Krause N, Deegan LJ, Brand RJ, Rudolph L. Physical workplace factors and return to work after compensated low back injury: a disability phase-specific analysis. J Occup Environ Med. 2000;42(3):323-33.

50. Krause N, Rugulies R, Ragland DR, Syme SL. Physical workload, ergonomic problems, and incidence of low back injury: a 7.5-year prospective study of San Francisco transit operators. Am J Ind Med. 2004;46(6):570-85.

51. Bernstein E, Carey TS, Garrett JM. The use of muscle relaxant medications in acute low back pain. Spine (Phila Pa 1976). 2004;29(12):1346-51.

52. Chang HY, Daubresse M, Kruszewski SP, Alexander GC. Prevalence and treatment of pain in EDs in the United States, 2000 to 2010. Am J Emerg Med. 2014;32(5):421-31.

53. Cherkin DC, Wheeler KJ, Barlow W, Deyo RA. Medication use for low back pain in primary care. Spine (Phila Pa 1976). 1998;23(5):607-14.

54. Dobscha SK, Morasco BJ, Duckart JP, Macey T, Deyo RA. Correlates of prescription opioid initiation and long-term opioid use in veterans with persistent pain. Clin J Pain. 2013;29(2):102-8.

55. Morasco BJ, Duckart JP, Carr TP, Deyo RA, Dobscha SK. Clinical characteristics of veterans prescribed high doses of opioid medications for chronic non-cancer pain. Pain. 2010;151(3):625-32.

56. Sullivan MD, Edlund MJ, Zhang L, Unutzer J, Wells KB. Association between mental health disorders, problem drug use, and regular prescription opioid use. Arch Intern Med. 2006;166(19):2087-93.

57. Breckenridge J, Clark JD. Patient characteristics associated with opioid versus nonsteroidal anti-inflammatory drug management of chronic low back pain. Journal of Pain4(6):344-50. 2003.

58. Krebs EE, Lurie JD, Fanciullo G, Tosteson TD, Blood EA, Carey TS, et al. Predictors of long-term opioid use among patients with painful lumbar spine conditions. J Pain. 2010;11(1):44-52.

59. Luo X, Pietrobon R, Curtis LH, Hey LA. Prescription of nonsteroidal anti-inflammatory drugs and muscle relaxants for back pain in the United States. Spine (Phila Pa 1976). 2004;29(23):E531-7.

60. Prunuske JP, St Hill CA, Hager KD, Lemieux AM, Swanoski MT, Anderson GW, et al. Opioid prescribing patterns for non-malignant chronic pain for rural versus non-rural US adults: a population-based study using 2010 NAMCS data. BMC Health Serv Res. 2014;14:563.

172

61. Rhee Y, Taitel MS, Walker DR, Lau DT. Narcotic drug use among patients with lower back pain in employer health plans: a retrospective analysis of risk factors and health care services. Clin Ther. 2007;29 Suppl:2603-12.

62. Langley PC. The prevalence, correlates and treatment of pain in the European Union. Curr Med Res Opin. 2011;27(2):463-80.

63. Tsang A, Von Korff M, Lee S, Alonso J, Karam E, Angermeyer MC, et al. Common chronic pain conditions in developed and developing countries: gender and age differences and comorbidity with depression-anxiety disorders. J Pain. 2008;9(10):883-91.

64. Gerteis J, Izrael D, Deitz D, LeRoy L, Ricciardi R, Miller T, et al. Multiple Chronic Conditions Chartbook. Rockville, MD: Agency for Healthcare Research and Quality; 2014.

65. Deyo RA, Smith DH, Johnson ES, Donovan M, Tillotson CJ, Yang X, et al. Opioids for back pain patients: primary care prescribing patterns and use of services. J Am Board Fam Med. 2011;24(6):717-27.


67. Bernacki EJ, Yuspeh L, Lavin R, Tao XG. Increases in the use and cost of opioids to treat acute and chronic pain in injured workers, 1999 to 2009. J Occup Environ Med. 2012;54(2):216-23.


69. Ireland J, Young B, Swedlow A. Changes in Schedule II & Schedule III Opioid Prescriptions and Payments in California Workers' Compensation. Oakland, California: California Workers' Compensation Institute; 2012.


71. Olsen Y, Daumit GL, Ford DE. Opioid prescriptions by U.S. primary care physicians from 1992 to 2001. J Pain. 2006;7(4):225-35.

72. Barnett ML, Olenski AR, Jena AB. Opioid-Prescribing Patterns of Emergency Physicians and Risk of Long-Term Use. N Engl J Med. 2017;376(7):663-73.

73. Ringwalt C, Gugelmann H, Garrettson M, Dasgupta N, Chung AE, Proescholdbell SK, et al. Differential prescribing of opioid analgesics according to physician specialty for Medicaid patients with chronic noncancer pain diagnoses. Pain Res Manag. 2014;19(4):179-85.

173

74. Mosher HJ, Krebs EE, Carrel M, Kaboli PJ, Weg MW, Lund BC. Trends in prevalent and incident opioid receipt: an observational study in Veterans Health Administration 2004-2012. J Gen Intern Med. 2015;30(5):597-604.

75. Volkow ND, McLellan TA, Cotto JH, Karithanom M, Weiss SR. Characteristics of opioid prescriptions in 2009. JAMA. 2011;305(13):1299-301.

76. Wenghofer EF, Wilson L, Kahan M, Sheehan C, Srivastava A, Rubin A, et al. Survey of Ontario primary care physicians' experiences with opioid prescribing. Can Fam Physician. 2011;57(3):324-32.

174

Table 6-1 Baseline characteristics of the overall cohort of workers’ compensation claimants with short-term disability claims for injuries to the low back between 1998 and 2009 (n=95,413)

Variable n ( % )

Sex

Women 36,023 (37.8) Men 59,390 (62.2)

Age at injury

18-24 10,016 (10.5) 25-34 21,080 (22.1) 35-44 29,368 (30.8) 45-54 24,965 (26.2) ≥55 9,984 (10.5)

Neighbourhood income quintile

1 (lowest) 21,155 (22.2) 2 22,276 (23.3) 3 20,853 (21.9) 4 18,214 (19.1) 5 (highest) 12,915 (13.5)

Regional health authority


Year of injury

1998/99 18,041 (18.9) 2000/01 16,616 (17.4) 2002/03 15,568 (16.3) 2004/05 15,500 (16.2) 2006/07 15,612 (16.4) 2008/09 14,076 (14.8)

ICD-9 diagnosis on claim

722 – Intervertebral disc disorders 2,877 ( 3.0) 724 – Other and unspecified disorders of back 9,519 (10.0) 846 – Sprains and strains of sacroiliac region 20,638 (21.6) 847 – Sprains and strains lumbar, sacrum, coccyx, unspecified site of back 62,379 (65.4)

Occupation at time of injury involving heavy physical strength requirements (handle loads ˃20kg) and/or working near or with equipment, instruments, machinery, or power/hand tools that may be a potential source of accident or injury

No 42,164 (44.2) Yes 53,249 (55.8)

Number of prior workers’ compensation claims in the 2 years before injury1

0 64,301 (67.4) 1 21,610 (22.6) ≥2 9,502 (10.0)

175

Variable n ( % ) Comorbidities in the two years before injury

Back and/or neck pain 31,638 (33.2) Arthritis/rheumatism and other joint pain 16,696 (17.5) Neuropathic pain 1,118 ( 1.2) Headache disorders 7,687 ( 8.1) Mood and anxiety disorders 15,825 (16.6) Suicide attempt/self-harm history 165 ( 0.2) Substance use disorder 1,297 ( 1.4) Diabetes 3,027 ( 3.2) Hypertension 5,929 ( 6.2) Heart and cerebrovascular disease 2,317 ( 2.4) Chronic respiratory disorders 2,934 ( 3.1) Chronic liver disease 711 ( 0.7) Chronic kidney disease 156 ( 0.2) Thyroid disease 1,282 ( 1.3) Upper gastrointestinal disorders 3,034 ( 3.2) Inflammatory bowel disease 395 ( 0.4) Insomnia and sleep apnea 512 ( 0.5) Seizure disorders 308 ( 0.3) Traumatic brain injury 527 ( 0.6)

Number of hospitalizations in the year before injury

0 88,084 (92.3) 1 6,300 ( 6.6) ≥2 1,029 ( 1.1)

Number of general practitioner visits in the year before injury

0 10,974 (11.5) 1-2 22,179 (23.2) 3-4 18,656 (19.6) 5-7 19,280 (20.2) ≥8 24,324 (25.5)

Number of pain-related physician specialist visits in the year before injury

0 79,069 (82.9) 1 9,278 ( 9.7) 2 3,577 ( 3.7) ≥3 3,489 ( 3.7)

Number of other physician specialist visits in the year before injury

0 73,992 (77.5) 1 11,799 ( 12.4) 2 4,621 ( 4.8) ≥3 5,001 ( 5.2)

Spine surgery in the two years before injury

No 95,202 (99.8) Yes 211 ( 0.2)

176

Variable n ( % ) At least one spine x-ray in the year before injury

No 88,751 (93.0) Yes 6,662 ( 7.0)

Cumulative days supplied with opioids in the year before injury

0 76,597 (80.3) 1-14 12,978 (13.6) ≥15 5,838 ( 6.1)

Cumulative days supplied with NSAIDs in the year before injury

0 74,729 (78.3) 1-14 8,561 ( 9.0) ≥15 12,123 (12.7)

Cumulative days supplied with SMRs in the year before injury

0 89,306 (93.6) 1-14 3,422 ( 3.6) ≥15 2,685 ( 2.8)

At least one antidepressant dispense in the year before injury

No 90,573 (94.9) Yes 4,840 ( 5.1)

At least one anticonvulsant dispense in the year before injury

No 92,920 (97.4) Yes 2,493 ( 2.6)

At least one sedative hypnotic/anxiolytic dispense in the year before injury

No 86,832 (91.0) Yes 8,581 ( 9.0)

Abbreviations: ICD-9=International Statistical Classification of Diseases and Related Health Problems, 9th Revision; NSAIDs=nonsteroidal anti-inflammatory drugs; SMRs=skeletal muscle relaxants 1Includes all types of claims (short-term disability, long-term disability, health care only, vocational rehabilitation) for any diagnosis (not limited to low back pain claims).

177

Table 6-2 Multinomial logistic regression analysis of factors associated with drug class(es) dispensed within eight weeks of injury (n=95,413) (no drugs as the reference, n=39,842)

Characteristic n (%)

Crude OR (95% CI) Multivariable OR (95% CI)

NSAIDs and/or SMRs only (n=29,104)

Opioids Only (n=7,730)

Opioids with NSAIDs and/or

SMRs (n=18,737)




SMRs (n=18,737)

Sex

Men 59,390 (62.2) 1.00*** 1.00 1.00 1.00*** 1.00 1.00 Women 36,023 (37.8) 1.00 (0.97-1.03) 0.89 (0.85-0.94) 0.82 (0.79-0.85) 1.00 (0.97-1.04) 0.83 (0.78-0.88) 0.81 (0.77-0.84)

Age at Injury

18-24 10,016 (10.5) 1.00*** 1.00 1.00 1.00*** 1.00 1.00 25-34 21,080 (22.1) 1.23 (1.16-1.30) 1.41 (1.27-1.55) 1.76 (1.64-1.89) 1.23 (1.16-1.30) 1.33 (1.20-1.47) 1.70 (1.58-1.84) 35-44 29,368 (30.8) 1.41 (1.33-1.48) 1.81 (1.65-1.99) 2.20 (2.05-2.35) 1.40 (1.33-1.48) 1.63 (1.48-1.79) 2.05 (1.91-2.21) 45-54 24,965 (26.2) 1.32 (1.25-1.39) 1.86 (1.69-2.05) 2.20 (2.05-2.36) 1.29 (1.22-1.37) 1.63 (1.48-1.80) 1.98 (1.84-2.13) ≥55 9,984 (10.5) 1.30 (1.22-1.39) 2.03 (1.82-2.27) 1.96 (1.80-2.12) 1.25 (1.16-1.34) 1.69 (1.50-1.89) 1.66 (1.52-1.82)

Neighbourhood Income quintile

1 (lowest) 21,155 (22.2) 1.00*** 1.00 1.00 1.00*** 1.00 1.00 2 22,276 (23.3) 1.04 (1.00-1.09) 0.96 (0.90-1.03) 1.02 (0.97-1.08) 1.04 (0.99-1.08) 0.98 (0.91-1.05) 1.02 (0.97-1.08) 3 20,853 (21.9) 0.92 (0.88-0.97) 0.87 (0.81-0.94) 0.90 (0.86-0.95) 0.92 (0.88-0.96) 0.90 (0.83-0.97) 0.91 (0.86-0.96) 4 18,214 (19.1) 0.89 (0.85-0.93) 0.82 (0.76-0.88) 0.85 (0.81-0.90) 0.89 (0.85-0.94) 0.86 (0.79-0.92) 0.87 (0.82-0.92) 5 (highest) 12,915 (13.5) 0.87 (0.83-0.92) 0.80 (0.73-0.87) 0.84 (0.79-0.89) 0.90 (0.85-0.95) 0.83 (0.76-0.91) 0.87 (0.82-0.93)


Fraser 40,865 (42.8) 1.00*** 1.00 1.00 1.00*** 1.00 1.00 Vancouver Coastal 18,574 (19.5) 0.85 (0.81-0.88) 1.04 (0.97-1.11) 0.74 (0.70-0.78) 0.84 (0.81-0.88) 1.06 (0.98-1.13) 0.76 (0.72-0.80) Interior 13,735 (14.4) 0.86 (0.82-0.90) 1.31 (1.22-1.41) 1.11 (1.05-1.17) 0.87 (0.82-0.91) 1.30 (1.20-1.40) 1.10 (1.04-1.16) Vancouver Island 17,156 (18.0) 0.63 (0.61-0.66) 1.06 (0.99-1.14) 0.77 (0.74-0.81) 0.66 (0.63-0.69) 1.10 (1.02-1.17) 0.81 (0.77-0.86) Northern 5,083 (5.3) 1.09 (1.01-1.17) 1.37 (1.23-1.54) 1.42 (1.31-1.53) 1.11 (1.04-1.20) 1.41 (1.25-1.58) 1.45 (1.34-1.57)

178






SMRs (n=18,737)




SMRs (n=18,737)

Year of Injury 1998/1999 18,041 (18.9) 1.00*** 1.00 1.00 1.00*** 1.00 1.00 2000/2001 16,616 (17.4) 1.11 (1.06-1.17) 0.99 (0.92-1.07) 1.12 (1.05-1.18) 1.09 (1.02-1.16) 0.98 (0.91-1.07) 1.09 (1.02-1.16) 2002/2003 15,568 (16.3) 1.09 (1.04-1.15) 0.96 (0.88-1.04) 1.25 (1.18-1.33) 1.05 (1.00-1.11) 0.97 (0.89-1.05) 1.24 (1.16-1.32) 2004/2005 15,500 (16.2) 1.19 (1.13-1.25) 1.06 (0.97-1.14) 1.33 (1.25-1.41) 1.11 (1.06-1.17) 1.06 (0.98-1.16) 1.29 (1.21-1.37) 2006/2007 15,612 (16.4) 1.23 (1.17-1.29) 1.08 (1.00-1.17) 1.46 (1.38-1.55) 1.15 (1.09-1.21) 1.08 (0.99-1.18) 1.40 (1.31-1.49) 2008/2009 14,076 (14.8) 1.35 (1.28-1.43) 1.01 (0.93-1.10) 1.63 (1.54-1.74) 1.27 (1.20-1.34) 0.98 (0.90-1.08) 1.53 (1.44-1.64)


722 - Intervertebral disc disorders 2,877 (3.0) 1.81 (1.61-2.03) 3.51 (3.06-4.04) 5.30 (4.78-5.88) 1.82 (1.62-2.05) 3.28 (2.85-3.78) 5.04 (4.52-5.60)

724 - Other and unspecified disorders of back 9,519 (10.0) 0.78 (0.74-0.82) 1.04 (0.97-1.13) 0.81 (0.76-0.86) 0.81 (0.77-0.85) 1.03 (0.95-1.12) 0.83 (0.78-0.88)

846 – Sprains and strains of sacroiliac region 20,638 (21.6) 0.95 (0.91-0.98) 0.93 (0.87-0.98) 0.89 (0.85-0.93) 0.96 (0.93-1.00) 0.93 (0.87-0.99) 0.92 (0.88-0.96)

847 – Sprains and strains lumbar, sacrum, coccyx, unspecified site of back 62,379 (65.4) 1.00*** 1.00 1.00 1.00*** 1.00 1.00


No 42,164 (44.2) 1.00*** 1.00 1.00 1.00*** 1.00 1.00 Yes 53,249 (55.8) 1.11 (1.08-1.14) 1.17 (1.11-1.23) 1.31 (1.27-1.36) 1.12 (1.08-1.16) 1.16 (1.09-1.23) 1.24 (1.19-1.29)

179






SMRs (n=18,737)




SMRs (n=18,737)

Number of prior workers’ compensation claims in the 2 years before injury

0 64,301 (67.4) 1.00*** 1.00 1.00 1.00*** 1.00 1.00 1 21,610 (22.6) 0.93 (0.90-0.97) 0.99 (0.93-1.05) 0.97 (0.93-1.01) 0.88 (0.84-0.91) 0.87 (0.82-0.93) 0.83 (0.79-0.87) ≥2 9,502 (10.0) 0.89 (0.85-0.94) 1.01 (0.93-1.10) 1.03 (0.97-1.09) 0.81 (0.76-0.85) 0.79 (0.72-0.86) 0.77 (0.72-0.82)

Comorbidities in the 2 years before injury

Back and/or neck pain No 63,775 (66.8) 1.00*** 1.00 1.00 1.00*** 1.00 1.00 Yes 31,638 (33.2) 0.89 (0.86-0.92) 1.14 (1.09-1.20) 1.05 (1.01-1.09) 0.81 (0.78-0.84) 0.88 (0.83-0.94) 0.81 (0.77-0.84)

Arthritis/rheumatism No 78,717 (82.5) 1.00*** 1.00 1.00 1.00 1.00 1.00 Yes 16,696 (17.5) 1.17 (1.12-1.22) 1.42 (1.34-1.51) 1.42 (1.36-1.49) 1.03 (0.99-1.08) 1.06 (0.99-1.13) 1.05 (1.00-1.10)

Neuropathic pain No 94,295 (98.8) 1.00*** 1.00 1.00 1.00* 1.00 1.00 Yes 1,118 (1.2) 1.39 (1.20-1.62) 1.68 (1.36-2.09) 1.87 (1.60-2.19) 1.24 (1.06-1.45) 1.06 (0.84-1.33) 1.27 (1.08-1.50)

Headache disorders No 87,726 (91.9) 1.00*** 1.00 1.00 1.00*** 1.00 1.00 Yes 7,687 (8.1) 1.12 (1.06-1.18) 1.56 (1.43-1.69) 1.44 (1.35-1.53) 1.06 (1.00-1.12) 1.15 (1.06-1.26) 1.16 (1.08-1.24)

Mood & anxiety disorders No 79,588 (83.4) 1.00*** 1.00 1.00 1.00*** 1.00 1.00 Yes 15,825 (16.6) 0.98 (0.94-1.02) 1.31 (1.23-1.40) 1.18 (1.13-1.24) 0.90 (0.85-0.94) 0.99 (0.92-1.06) 0.94 (0.89-1.00)

Substance use disorders No 94,116 (98.6) 1.00*** 1.00 1.00 1.00 1.00 1.00 Yes 1,297 (1.4) 1.05 (0.92-1.21) 1.83 (1.53-2.19) 1.51 (1.31-1.74) 0.96 (0.83-1.11) 1.14 (0.93-1.38) 0.95 (0.81-1.12)

Prior self-inflicted harm or suicide attempt

No 95,248 (99.8) 1.00*** 1.00 1.00 1.00** 1.00 1.00 Yes 165 (0.2) 1.21 (0.81-1.81) 1.32 (0.72-2.42) 2.35 (1.61-3.43) 1.34 (0.89-2.02) 0.97 (0.51-1.84) 2.15 (1.41-3.27)

180






SMRs (n=18,737)




SMRs (n=18,737)

Diabetes No 92,386 (96.8) 1.00*** 1.00 1.00 1.00*** 1.00 1.00 Yes 3,027 (3.2) 1.17 (1.07-1.28) 1.74 (1.53-1.97) 1.64 (1.50-1.81) 1.05 (0.95-1.15) 1.34 (1.18-1.54) 1.32 (1.19-1.46)

Hypertension No 89,484 (93.8) 1.00*** 1.00 1.00 1.00 1.00 1.00 Yes 5,929 (6.2) 1.09 (1.02-1.16) 1.37 (1.25-1.51) 1.37 (1.28-1.47) 0.97 (0.90-1.04) 1.00 (0.90-1.11) 1.08 (1.00-1.17)

Heart & Cerebrovascular Disease

No 93,096 (97.6) 1.00*** 1.00 1.00 1.00 1.00 1.00 Yes 2,317 (2.4) 0.99 (0.89-1.10) 1.60 (1.39-1.84) 1.29 (1.16-1.44) 0.96 (0.86-1.07) 1.17 (1.01-1.36) 1.06 (0.94-1.19)

Upper GI disorders No 92,379 (96.8) 1.00*** 1.00 1.00 1.00*** 1.00 1.00 Yes 3,034 (3.2) 1.27 (1.16-1.39) 1.63 (1.43-1.85) 1.48 (1.35-1.63) 1.19 (1.08-1.31) 1.31 (1.15-1.50) 1.26 (1.13-1.39)

Inflammatory bowel disorders

No 95,018 (99.6) 1.00** 1.00 1.00 1.00 1.00 1.00 Yes 395 (0.4) 0.91 (0.71-1.17) 1.73 (1.27-2.38) 1.17 (0.90-1.52) 0.91 (0.71-1.18) 1.25 (0.90-1.75) 0.93 (0.70-1.23)

Chronic kidney disease No 95,257 (99.8) 1.00** 1.00 1.00 1.00* 1.00 1.00 Yes 156 (0.2) 0.96 (0.65-1.41) 2.05 (1.29-3.26) 0.81 (0.51-1.30) 0.98 (0.66-1.45) 1.39 (0.85-2.27) 0.61 (0.37-1.00)

Chronic liver disease No 94,702 (99.3) 1.00** 1.00 1.00 1.00 1.00 1.00 Yes 711 (0.7) 1.05 (0.87-1.26) 1.57 (1.22-2.01) 1.25 (1.02-1.52) 0.97 (0.80-1.17) 1.15 (0.89-1.50) 0.98 (0.79-1.21)

Thyroid disorders No 94,131 (98.7) 1.00* 1.00 1.00 1.00 1.00 1.00 Yes 1,282 (1.3) 1.19 (1.05-1.36) 0.97 (0.78-1.21) 1.08 (0.93-1.26) 1.11 (0.97-1.27) 0.86 (0.69-1.09) 1.00 (0.85-1.17)

Chronic respiratory disorders No 92,479 (96.9) 1.00*** 1.00 1.00 1.00* 1.00 1.00 Yes 2,934 (3.1) 1.12 (1.03-1.23) 1.46 (1.28-1.66) 1.25 (1.13-1.38) 1.09 (1.00-1.20) 1.20 (1.05-1.38) 1.12 (1.01-1.24)

Seizure disorders No 95,105 (99.7) 1.00* 1.00 1.00 1.00 1.00 1.00 Yes 308 (0.3) 0.83 (0.62-1.10) 1.53 (1.06-2.19) 1.04 (0.77-1.40) 0.75 (0.56-1.01) 0.98 (0.66-1.45) 0.73 (0.52-1.01)

181






SMRs (n=18,737)




SMRs (n=18,737)

Sleep apnea and insomnia No 94,901 (99.5) 1.00** 1.00 1.00 1.00 1.00 1.00 Yes 512 (0.5) 1.15 (0.93-1.43) 1.46 (1.07-1.99) 1.51 (1.20-1.89) 1.08 (0.86-1.34) 1.01 (0.73-1.39) 1.13 (0.89-1.43)

Traumatic brain injury and other head injuries

No 94,886 (99.4) 1.00 1.00 1.00 1.00 1.00 1.00 Yes 527 (0.6) 0.92 (0.75-1.14) 1.06 (0.76-1.46) 1.22 (0.98-1.52) 0.94 (0.76-1.17) 0.81 (0.58-1.14) 0.98 (0.77-1.24)


0 88,084 (92.3) 1.00*** 1.00 1.00 1.00*** 1.00 1.00 1 6,300 (6.6) 1.05 (0.99-1.12) 1.33 (1.21-1.46) 1.24 (1.16-1.32) 0.97 (0.91-1.04) 0.83 (0.75-0.92) 0.83 (0.77-0.90) ≥2 1,029 (1.1) 1.01 (0.87-1.19) 1.88 (1.54-2.30) 1.54 (1.31-1.81) 0.90 (0.76-1.07) 0.77 (0.61-0.96) 0.78 (0.65-0.94)

Number of GP visits in the year before injury

0 10,974 (11.5) 1.00*** 1.00 1.00 1.00 1.00 1.00 1-2 22,179 (23.2) 1.07 (1.01-1.13) 1.05 (0.95-1.15) 1.06 (1.00-1.13) 1.08 (1.02-1.14) 1.06 (0.96-1.16) 1.07 (1.00-1.14) 3-4 18,656 (19.6) 1.09 (1.03-1.15) 1.13 (1.03-1.24) 1.08 (1.01-1.15) 1.09 (1.02-1.15) 1.09 (0.99-1.20) 1.03 (0.96-1.11) 5-7 19,280 (20.2) 1.15 (1.08-1.21) 1.18 (1.07-1.30) 1.19 (1.11-1.27) 1.10 (1.04-1.17) 1.04 (0.94-1.15) 1.04 (0.96-1.11) ≥8 24,324 (25.5) 1.27 (1.21-1.35) 1.83 (1.68-2.00) 1.66 (1.56-1.76) 1.15 (1.08-1.23) 1.19 (1.07-1.33) 1.07 (0.99-1.16)

Number of pain-related specialist visits in the year before injury

0 79,069 (82.9) 1.00*** 1.00 1.00 1.00 1.00 1.00 1 9,278 (9.7) 0.97 (0.92-1.02) 1.27 (1.17-1.37) 1.15 (1.08-1.22) 0.89 (0.85-0.94) 1.03 (0.95-1.12) 0.95 (0.89-1.01) 2 3,577 (3.7) 1.00 (0.92-1.08) 1.43 (1.27-1.61) 1.30 (1.19-1.42) 0.88 (0.81-0.96) 0.99 (0.87-1.13) 0.94 (0.86-1.04) ≥3 3,534 (3.7) 1.06 (0.97-1.16) 1.93 (1.73-2.16) 1.49 (1.36-1.63) 0.90 (0.82-0.99) 1.14 (1.00-1.29) 0.94 (0.85-1.05)

182






SMRs (n=18,737)




SMRs (n=18,737)

Number of other specialist visits in the year before injury

0 73,992 (77.5) 1.00*** 1.00 1.00 1.00 1.00 1.00 1 11,799 (12.4) 1.04 (0.99-1.09) 1.09 (1.02-1.18) 1.05 (0.99-1.10) 1.00 (0.95-1.05) 0.96 (0.89-1.03) 0.95 (0.90-1.01) 2 4,621 (4.8) 1.08 (1.01-1.16) 1.10 (0.98-1.23) 1.10 (1.02-1.20) 1.02 (0.94-1.09) 0.91 (0.80-1.02) 0.97 (0.89-1.06) ≥3 5,001 (5.2) 1.05 (0.98-1.13) 1.31 (1.18-1.45) 1.25 (1.16-1.35) 0.98 (0.91-1.06) 0.98 (0.87-1.10) 1.02 (0.94-1.11)

Spine surgery in the 2 years before injury

No 95,202 (99.8) 1.00*** 1.00 1.00 1.00 1.00 1.00 Yes 211 (0.2) 1.20 (0.83-1.72) 3.37 (2.27-5.00) 1.76 (1.22-2.54) 1.08 (0.75-1.57) 1.51 (0.98-2.31) 0.75 (0.50-1.13)

At least one spine x-ray visit in the year before injury

No 88,751 (93.0) 1.00*** 1.00 1.00 1.00 1.00 1.00 Yes 6,662 (7.0) 1.19 (1.12-1.27) 1.43 (1.30-1.56) 1.55 (1.46-1.66) 1.01 (0.95-1.08) 1.00 (0.90-1.10) 1.02 (0.95-1.10)


0 76,597 (80.3) 1.00*** 1.00 1.00 1.00*** 1.00 1.00 1-14 12,978 (13.6) 1.14 (1.09-1.20) 1.77 (1.65-1.89) 1.91 (1.82-2.00) 1.03 (0.99-1.09) 1.74 (1.62-1.87) 1.74 (1.65-1.84) ≥15 5,838 (6.1) 1.28 (1.17-1.41) 8.38 (7.69-9.15) 6.74 (6.25-7.27) 1.00 (0.91-1.10) 6.85 (6.22-7.55) 4.91 (4.52-5.34)


0 74,729 (78.3) 1.00*** 1.00 1.00 1.00*** 1.00 1.00 1-14 8,561 (9.0) 1.54 (1.46-1.63) 1.31 (1.20-1.43) 1.63 (1.53-1.73) 1.52 (1.44-1.61) 1.08 (0.99-1.18) 1.38 (1.29-1.47) ≥15 12,123 (12.7) 2.09 (1.99-2.19) 1.78 (1.65-1.92) 2.85 (2.70-3.00) 2.02 (1.91-2.13) 1.11 (1.02-1.21) 1.97 (1.86-2.09)

183






SMRs (n=18,737)




SMRs (n=18,737)


0 89,306 (93.6) 1.00*** 1.00 1.00 1.00*** 1.00 1.00 1-14 3,422 (3.6) 1.44 (1.32-1.57) 1.41 (1.23-1.61) 2.09 (1.91-2.28) 1.23 (1.13-1.35) 1.00 (0.87-1.16) 1.42 (1.29-1.57) ≥15 2,685 (2.8) 1.91 (1.72-2.12) 2.33 (2.02-2.70) 3.48 (3.14-3.86) 1.47 (1.32-1.65) 1.10 (0.94-1.29) 1.62 (1.45-1.82)


No 90,573 (94.9) 1.00*** 1.00 1.00 1.00*** 1.00 1.00 Yes 4,840 (5.1) 1.25 (1.16-1.34) 2.03 (1.84-2.25) 1.91 (1.77-2.05) 1.13 (1.05-1.23) 1.27 (1.14-1.41) 1.23 (1.13-1.34)


No 92,920 (97.4) 1.00*** 1.00 1.00 1.00*** 1.00 1.00 Yes 2,493 (2.6) 1.26 (1.13-1.40) 2.31 (2.02-2.63) 2.20 (1.98-2.44) 1.15 (1.03-1.29) 1.33 (1.15-1.55) 1.35 (1.20-1.52)


No 86,832 (91.0) 1.00*** 1.00 1.00 1.00*** 1.00 1.00 Yes 8,581 (9.0) 1.16 (1.10-1.23) 2.07 (1.92-2.23) 1.86 (1.76-1.97) 1.08 (1.01-1.15) 1.31 (1.20-1.43) 1.28 (1.20-1.37)

Abbreviations: CI=confidence interval; GP=general practitioner; ICD-9=International Classification of Diseases 9th Revision; NSAIDs=nonsteroidal anti-inflammatory drugs; OR=odds ratio; SMRs=skeletal muscle relaxants

*Significant at p<0.05 **Significant at p<0.01 ***Significant at p<0.001. Refers to significance of variable overall.

184

Table 6-3 Multilevel multinomial logistic regression analysis of factors associated with drug class(es) at first prescription within 8 weeks of injury (n=54,197) (NSAIDs and/or SMRs only as the reference, n=32,846)


Crude OR (95% CI) Model 2 OR (95% CI) Model 3 OR (95% CI)



SMRs (n=10,634) Opioids Only (n=10,717)


SMRs (n=10,634)



SMRs (n=10,634)

Fixed Effects Intercept (SE) -1.32 (0.03)***, 1 -1.39 (0.02)***, 1 -1.83 (0.09)*** -2.01 (0.08)*** -1.82 (0.11)*** -2.03 (0.10)*** Sex

Men 34,177 (63.1) 1.00*** 1.00 1.00*** 1.00 1.00*** 1.00 Women 20,020 (36.9) 0.94 (0.89-0.98) 0.68 (0.65-0.71) 0.89 (0.84-0.95) 0.68 (0.64-0.73) 0.92 (0.86-0.98) 0.69 (0.65-0.74)

Age at Injury

18-24 4,709 (8.7) 1.00*** 1.00 1.00*** 1.00 1.00*** 1.00 25-34 11,578 (21.4) 1.17 (1.07-1.29) 1.35 (1.23-1.49) 1.08 (0.97-1.21) 1.42 (1.28-1.58) 1.09 (0.98-1.22) 1.42 (1.28-1.57) 35-44 17,433 (32.2) 1.36 (1.25-1.49) 1.45 (1.33-1.59) 1.17 (1.05-1.30) 1.57 (1.42-1.73) 1.18 (1.07-1.31) 1.57 (1.42-1.73) 45-54 14,677 (27.1) 1.49 (1.36-1.63) 1.50 (1.37-1.64) 1.29 (1.16-1.43) 1.61 (1.46-1.79) 1.30 (1.17-1.45) 1.62 (1.46-1.79) ≥55 5,800 (10.7) 1.61 (1.45-1.78) 1.30 (1.17-1.45) 1.35 (1.19-1.53) 1.37 (1.21-1.54) 1.36 (1.20-1.54) 1.37 (1.21-1.54)

Neighbourhood Income Quintile

1 (lowest) 12,366 (22.8) 1.00* 1.00 1.00 1.00 1.00 1.00 2 13,060 (24.1) 0.96 (0.90-1.02) 1.01 (0.95-1.08) 0.97 (0.90-1.04) 1.02 (0.95-1.10) 0.97 (0.90-1.04) 1.02 (0.95-1.10) 3 11,711 (21.6) 0.97 (0.91-1.03) 1.05 (0.98-1.12) 0.97 (0.89-1.05) 1.01 (0.94-1.09) 0.97 (0.89-1.05) 1.01 (0.94-1.09) 4 10,022 (18.5) 0.93 (0.87-0.99) 1.04 (0.97-1.11) 0.93 (0.86-1.01) 1.00 (0.93-1.08) 0.93 (0.86-1.01) 1.00 (0.93-1.09) 5 (highest) 7,038 (13.0) 0.93 (0.87-1.01) 1.10 (1.02-1.19) 0.91 (0.83-1.00) 1.05 (0.96-1.14) 0.92 (0.84-1.01) 1.05 (0.97-1.15)

Regional Health Authority

Fraser 23,999 (44.3) 1.00*** 1.00 1.00*** 1.00 1.00*** 1.00 Vancouver Coastal 10,083 (18.6) 1.23 (1.15-1.30) 0.85 (0.80-0.91) 1.06 (0.97-1.17) 0.89 (0.82-0.98) 1.10 (1.00-1.20) 0.91 (0.83-0.99) Interior 8,120 (15.0) 1.52 (1.42-1.62) 1.35 (1.27-1.44) 1.15 (1.02-1.30) 1.27 (1.14-1.2) 1.22 (1.08-1.37) 1.27 (1.14-1.41) Vancouver Island 8,745 (16.1) 1.58 (1.49-1.68) 1.30 (1.22-1.38) 1.31 (1.16-1.48) 1.25 (1.12-1.39) 1.34 (1.19-1.51) 1.25 (1.13-1.39) Northern 3,250 (6.0) 1.24 (1.13-1.37) 1.52 (1.39-1.66) 1.04 (0.88-1.24) 1.41 (1.22-1.63) 1.13 (0.95-1.34) 1.41 (1.22-1.63)

185







SMRs (n=10,634)



SMRs (n=10,634) Year of Injury

1998/1999 9,598 (17.7) 1.00*** 1.00 1.00*** 1.00 1.00*** 1.00 2000/2001 9,163 (16.9) 0.90 (0.83-0.96) 1.03 (0.96-1.11) 0.88 (0.80-0.96) 1.04 (0.95-1.13) 0.88 (0.81-0.97) 1.04 (0.95-1.13) 2002/2003 8,710 (16.1) 0.91 (0.85-0.98) 1.17 (1.08-1.26) 1.00 (0.91-1.09) 1.21 (1.11-1.32) 1.02 (0.93-1.12) 1.21 (1.11-1.32) 2004/2005 8,966 (16.5) 0.90 (0.84-0.97) 1.14 (1.05-1.23) 1.03 (0.94-1.13) 1.20 (1.10-1.31) 1.06 (0.97-1.16) 1.20 (1.10-1.31) 2006/2007 9,208 (17.0) 0.89 (0.83-0.96) 1.19 (1.10-1.28) 1.00 (0.92-1.10) 1.26 (1.15-1.38) 1.04 (0.95-1.14) 1.26 (1.15-1.38) 2008/2009 8,552 (15.8) 0.79 (0.73-0.85) 1.18 (1.09-1.27) 0.84 (0.76-0.92) 1.26 (1.16-1.39) 0.88 (0.80-0.97) 1.26 (1.15-1.38)


722 - Intervertebral disc disorders 2,283 (4.2) 1.78 (1.61-1.98) 1.82 (1.64-2.02) 1.69 (1.49-1.92) 1.74 (1.54-1.95) 1.68 (1.48-1.90) 1.72 (1.53-1.94) 724 - Other and unspecified

disorders of back 5,000 (9.2) 1.18 (1.10-1.28) 1.25 (1.16-1.35) 1.15 (1.05-1.26) 1.20 (1.10-1.31) 1.16 (1.05-1.27) 1.20 (1.10-1.31) 846 – Sprains and strains of sacroiliac

region 11,384 (21.0) 0.96 (0.90-1.01) 1.02 (0.96-1.08) 0.92 (0.85-0.98) 1.05 (0.98-1.12) 0.91 (0.85-0.97) 1.05 (0.98-1.12) 847 – Sprains and strains lumbar,

sacrum, coccyx, unspecified site of back 35,530 (65.6) 1.00*** 1.00 1.00*** 1.00 1.00*** 1.00


No 23,053 (42.5) 1.00*** 1.00 1.00*** 1.00 1.00*** 1.00 Yes 31,144 (57.5) 1.07 (1.02-1.12) 1.29 (1.23-1.35) 1.10 (1.03-1.17) 1.12 (1.06-1.19) 1.10 (1.04-1.17) 1.12 (1.06-1.19)

186







SMRs (n=10,634)



SMRs (n=10,634) Number of prior workers’ compensation claims in the 2 years before injury

0 36,787 (67.9) 1.00* 1.00 1.00* 1.00 1.00* 1.00 1 12,093 (22.3) 1.03 (0.97-1.08) 1.02 (0.97-1.08) 0.92 (0.87-0.99) 0.96 (0.91-1.02) 0.92 (0.87-0.99) 0.96 (0.91-1.02) ≥2 5,317 (9.8) 1.13 (1.05-1.22) 1.14 (1.06-1.22) 0.91 (0.83-1.00) 1.05 (0.96-1.15) 0.91 (0.83-1.00) 1.05 (0.96-1.15)

Comorbidities in the 2 years before injury

Back and/or neck pain No 36,399 (67.2) 1.00*** 1.00 1.00** 1.00 1.00* 1.00 Yes 17,798 (32.8) 1.29 (1.23-1.35) 1.02 (0.98-1.07) 1.09 (1.03-1.16) 0.98 (0.93-1.04) 1.09 (1.02-1.16) 0.98 (0.93-1.04)

Arthritis/rheumatism No 43,929 (81.1) 1.00*** 1.00 1.00 1.00 1.00 1.00 Yes 10,268 (18.9) 1.26 (1.20-1.33) 1.00 (0.95-1.06) 1.02 (0.95-1.09) 0.96 (0.90-1.03) 1.01 (0.94-1.09) 0.96 (0.90-1.03)

Neuropathic pain No 53,444 (98.6) 1.00 1.00 1.00 1.00 1.00 1.00 Yes 753 (1.4) 1.18 (0.99-1.42) 1.15 (0.96-1.38) 0.81 (0.65-1.02) 1.04 (0.84-1.28) 0.81 (0.65-1.01) 1.03 (0.84-1.27)

Headache disorders No 49,427 (91.2) 1.00*** 1.00 1.00* 1.00 1.00* 1.00 Yes 4,770 (8.8) 1.38 (1.28-1.48) 0.98 (0.90-1.06) 1.13 (1.03-1.24) 1.01 (0.92-1.11) 1.13 (1.03-1.25) 1.01 (0.92-1.11)

Mood and anxiety disorders No 44,974 (83.0) 1.00*** 1.00 1.00 1.00 1.00 1.00 Yes 9,223 (17.0) 1.34 (1.27-1.41) 1.01 (0.95-1.07) 1.06 (0.99-1.15) 1.06 (0.98-1.15) 1.07 (0.99-1.15) 1.06 (0.99-1.15)

Substance use disorders No 53,384 (98.5) 1.00*** 1.00 1.00 1.00 1.00 1.00 Yes 813 (1.5) 1.65 (1.41-1.94) 1.06 (0.88-1.28) 1.05 (0.86-1.30) 0.81 (0.65-1.00) 1.05 (0.85-1.29) 0.80 (0.65-1.00)

Prior self-inflicted harm or suicide attempt

No 54,087 (99.8) 1.00 1.00 1.00 1.00 1.00 1.00 Yes 110 (0.2) 1.56 (1.01-2.42) 1.10 (0.67-1.81) 1.08 (0.62-1.88) 1.22 (0.69-2.17) 1.10 (0.63-1.91) 1.24 (0.70-2.21)

187







SMRs (n=10,634)



SMRs (n=10,634)

Diabetes No 52,247 (96.4) 1.00*** 1.00 1.00*** 1.00 1.00*** 1.00 Yes 1,950 (3.6) 1.47 (1.32-1.64) 1.15 (1.03-1.30) 1.33 (1.16-1.52) 1.10 (0.96-1.27) 1.33 (1.16-1.53) 1.11 (0.97-1.27)

Hypertension No 50,554 (93.3) 1.00*** 1.00 1.00 1.00 1.00 1.00 Yes 3,643 (6.7) 1.25 (1.15-1.36) 1.09 (1.00-1.19) 1.07 (0.96-1.19) 1.11 (1.00-1.23) 1.07 (0.96-1.19) 1.11 (1.00-1.23)

Heart and Cerebrovascular Disease No 52,795 (97.4) 1.00*** 1.00 1.00* 1.00 1.00* 1.00 Yes 1,402 (2.6) 1.56 (1.38-1.77) 1.09 (0.95-1.26) 1.22 (1.04-1.43) 1.04 (0.88-1.22) 1.22 (1.04-1.43) 1.04 (0.88-1.22)

Upper gastrointestinal disorders No 52,257 (96.4) 1.00*** 1.00 1.00* 1.00 1.00* 1.00 Yes 1,940 (3.6) 1.29 (1.16-1.44) 0.98 (0.87-1.10) 1.18 (1.03-1.36) 0.94 (0.82-1.08) 1.18 (1.03-1.36) 0.94 (0.81-1.08)

Inflammatory bowel disorders No 53,963 (99.6) 1.00*** 1.00 1.00** 1.00 1.00** 1.00 Yes 234 (0.4) 1.77 (1.33-2.36) 0.83 (0.57-1.21) 1.67 (1.17-2.38) 0.74 (0.49-1.13) 1.69 (1.19-2.42) 0.75 (0.49-1.14)

Chronic kidney disease No 54,105 (99.8) 1.00** 1.00 1.00 1.00 1.00 1.00 Yes 92 (0.2) 2.05 (1.31-3.21) 0.77 (0.41-1.45) 1.41 (0.81-2.45) 0.68 (0.34-1.35) 1.37 (0.79-2.38) 0.68 (0.34-1.35)

Chronic liver disease No 53,764 (99.2) 1.00* 1.00 1.00 1.00 1.00 1.00 Yes 433 (0.8) 1.37 (1.10-1.72) 1.01 (0.79-1.31) 1.11 (0.84-1.48) 1.01 (0.76-1.35) 1.12 (0.84-1.49) 1.01 (0.76-1.35)

Thyroid disorders No 53,436 (98.6) 1.00** 1.00 1.00* 1.00 1.00* 1.00 Yes 761 (1.4) 0.92 (0.76-1.10) 0.71 (0.58-0.87) 0.88 (0.70-1.09) 0.76 (0.60-0.95) 0.87 (0.69-1.08) 0.75 (0.60-0.94)

Chronic respiratory disorders No 52,401 (96.7) 1.00*** 1.00 1.00 1.00 1.00 1.00 Yes 1,796 (3.3) 1.26 (1.13-1.41) 0.88 (0.77-1.00) 1.09 (0.95-1.26) 0.92 (0.79-1.06) 1.10 (0.96-1.27) 0.92 (0.80-1.07)

Seizure disorders No 54,022 (99.7) 1.00* 1.00 1.00 1.00 1.00 1.00 Yes 175 (0.3) 1.64 (1.16-2.31) 1.02 (0.68-1.53) 1.16 (0.75-1.80) 0.89 (0.55-1.42) 1.14 (0.73-1.77) 0.89 (0.56-1.42)

188







SMRs (n=10,634)



SMRs (n=10,634)

Sleep apnea and insomnia No 53,872 (99.4) 1.00 1.00 1.00 1.00 1.00 1.00 Yes 325 (0.6) 1.33 (1.02-1.74) 1.09 (0.82-1.45) 0.96 (0.69-1.34) 1.06 (0.77-1.47) 0.98 (0.70-1.36) 1.07 (0.77-1.49)

Traumatic brain injury and other head injuries

No 53,891 (99.4) 1.00 1.00 1.00 1.00 1.00 1.00 Yes 306 (0.6) 1.18 (0.89-1.57) 1.25 (0.94-1.65) 0.93 (0.66-1.33) 1.21 (0.88-1.67) 0.95 (0.67-1.35) 1.22 (0.89-1.68)


0 49,773 (91.8) 1.00*** 1.00 1.00** 1.00 1.00* 1.00 1 3,775 (7.0) 1.22 (1.13-1.33) 1.01 (0.93-1.10) 0.83 (0.75-0.93) 0.87 (0.79-0.97) 0.84 (0.76-0.94) 0.88 (0.79-0.98) ≥2 649 (1.2) 1.79 (1.50-2.14) 1.05 (0.85-1.29) 0.89 (0.70-1.12) 0.94 (0.73-1.20) 0.92 (0.73-1.17) 0.95 (0.74-1.21)

Number of GP visits in the year before injury

0 5,816 (10.7) 1.00*** 1.00 1.00*** 1.00 1.00*** 1.00 1-2 12,118 (22.4) 1.01 (0.93-1.10) 0.93 (0.86-1.01) 1.02 (0.92-1.13) 0.95 (0.87-1.04) 1.03 (0.93-1.14) 0.95 (0.87-1.04) 3-4 10,305 (19.0) 1.06 (0.98-1.16) 0.88 (0.81-0.95) 1.01 (0.91-1.13) 0.91 (0.83-1.00) 1.01 (0.91-1.13) 0.91 (0.83-1.00) 5-7 10,918 (20.1) 1.08 (0.99-1.18) 0.85 (0.78-0.92) 1.00 (0.90-1.12) 0.87 (0.79-0.96) 1.00 (0.90-1.12) 0.87 (0.79-0.96) ≥8 15,040 (27.8) 1.47 (1.36-1.60) 0.87 (0.80-0.94) 1.11 (0.99-1.25) 0.82 (0.74-0.91) 1.11 (0.99-1.25) 0.82 (0.74-0.91)

Number of pain-related specialist visits in the year before injury

0 44,519 (82.1) 1.00*** 1.00 1.00** 1.00 1.00** 1.00 1 5,365 (9.9) 1.26 (1.18-1.36) 1.06 (0.98-1.14) 1.12 (1.03-1.23) 1.08 (0.99-1.18) 1.11 (1.02-1.22) 1.08 (0.99-1.18) 2 2,127 (3.9) 1.43 (1.29-1.59) 1.12 (1.00-1.25) 1.15 (1.01-1.32) 1.12 (0.98-1.28) 1.15 (1.00-1.31) 1.12 (0.98-1.27) ≥3 2,186 (4.0) 1.70 (1.54-1.88) 0.98 (0.87-1.11) 1.20 (1.05-1.37) 0.93 (0.81-1.07) 1.15 (1.00-1.32) 0.92 (0.79-1.06)

189







SMRs (n=10,634)



SMRs (n=10,634) Number of other specialist visits in the year before injury

0 41,748 (77.0) 1.00*** 1.00 1.00 1.00 1.00 1.00 1 6,785 (12.5) 1.05 (0.98-1.12) 0.92 (0.86-0.99) 0.94 (0.87-1.02) 0.97 (0.90-1.05) 0.93 (0.86-1.01) 0.97 (0.90-1.05) 2 2,701 (5.0) 1.01 (0.92-1.12) 0.89 (0.80-0.98) 0.85 (0.75-0.96) 0.95 (0.84-1.07) 0.83 (0.73-0.94) 0.95 (0.84-1.07) ≥3 2,963 (5.5) 1.29 (1.18-1.41) 0.86 (0.78-0.96) 0.96 (0.85-1.08) 0.91 (0.81-1.03) 0.92 (0.82-1.04) 0.91 (0.81-1.03)

Spine surgery in the 2 years before injury

No 54,050 (99.7) 1.00*** 1.00 1.00* 1.00 1.00* 1.00 Yes 147 (0.3) 2.61 (1.83-3.73) 1.17 (0.74-1.86) 1.60 (1.02-2.51) 0.77 (0.46-1.29) 1.64 (1.04-2.58) 0.79 (0.47-1.32)

At least one spine x-ray visit in the year before injury

No 50,003 (92.3) 1.00*** 1.00 1.00 1.00 1.00 1.00 Yes 4,194 (7.7) 1.20 (1.11-1.30) 1.09 (1.01-1.18) 0.94 (0.85-1.04) 0.99 (0.90-1.09) 0.94 (0.85-1.04) 0.99 (0.90-1.09)


0 41,441 (76.5) 1.00*** 1.00 1.00*** 1.00 1.00*** 1.00 1-14 8,093 (14.9) 1.52 (1.43-1.61) 1.39 (1.31-1.48) 1.62 (1.50-1.74) 1.49 (1.39-1.60) 1.62 (1.51-1.75) 1.49 (1.39-1.60) ≥15 4,663 (8.6) 4.54 (4.23-4.88) 2.46 (2.27-2.67) 5.10 (4.64-5.62) 2.82 (2.55-3.11) 5.13 (4.66-5.65) 2.82 (2.56-3.12)


0 40,091 (74.0) 1.00* 1.00 1.00*** 1.00 1.00*** 1.00 1-14 5,439 (10.0) 0.90 (0.84-0.97) 0.96 (0.89-1.04) 0.81 (0.73-0.88) 0.94 (0.87-1.03) 0.81 (0.74-0.89) 0.94 (0.86-1.03) ≥15 8,667 (16.0) 1.04 (0.98-1.10) 1.05 (0.99-1.12) 0.73 (0.68-0.79) 0.91 (0.84-0.98) 0.73 (0.67-0.79) 0.91 (0.84-0.98)

190







SMRs (n=10,634)



SMRs (n=10,634) Cumulative days supplied with SMRs in the year before injury

0 49,898 (92.1) 1.00*** 1.00 1.00*** 1.00 1.00** 1.00 1-14 2,296 (4.2) 1.05 (0.94-1.17) 1.25 (1.13-1.39) 0.89 (0.78-1.02) 1.14 (1.01-1.29) 0.90 (0.79-1.03) 1.14 (1.00-1.28) ≥15 2,003 (3.7) 1.37 (1.23-1.53) 1.38 (1.24-1.54) 0.95 (0.82-1.10) 1.22 (1.07-1.40) 0.96 (0.83-1.11) 1.22 (1.07-1.40)


No 50,979 (94.1) 1.00*** 1.00 1.00* 1.00 1.00* 1.00 Yes 3,218 (5.9) 1.62 (1.49-1.76) 1.13 (1.03-1.24) 1.15 (1.03-1.29) 0.99 (0.88-1.11) 1.16 (1.04-1.30) 0.99 (0.88-1.11)


No 52,494 (96.9) 1.00*** 1.00 1.00*** 1.00 1.00*** 1.00 Yes 1,703 (3.1) 1.95 (1.74-2.18) 1.27 (1.12-1.45) 1.37 (1.18-1.59) 1.13 (0.96-1.32) 1.38 (1.19-1.60) 1.13 (0.96-1.32)


No 48,592 (89.7) 1.00*** 1.00 1.00*** 1.00 1.00*** 1.00 Yes 5,605 (10.3) 1.73 (1.62-1.85) 1.14 (1.06-1.22) 1.21 (1.10-1.32) 1.03 (0.94-1.13) 1.20 (1.10-1.31) 1.03 (0.94-1.13)

Prescriber Sex, n (%)

Men 44,167 (81.5) 1.00*** 1.00 1.00*** 1.00 Women 10,030 (18.5) 0.61 (0.57-0.65) 0.80 (0.75-0.84) 0.57 (0.51-0.65) 0.80 (0.73-0.89)

Prescriber Birth Year, n (%)

<1945 7,283 (13.4) 1.00*** 1.00 1.00*** 1.00 1945-1954 16,666 (30.8) 1.04 (0.97-1.11) 0.87 (0.81-0.93) 1.15 (0.98-1.35) 0.93 (0.81-1.08) 1955-1964 18,860 (34.8) 0.85 (0.79-0.91) 1.00 (0.93-1.07) 0.98 (0.84-1.15) 1.14 (0.99-1.31) ≥1965 11,388 (21.0) 0.66 (0.61-0.72) 1.00 (0.93-1.08) 0.76 (0.64-0.90) 1.09 (0.95-1.27)

191







SMRs (n=10,634)



SMRs (n=10,634) Prescriber Specialty, n (%)

General Practitioner 52,103 (96.1) 1.00*** 1.00 1.00*** 1.00 Pain-Related Specialist 1,475 (2.7) 3.96 (3.51-4.46) 2.05 (1.78-2.35) 4.94 (3.82-6.39) 1.77 (1.37-2.29) Other Medical Specialist 552 (1.0) 2.34 (1.95-2.80) 0.79 (0.61-1.03) 5.92 (4.17-8.40) 0.91 (0.61-1.37) Dental Surgeon 67 (0.1) 7.16 (4.22-12.15) 0.47 (0.14-1.58) 11.28 (5.29-24.06) 0.46 (0.12-1.69)

Intraclass Correlation 0.361 0.251 0.36 0.24 0.34 0.24 Model Fit -2LL 93,097.15 90,133.45 89,639.68 Between-Prescriber Variability (SE) 1.82 (0.07)***,1 1.08 (0.05)***,1 1.84 (0.07)*** 1.06 (0.05)*** 1.68 (0.07)*** 1.04 (0.05)***

Abbreviations: CI=confidence interval; GP=general practitioner; ICD-9=International Classification of Diseases 9th Revision; NSAIDs=nonsteroidal anti-inflammatory drugs; OR=odds ratio; SE=standard error; SMRs=skeletal muscle relaxants

*Significant at p<0.05 **Significant at p<0.01 ***Significant at p<0.001. Refers to significance of variable overall. 1For Model 1 (intercept-only model)

192

CHAPTER 7 THE RELATIONSHIP BETWEEN EARLY PRESCRIPTION DISPENSING

PATTERNS AND WORK DISABILITY IN A COHORT OF LOW BACK PAIN WORKERS’ COMPENSATION CLAIMANTS

Nancy Carnide, Sheilah Hogg-Johnson, Mieke Koehoorn, Andrea Furlan, Pierre Côté

Abstract

Objectives To examine the association between prescription opioids dispensed in the first eight

weeks following a compensated work-related low back injury and their relationship with work

disability over one year compared to prescription nonsteroidal anti-inflammatory drugs (NSAID)

and skeletal muscle relaxants (SMR). To also examine the association between strength of

opioids received, cumulative days supply of each drug class, and average daily opioid dose with

work disability.

Study Design Analysis of administrative data from an historical cohort.

Population Workers in British Columbia, Canada filing a new, short-term disability claim with

the workers’ compensation system for a low back injury and having at least one compensated

lost workday.

Methods Five exposure variables were constructed using prescription dispensing data from the

first eight weeks after injury: drug class(es) dispensed, drug class(es) at first dispense,

cumulative days supply, strength of opioids dispensed, and average daily morphine-equivalent

dose (MED). The relationship between these exposures and two administrative work disability

outcomes (time on compensation benefits, any benefit receipt) after eight weeks and up to one

193

year were analyzed using a zero-inflated negative binomial analysis. Potential confounders

included sociodemographics, occupation, and injury diagnosis, as well as pre-injury health care

and dispensing.

Results When compared to claimants receiving NSAIDs and/or SMRs only in the first eight

weeks, the incidence rate ratio (IRR) of days on benefits for claimants dispensed opioids only

was 1.09 (95%CI 1.04-1.14) and 1.26 (95%CI 1.22-1.30) for claimants dispensed opioids with

NSAIDs and/or SMRs. The IRR for claimants dispensed strong opioids only or strong and weak

opioids combined was 1.21 (95%CI 1.12-1.30) and 1.29 (95%CI 1.20-1.39), respectively,

compared to weak opioids only. Similar findings were seen for the outcome of benefit receipt,

but with larger effect sizes for claimants receiving more complex drug combinations. The rate of

days on benefits associated with each seven-day increase in days supplied of opioids was 10%,

compared to 3% for SMRs and 4% for NSAIDs. Effect sizes associated with the outcome of

benefit receipt were larger than seen for time on benefits. Each 30mg/day increase in daily

average morphine-equivalent dose resulted in a 4% increase in the incident rate of days on

benefits.

Conclusion Our findings suggest early opioid provision has an adverse effect on work disability.

However, residual confounding likely still accounts for some of the findings. Further research is

needed that better elucidates the mechanism by which opioids may contribute to prolonged

disability and that accounts for prescriber, system, and workplace factors in the relationship.

194

7.1 Introduction

The use of prescription opioids among injured workers in North America has been a significant

source of concern for more than a decade. From approximately 2000 until 2010, opioid

prescriptions paid by workers’ compensation organizations rose steadily.(1-4) In response, a

number of workers’ compensation organizations introduced various policies and guidelines

aimed at curbing inappropriate and excessive use of opioids, as well as their reimbursement.(5-

9) While use is now on the decline,(1, 10-12) opioids remain some of the most commonly

reimbursed prescriptions.(13)

There has been particular apprehension about what is considered inappropriate early

treatment with opioids for musculoskeletal disorders.(14) Back pain, namely sprains and

strains, has been shown in some jurisdictions to be the most common diagnosis for which

injured workers receive opioids.(15, 16) Further, early use of opioids for claimants with work-

related low back pain (LBP) has been broadly documented.(17-23)

We conducted a systematic review to determine whether prescription opioids provided early in

the course of a musculoskeletal disorder are associated with future work outcomes.(24) Most

studies examined workers’ compensation claimants with LBP injuries and findings were

generally consistent, demonstrating an association between early opioids and prolonged work

disability. However, the internal validity of these studies was jeopardized by: 1) exposure

measurement bias, due to incomplete prescription data, the definition of ‘early’, and immortal

time bias; and 2) confounding bias, namely confounding by indication and residual confounding

related to pre-injury and concomitant health care.

195

Two additional studies of LBP claimants have been published since our systematic review.(17,

21) Busse et al. found opioid reimbursement in the first four weeks of claim was associated with

delayed claim closure.(17) In contrast, Lee et al. found no significant association between

prescription opioids filled within two days after the first emergency department visit and the

outcome of disability duration.(21) Many of the limitations identified in our review continue to

be problematic in these studies.

One aspect that has yet to be addressed is the comparative effect of early opioids relative to

other LBP-relevant medications on work disability. Our previous analyses suggest that these

drugs are commonly prescribed to workers’ compensation claimants with LBP in British

Columbia (BC): 27.7%, 40.7%, and 24.8% of claimants, respectively, were dispensed opioids,

nonsteroidal anti-inflammatory drugs (NSAIDs), and skeletal muscle relaxants (SMRs) in the first

eight weeks after injury.(25) In the same cohort, we also found that pre-injury prescriptions

and health care were positively associated with early dispensing,(26) confirming the need to

consider pre-injury factors in this relationship.

The purpose of this study is to determine whether dispensing patterns of prescription opioids,

NSAIDs, and SMRs in the first eight weeks following a compensated LBP injury are associated

with work disability over one year in a cohort of BC workers’ compensation claimants. Our

study aims to minimize the limitations of the previous literature by capturing all prescriptions

dispensed irrespective of payment source, accounting for pre-injury prescriptions and health

care, and improving the temporality of the analysis by using an event-based cohort approach to

our analyses.(27) To partially address confounding by indication and severity, we limit our

196

sample to workers with at least one dispense and, hence, an indication for a LBP-relevant

prescription medication.

7.2 Methods

The University of Toronto Health Sciences Research Ethics Board provided ethics approval for

this study (protocol reference number 26885).


We conducted a historical cohort study using administrative health data. The source population

includes workers’ compensation claimants filing new short-term disability claims for LBP injuries

to WorkSafeBC between 1998 and 2009. Details of the study have been previously

described.(25, 26) WorkSafeBC is the provincial workers’ compensation organization in BC,

providing no-fault single payer insurance to workers with work-related injuries and diseases.

Workers excluded from coverage include the self-employed and other smaller segments of the

labour force.(28) The portion of the workforce eligible for coverage by WorkSafeBC during our

study period ranged from 92.5% to 94%.(29)

WorkSafeBC Claims and Firm Level Files (30) were linked with administrative health data from

five additional datasets capturing prescription dispenses (PharmaNet), outpatient health care

(Medical Services Plan [MSP] Payment Information File), inpatient hospitalizations (Discharge

Abstract Database [DAD]), and practitioner (MSP Practitioner File) and claimant (MSP

Consolidation File) demographics (31-35) for the years 1996 through 2010. PharmaNet is a

197

province-wide system that captures detailed information on all prescription medications

dispensed from community and hospital outpatient pharmacies in BC.

Medical Services Plan (MSP) is the universal health insurance program in BC, providing coverage

for 96% of residents for all medically required outpatient services provided by fee-for-service

practitioners, including examinations and consultations and laboratory and diagnostic

procedures. The Discharge Abstract Database (DAD) file contains information on hospital

discharges (deaths, sign-outs, transfers) of inpatients and day surgery patients from acute care

hospitals in BC.

The MSP Practitioner File contains basic demographic data on practitioners enrolled in MSP.

The MSP Registration and Premium Billing file is the central demographics dataset for

individuals registered to receive health services in BC. The BC Ministry of Health, WorkSafeBC,

PharmaNet, and the College of Physicians and Surgeons of BC approved access to and use of

the data facilitated by Population Data BC in this study.

7.2.2 Study Population

Data were extracted for workers who filed at least one new workers’ compensation short-term

disability claim for LBP between January 1, 1998 and December 31, 2009. Population Data BC

delivered claims for these workers dating back to 1991. Claim-level inclusion/exclusion criteria

were then applied. Claims had to be accepted short-term disability claims for a nonspecific LBP

disorder identified using four digit International Classification of Diseases, 9th Revision (ICD-9)

codes (36) (Appendix B, Table B-1) and with at least one day of wage replacement benefits in

the first eight weeks following injury. Consolidated claims (e.g., duplicate claims created for the

198

same worker and injury event) were ineligible. The minimum eligible age at the time of injury

was 18 years. Claims with a LBP-related hospitalization and/or serious LBP-related outpatient

service within five days after injury (Appendix B, Table B-2) were excluded. Claimants had to be

BC residents continuously eligible for health services for two years before and one year after

injury date. Finally, injury date had to be equal to or precede claim registration date.

From this group of eligible claims, one index claim per claimant was selected. To be an index

claim, the claim had to be the earliest LBP short-term disability claim where no other allowed

claims occurred in the year after injury. In addition, the total benefits paid in the year post-

injury had to be greater than zero. Following selection of the index claim, we then excluded

claimants who had at least one hospitalization and/or two outpatient health care billings for

cancer (ICD-9 three-digit codes 140-209 and 235-239; International Statistical Classification of

Diseases, Tenth Revision, Canadian (ICD-10-CA) (37) three-digit codes C00-C97 and D37-D48)

two years before and one year after injury.

7.2.3 Exposures

Our exposures were constructed using detailed dispensing data from the following drug classes:

opioids (American Hospital Formulary Service [AHFS codes] (38) 28:08.08, 28:08.12), NSAIDs

(28:08.04.08, 28:08.04.24, 28:08.04.92), and SMRs (12.20.04, 12.20.08, 12.20.12). A complete

list of active ingredients included by drug class are provided in Appendix B (Table B-3).(26) We

used data from the first eight weeks after injury. During the study period, WorkSafeBC policy

limited reimbursement of opioids to the first eight weeks after injury or surgery for most

claims.(39) However, in 2015, this policy was revised to limit payment to the first four weeks.(6)

199

Therefore, as a sensitivity analysis, we also constructed our exposures using dispensing data

from only the first four weeks after injury.

For each exposure period, we constructed five prescription dispensing exposures:

i. Drug class(es) dispensed: This exposure captures which drug class(es) were dispensed

during the exposure window. For each drug class, the presence of at least one

dispensing record in the exposure window defined a dispense. This variable was

categorized as: dispensed at least one NSAID and/or SMR only; dispensed at least one

opioid only; and dispensed at least one opioid and at least one NSAID and/or SMR.

ii. Drug class(es) dispensed at the first prescription dispense post-injury: This exposure

captures the drug class(es) dispensed at the first post-injury prescription in the relevant

exposure window. It was constructed in a similar manner to the first exposure and uses

the same categories.

iii. Cumulative days supply: This exposure aims to measure the cumulative duration of

exposure to each of the drug classes in a given exposure window. This variable was

calculated by summing the days supply across all prescriptions for a given drug class.

This was done separately for opioids, NSAIDs, and SMRs and was calculated among

claimants with at least one days supply in the respective exposure period.

iv. Strength of opioid(s) dispensed: This exposure identifies the strength of opioid(s)

dispensed to claimants receiving at least one opioid dispense in the respective exposure

period. In our data, anileridine, codeine, meperidine, pentazocine, propoxyphene, and

tramadol were considered weak opioids, while buprenorphine, butorphanol, fentanyl,

hydromorphone, methadone, morphine, and oxycodone were considered strong

200

opioids.(40-47) This variable was categorized as: dispensed only weak opioids; dispensed

only strong opioids; and dispensed at least one weak and one strong opioid.

v. Average daily morphine-equivalent dose (MED): The purpose of this exposure is to

measure the average daily dose of opioids received among claimants dispensed at least

one opioid prescription in the exposure period. For oral formulations, daily dose ((drug

strength*quantity dispensed)/days supply) was calculated for each day supplied with an

opioid or an SMR combination containing codeine (approximately 5% of all records used

in the calculation). A similar calculation was conducted for transdermal formulations:

(drug strength in microgram/hour * hours supplied)/days supply. Daily doses were then

converted into a morphine-equivalent dose (MED) using previously published

conversion ratios (Appendix B, Table B-4).(40-47) For transdermal formulations, the

result was divided by 1,000 to convert from micrograms to milligrams. For each

claimant, the average daily MED was then calculated as the sum of the daily MED over

the total days supplied with opioids in the respective time period. This variable was

calculated for oral and transdermal formulations only, as the unit of measure in the

dispensed quantity field for all other routes was not clear. Opioids by these routes

represented 0.6% of opioid records.

7.2.4 Outcomes

The outcome of work disability was constructed using data on wage replacement benefits paid

by WorkSafeBC up to the first year following injury. The main outcome of interest, cumulative

duration on short-term disability benefits, was calculated as the total number of days receiving

201

short-term disability benefits more than eight weeks post-injury and up to 52 weeks post-injury

(i.e., the number of days after the eight week exposure window). In sensitivity analyses, we also

calculated the number of days of short-term disability benefits received after four weeks and

up to 52 weeks post-injury (i.e., to correspond with the four week exposure window). The zero-

inflated negative binomial model used for the analysis of this outcome also allowed an

examination of receipt of short-term disability benefits (yes/no) in these two outcome windows.

7.2.5 Potential Confounders

7.2.5.1 Sociodemographic and Work-Related Factors

Data on sex, regional health authority (Fraser Health, Vancouver Coastal Health, Vancouver

Island Health, Northern Health, and Interior Health), age at injury date (18-24, 25-34, 35-44, 45-

54, ≥55 years), and neighbourhood income quintile were obtained from the MSP Consolidation

File. Claimant occupation at the time of injury, obtained from the WorkSafeBC Claim File, was

linked to data from The National Occupational Classification Career Handbook (48) using

Statistics Canada’s Standard Occupational Classification.(48) For each code, this handbook

provides global ratings to define skills, worker characteristics, and other indicators, including

physical activities and environmental conditions. For this study, two occupational

characteristics were considered: physical demands of the occupation and whether the job

involved working near or with equipment, instruments, machinery, or power/hand tools that

may be a potential source of accident or injury. A dichotomous yes/no variable was created

that identified whether claimant occupation involved heavy physical strength requirements (i.e.,

handling loads of more than 20kg) and/or involved use of equipment/machinery/instruments.

202

7.2.5.2 Injury-Related Factors

Year of injury (categorized into two-year groups from 1998 to 2009), ICD-9 diagnosis code

associated with the claim, and number of prior workers’ compensation claims (0, 1, ≥2) were

obtained from the WorkSafeBC Claims File. ICD-9 diagnosis was re-categorized based on the

first three digits: 722 (Intervertebral Disc Disorders), 724 (Other and Unspecified Disorders of

Back), 846 (Sprains and Strains of Sacroiliac Region), and 847 (Sprains and Strains of Other and

Unspecified Parts of Back).

7.2.5.3 Comorbidities

Pre-existing health conditions were identified using MSP outpatient and DAD hospitalization

data. These included pain-related conditions, mental health and substance use disorders, and

other chronic conditions. Claimants with one unique hospital record and/or two unique

outpatient visits in the two years before injury with at least one relevant ICD-9 or ICD-10-CA

diagnosis (Appendix B, Table B-5) were considered to have the condition of interest. Case

definitions were informed by the findings of prior validation studies (references in Appendix B,

Table B-5). The combination of claimant ID and date of service or separation date was used to

identify unique visits.

7.2.5.4 Health Care Utilization

Spinal x-rays in the year before injury (yes/no) and concurrent spinal x-rays in the exposure

window (four or eight weeks, yes/no) were defined as at least one MSP outpatient billing record

with a relevant fee item code or one DAD hospital separation record with a relevant Canadian

203

Classification of Diagnostic, Therapeutic, and Surgical Procedures system (CCP) (49) or Canadian

Classification of Health Interventions (CCI) (50) procedure code (Appendix B, Table B-6). Fee

item codes are MSP codes used to identify the insured medical service provided. Similarly, spine

surgery in the two years before injury (yes/no) and concurrent spine surgery in the exposure

window (yes/no) were defined as the presence of one hospital separation record in the DAD

with an applicable CCP or CCI procedure code (Appendix B, Table B-7). Hospitalization in the

year before injury (yes/no) and concurrent hospitalization in the exposure window (yes/no)

were defined as at least one unique hospital separation in the DAD with any diagnosis. The

unique combination of most responsible physician ID, claimant ID, and separation date was

used to define a hospitalization.

Number of outpatient physician health care visits in the year before injury was determined for

three groups of physicians: general practitioners (GPs), medical specialists commonly seen for

pain-related complaints (herein referred to as pain-related specialists), and other medical

specialists. Up to three specialty fields were available. Physicians were assigned a specialty of

GP if no other specialty was listed. Physicians with at least one specialty field that included

physiatry, neurology, neuropsychiatry, neurosurgery, orthopaedic surgery, internal medicine,

anesthesiology, occupational medicine, emergency medicine, osteopathic medicine, or

rheumatology were assigned a designation of ‘pain-related specialist’. If no pain-related

specialty was listed, but at least one specialty field contained another medical specialty, the

physician was designated as ‘other medical specialist’. The unique combination of physician ID,

claimant ID, and service date defined one visit. Number of GP visits was categorized as 0, 1-2, 3-

4, 5-7, and ≥8, while number of pain-related and other medical specialist visits was categorized

204

as 0, 1, 2, and ≥3. A dichotomous (yes/no) variable describing the presence of at least one pain-

related specialist visit in the exposure window was also constructed in a similar manner.

For claimants with injury years 1998 to 2001, we also constructed separate variables for the

number of outpatient supplementary care visits to physiotherapists, chiropractors, and massage

therapists in the year before injury (0, 1-5, ≥6) using a similar procedure with MSP outpatient

data. Separate dichotomous (yes/no) variables describing the presence of at least one

physiotherapist, chiropractor, or massage therapy visit in the exposure window were also

constructed. These variables were restricted to claimants with injury years 1998 to 2001 as MSP

ceased coverage for most supplemental health care services in 2002.

7.2.5.5 Other Prescription Dispensing

Using PharmaNet data, the cumulative number of days supply in receipt of dispensed opioid,

NSAID, and SMR prescriptions (0, 1-14, ≥15) in the year before injury was calculated by

summing the days supply across all dispensed prescriptions for a given drug class. When

dispensing records overlapped in time, days were counted only once. This was also done for

select antidepressants (AHFS 28:16.04.16, 28:16.04.24, 28:16.04.28), anticonvulsants (28:12.08,

28:12.12, 28:12.92), and sedative hypnotics/anxiolytics (28:24.08, 28:24.92), as these drug

classes may be used as adjuvants in management of LBP and may be associated with the

outcome of work disability (Appendix B, Table B-1).(51) Separate dichotomous (yes/no)

variables describing the presence of at least one antidepressant, anticonvulsant, and sedative

hypnotic/anxiolytic dispense in the exposure window were also constructed.

205


In order to model the count data for the outcome of days on benefits, a number of potential

models were initially considered and estimated, including ordinary Poisson with and without a

dispersion parameter and a negative binomial model, but overdispersion remained evident.

Due to the excess zeroes in this outcome, zero-inflated Poisson and negative binomial models

were then estimated and compared using a likelihood ratio test of alpha=0. The Vuong test was

also used to compare the ordinary negative binomial with the zero-inflated model. The zero-

inflated negative binomial model was found to be the most appropriate model based on these

model fit indices.

Using a zero-inflated negative binomial model, we examined the association between the five

dispensing exposures in the first eight weeks after injury and number of days on benefits after

the eight-week exposure window (i.e., >8 weeks, up to one year). Crude and adjusted incidence

rate ratios (IRR) and corresponding 95% confidence intervals (CI) were estimated. This type of

analysis also allowed estimation of the crude and adjusted odds ratios (OR) and 95% confidence

intervals for the association between the five exposures and benefit status after the exposure

window (i.e., receipt of at least one day of benefits after eight weeks, yes/no).

For the exposures drug class(es) dispensed and drug class(es) dispensed at first prescription, the

reference category was NSAIDs and/or SMRs only, while the reference category for the

exposure strength of opioid(s) received was weak opioids only. Cumulative days supply was

retained as a count variable and seven days was chosen as the unit of measure for analyses.

Seven days was chosen as opioids and SMRs are typically recommended to be provided as short

courses and one week increments within the eight week exposure window seemed a natural

206

cut point to use. This was also informed by our prior research where the median cumulative

days supply for these drug classes in the first eight weeks ranged from 8 to 17 days.(25) For

average daily MED, 30mg/day was chosen as the unit of measure, as low doses of opioids are

typically recommended for initial prescriptions. Our prior research also demonstrated that

average daily MED in the first eight weeks was approximately 30mg/day.(25)

For each analysis, models were adjusted for age, sex, neighbourhood income quintile, and year

of injury as a base model. Potential confounding by other variables was then assessed using the

change-in-estimate criterion (52) in the following manner. First, all potential confounders were

entered into a model with the exposure and the base model variables. Each potential

confounder was then removed one by one to determine the change in the effect measure and

beta estimate of the exposure. If removal resulted in at least a ten per cent change in either

estimate, the variable was retained for the final model.

Specific to the exposure strength of opioid dispensed, the potential confounder of cumulative

days of opioids in the prior year was separated based on strength of opioids, as it was

hypothesized that the strength of opioids received previously may influence the strength of

opioids received following injury. For the analysis of the exposure average daily MED, we also

considered potential confounding by concurrent NSAID and SMR dispenses in the exposure

window as dichotomous variables (yes/no). Finally, for the analysis of cumulative number of

days supply as an exposure, we considered cumulative days supply of the other two drug

classes in the exposure window as potential confounders (e.g., cumulative days supplied with

NSAIDs and SMRs for the independent variable cumulative days of opioids). In these latter two

207

cases, these variables were included in order to adjust for the potential confounding effect of

concomitant medications.

In sensitivity analyses, we assessed the impact of varying the exposure window to four weeks

and the corresponding outcome window to greater than four weeks and up to 52 weeks post-

injury. We also examined the impact of prior and concurrent supplementary health care as

potential confounders. These analyses were limited to claimants with injury years 1998 through

2001. The data analysis for this paper was generated using SAS software Version 9.3. Copyright

© 2000-2010 SAS Institute Inc. SAS and all other SAS Institute Inc. product or service names are

registered trademarks or trademarks of SAS Institute Inc., Cary, NC, USA.

7.3 Results

7.3.1 Study Sample

We received data on 916,804 claims. After applying our claim-level inclusion criteria, a total of

142,993 claims were considered eligible, linked to 113,434 claimants. Selection of one eligible

index claim per claimant resulted in 99,233 claimants, from which 2,109 were excluded due to a

pre-injury history of cancer, leaving 97,124 claimants. Data on sex (n=1), neighbourhood

income (n=1,622), and health authority (n=96) were missing for a total of 1,711 claimants,

resulting in 95,413 claimants. For the purposes of this analysis, we also limited our sample to

claimants with at least one dispensed prescription for an opioid, NSAID, and/or SMR in the first

eight weeks after injury (n=55,571).

208


Among claimants receiving at least one dispense in the first eight weeks after injury (n=55,571),

almost two thirds were men (63.1%) between the ages of 35 and 54 years (59.2%) (Table 7-1).

Claimants from the Fraser regional health authority (44.3%) and those with sprains and strains

(86.6%) made up the largest proportion of the cohort. Prior workers’ compensation claims were

common (32.1%) and just over half worked at a job at the time of injury that involved heavy

physical strength requirements and/or the use of equipment or machinery (57.5%). Pre-existing

pain-related comorbidities were prevalent, namely back and/or neck pain (32.9%) and arthritis

and other joint pain (19.0%), as were mood and anxiety disorders (17.1%).

Most claimants had at least one general practitioner visit in the year before injury (89.3%),

while visits to pain-related specialists (17.8%) and other medical specialists (23%) were less

common. A small proportion of claimants had a hospitalization in the year before injury (8.2%)

or first eight weeks after injury (2.1%). Less than 1% of claimants had spine surgery either in the

two years before injury or in the first eight weeks. Spine x-rays were common in the first eight

weeks (25.8%).

Most claimants did not have any days supplied of opioids (76.3%), NSAIDs (73.9%), or SMRs

(92.1%) in the year before injury. Antidepressant, anticonvulsant, and sedative-

hypnotic/anxiolytic dispenses in the year before injury or within the first eight weeks were also

not common.

The proportion of claimants receiving benefits after the first four weeks and after the first eight

weeks up to one year post-injury were 65.8%, and 42.7%, respectively. The median

209

(interquartile range, IQR) number of days on benefits after the first eight weeks and up to one

year post-injury was 0 (IQR 0-24) days. The corresponding value after the first four weeks was

10 (IQR 0-44) days.

7.3.3 Time on Benefits After the Eight-Week Exposure Period

In crude analyses, the incidence rate ratio (IRR) of days on benefits for claimants dispensed

opioids only or opioids with SMRs and/or NSAIDs in the first eight weeks was 1.22 (95%CI 1.17-

1.28) and 1.40 (95%CI 1.35-1.45), respectively, when compared to claimants receiving NSAIDs

and/or SMRs only. After adjusting for confounding, the IRR decreased to 1.09 (95%CI 1.04-1.14)

for claimants receiving opioids only and 1.26 (95%CI 1.22-1.30) for claimants receiving opioids

with NSAIDs and/or SMRs (Table 7-2).

Similar, but attenuated findings were seen for drug class at first prescription. In adjusted

analyses, the IRR of days on benefits for claimants receiving opioids only compared to claimants

receiving NSAIDs and/or SMRs only was 1.07 (95%CI 1.03-1.12) and for claimants receiving

opioids with NSAIDs and/or SMRs, the IRR was 1.12 (95%CI 1.08-1.16).

The adjusted IRR of days on benefits after eight weeks comparing claimants dispensed strong

opioids only to those dispensed weak opioids only in the first eight weeks (IRR 1.21, 95%CI 1.12-

1.30) was similar to that seen for claimants dispensed a combination of weak and strong

opioids (IRR 1.29, 95%CI 1.20-1.39) (Table 7-2).

There was a 10% increase in the rate of days on benefits after the first eight weeks associated

with a seven day increase in cumulative days supplied for opioids within eight weeks after

210

injury (95%CI 1.09-1.11). The increase was smaller for NSAIDs (IRR 1.04, 95%CI 1.03-1.05) and

SMRs (IRR 1.03, 95%CI 1.01-1.04) (Table 7-3). For every 30mg/day increase in daily average

MED in the first eight weeks, the incident rate of days on benefits after this period increased by

4% (95%CI 1.02-1.07) (Table 7-3).

7.3.4 Receipt of at Least One Day of Benefits After the Eight-Week Exposure Period

When examining whether claimants receive at least one day of benefits after the eight week

period, the findings suggest there was no significant difference between claimants receiving

opioids only and claimants receiving NSAIDS and/or SMRs only in the first eight weeks (OR 0.99,

95%CI 0.93-1.05), but claimants receiving opioids with NSAIDS and/or SMRs had 61% higher

odds (95%CI 1.54-1.69) of receiving benefits after the eight week exposure period (Table 7-2).

The odds of receiving benefits after eight weeks were 14% higher (95%CI 1.08-1.19) for

claimants receiving opioids only as their first prescription and 11% higher (95%CI 1.05-1.16) for

claimants receiving opioids with NSAIDS and/or SMRs compared to NSAIDS and/or SMRs only

(Table 7-2). Claimants dispensed strong opioids only (OR 1.18, 95%CI 1.05-1.32) or weak and

strong opioids (OR 2.27, 95%CI 2.00-2.63) also had higher odds of receiving benefits after eight

weeks (Table 7-2).

Every seven day increase in days supplied with opioids in the first eight weeks resulted in a 35%

increase in the odds of being on benefits after that period (95%CI 1.33-1.39), while for NSAIDs

and SMRs, the increase was 25% (95% 1.23-1.28) and 23% (95%CI 1.20-1.27), respectively

(Table 7-3). The odds of being on benefits after eight weeks associated with an increase of

211

30mg/day in average daily MED within the first eight weeks post-injury were not significant (OR

1.02, 95%CI 0.99-1.04) (Table 7-3).

7.3.5 Sensitivity Analyses Examining a Four-Week Exposure Window

We repeated our analyses using a four-week exposure window, limiting the sample to claimants

with at least one dispensed prescription for an opioid, NSAID, and/or SMR in the first four

weeks after injury (n=52,723).

For the exposures of drug class(es) dispensed and drug class(es) dispensed at the first

prescription in the first four weeks, the results for time on benefits as an outcome were similar

to those seen for eight weeks (Appendix E, Table E1). However, for the outcome of benefit

status, a stronger association was seen for the drug class(es) at first prescription: an OR of 1.26

(95% 1.19-1.33) for opioids only compared to NSAIDs and/or SMRs only and an OR of 1.23

(95%CI 1.16-1.30) for opioids with NSAIDs and/or SMRs. For drug class(es) dispensed in the first

four weeks, the resulting effect size estimates for claimants dispensed opioids with NSAIDs

and/or SMRs was similar to that seen for eight week exposure, but unlike the eight week

estimates, claimants receiving opioids only were at an 11% increase in odds of being on benefits

after four weeks (95%CI 1.05-1.18).

For strength of opioids received, for both time on benefits and benefit status outcomes, the

point estimates for four week exposure were higher than seen for eight weeks (Appendix E,

Table E1). However, confidence intervals overlapped in most cases, with the exception of the

comparison of claimants receiving strong opioids only to those receiving weak opioids only and

the outcome of time on benefits.

212

The results pertaining to the four week exposures of cumulative days supply (for any drug class)

and average daily MED were equivalent to that seen for eight weeks for both work disability

outcomes were equivalent (Appendix E, Table E2).

7.3.6 Sensitivity Analyses Incorporating Supplementary Health Care Visits as Potential Confounders

In sub-analyses of claimants with injury years 1998 through 2001, including the number of

physiotherapy, chiropractic, and massage therapy visits in the year before injury, as well as

dichotomous measures of visits to each of these practitioners in the exposure window

(concurrent therapy) in the models did not have any impact on the findings (Appendix E, Tables

E3 and E4).

7.4 Discussion

We examined the relationship between early dispensing of opioids, NSAIDs, and SMRs and

subsequent work disability outcomes after a compensated work-related LBP injury. Overall, our

results suggest that workers receiving early opioids are at a higher risk to subsequently receive

wage replacement benefits compared to workers receiving only NSAIDs and/or SMRs,

particularly those receiving opioids with NSAIDs and/or SMRs. Claimants receiving strong

opioids, and in particular, strong and weak opioids combined, were also more likely to receive

benefits than those receiving only weak opioids. Furthermore, increasing days supply for all

three drug classes was also associated with receiving benefits in the post-exposure period.

Increasing opioid dose in the early weeks, however, was not a strong risk factor for work

disability.

213

Our findings are generally consistent with most prior studies that found an increase in work

disability associated with insurer-reimbursed opioid prescriptions early after compensated LBP

injuries.(17, 24) However, our effect size estimates are generally smaller than previously

reported, particularly for time on benefits. This may be due to differences in our methodology.

In the current study, we attempted to address one of the key limitations of previous studies:

confounding by indication and severity. We accounted for a wide variety of potential

confounders that could be risk factors for dispensing and work disability, including pre-injury

and co-occurring prescriptions, health care utilization, and pre-existing health conditions.

Adjustment for confounding greatly reduced the magnitude of the crude estimates. We also

limited our analyses to workers with at least one prescription dispense. This was done to

increase the likelihood that claimants had an indication for one of the three drug classes and to

create a more homogeneous sample of workers with respect to severity. As a result, we were

able to conduct comparative analyses between workers dispensed opioids and an active

reference group dispensed other medications. Finally, we only considered benefit days paid

after the exposure window.

We observed that claimants dispensed opioids, over the exposure window and at first dispense,

had a significantly higher rate of days on benefits and greater risk of receiving at least one day

of benefits in the outcome window than claimants receiving only NSAIDs and/or SMRs. For

example, the incidence rate of days on benefits for claimants who received opioids only in the

first eight weeks after injury were 9% higher than claimants receiving NSAIDS and/or SMRs only,

while the incidence rate was 26% higher for claimants receiving opioids with NSAIDs and/or

SMRs. When looking at the outcome of receipt of benefits (yes/no), there was no association

214

seen for claimants receiving opioids only, but the odds of receiving at least one day of benefits

was 61% higher for claimants receiving opioids with NSAIDs and/or SMRs. If opioid dispensing

alone was responsible for prolonged disability, we might anticipate effect sizes to be similar

whether opioids were provided alone or in combination with other drug classes.

Similarly, when compared to workers receiving weak opioids only, workers dispensed both

weak and strong opioids in the early weeks after injury demonstrated a stronger association

with our work outcomes than workers only receiving strong opioids. Again, if exposure to

strong opioids were most detrimental, effect sizes seen for claimants taking strong opioids only

should not differ from estimates for weak and strong opioids to the extent seen. This disparity

in the strength of the relationship seen between the groups suggests there continues to be

other underlying factors we could not account for that, at least in part, are driving the

relationship. Claimants receiving various drug combinations may be a subgroup who are

struggling the most with their injuries or have more complicated and severe injuries and

physicians may be attempting various types of medications to manage their pain.

Further, we also observed an increase in the risk of receiving at least one day of benefits

associated with increasing cumulative days supplied, not just for opioids, but all drug classes,

albeit the strength of the relationship was modestly higher for opioids. A much more

attenuated association was seen for time on benefits. There was also little difference in the

effect estimates seen for NSAIDs and SMRs, despite the fact that SMRs are also associated with

central nervous system effects similar to opioids and are cautiously recommended for

treatment of early LBP episodes.(53-55) Regardless of the medication, those receiving longer

supplies may be experiencing greater difficulty in their recovery and may differ in important

215

ways, such as injury severity or the capacity to cope with their symptoms. Gross et al. also

found a relationship between early receipt of NSAIDs and SMRs with increasing time on

benefits compared to not receiving these drugs.(20) Taken together, the pattern of results

appears to suggest residual confounding by indication is likely, in part, accounting for the

findings in our study.

Prior studies have found an increase in the risk of disability associated with increasing MED.(19,

23) In these studies, the comparison group was claimants with no reimbursed opioids and total

MED was examined. In contrast, we examined average daily MED among claimants with at least

one day supply of opioids. Our results suggest that, among claimants who have dispensed

opioids, each 30mg/day increase in dose is associated with a small, but statistically significant

increased risk of prolonged disability. It should be noted that the median (IQR) average daily

MED in our sample was 27mg/day (18.0-38.6) and 90% of claimants had an average daily MED

of 56mg/day or less. Due to this lack of exposure variability, we may have been unable to detect

a stronger relationship, particularly if the effect was the same within the range of MED used by

claimants in our study. It is also important to note that, across all outcomes, the largest effect

sizes were seen for strength of opioid(s) received among claimants with at least one opioid. As

previously suggested, there may be underlying differences between individuals receiving weak

opioids only compared to those receiving stronger opioids that we could not account for, but

our observations should be confirmed in other samples.

Across all analyses, a number of factors acted frequently as confounders, namely ICD-9

diagnosis, prior dispensing history, co-occurring medications, early visits to pain specialists, and

early ordering of spine x-rays, all of which may represent potential markers for severity or

216

indication. However, our use of administrative data limited our ability to fully address sources

of confounding. Importantly, we did not have access to clinical data on injury severity, pain

intensity, or psychosocial factors, such as recovery expectations, which research has shown to

be associated with prescribing and/or return to work after injury.(56-59)

We cannot exclude the possibility that early exposure to opioids after a LBP injury, and in

particular, strong opioids, is causally related to work disability. However, the underlying

mechanism driving the potential relationship is not yet known. It is possible that the side effects

associated with opioid use could contribute to reduced work capacity. The provision of early

opioids could also lead to iatrogenic disability, as suggested by studies of early imaging.(60-62)

Early use may also lead to prolonged use and potentially abuse and addiction. However, none

of these theories have been examined to date. Future research is needed to better understand

how opioids may contribute to prolonged disability.

In general, the findings in our study were similar whether using an eight or four week exposure

window. In some instances, there was a trend toward stronger effect sizes with four week

exposure. This was seen for the exposure of strength of opioid(s) received (across both

outcomes) and for drug class(es) dispensed in the first eight weeks and at first dispense for the

outcome of receipt of benefits (yes/no). Previously described explanations could also account

for these findings. For instance, the stronger the drugs received more quickly after injury,

perhaps the greater degree of injury severity and stronger medications may simply be a red flag

for a claimant at risk of prolonged disability. This pattern of findings could also be indicative of

iatrogenic disability arising from early intensive treatment. However, for the exposure strength

217

of opioid(s) received, the confidence intervals overlapped and results may not represent true

differences.

To date, no study has demonstrated any benefits of the early provision of opioids or other

related medications on work disability after a LBP injury, including our study. Given the

preponderance of research demonstrating the serious harms associated with opioid use and

the unclear mechanism linking exposure with work disability, prescribing of opioids for early

management of LBP should be done with an abundance of caution and as conservatively as

possible.

Our study has a number of strengths. In particular, we ensured separation of exposure and

outcome windows, our prescription data were comprehensive, and we were able to account for

a number of health care and prescription factors, namely in the pre-injury period, that could

not be addressed in previous studies. Our analytic sample was also restricted to claimants who

received at least one dispensed prescription, minimizing the potential for confounding by

indication and severity. We were also able to do a comparative analysis of claimants receiving

opioids to those receiving other drugs commonly prescribed for LBP.

In addition to residual confounding by indication and severity (previously discussed), other

potential sources of residual confounding may have affected our findings. For instance, we

lacked data on system and workplace factors that could influence both prescribing and work

disability, such as the availability of work accommodations after injury or access to non-

pharmacologic care. In our analyses, we also could not account for the impact of the physician

responsible for managing the claim. In a previous paper, we demonstrated the importance of

218

prescriber variation in dispensing after injury (26) and physicians also likely vary in their

management of worker’s compensated injuries, potentially influencing the work disability

trajectory of an injured worker.

This study has additional limitations. Our work outcomes are based on administrative measures

of wage replacement benefits that are not necessarily indicative of return-to-work, as benefits

can cease for other reasons. The structure of the workers’ compensation payment data was

also such that we only had information on the number of days paid by month and year. For

months in which the exposure/outcome window cut-off date fell, days were assigned

consecutively to each workday of the month, beginning at the start of the month. This may

have led to non-differential misclassification of our time on benefits outcome.

We do not know whether prescriptions were actually consumed as dispensed, which may have

led to an underestimation of the effect. We did not have data on over-the-counter medication

use and therefore, could not account for additional medications used that were not prescribed.

However, given the comparative nature of our analyses, the direction of potential bias would

be difficult to predict. The prescription-related measures were also based on prescription

dispensing records irrespective of indication and prescriptions cannot necessarily be attributed

to the injury. While this is a limitation of our data, we believe it is still important to identify the

drugs claimants are exposed to, whether they are for LBP or not, particularly if they have a

causal role to play in a claimant’s work disability trajectory. Our data also lacked information on

date of claim closure. Therefore, it is possible that for some claimants, some degree of

exposure was captured after the claim was closed, potentially resulting in an underestimate of

the association. The generalizability of our findings may also be limited with respect to non-

219

claimants or workers with no initial lost time due to injury. Finally, the odds ratios obtained for

the outcome of benefit status after exposure likely overestimated the relative risk, as this

outcome was not rare. However, this would not have had an impact on the significance of the

findings.

7.5 Conclusion

Interpreting the relationship between early opioid prescriptions and work disability outcomes is

complex. While our findings suggest early opioid provision has an adverse effect on work

disability, effect estimates were much attenuated compared to those of previous studies.

Despite our efforts, we still believe residual confounding by indication and severity may have

accounted for some of our results. Using these findings and those of previous studies to inform

policy should be done cautiously until further research is conducted that better elucidates the

mechanism by which opioids may contribute to prolonged disability and that accounts for the

influence of prescriber, system, and workplace factors. Nevertheless, our data also does not

demonstrate a benefit to opioid analgesics with respect to work disability. Insurers and

clinicians should remain cautious in the management of injured workers with LBP, while

continuing to ensure pain is adequately treated.

220

7.6 References




4. Dunn C, Rothfels P, Nguyen T, Martin C, Sambandam R, Noertjojo K, et al. Developing Practice Directive on Claims With Opioids Prescribed at WorkSafeBC. 2009.

5. California Division of Workers' Compensation. Medical Treatment Utilization Schedule (MTUS) Opioids Treatment Guidelines. Part 1: Executive Summary, Introduction, and Recommendations. California. 2016.

6. Department WCPQ. Practice Directive #C10-1. Claims With Opioids, Sedative/Hypnotics or Other Prescribed Potentially Addictive Drugs. Richmond, BC: WorkSafeBC; 2015.

7. Office of the Medical Director. Guideline for Prescribing Opioids to Treat Pain in Injured Workers. Olympia, Washington: Washington State Department of Labor & Industries; 2013.

8. Workplace Safety and Insurance Board (WSIB) of Ontario. Enhanced narcotics management for injured workers. 2010; Available from: http://www.wsib.on.ca/WSIBPortal/faces/WSIBDetailPage?cGUID=WSIB014326&rDef=WSIB_RD_ARTICLE&_afrLoop=2075015899742000&_afrWindowMode=0&_afrWindowId=null#%40%3FcGUID%3DWSIB014326%26_afrWindowId%3Dnull%26_afrLoop%3D2075015899742000%26rDef%3DWSIB_RD_ARTICLE%26_afrWindowMode%3D0%26_adf.ctrl-state%3D18vkgcbv2w_4. Last Accessed: December 20, 2016.

9. Ohio Bureau of Workers' Compensation. Opioid prescribing rule and guidelines. 2016; Available from: https://www.bwc.ohio.gov/provider/brochureware/opioideducate/default.asp. Last Accessed: December 20, 2016.


11. Garg RK, Fulton-Kehoe D, Turner JA, Bauer AM, Wickizer T, Sullivan MD, et al. Changes in opioid prescribing for Washington workers' compensation claimants after implementation of an opioid dosing guideline for chronic noncancer pain: 2004 to 2010. J Pain. 2013;14(12):1620-8.






https://www.bwc.ohio.gov/provider/brochureware/opioideducate/default.asp

221

12. Thumula V, Wang D, Liu T-C. Interstate Variations in Use of Opioids, 3rd Edition. Cambridge, Massachusetts: Workers Compensation Research Institute; 2016.

13. Lipton B, Colon D. Workers Compensation and Prescription Drugs: 2016 Update. Florida: National Council on Compensation Insurance; 2016.

14. Deyo RA, Mirza SK, Turner JA, Martin BI. Overtreating chronic back pain: time to back off? J Am Board Fam Med. 2009;22(1):62-8.

15. Lipton B, Laws C, Li L. Narcotics in Workers Compensation. Florida: National Council on Compensation Insurance; 2009.

16. Swedlow A, Ireland J, Johnson G. Prescribing Patterns of Schedule II Opioids in California Workers' Compensation. Oakland, California: California Workers' Compensation Institute; 2011.








222

24. Carnide N, Hogg-Johnson S, Côté P, Irvin E, Van Eerd D, Koehoorn M, et al. Early Prescription Opioid Use for Musculoskeletal Disorders and Work Outcomes: A Systematic Review of the Literature. Clin J Pain. 2017;33(7):647-658.

25. Carnide N, Hogg-Johnson S, Furlan AD, Côté P, Koehoorn M. Prescription dispensing patterns before and after a workers' compensation claim: An historical cohort study of workers with low back pain injuries in British Columbia. In preparation. 2017.

26. Carnide N, Hogg-Johnson S, Côté P, Koehoorn M, Furlan AD. The factors associated with early post-injury prescription dispensing after a work-related low back pain injury. In preparation. 2017.













223





41. Washington State Agency Medical Directors’ Group (AMDG). Interagency Guideline on Prescribing Opioids for Pain. 2015; Available from: http://www.agencymeddirectors.wa.gov/Files/2015AMDGOpioidGuideline.pdf. Last Accessed: September 2, 2015.

42. American Pain Society. Principles of Analgesic Use in the Treatment of Acute Pain and Cancer Pain. 6th Edition ed. Glenview, IL. 2008.






48. Canada HRD. National Occupational Classification Career Handbook. Ottawa, ON: Government of Canada; 2011.





224



52. Rothman KJ, Greenland S, Lash TL. Modern Epidemiology. Philadelphia, PA: Lippincott Williams & Wilkins; 2008.




56. Bernstein E, Carey TS, Garrett JM. The use of muscle relaxant medications in acute low back pain. Spine (Phila Pa 1976). 2004;29(12):1346-51.

57. Cancelliere C, Donovan J, Stochkendahl MJ, Biscardi M, Ammendolia C, Myburgh C, et al. Factors affecting return to work after injury or illness: best evidence synthesis of systematic reviews. Chiropr Man Therap. 2016;24(1):32.

58. Cherkin DC, Wheeler KJ, Barlow W, Deyo RA. Medication use for low back pain in primary care. Spine (Phila Pa 1976). 1998;23(5):607-14.

59. Steenstra I, Irvin E, Mahood Q, Hogg-Johnson S, Heymans MW. Systematic Review of Prognostic Factors for Workers’ Time Away from Work Due to Acute Low-Back Pain: An Update of a Systematic Review. Final Report to Workers Compensation Board of Manitoba. Toronto, ON: Institute for Work & Health; 2011.

60. Graves JM, Fulton-Kehoe D, Jarvik JG, Franklin GM. Early imaging for acute low back pain: one-year health and disability outcomes among Washington State workers. Spine (Phila Pa 1976). 2012;37(18):1617-27.

61. Webster BS, Cifuentes M. Relationship of early magnetic resonance imaging for work-related acute low back pain with disability and medical utilization outcomes. J Occup Environ Med. 2010;52(9):900-7.

225

62. Webster BS, Bauer AZ, Choi Y, Cifuentes M, Pransky GS. Iatrogenic consequences of early magnetic resonance imaging in acute, work-related, disabling low back pain. Spine (Phila Pa 1976). 2013;38(22):1939-46.

226

Table 7-1 Baseline characteristics of the cohort of workers’ compensation claimants with short-term disability claims for injuries to the low back from 1998-2009 and with at least one prescription opioid, NSAID, or SMR dispense in the first eight weeks after injury (n=55,571)

Variable n ( % )

Sex Women 20,517 (36.9) Men 35,054 (63.1)

Age at injury

18-24 4,838 (8.7) 25-34 11,911 (21.4) 35-44 17,871 (32.2) 45-54 15,019 (27.0) ≥55 5,932 (10.7)

Neighbourhood income quintile

1 (lowest) 12,658 (22.8) 2 13,452 (24.2) 3 11,992 (21.6) 4 10,259 (18.5) 5 (highest) 7,210 (13.0)



Year of injury

1998/99 9,833 (17.7) 2000/01 9,419 (17.0) 2002/03 8,920 (16.1) 2004/05 9,179 (16.5) 2006/07 9,439 (17.0) 2008/09 8,781 (15.8)


722 – Intervertebral disc disorders 2,353 ( 4.2) 724 – Other and unspecified disorders of back 5,117 ( 9.2) 846 – Sprains and strains of sacroiliac region 11,658 (21.0) 847 – Sprains and strains lumbar, sacrum, coccyx, unspecified site of back 36,443 (65.6)

Occupation at time of injury involving heavy physical strength requirements (handle loads ˃20kg) and/or working near/with equipment, instruments, machinery, or power/hand tools that may be a source of accident or injury

No 23,602 (42.5) Yes 31,969 (57.5)

Number of prior workers’ compensation claims in the 2 years before injury

0 37,699 (67.8) 1 12,406 (22.3) ≥2 5,466 (9.8)

227

Variable n ( % ) Comorbidities in the two years before injury

Back and/or neck pain 18,303 (32.9) Arthritis/rheumatism and other joint pain 10,554 (19.0) Neuropathic pain 770 ( 1.4) Headache disorders 4,898 ( 8.8) Mood and anxiety disorders 9,490 (17.1) Suicide attempt/self-harm history 114 ( 0.2) Substance use disorder 838 ( 1.5) Diabetes 1,998 ( 3.6) Hypertension 3,718 ( 6.7) Heart and cerebrovascular disease 1,437 ( 2.6) Chronic respiratory disorders 1,840 ( 3.3) Chronic liver disease 443 ( 0.8) Chronic kidney disease 93 ( 0.2) Thyroid disease 1,282 ( 1.3) Upper gastrointestinal disorders 1,995 ( 3.6) Inflammatory bowel disease 395 ( 0.4) Insomnia and sleep apnea 331 ( 0.6) Seizure disorders 179 ( 0.3) Traumatic brain injury 312 ( 0.6)


0 51,046 (91.9) 1 3,863 ( 7.0) ≥2 662 ( 1.2)

Hospitalization in the first eight weeks after injury

No 54,392 (97.9) Yes 1,179 ( 2.1)

Number of general practitioner visits in the year before injury 0 5,963 (10.7) 1-2 12,396 (22.3) 3-4 10,541 (19.0) 5-7 11,195 (20.2) ≥8 15,476 (27.9)

Number of pain-related physician specialist visits in the year before injury

0 45,636 (82.2) 1 5,505 ( 9.9) 2 2,188 ( 3.9) ≥3 2,242 ( 4.0)

Number of other physician specialist visits in the year before injury

0 42,778 (77.0) 1 6,960 ( 12.5) 2 2,769 ( 5.0) ≥3 3,064 ( 5.5)

228

Variable n ( % ) Pain-related physician specialist visit in the first eight weeks after injury

No 50,554 (91.0) Yes 5,017 ( 9.0)

Spine surgery in the two years before injury

No 55,423 (99.7) Yes 148 ( 0.3)

Spine surgery in the first eight weeks after injury

No 55,490 (99.9) Yes 81 ( 0.2)

At least one spine x-ray in the year before injury

No 51,257 (92.2) Yes 4,314 ( 7.8)

At least one spine x-ray in the first eight weeks after injury

No 41,239 (74.2) Yes 14,332 (25.8)


0 42,413 (76.3) 1-14 8,341 (15.0) ≥15 4,817 ( 8.7)


0 41,066 (73.9) 1-14 5,580 ( 10.0) ≥15 8,925 (16.1)


0 51,151 (92.1) 1-14 2,355 ( 4.2) ≥15 2,065 ( 3.7)


No 52,269 (94.1) Yes 3,302 ( 5.9)

At least one antidepressant dispense in the eight weeks after injury

No 53,575 (96.4) Yes 1,996 ( 3.6)


No 53,820 (96.9) Yes 1,751 ( 3.2)

At least one anticonvulsant dispense in the eight weeks after injury

No 54,234 (97.6) Yes 1,337 ( 2.4)

229

Variable n ( % ) At least one sedative hypnotic/anxiolytic dispense in the year before injury

No 49,817 (89.7) Yes 5,754 (10.4)

At least one sedative hypnotic/anxiolytic dispense in the eight weeks after injury No 51,266 (92.3) Yes 4,305 (7.8)

Abbreviations: ICD-9=International Statistical Classification of Diseases and Related Health Problems, 9th Revision; NSAID=nonsteroidal anti-inflammatory drugs; SMR=skeletal muscle relaxants 1Includes all types of claims (short-term disability, long-term disability, health care only, vocational rehabilitation) for any diagnosis (not limited to low back pain claims).

230

Table 7-2 Results of a zero-inflated negative binomial regression analysis of the association between early drug class(es) dispensed, drug class(es) at first prescription, and strength of opioid received in the first eight weeks after injury and time on benefits and benefit status after the eight week exposure window up to 1 year post-injury

Predictor

Time on Benefits After Eight Week Exposure Window On Benefits After Eight Week Exposure Window # of days on

benefits after 8-week exposure window, mean (SD), median

Receiving benefits

after 8-week exposure window,

n (%)

IRR (95% CI) OR (95% CI)

Unadjusted Adjusted Unadjusted Adjusted

Drug class(es) dispensed in first 8 weeks (n=55,571) NSAIDS and/or SMRs only (n=29,104) 19.5 (46.9), 0.0 1.00 1.001 10,817 (37.2) 1.00 1.001 Opioids only (n=7,730) 24.9 (56.3), 0.0 1.22 (1.17-1.28)*** 1.09 (1.04-1.14)*** 3,030 (39.2) 1.08 (1.02-1.14)** 0.99 (0.93-1.05) Opioids with NSAIDS and/or SMRs (n=18,737) 38.3 (67.0), 3.0 1.40 (1.35-1.45)*** 1.26 (1.22-1.30)*** 9,889 (52.8) 1.89 (1.82-1.96)*** 1.61 (1.54-1.69)*** Drug class(es) dispensed at first prescription in first 8 weeks (n=55,571) NSAIDS and/or SMRs only (n=33,343) 23.4 (52.2), 0.0 1.00 1.002 13,563 (40.7) 1.00 1.002 Opioids only (n=11,028) 31.8 (62.4), 0.0 1.19 (1.15-1.24)*** 1.07 (1.03-1.12)*** 5,147 (46.7) 1.28 (1.22-1.33)*** 1.14 (1.08-1.19)*** Opioids with NSAIDS and/or SMRs (n=11,200) 30.8 (61.3), 0.0 1.20 (1.16-1.25)*** 1.12 (1.08-1.16)*** 5,026 (44.9) 1.18 (1.14-1.23)*** 1.11 (1.05-1.16)*** Strength of opioids dispensed in first 8 weeks (n=26,467) Weak opioid(s) only (n=23,271) 31.1 (60.7), 0.0 1.00 1.003 10,903 (46.9) 1.00 1.003 Strong opioid(s) only (n=1,711) 47.0 (75.8), 6.0 1.30 (1.20-1.40)*** 1.21 (1.12-1.30)*** 941 (55.0) 1.39 (1.25-1.54)*** 1.18 (1.05-1.32)** Weak and strong opioids (n=1,485) 70.5 (87.3), 32.0 1.48 (1.37-1.59)*** 1.29 (1.20-1.39)*** 1,075 (72.4) 3.13 (2.70-3.45)*** 2.27 (2.00-2.63)***

Abbreviations: CI=confidence interval; IRR=incidence rate ratio; NSAIDs=nonsteroidal anti-inflammatory drugs; OR=odds ratio; SD=standard deviation; SMRs=skeletal muscle relaxants *Significant at p<0.05 **Significant at p<0.01 ***Significant at p<0.001

231

1Adjusted for: sex, age, year of injury, income, diagnosis code, health authority, pain specialist visit within first 8 weeks, cumulative days of opioids in previous year, sedative hypnotic

dispense within first 8 weeks, spinal x-ray within first 8 weeks, hospital contact within first 8 weeks, back/neck pain in prior 2 years, cumulative days of NSAIDs in previous year 3Adjusted for: sex, age, year of injury, income, diagnosis code, cumulative days supplied of opioids in prior year, pain specialist visit within first 8 weeks, health authority, sedative hypnotic

dispense within first 8 weeks, spinal x-ray within first 8 weeks 5Adjusted for: sex, age, year of injury, income, diagnosis code, pain specialist visit within first 8 weeks, sedative hypnotic dispense within first 8 weeks, back/neck pain comorbidity within

prior 2 years

232

Table 7-3 Results of a zero-inflated negative binomial regression analysis of the association between cumulative days supplied and average daily MED in the first eight weeks after injury and time on benefits and benefit status after the eight week exposure window up to 1 year post-injury

Predictor

Time on Benefits After Exposure Window On Benefits After Exposure Window IRR (95% CI) OR (95% CI)


Cumulative days supply in first 8 weeks Opioids (n=26,673)

For every 7 day increase 1.10 (1.09-1.11)*** 1.10 (1.09-1.11)1, *** 1.32 (1.30-1.35)*** 1.35 (1.33-1.39)1, *** NSAIDs (n=39,571)

For every 7 day increase 1.07 (1.06-1.08)*** 1.04 (1.03-1.05)2, *** 1.37 (1.35-1.39)*** 1.25 (1.23-1.28)2, *** SMRs (n=23,837)

For every 7 day increase 1.07 (1.05-1.08)*** 1.03 (1.01-1.04)3, *** 1.41 (1.37-1.43)*** 1.23 (1.20-1.27)3, *** Average daily MED in first 8 weeks (n=29,818)

For every 30mg/day increase 1.06 (1.04-1.09)*** 1.04 (1.02-1.07)4, ** 1.02 (0.99-1.05) 1.02 (0.99-1.04)4

Abbreviations: CI=confidence interval; IRR=incidence rate ratio; NSAIDs=nonsteroidal anti-inflammatory drugs; OR=odds ratio; SMRs=skeletal muscle relaxants *Significant at p<0.05 **Significant at p<0.01 ***Significant at p<0.001 1Adjusted for: sex, age, year of injury, income, cumulative days of opioids in previous year 2Adjusted for: sex, age, year of injury, income, cumulative days of opioids within 8 weeks, diagnosis code, cumulative days of SMRs within 8 weeks, spine x-ray within first 8 weeks 3Adjusted for: sex, age, year of injury, income, cumulative days of opioids within 8 weeks, cumulative days of NSAIDs within 8 weeks, spine x-ray within first 8 weeks 7Adjusted for: sex, age, year of injury, income, diagnosis code, spine x-ray within first 8 weeks, pain specialist visit within first 8 weeks, health authority, NSAID dispense within first 8

weeks, cumulative days of opioids in previous year, anticonvulsant dispense within first 8 weeks

233

CHAPTER 8 DISCUSSION

8.1 Major Findings and Contribution to the Literature

The overall objective of this dissertation was to better understand the relationship between the

provision of opioids early after a compensated work-related low back pain (LBP) injury and

work disability. I explored this objective across four manuscripts that progressively built upon

one another.

8.1.1 Systematic Review (Manuscript 1)

The first step involved conducting a systematic review and critical appraisal of the existing

literature examining early opioids and work outcomes (Chapter 3). This review was recently

published in the Clinical Journal of Pain.(1) The eligible population for this review was more

broadly defined as individuals with musculoskeletal conditions that were not required to be

work-related in nature. Only five studies were identified that examined this relationship, all

based on samples of individuals claiming compensation for their injuries and most examining

workers with LBP injuries in a workers’ compensation setting. Findings in four studies

consistently demonstrated an association between the provision of early opioids and prolonged

work disability. Increasing dose and duration of use were also associated with longer disability.

In a fifth study, shorter time between early opioid prescriptions was associated with a shorter

time on disability benefits.

I set out to do a best evidence synthesis of these studies, where only studies with a low risk of

bias would be eligible for inclusion. Importantly, however, this was not possible due to the high

234

risk of bias affecting each of the studies. As a result, I changed course and instead, conducted a

thorough critical appraisal of the literature to highlight the key sources of bias and inform

future research in this area. Exposure measurement bias and confounding bias was particularly

prominent. I concluded that, given these methodological limitations, the true relationship

between early opioid provision and work disability is unknown. This systematic review was not

only comprehensive, but it also made an important methodological contribution to our

interpretation of the existing literature. This was crucially needed for this evidence base as

policies have been developed, in part, based on the findings of these studies and interpreted to

assume the relationship between opioids and work disability is causal.

8.1.2 Administrative Data Analyses (Manuscripts 2 to 4)

The findings of the systematic review informed the analyses presented in chapters 5 to 7, which

attempted to address some of the methodological limitations of the previous research and

provide a better understanding of prescription dispensing patterns, their determinants, and

their effect on work disability in a sample of workers’ compensation claimants with LBP claims. I

studied a large sample of workers’ compensation claimants from British Columbia (BC) using

individual-level administrative data on workers’ compensation claims, outpatient health care

visits, hospitalizations, pharmacy dispensing, and sociodemographics with minimal missing

data. My study is the first to include data on pre-injury factors that could influence both post-

injury dispensing patterns and the work disability trajectory of a claimant. Importantly, the

prescription data for these analyses came from a comprehensive prescription monitoring

235

program where all prescriptions dispensed in community pharmacies are required to be

recorded, irrespective of the payment source.

8.1.2.1 Manuscript 2

In Chapter 5, I described dispensing patterns of prescription opioids, nonsteroidal anti-

inflammatory drugs (NSAIDs), and skeletal muscle relaxants (SMRs) one year before and after

injury. Each of the three drug classes were commonly dispensed in both periods, but

particularly post-injury. The first prescription was dispensed quickly after injury for all three

drug classes. Prescriptions were typically dispensed for short periods and opioids were

generally provided in low doses as weak opioids and short-acting formulations.

Prescription dispensing was generally stable prior to the injury. Dispensing for all drug classes

peaked in the first eight weeks post-injury and declined substantially thereafter, suggesting for

the majority of claimants, prescriptions were generally “early” and treatment was acute in

nature. However, there was a small sustained increase in post-injury dispensing compared to

pre-injury attributable to a small proportion of claimants. There was also an increase in

dispensing of strong and long-acting opioids, as well as an increase in average days supply

across all drug classes in the post-injury period. The findings suggest a small proportion of

claimants may be struggling with recovery and will persist in their use of these medications over

a year after injury.

236


In Chapter 6, I explored the relationship between pre-injury and baseline factors and dispensing

patterns in the first eight weeks after injury. The factors found to be consistently associated

with drug class dispensed in the first eight weeks and first drug class dispensed were prior

cumulative days supplied with one of the three drug classes, as well as antidepressants,

anticonvulsants, and sedative-hypnotic/anxiolytics, ICD-9 diagnosis on the claim, age at injury,

pre-existing comorbidities, prior physician visits and hospitalizations, year of injury, sex, health

authority, occupation, and prior workers’ compensation claims. Prescriber sex, birth year, and

specialty were also associated with drug class dispensed at first post-injury prescription, but did

not account for the observed prescriber variation. This paper revealed the importance of pre-

injury factors and possible injury severity in influencing early post-injury dispensing. In addition,

this was the first study, to our knowledge, to demonstrate the importance of prescriber

variation in explaining the first drug class dispensed after work-related injury. The factors that

were identified could also, in theory, influence work disability trajectories post-injury and would

be important to consider in an analysis examining the association between early dispensing

after injury and work outcomes.


In Chapter 7, I looked at the relationship between dispensed opioids, NSAIDs, and SMRs in the

first eight weeks after injury and work disability. I considered several exposure variables in this

period: drug class(es) dispensed, drug class(es) dispensed at the first post-injury dispense,

cumulative days supplied for each of the three drug classes, strength of opioid(s) received, and

237

average daily morphine-equivalent dose (MED). I also examined two exposure periods (four and

eight weeks) and considered two outcomes: time on benefits after the early exposure period

and short-term disability benefit status after the early exposure period.

Claimants dispensed opioids, alone or in combination with NSAIDs and/or SMRs, had a

significantly higher rate of days on benefits and greater risk of receiving at least one day of

benefits in the outcome window than claimants receiving only NSAIDs and/or SMRs. However,

the effect size for the association among claimants receiving opioids with NSAIDs and/or SMRs

was larger than that seen for opioids only.

Overall, our results suggest that workers receiving early opioids are more likely to subsequently

receive wage replacement benefits compared to workers receiving only NSAIDs and/or SMRs.

This was seen for both work outcomes and for both drug class(es) dispensed in the first eight

weeks, as well as at first dispense. Similarly, among claimants receiving opioids, those receiving

strong opioids were more likely to receive benefits than those receiving weak opioids only.

However, in both cases, effect sizes were strongest for workers receiving more complex

combinations of drugs (i.e., opioids with NSAIDs and/or SMRs and strong and weak opioids

combined). These findings may suggest these claimants, who require various drug

combinations, are struggling the most with their injuries and are more likely to remain on

disability for longer periods. Interestingly, work disability outcomes were more common with

increasing cumulative days supplied for all drug classes and there was little evidence of

prolonged disability associated with daily opioid dose.

This study adds to the existing literature because it addresses, at least in part, some of the

238

limitations identified in my systematic review. In particular, all analyses were limited to workers

with at least one prescription dispense in order to increase the likelihood that claimants had an

indication for a prescription prescribed for LBP. I also had access to a wide variety of potential

confounders, including pre-injury factors. Adjustment for these confounders dampened the

magnitude of the observed associations. I also ensured separation of the exposure and

outcome windows to avoid immortal time bias. In addition, as previously mentioned, data on

drug exposure was comprehensively captured, minimizing the chance that workers would be

misclassified as unexposed. This study is also the first, to my knowledge, to examine the

comparative effect of opioids on work disability as compared to other drugs commonly used for

LBP. The findings suggest that claimants on opioids alone do not differ on work disability

benefit outcomes to the extent previously seen in other studies when the comparison group

was no opioids.

8.2 External and Internal Validity

Within each individual chapter, the limitations are briefly described. Many of these limitations

are as a result of using administrative data not collected for research purposes. Here I review

and elaborate upon some of the key limitations that cut across all or most of the chapters.

8.2.1 External Validity

As described in the systematic review in Chapter 3 (Manuscript 1), prior studies examining the

association between early opioid dispensing and work outcomes have examined insurer-based

populations, namely from workers’ compensation. The focus of my administrative data analyses

239

has also been on a workers’ compensation population. Underreporting of work injuries is a

common phenomenon (2, 3) and workers with a shorter work absence have also been shown to

be less likely to file a claim.(4, 5) It is not clear how workers who do not enter the compensation

system use these medications, nor the potential effect of these medications on work disability

in this population. Other researchers have successfully used emergency department visit data

to more comprehensively identify work-related injuries (6) and this method of case

identification could be considered in future studies. However, this does not negate the fact that

workers’ compensation claimants are an important population to consider with regards to

opioid use.

Our study and prior studies have also included only claimants with lost work time, which may

represent a more severe population requiring stronger medications. The extent to which

workers with work-related injuries with no lost work time receive these drug classes is not

known. In BC, approximately 50% of all claims are health care only claims.(7) In WorkSafeBC

data, there are no diagnoses attached to these claims and, therefore, it is not possible to use

these data to create a sample of LBP claimants who only receive compensation for health care.

However, this does represent an important avenue for future research. Claimants may receive

opioid prescriptions, but experience no work disability and our judgement of these matters may

be clouded by the fact that the research to date has only considered lost time claimants. It

would be advantageous to our understanding of these issues to assess to what extent claimants

with no lost time are also receiving these medications. Still, the use of a lost time population to

investigate outcomes, such as work loss, is advantageous and not easily accomplished in a no

loss time sample.

240

Further, this thesis used data from workers’ compensation claimants with injury years dating

from 1998 through 2009. This decade represented a period of increasingly prolific opioid

prescribing for non-cancer pain. In several general population (8, 9) and workers’ compensation

studies,(10-12) there is evidence of a recent trend toward a decrease in prescribing of opioids in

the last five years. While use of this older dataset is unlikely to have affected the findings in

Chapter 7 (Manuscript 4), dispensing patterns as described in Chapter 5 (Manuscript 2) may

differ from current trends and the factors identified in Chapter 6 (Manuscript 3) that influence

early dispensing patterns may also have changed given the increasing awareness of the harms

associated with opioids and changing perception around appropriateness of use. Further

studies using more updated data are certainly warranted.

8.2.2 Internal Validity

8.2.2.1 Misclassification of Comorbidities

Health care utilization data were used to identify a number of pre-existing comorbidities. Case

definitions were informed by the findings of prior validation studies and comorbidities were

considered present if the claimant had one unique hospital record and/or two outpatient billing

records with a relevant diagnostic code in the two years before injury. The combination of

hospitalization and outpatient records has been shown to improve agreement for some

conditions, but the need for two outpatient records may have decreased sensitivity for some

conditions (e.g., arthritis) at the expense of improving specificity.(13) I was also limited to a two

year time window before injury based on data availability, which is a time period often used in

identifying chronic conditions in administrative data.(13, 14) However, I recognize that for some

241

conditions (e.g., self-harm), a longer time period may have been more sensitive in identifying

cases. Therefore, our definition only captures more recent and treated conditions and may

have led to non-differential misclassification.

8.2.2.2 Misclassification of Supplementary Health Care

To a limited extent, I was able to assess the influence of supplementary health care (i.e.,

chiropractic, massage therapy, physiotherapy) on early dispensing patterns and on the

relationship between early dispensing and work disability. However, this analysis was limited to

a small subset of our cohort based on earlier injury years. Up until 2002, supplementary health

care benefits were provided through the universal health care system up to a maximum of 12

visits per year per practitioner for individuals less than 65 and 15 visits for those 65 and older.

As of January 1, 2002, these services were only made insured through the provincial Medical

Services Plan in medically necessary cases for lower income individuals. Some degree of non-

differential misclassification is, therefore, possible as I was not able to capture visits that may

have been paid through other means, including out-of-pocket and employer-based private

health insurance.

8.2.2.3 Misclassification of Work Outcomes

In Chapter 7, the association between early dispensing patterns and work disability was

examined. Outcomes were derived from workers’ compensation benefit data. There were

limitations to the payment data provided, as only information on the number of days paid by

month and year was available. The cut-off date used to separate the exposure period from the

242

outcome period could fall at different points within a month, depending on the injury date.

Therefore, a decision had to be made as to how to distribute the days paid between the

exposure and outcome windows. Based on previous team experience with these data, days

paid in the early weeks after injury are typically paid consecutively at the front end of the

period. Therefore, paid days were assigned consecutively to each work day of the cut-off

month, beginning on the first work day of the month, rather than distributing them equally

across the month. This may have led to non-differential misclassification of our time on benefits

outcome.

8.2.2.4 Misclassification of Prescription Exposure

The prescription exposure measures derived in these analyses were based on prescription

dispensing records from community and hospital outpatient pharmacies. The data did not

capture prescriptions provided in hospital. Approximately 7.7% of individuals in our sample

were hospitalized in the year prior to the injury and 10.5% in the year after injury. However,

individuals with a LBP-related hospitalization in the first five days after injury were excluded,

which was the period of greatest dispensing activity and was the focus of Chapters 6 and 7. In

those first eight weeks, only 1.7% of workers were hospitalized. Therefore, any underestimation

is likely to be minor.

The degree of exposure to these medications may also have been overestimated as actual

consumption patterns are not known. This may be particularly problematic for these drug

classes as many claimants may only take them on an as-needed basis. The results can only be

interpreted in the light of intended and prescribed instructions for use. For the purposes of

243

understanding prescription drug exposure in this population and its related factors (Chapters 5

and 6), this was likely less problematic as dispensing data does describe, to an extent, the

prescribing practices for the injured worker population. In Chapter 7, on the other hand, it is

possible that observed associations were underestimated.

Further, dispensing records in PharmaNet do not include an indicator of the diagnosis for which

the drug is prescribed and, therefore, post-injury prescriptions could not necessarily be

attributed to the work-related low back injury. Initially, I explored various coding rules in an

attempt to link proximal LBP diagnoses in health care utilization data to subsequent dispensed

prescriptions. However, it became increasingly apparent that this would not be straightforward

and a number of issues complicated this task, including determining the time period used to

define ‘proximal’, the presence of other nearby diagnoses that could act as indications for these

drug classes, and instances where a refill may have been dispensed without any recent health

care utilization. The potential misclassification would have been of unknown magnitude and

direction and, therefore, the decision was made to include all prescriptions in the analyses. As a

result, I may have overestimated exposure to injury-related prescriptions. Workers who are

receiving prescriptions for other conditions may be more likely to have poor health, which

could have an impact on their work disability trajectory. As such, the relationship specifically

between prescriptions for the LBP injury and work disability may have been overestimated.

However, the general stability in pre-injury patterns followed by the sudden dramatic peak in

the initial post-injury days displayed in this sample suggest, that for most claimants,

prescriptions at least in the early post-injury period were new and likely attributable to the

injury. Presumably clinicians also do not approach treatment in compartmentalized ways when

244

an individual presents with health conditions for which there is an overlap in drug indication.

For example, at the time of the work-related LBP injury, a claimant may already be taking an

NSAID for an unrelated condition and finding it to be effective. The physician in this instance

may suggest that the claimant use their current NSAID for the injury as well. In such situations,

attempting to label a prescription with a specific indication may be artificial. Ultimately, it is

important to identify exposure to these drugs, whether they are for the LBP injury or not,

particularly if they have a causal role to play in work disability.

This database also does not comprehensively capture over-the-counter medications. At their

own discretion, pharmacists may enter such medications to the patient’s file to improve drug

monitoring, but this is not a requirement. In Canada, a number of drugs from the main three

drug classes are available over-the-counter, namely NSAIDs and SMRs. Low dose codeine in

combination with acetaminophen is also available behind the pharmacist’s counter without a

prescription. In our data, only 1.8% of all dispensing records captured in the year before and

after injury were scheduled as over-the-counter. This information would have been important

for a more comprehensive understanding of the degree of exposure to these drug classes in this

population, irrespective of whether a prescription is needed. Therefore, the findings can only

be generalized to prescription drugs. The direction of bias due to this lack of data, however, is

difficult to estimate. For instance, we might assume workers with less severe pain and

potentially less likely to be disabled may successfully manage their symptoms with over-the-

counter medications alone. However, it may also be possible that workers experiencing more

difficulty may use over-the-counter medications in addition to those prescribed to manage their

245

pain. The use of active comparators in our analyses further complicate the determination of the

direction of this bias.

Our data also lacked information on date of claim closure. Prior research has shown that 50-

60% of LBP claimants receive compensation for four weeks or less.(15) Therefore, it is possible

that for some claimants, some degree of exposure was captured after the claim was closed. It is

reassuring to note that the majority of claimants received prescriptions in the first four weeks

only (73.6% for opioids, 80.2% for NSAIDs, 85.4% for SMRs). However, a quarter of all claimants

dispensed opioids received them after the first four weeks. While resumption of benefits can

occur after claim closure,(16) it is possible we have underestimated the association. This may

also partially explain the differences seen in effect sizes between the four and eight week

window.

8.2.2.5 Residual Confounding

Confounding by indication and severity is likely the most important source of residual

confounding in this dissertation. Our analyses and those of previous studies have been limited

to variables found within an administrative database. I exploited the data to the extent possible

to account for a wide variety of factors that could influence dispensing, as well as act as

confounders in the relationship between early opioid dispensing and work disability. I also

limited my sample in Chapter 7 to workers with at least one prescription dispense and,

therefore, an indication for one of the three drug classes. ICD-9 diagnosis was strongly

associated with early dispensing in Chapter 6 and also frequently acted as a confounder across

all analyses in Chapter 7. This may suggest severity of the injury plays a crucial role in

246

influencing early dispensing patterns and work disability. However, the use of diagnosis codes

as a surrogate marker for severity is controversial and clinical data on pain intensity, function,

distress, and recovery expectations were not available. As a result, despite the methodological

attempts implemented to minimize bias due to underlying differences in indication and

severity, the results in Chapter 7 suggest confounding by indication and severity may still

explain some of our findings and likely led to an overestimation of the observed associations.

8.3 Implications for Future Research

8.3.1 Consideration of Additional Work Outcomes

There is a pressing need to expand to other relevant work disability outcomes in this area of

research. The outcomes assessed in our study and those of previous studies may be important

for insurers, but use of benefit data has been shown to underestimate the true extent of work

disability.(17) Primary data collection on return-to-work and at-work functioning would

contribute greatly to our understanding of how opioids influence work outcomes.

Another important outcome to be considered in future studies is the potential safety risks

associated with use of these sedating medications in the workplace, namely opioids and SMRs.

In Chapter 6, I found some evidence to suggest that claimants with a job involving heavy

physical demands and/or equipment and machinery were at increased risk of receiving drug

class combinations including SMRs and opioids. The American College of Occupational and

Environmental Medicine recently released guidelines regarding the use of opioids among

individuals in safety-sensitive positions, but the literature cited to support the guideline

regarding potential risks was not workplace-specific.(18) Surprisingly, little is known about the

247

risk of workplace injuries, deaths, and near misses associated with the use of sedating

medications in the workplace. The findings here and those of prior literature suggest a high

prevalence of use of opioids and SMRs among injured workers. Prior research has also shown

that a return-to-work does not equate to complete recovery from a musculoskeletal disorder

and workers may continue to experience residual pain and disability even after a return to

work.(19, 20) Therefore, it would not be improbable to assume workers may continue to take

these medications even after returning to work. This is a question of increasing interest for

policy makers and is critically important to examine in future studies.

8.3.2 Advancing Our Understanding of Prescribing for LBP Among Injured Workers

While there was some evidence in this dissertation to suggest physicians may have not been

adhering to clinical practice guidelines around the pharmacologic management of LBP, our lack

of data on the indication of these prescriptions, as well as over-the-counter medications

prevents certainty around this issue. Future studies are needed to understand whether

pharmacologic care among claimants is in accordance with clinical guideline recommendations.

Other more overt indicators of inappropriate prescribing, such as the provision of long-acting

opioids among opioid naïve claimants, should also be examined in future research.

The analyses undertaken in this thesis were also not intended to identify patterns of claimant

misuse and abuse of opioids, such as doctor shopping. There is certainly a great deal of concern

about the potential for this issue in this population, but it not yet clear to what extent these

concerns have borne out in reality. A few studies have documented opioid-related accidental

248

overdoses among workers’ compensation claimants,(21-23) but less is known about misuse,

abuse, and addiction. In a recent study of workers’ compensation claimants with various

injuries, 0.5% of opioid-treated claimants between 2004 and 2013 had at least one medical

claim for opioid abuse, dependence, or poisoning within 12 months of claim onset.(24) Another

small exploratory study of 100 claimants found 8% required intervention for an opioid-related

disorder and 46% screened positive for misuse.(25) Further research in this area is critically

needed to better understand the extent of aberrant opioid use in this population and its

determinants in order to be able to identify claimants at highest risk and implement

appropriate preventative measures.

The focus of this dissertation was mainly on dispensing in the early period after injury.

However, the findings of Manuscript 2 (Chapter 5) demonstrated that a small group of

claimants continue to receive dispenses for at least one of these three drug classes almost a

year after injury. Little research has been done to examine the determinants of persistent

exposure after a work-related injury and all have focused on opioids.(26-29) In one study, early

opioid provision was been found to be associated with long-term use,(29) but another study

found no association.(27) Long-term users may be particularly vulnerable to opioid-related

harms and identifying the factors that may contribute to persistent use among claimants merits

greater research attention.

249

8.3.3 Additional Factors That Could Influence Prescribing and Work Disability Trajectory

8.3.3.1 Claimant-Level Factors

Despite the fact that this and other studies have focused primarily on individual factors as

potential determinants of early dispensing, there remains an opportunity to examine additional

claimant contributors to drug use patterns that could also act as confounders in the association

between prescription opioid use and work disability. As previously described, confounding by

indication and severity is a crucial source of confounding in this dissertation, as well as previous

studies. More detailed data on the pain experience is needed for future studies. In addition to

severity and pain measures, other important injury-related psychosocial factors should be

considered, including recovery expectations and self-efficacy. These have all been shown to be

associated with prolonged work disability (30, 31) and endorsed by clinicians as contributing

factors to opioid utilization in injured workers.(32) However, these have not been examined

empirically in this population.

Our results suggest that workers who have a history of seeking alternative health care may be

less likely to fill prescriptions, including opioid prescriptions, for their work injuries. Researchers

of a previous study using insurance claim data found that early LBP care predominantly

provided by a chiropractor led to reduced use of all types of medications, including opioids,

while care predominantly provided by a physiotherapist led to higher rates of use of all three

drug classes.(33) Another recent study of LBP claimants in Ontario who sought initial care from

a chiropractor were at a lower risk of prolonged work disability than those who first saw a

physician and were less likely to seek other care within a month after injury.(34) However, this

same study also found individuals seeing physiotherapists as the first health care provider

250

experienced longer episodes of disability. Future studies should account for this type of care

using more comprehensive data and clarify how care may influence prescribing and subsequent

disability, as this may represent an important avenue for workers’ compensation organizations

to pursue in order to reduce non-evidence based and potentially harmful pharmacologic use in

injured workers.

The use of high dimensional propensity scores to adjust for claimant-level confounding may

also be considered in future studies using administrative data.(35) This methodology can be

used to essentially mine administrative health claims data to identify a set of proxies that

indirectly measure health status. This method is advantageous as it considers all health claims

data to define health status and is not limited to a subset of predefined covariates, thereby

potentially identifying confounding that is unknown to the investigator. The use of high

dimensional propensity score analysis has been previously shown to result in improved effect

estimates compared to adjustment limited to investigator-defined covariates.(35-38)

8.3.3.2 Prescriber-Level Factors

In my analysis of the determinants of early prescription dispensing (Chapter 6), I found that the

prescriber accounts for anywhere between 25% to 36% of the variability in the odds that the

first dispense after injury includes an opioid. Certain prescriber characteristics were also strong

risk factors of the drug class at first dispense, namely prescriber sex, birth year, and specialty.

However, adding these prescriber-level factors to the model did little to explain between-

prescriber variation. Clearly, much of the choice in the first dispense is due to the prescriber

and to factors that could not be captured in these data. Further research is needed to better

251

understand the prescriber factors that act as determinants of early post-injury prescribing in

this population. Prescriber comfort, satisfaction, and prior experience with prescribing, as well

as prescriber perceptions about risks have been previously shown to influence prescribing

behaviour in a primary care population,(39) but were not available in these data. Some prior

research has also pointed to the influence of interactions between prescriber and patient

characteristics in the prescribing of opioids. In one study of emergency department patients,

male physicians were more likely to prescribe opioids to male patients, and vice versa for

women.(40) In another study of Medicaid beneficiaries with chronic non-cancer pain, a

significant interaction between prescriber specialty and patient race was identified.(41)

The role of the treating physician in the relationship between prescription opioids and

prolonged work disability has also not yet been investigated. In addition to providing medical

treatment, physicians often play the role of gatekeeper in a workers’ compensation context and

provide recommendations to workers and the insurer regarding a worker’s capacity to return to

work. Similar to prescribing, physicians likely vary in how they manage workers’ compensation

patients and how cautious they are in their recommendations for return to work and activity.

On one hand, it may be possible that a physician who is aggressive with treatment may also be

aggressive with recommending an early return-to-work. On the other hand, a physician who

feels the pain of a LBP injury necessitates opioids may also be more inclined to recommend a

greater need for recovery time. Future research in this area would benefit from considering the

role of the physicians involved in the prescribing and management of worker’s compensated

injuries in influencing the work disability trajectory of a claimant.

252

8.3.3.3 System- and Workplace-Level Factors

The data for this dissertation did not include information on important system or workplace

factors that could influence pharmacologic treatment patterns and work disability in this

population. One previous study examining early opioids and work disability adjusted for

whether the employee worked for an employer with a return-to-work program, but data on

how the employee received this program were not available.(42)

The findings of several qualitative studies begin to provide some indication of what factors may

be important. In one qualitative study of physicians and pharmacists in Ontario, participants

were asked to identify the factors influencing opioid utilization among individuals with work-

related injuries and illnesses.(32) Among the factors highlighted are those that could be

considered unique or magnified in the care of workers’ compensation patients. This included

systemic-third party factors, such as claim-related delays and pressure applied by the

compensation system to the worker (whether real or perceived). Work environment factors

included unsupportive employer attitudes and a lack of workplace accommodations. Some of

these factors have also been shown to be associated with work disability.(30, 31)

Physicians also cited time pressures and an inability to obtain non-pharmacologic treatments as

important determinants. A lack of time and resources has previously been shown to be a major

barrier to guideline-concordant opioid management practices in non-compensation populations

(43) and physicians have often reported on the additional administrative burden associated

with care of workers’ compensation claimants.(44, 45) Several qualitative studies have also

shown that many systemic barriers exist with respect to timely access of alternative treatments

253

and receiving quality care from practitioners offering alternative therapies within the context of

workers’ compensation.(44, 46, 47)

Other qualitative research has also suggested a potential role for return-to-work policies in

medication use among injured workers. During the study period, a number of workers’

compensation organizations had implemented early return-to-work policies. WorkSafeBC had

their own guideline entitled “RTW is Good Medicine”, promoting an early return to safe and

modified duties.(48) However, in MacEachen et al.’s study of injured workers experiencing

difficulties with return-to-work, researchers found workers often mentioned using excessive

medication, namely opioids, to comply and cope with early return-to-work requirements.(49)

These studies highlight the critical need to expand beyond claimant-level factors to identify

some of the workplace and system-level factors that may play a role in a physician’s decision-

making process around opioid treatment, as these factors may also contribute to work

disability.

8.3.4 Causal Mechanism Linking Opioids and Work Disability

As with previous studies in this area, it was not possible to explore the potential underlying

causal mechanism linking opioids and work disability. To date, there have only been theories

postulated for how opioids may negatively influence work outcomes among injured workers.

Some prior research has shown an association between early imaging after a workplace injury

and prolonged disability, with the inference that early imaging may be an indicator of a pattern

of aggressive treatment leading to iatrogenic disability.(50-52) The early provision of opioids

could certainly be part and parcel with this pattern. If physicians are willing to prescribe a

254

strong analgesic, such as an opioid, claimants may perceive this as an indication of the severity

of their injuries. This could increase sick role behaviours, leading claimants to be more cautious

in returning to their normal activities and thereby promoting prolonged disability.

Opioid-related side effects, particularly when a worker is opioid naïve, could also impact a

worker’s capacity for working. Use in the early stages could also lead to prolonged use, as seen

by at least one group of researchers,(29) making claimants potentially more susceptible to

tolerance and physical dependence, escalating doses to achieve the same degree of analgesia,

and eventually opioid abuse and addiction,(53) which could certainly influence a claimant’s

work trajectory. As previously mentioned, little is known about abuse and addiction among

workers’ compensation claimants, but based on the few studies to date,(24, 25) this is likely to

account for a very small group of claimants. Future research is needed to delve further into

these theories in order to clarify the underlying mechanism driving any potential relationship.

8.4 Conclusions and Implications for Policy and Practice

To date, opioid policies in workers’ compensation have been largely informed by the findings of

previous studies demonstrating an association between early opioid use and work disability,

which I have shown to be problematic. This work, which addressed many of the preceding

problems, demonstrated similar findings. However, the strength of the associations seen was

attenuated in my comparative analyses where only claimants with at least one dispense were

included. The findings also appear to suggest that the association may be partially explained by

other factors, such as severity of injury, pre-injury drug exposure, and co-occurring prescription

drugs and health care. However, residual confounding by indication and severity likely still

255

accounts for some of the relationships seen.

Given the findings of this dissertation, true cause and effect between the use of early

prescription opioids and prolonged work disability cannot yet be assumed. There is a need for

further high quality research that addresses confounding by indication and severity and

identifies the possible mechanisms linking opioid exposure to work disability. Tackling the issue

of opioid use among injured workers and understanding its impacts can also not be

accomplished without a greater understanding of the contribution of system, workplace, and

prescriber factors.

Until such knowledge is acquired, insurers and clinicians must remain vigilant and cautious in

prescribing opioids to injured workers with LBP, while continuing to ensure pain is adequately

treated. Policies aimed at reducing opioid use among injured workers have been implemented

with the best of intentions, but often with no parallel change in policy to improve access to

other non-pharmacologic alternatives, such as chiropractic or massage therapy. Efforts to

combat the opioid crisis currently plaguing workers’ compensation cannot be focused solely on

reducing opioid exposure, but must also ensure timely access to alternative pain-relief

measures.

256

8.5 References

1. Carnide N, Hogg-Johnson S, Côté P, Irvin E, Van Eerd D, Koehoorn M, et al. Early Prescription Opioid Use for Musculoskeletal Disorders and Work Outcomes: A Systematic Review of the Literature. Clin J Pain. 2017;33(7):647-658.

2. Boden LI, Ozonoff A. Capture-recapture estimates of nonfatal workplace injuries and illnesses. Ann Epidemiol. 2008;18(6):500-6.

3. Fan ZJ, Bonauto DK, Foley MP, Silverstein BA. Underreporting of work-related injury or illness to workers' compensation: individual and industry factors. J Occup Environ Med. 2006;48(9):914-22.

4. Rosenman KD, Gardiner JC, Wang J, Biddle J, Hogan A, Reilly MJ, et al. Why most workers with occupational repetitive trauma do not file for workers' compensation. Journal of occupational and environmental medicine / American College of Occupational and Environmental Medicine. 2000;42(1):25-34.

5. Stock S, Nicolakakis N, Raiq H, Messing K, Lippel K, Turcot A. Underreporting work absences for nontraumatic work-related musculoskeletal disorders to workers' compensation: results of a 2007-2008 survey of the Quebec working population. Am J Public Health. 2014;104(3):e94-e101.

6. Mustard CA, Chambers A, McLeod C, Bielecky A, Smith PM. Comparison of data sources for the surveillance of work injury. Occup Environ Med. 2012;69(5):317-24.

7. WorkSafeBC. WorkSafeBC 2015 Statistics. Richmond, BC: Workers Compensation Board Province of British Columbia; 2015.





12. Thumula V, Wang D, Liu T-C. Interstate Variations in Use of Opioids, 3rd Edition. Cambridge, Massachusetts: Workers Compensation Research Institute; 2016.

257

13. Lix LM, Yogendran MS, Shaw SY, Burchill C, Metge C, Bond R. Population-based data sources for chronic disease surveillance. Chronic Dis Can. 2008;29(1):31-8.

14. Tonelli M, Wiebe N, Fortin M, Guthrie B, Hemmelgarn BR, James MT, et al. Methods for identifying 30 chronic conditions: application to administrative data. BMC Med Inform Decis Mak. 2015;15:31.

15. Frank JW, Brooker AS, DeMaio SE, Kerr MS, Maetzel A, Shannon HS, et al. Disability resulting from occupational low back pain. Part II: What do we know about secondary prevention? A review of the scientific evidence on prevention after disability begins Spine. 1996;21(24):2918-29.

16. Wasiak R, Kim J, Pransky G. Work disability and costs caused by recurrence of low back pain: longer and more costly than in first episodes. Spine (Phila Pa 1976). 2006;31(2):219-25.

17. Pole JD, Franche RL, Hogg-Johnson S, Vidmar M, Krause N. Duration of work disability: a comparison of self-report and administrative data. Am J Ind Med. 2006;49(5):394-401.

18. Hegmann KT, Weiss MS, Bowden K, Branco F, DuBrueler K, Els C, et al. ACOEM practice guidelines: opioids and safety-sensitive work. J Occup Environ Med. 2014;56(7):e46-53.

19. Bültmann U, Franche RL, Hogg-Johnson S, Côté P, Lee H, Severin C, et al. Health status, work limitations, and return-to-work trajectories in injured workers with musculoskeletal disorders. Qual Life Res. 2007;16(7):1167-78.

20. Lötters F, Franche RL, Hogg-Johnson S, Burdorf A, Pole JD. The prognostic value of depressive symptoms, fear-avoidance, and self-efficacy for duration of lost-time benefits in workers with musculoskeletal disorders. Occupational & Environmental Medicine. 63(12):794-801. 2006.

21. Cheng M, Sauer B, Johnson E, Porucznik C, Hegmann K. Comparison of opioid-related deaths by work-related injury. Am J Ind Med. 2013;56(3):308-16.


23. Fulton-Kehoe D, Garg RK, Turner JA, Bauer AM, Sullivan MD, Wickizer TM, et al. Opioid poisonings and opioid adverse effects in workers in Washington state. Am J Ind Med. 2013;56(12):1452-62.

24. Johnston SS, Alexander AH, Masters ET, Mardekian J, Semel D, Malangone-Monaco E, et al. Costs and Work Loss Burden of Diagnosed Opioid Abuse Among Employees on Workers Compensation or Short-term Disability. J Occup Environ Med. 2016;58(11):1087-97.

258

25. Parhami I, Hyman M, Siani A, Lin S, Collard M, Garcia J, et al. Screening for addictive disorders within a workers' compensation clinic: an exploratory study. Subst Use Misuse. 2012;47(1):99-107.


27. Heins SE, Feldman DR, Bodycombe D, Wegener ST, Castillo RC. Early opioid prescription and risk of long-term opioid use among US workers with back and shoulder injuries: a retrospective cohort study. Inj Prev. 2016;22(3):211-5.

28. Shafer LA, Raymond C, Ekuma O, Kraut A. The impact of opioid prescription dose and duration during a workers compensation claim, on post-claim continued opioid use: A retrospective population-based study. Am J Ind Med. 2015;58(6):650-7.


30. Cancelliere C, Donovan J, Stochkendahl MJ, Biscardi M, Ammendolia C, Myburgh C, et al. Factors affecting return to work after injury or illness: best evidence synthesis of systematic reviews. Chiropr Man Therap. 2016;24(1):32.

31. Steenstra I, Irvin E, Mahood Q, Hogg-Johnson S, Heymans MW. Systematic Review of Prognostic Factors for Workers’ Time Away from Work Due to Acute Low-Back Pain: An Update of a Systematic Review. Final Report to Workers Compensation Board of Manitoba. Toronto, ON: Institute for Work & Health; 2011.

32. Parsons JA, Mamdani M, Bhattacharyya O, Fortin CM, Melo M, Salmon C, et al. Narcotic analgesic utilization amongst injured workers: using concept mapping to understand current issues from the perspectives of physicians and pharmacists. BMC Health Serv Res. 2011;11:280.

33. Allen H, Wright M, Craig T, Mardekian J, Cheung R, Sanchez R, et al. Tracking low back problems in a major self-insured workforce: toward improvement in the patient's journey. J Occup Environ Med. 2014;56(6):604-20.

34. Blanchette MA, Rivard M, Dionne CE, Hogg-Johnson S, Steenstra I. Association Between the Type of First Healthcare Provider and the Duration of Financial Compensation for Occupational Back Pain. J Occup Rehabil. 2016. E-pub ahead of print.

35. Schneeweiss S, Rassen JA, Glynn RJ, Avorn J, Mogun H, Brookhart MA. High-dimensional propensity score adjustment in studies of treatment effects using health care claims data. Epidemiology. 2009;20(4):512-22.

259

36. Guertin JR, Rahme E, Dormuth CR, LeLorier J. Head to head comparison of the propensity score and the high-dimensional propensity score matching methods. BMC Med Res Methodol. 2016;16:22.

37. Guertin JR, Rahme E, LeLorier J. Performance of the high-dimensional propensity score in adjusting for unmeasured confounders. Eur J Clin Pharmacol. 2016;72(12):1497-505.

38. Garbe E, Kloss S, Suling M, Pigeot I, Schneeweiss S. High-dimensional versus conventional propensity scores in a comparative effectiveness study of coxibs and reduced upper gastrointestinal complications. Eur J Clin Pharmacol. 2013;69(3):549-57.

39. Wenghofer EF, Wilson L, Kahan M, Sheehan C, Srivastava A, Rubin A, et al. Survey of Ontario primary care physicians' experiences with opioid prescribing. Can Fam Physician. 2011;57(3):324-32.

40. Safdar B, Heins A, Homel P, Miner J, Neighbor M, DeSandre P, et al. Impact of physician and patient gender on pain management in the emergency department--a multicenter study. Pain Med. 2009;10(2):364-72.

41. Ringwalt C, Gugelmann H, Garrettson M, Dasgupta N, Chung AE, Proescholdbell SK, et al. Differential prescribing of opioid analgesics according to physician specialty for Medicaid patients with chronic noncancer pain diagnoses. Pain Res Manag. 2014;19(4):179-85.


43. Krebs EE, Bergman AA, Coffing JM, Campbell SR, Frankel RM, Matthias MS. Barriers to guideline-concordant opioid management in primary care--a qualitative study. J Pain. 2014;15(11):1148-55.

44. Kosny A, MacEachen E, Ferrier S, Chambers L. The role of health care providers in long term and complicated workers' compensation claims. J Occup Rehabil. 2011;21(4):582-90.

45. Russell G, Brown JB, Stewart M. Managing injured workers: family physicians' experiences. Can Fam Physician. 2005;51:78-9.

46. Hudon A, Laliberte M, Hunt M, Feldman DE. Quality of physiotherapy services for injured workers compensated by workers' compensation in Quebec: a focus group study of physiotherapy professionals. Healthc Policy. 2015;10(3):32-47.

47. Beardwood BA, Kirsh B, Clark NJ. Victims twice over: perceptions and experiences of injured workers. Qual Health Res. 2005;15(1):30-48.

48. WorkSafeBC. RTW is Good Medicine. Richmond, BC: WorkSafeBC.

260

49. MacEachen E, Ferrier S, Kosny A, Chambers A. A deliberation on 'hurt versus harm' logic in early-return-to-work policy. Policy and Practice in Health and Safety. 2007;5(2):41-62.

50. Graves JM, Fulton-Kehoe D, Jarvik JG, Franklin GM. Early imaging for acute low back pain: one-year health and disability outcomes among Washington State workers. Spine (Phila Pa 1976). 2012;37(18):1617-27.

51. Webster BS, Cifuentes M. Relationship of early magnetic resonance imaging for work-related acute low back pain with disability and medical utilization outcomes. J Occup Environ Med. 2010;52(9):900-7.

52. Webster BS, Bauer AZ, Choi Y, Cifuentes M, Pransky GS. Iatrogenic consequences of early magnetic resonance imaging in acute, work-related, disabling low back pain. Spine (Phila Pa 1976). 2013;38(22):1939-46.

53. Chou R, Turner JA, Devine EB, Hansen RN, Sullivan SD, Blazina I, et al. The effectiveness and risks of long-term opioid therapy for chronic pain: a systematic review for a National Institutes of Health Pathways to Prevention Workshop. Ann Intern Med. 2015;162(4):276-86.

261

APPENDIX A SUPPLEMENTARY TABLES FOR MANUSCRIPT 1 (CHAPTER 3)

262

Appendix Table A-1 Examples of work outcome measures eligible for inclusion in the systematic review

Work Outcome Category Examples

Generic Work Outcomes1 work disability functional ability

functional limitations functional capacity functional impairment physical functioning functional status physical functioning scales with specific work components (e.g., The Disabilities of the Arm, Shoulder, and Hand (DASH) Scale)

Absenteeism-Related and Work work status time to return-to-work

Status Outcomes duration of

unemployment occurrence of work/sickness

absence work status transitions duration of work/sickness absence

return-to-work

trajectories recurrences of work/sickness

absence return-to-work status

Disability Payment Outcomes Receipt of (status of), duration of, recurrences of:

disability benefits long-term disability short-term disability sickness benefits

wage replacement

benefits disability pension

Work Functioning Outcomes Occurrence of, duration of, changes in (improvements, decrements):

work limitations work functioning presenteeism work capacity work performance work interference work productivity work impairment

1Applicable only if the outcome refers to some aspect of work.

263

Appendix Table A-2 Sample search strategy for the systematic review

1 exp Analgesics, Opioid/ 2 exp Narcotics/ 3 opioid$.mp. 4 opiate$.mp. 5 narcotic$.mp. 6 buprenorphine/ 7 buprenorphine.mp. 8 codeine/ 9 codeine.mp. 10 fentanyl/ 11 fentanyl.mp. 12 hydrocodone/ 13 hydrocodone.mp. 14 hydromorphone/ 15 hydromorphone.mp. 16 dihydromorphinone.mp. 17 levorphanol/ 18 levorphanol.mp. 19 meperidine/ 20 meperidine.mp. 21 pethidine.mp. 22 methadone/ 23 methadone.mp. 24 morphine/ 25 morphine.mp. 26 oxycodone/ 27 oxycodone.mp. 28 oxymorphone/ 29 oxymorphone.mp. 30 pentazocine/ 31 pentazocine.mp. 32 dextropropoxyphene/ 33 dextropropoxyphene.mp. 34 propoxyphene.mp. 35 tramadol/ 36 tramadol.mp. 37 or/1-36 38 exp back/ 39 exp spine/ 40 thoracic vertebrae/ 41 cervical vertebrae/ 42 lumbar vertebrae/ 43 coccyx/ 44 intervertebral disk/ 45 sacrum/ 46 lumbosacral region/

47 exp neck/ 48 exp upper extremity/ 49 upper extremit$.mp. 50 upper limb$.mp. 51 Shoulder/ 52 Arm/ 53 Elbow/ 54 Forearm/ 55 Wrist/ 56 exp Hand/ 57 exp lower extremity/ 58 lower extremit$.mp. 59 lower limb$.mp. 60 hip/ 61 leg/ 62 thigh/ 63 knee/ 64 ankle/ 65 exp foot/ 66 exp musculoskeletal system/ 67 musculoskeletal.mp. 68 exp joints/ 69 exp cartilage/ 70 exp fascia/ 71 exp ligaments/ 72 exp muscles/ 73 exp "bone and bones"/ 74 exp tendons/ 75 pain/ 76 "Wounds and Injuries"/ 77 exp "Sprains and Strains"/ 78 sprain$.mp. 79 strain$.mp. 80 tear.mp. 81 torn.mp. 82 Athletic Injuries/ 83 athletic injur$.mp. 84 sports injur$.mp. 85 Back pain/ 86 Low back pain/ 87 exp back injuries/ 88 exp spinal injuries/ 89 Dorsalgia.mp. 90 Lumbago.mp. 91 coccydynia.mp. 92 exp neck pain/

264

93 exp neck injuries/ 94 shoulder pain/ 95 exp arm injuries/ 96 exp forearm injuries/ 97 wrist injuries/ 98 exp hand injuries/ 99 exp tendon injuries/ 100 exp hip injuries/ 101 exp leg injuries/ 102 exp knee injuries/ 103 ankle injuries/ 104 foot injuries/ 105 exp dislocations/ 106 dislocat$.mp. 107 fractures, bone/ 108 exp femoral fractures/ 109 fractures, closed/ 110 fractures, comminuted/ 111 fractures, compression/ 112 fractures, malunited/ 113 fractures, open/ 114 fractures, stress/ 115 fractures, ununited/ 116 humeral fractures/ 117 exp radius fractures/ 118 shoulder fractures/ 119 spinal fractures/ 120 tibial fractures/ 121 exp ulna fractures/ 122 fracture$.mp. 123 occupational diseases/ 124 occupational health/ 125 accidents, occupational/ 126 injured worker$.mp. 127 injured employee$.mp. 128 (work$ adj2 injur$).mp. 129 (work$ adj2 disorder$).mp. 130 (work$ adj2 accident$).mp. 131 (job adj2 injur$).mp. 132 (job adj2 disorder$).mp. 133 (job adj2 accident$).mp. 134 (occupation$ adj2 injur$).mp. 135 (occupation$ adj2 disorder$).mp. 136 (occupation$ adj2 accident$).mp. 137 (industrial adj2 injur$).mp. 138 (industrial adj2 disorder$).mp. 139 (industrial adj2 accident$).mp. 140 exp Arthralgia/

141 exp Bursitis/ 142 Brachial Plexus Neuritis/ 143 exp brachial plexus neuropathies/ 144 Carpal Tunnel Syndrome/ 145 exp complex regional pain syndromes/ 146 Cubital Tunnel Syndrome/ 147 exp Cumulative Trauma Disorders/ 148 De Quervain.mp. 149 dis? herniat$.mp. 150 dis? prolapse.mp. 151 epicondylitis.mp. 152 fasciitis/ 153 Ganglion Cysts/ 154 Hand-Arm Vibration Syndrome/ 155 herniated dis?.mp. 156 Intervertebral Disk Displacement/ 157 joint diseases/ 158 joint instability/ 159 metatarsalgia/ 160 exp mononeuropathies/ 161 musculoskeletal diseases/ 162 (musculoskeletal adj (disorder$ or injur$ or condition$ or pain)).mp. 163 exp nerve compression syndromes/ 164 Neuralgia/ 165 Neuritis/ 166 exp osteoarthritis/ 167 osteoarthr$.mp. 168 overuse syndrome.mp. 169 patellofemoral pain syndrome/ 170 plantar fasciitis/ 171 Polyradiculoneuropathy/ 172 polyradiculopathy/ 173 Radiculopathy/ 174 (repetitive adj (strain or stress or motion)).mp. 175 sciatica/ 176 Shoulder Impingement Syndrome/ 177 soft tissue injuries/ 178 (soft tissue adj (disorder$ or injur$)).mp. 179 spinal stenosis/ 180 spondylosis/ 181 spondylolysis/ 182 spondylolisthesis/ 183 synovial cysts/ 184 Synovitis/ 185 tarsal tunnel syndrome/ 186 Tendinopathy/

265

187 Tennis Elbow/ 188 exp Tenosynovitis/ 189 ten?ovaginitis.mp. 190 tend?nitis.mp. 191 exp Thoracic Outlet Syndrome/ 192 exp Ulnar Nerve Compression Syndromes/ 193 whiplash injuries/ 194 or/38-193 195 work/ 196 employment/ 197 employment.mp. 198 unemployment/ 199 unemploy$.mp. 200 occupation$.mp. 201 (labo?r force adj2 participation).mp. 202 (work$ adj2 participation).mp. 203 (job adj2 participation).mp. 204 (occupation$ adj2 participation).mp. 205 ((employee$ or employment) adj2 participation).mp. 206 work$ status.mp. 207 job status.mp. 208 occupational status.mp. 209 employment status.mp. 210 (work$ adj2 disab$).mp. 211 (job adj2 disab$).mp. 212 (occupation$ adj2 disab$).mp. 213 ((employee$ or employment) adj2 disab$).mp. 214 absenteeism/ 215 absenteeism.mp. 216 (work$ adj2 absen$).mp. 217 (job adj2 absen$).mp. 218 (occupation$ adj2 absen$).mp. 219 ((employee$ or employment) adj2 absen$).mp. 220 sick leave/ 221 sick$ leave$.mp. 222 sick$ absen$.mp. 223 sick$ list$.mp. 224 sick$ related leave$.mp. 225 sick$ related absen$.mp. 226 disability leave$.mp. 227 disability absen$.mp. 228 medical leave$.mp. 229 time los#.mp. 230 lost time.mp.

231 lost work$ day$.mp. 232 lost workday$.mp. 233 lost work$ time.mp. 234 lost worktime.mp. 235 missed work$ day$.mp. 236 missed workday$.mp. 237 lost work.mp. 238 missed work.mp. 239 work los#.mp. 240 job los#.mp. 241 loss of employment.mp. 242 loss of work.mp. 243 work$ day los#.mp. 244 workday los#.mp. 245 work$ time los#.mp. 246 worktime los#.mp. 247 return$ to work.mp. 248 return$ to employment.mp. 249 Return-to-work.mp. 250 RTW.mp. 251 reemployment.mp. 252 re-employment.mp. 253 (work$ adj2 resum$).mp. 254 (job adj2 resum$).mp. 255 (occupation$ adj2 resum$).mp. 256 ((employee$ or employment) adj2 resum$).mp. 257 (work$ adj2 (re-entry or re-enter$)).mp. 258 (work$ adj2 (reentry or reenter$)).mp. 259 (job adj2 (re-entry or re-enter$)).mp. 260 (job adj2 (reentry or reenter$)).mp. 261 (occupation$ adj2 (re-entry or re-enter$)).mp. 262 (occupation$ adj2 (reentry or reenter$)).mp. 263 ((employee$ or employment) adj2 (re-entry or re-enter$)).mp. 264 ((employee$ or employment) adj2 (reentry or reenter$)).mp. 265 (work$ adj2 integration).mp. 266 (work$ adj2 re-integration).mp. 267 (work$ adj2 reintegration).mp. 268 (job adj2 integration).mp. 269 (job adj2 re-integration).mp. 270 (job adj2 reintegration).mp. 271 (occupation$ adj2 integration).mp. 272 (occupation$ adj2 re-integration).mp. 273 (occupation$ adj2 reintegration).mp.

266

274 ((employee$ or employment) adj2 integration).mp. 275 ((employee$ or employment) adj2 re-integration).mp. 276 ((employee$ or employment) adj2 reintegration).mp. 277 ((labo?r market or labo?r force) adj2 (re-entry or re-enter$)).mp. 278 ((labo?r market or labo?r force) adj2 (reentry or reenter$)).mp. 279 (disability adj2 management).mp. 280 (disability adj2 prevention).mp. 281 workers' compensation/ 282 work$ compensation.mp. 283 time on benefit$.mp. 284 (benefit$ adj2 duration).mp. 285 wage replacement.mp. 286 exp Insurance, Disability/ 287 disability insurance.mp. 288 disability claim$.mp. 289 disability pension.mp. 290 disability compensation.mp. 291 disability payment$.mp. 292 (disability adj2 duration).mp. 293 short term disability.mp. 294 long term disability.mp. 295 disability benefit$.mp. 296 sick$ benefit$.mp. 297 functional status.mp. 298 (function$ adj2 limitation$).mp. 299 (function$ adj2 impairment$).mp. 300 functional capacity.mp. 301 functional abilit$.mp. 302 (functional abilit$ adj2 evaluation$).mp. 303 (functional capacity adj2 assessment$).mp. 304 (functional capacity adj2 evaluation$).mp. 305 (function$ adj2 assessment$).mp. 306 (function$ adj2 evaluation$).mp. 307 functional outcome$.mp. 308 (function$ adj2 restoration).mp. 309 physical capacity.mp. 310 physical function$.mp. 311 (work$ adj2 abilit$).mp. 312 (job adj2 abilit$).mp. 313 (occupation$ adj2 abilit$).mp.

314 ((employee$ or employment) adj2 abilit$).mp. 315 workability.mp. 316 ability to work.mp. 317 employability.mp. 318 (work$ adj2 capacity).mp. 319 (job adj2 capacity).mp. 320 (occupation$ adj2 capacity).mp. 321 ((employee$ or employment) adj2 capacity).mp. 322 (work$ adj2 capabilit$).mp. 323 (job adj2 capabilit$).mp. 324 (occupation$ adj2 capabilit$).mp. 325 ((employee$ or employment) adj2 capabilit$).mp. 326 Work Capacity Evaluation/ 327 work$ capacity evaluation$.mp. 328 exp Disability Evaluation/ 329 (work$ adj2 limitation$).mp. 330 (job adj2 limitation$).mp. 331 (occupation$ adj2 limitation$).mp. 332 ((employee$ or employment) adj2 limitation$).mp. 333 (work$ adj2 impairment$).mp. 334 (job adj2 impairment$).mp. 335 (occupation$ adj2 impairment$).mp. 336 ((employee$ or employment) adj2 impairment$).mp. 337 (work$ adj2 incapacity).mp. 338 (job adj2 incapacity).mp. 339 (occupation$ adj2 incapacity).mp. 340 ((employee$ or employment) adj2 incapacity).mp. 341 (work$ adj2 incapabilit$).mp. 342 (job adj2 incapabilit$).mp. 343 (occupation$ adj2 incapabilit$).mp. 344 ((employee$ or employment) adj2 incapabilit$).mp. 345 (work$ adj2 function$).mp. 346 (job adj2 function$).mp. 347 (occupation$ adj2 function$).mp. 348 ((employee$ or employment) adj2 function$).mp. 349 (work$ adj2 outcome$).mp. 350 (job adj2 outcome$).mp. 351 (occupation$ adj2 outcome$).mp. 352 (employment adj2 outcome$).mp. 353 presenteeism.mp.

267

354 (work$ adj2 performance).mp. 355 (job adj2 performance).mp. 356 (occupation$ adj2 performance).mp. 357 ((employee$ or employment) adj2 performance).mp. 358 efficiency/ 359 (work$ adj2 productivity).mp. 360 (job adj2 productivity).mp. 361 (occupation$ adj2 productivity).mp. 362 (employee$ adj2 productivity).mp. 363 (work$ adj2 adjustment$).mp. 364 (job adj2 adjustment$).mp. 365 (occupation$ adj2 adjustment$).mp. 366 ((employee$ or employment) adj2 adjustment$).mp. 367 (work$ adj2 adaptation$).mp. 368 (job adj2 adaptation$).mp. 369 (occupation$ adj2 adaptation$).mp. 370 ((employee$ or employment) adj2 adaptation$).mp.

371 worksite.mp. 372 work site.mp. 373 Workplace/ 374 modified work.mp. 375 modified dut$.mp. 376 (work$ adj2 accommodation$).mp. 377 (job adj2 accommodation$).mp. 378 (occupation$ adj2 accommodation$).mp. 379 ((employe$ or employment) adj2 accommodation$).mp. 380 light dut$.mp. 381 light work.mp. 382 graduated hours.mp. 383 flexible hours.mp. 384 flexible work.mp. 385 alternative work.mp. 386 graded work.mp. 387 or/195-386 388 37 and 194 and 387 389 limit 388 to english

268

Appendix Table A-3 Items considered in the risk of bias assessment Domain Item Question

Selection Bias

Source population Was the source population clearly defined and described?

Sampling and recruitment Were the sampling frame and recruitment methods clearly described and comparable for all exposure groups?

Inclusion/exclusion criteria Were sample inclusion and exclusion criteria clearly described and appropriate?

Study participation

Were participation rates and reasons for refusal to participate clearly reported? Were characteristics of participants and non-participants described and comparable?

Randomization sequence1 Was the method of generating the randomization sequence reported and adequate?

Allocation concealment1

Was random allocation to treatment groups adequately concealed from participants and investigators?

Attrition Bias

Attrition rates and reasons for withdrawals

Were attrition rates and reasons for study withdrawal clearly reported?

Factors associated with attrition

Were characteristics of those who completed the study and those who withdrew from the study clearly described and similar?

Measurement Bias – Opioid Exposure

Definition of opioid use Was the construct of opioid use clearly described and defined, including the method(s) of measurement?

Valid and reliable measurement

Was the construct of opioid use measured using valid and reliable methods?

Standard data collection procedures

Was data collection with respect to opioid use carried out equivalently for all participants?

Performance Bias1

Intervention description Were the interventions clearly described and defined?

Equal treatment Apart from the allocated intervention, were treatment groups treated equally?

Blinding Were participants blinded to treatment assignment?

Co-interventions Were co-interventions monitored and comparable for all treatment groups?

Contamination Was contamination between treatment groups monitored and described?

Compliance Was treatment compliance monitored and comparable for all treatment groups?

Measurement Bias – Work Outcome(s)

Definition of work outcome(s) Were the work outcomes clearly described and defined, including the method(s) of measurement?


Were the work outcomes measured using valid and reliable methods?

Blinded outcome assessment Was data collection with respect to work outcomes carried out blindly?


Was data collection with respect to work outcomes carried out equivalently for all participants?

Confounding

Selection of confounders Did the authors justify the potential confounders selected for the analysis?

Definition of potential confounders

Were the potential confounders clearly described and defined, including the method(s) of measurement?


Were the confounders measured using valid and reliable methods?

269

Domain Item Question


Was data collection with respect to confounders carried out equivalently for all participants?

Baseline characteristics Were baseline characteristics of study participants presented and comparable between exposure groups?

Controlling confounding Were all important confounders appropriately controlled?

Analysis

Appropriate statistical analyses

Were statistical analyses appropriate for the design of the study?

Reporting of results Were results appropriately reported?

Intention-to-treat1 Were intention-to-treat analyses carried out? 1 For randomized controlled trials only

270

APPENDIX B METHODS APPENDICES FOR MANUSCRIPTS 2 TO 4 (CHAPTERS 5 TO 7)

271

Appendix Table B-1 ICD-9 diagnoses used to identify LBP-related claims in the WorkSafeBC claim file

ICD-9 Codes1 ICD-9 Diagnosis

722 Intervertebral disc disorders

722.0 Displacement of cervical intervertebral disc without myelopathy

722.1 Displacement of thoracic or lumbar intervertebral disc without myelopathy

722.10 Displacement of lumbar intervertebral disc without myelopathy

722.11 Displacement of thoracic intervertebral disc without myelopathy

722.2 Displacement of intervertebral disc, site unspecified, without myelopathy

722.3 Schmorl’s nodes

722.30 Schmorl’s nodes, site unspecified

722.31 Schmorl’s nodes, thoracic region

722.32 Schmorl's nodes, lumbar region

722.39 Schmorl's nodes, other region

722.4 Degeneration of cervical intervertebral disc

722.5 Degeneration of thoracic or lumbar intervertebral disc

722.51 Degeneration of thoracic or thoracolumbar intervertebral disc

722.52 Degeneration of lumbar or lumbosacral intervertebral disc

722.6 Degeneration of intervertebral disc, site unspecified

722.7 Intervertebral disc disorder with myelopathy

722.70 Intervertebral disc disorder with myelopathy, site unspecified

722.71 Intervertebral disc disorder with myelopathy, cervical region

722.72 Intervertebral disc disorder with myelopathy, thoracic region

722.73 Intervertebral disc disorder with myelopathy, lumbar region

722.8 Postlaminectomy syndrome

722.80 Postlaminectomy syndrome, Unspecified region

722.81 Postlaminectomy syndrome, cervical region

722.82 Postlaminectomy syndrome, thoracic region

722.83 Postlaminectomy syndrome, lumbar region

272


722.9 Other and unspecified disc disorder

722.90 Other and unspecified disc disorder, site unspecified

722.91 Other and unspecified disc disorder, cervical region

722.92 Other and unspecified disc disorder, thoracic region

722.93 Other and unspecified disc disorder, lumbar region

724 Other and unspecified disorders of back

724.0 Spinal stenosis other than cervical

724.00 Spinal stenosis, unspecified region

724.01 Spinal stenosis, thoracic region

724.02 Spinal stenosis, Lumbar region, without neurogenic claudication

724.03 Spinal stenosis, lumbar region, with neurogenic claudication

724.09 Spinal stenosis, other

724.1 Pain in thoracic spine

724.2 Lumbago

724.3 Sciatica

724.4 Thoracic or lumbosacral neuritis or radiculitis, unspecified

724.5 Backache, unspecified

724.6 Disorders of sacrum

724.7 Disorders of coccyx

724.70 Unspecified disorder of coccyx

724.71 Hypermobility of coccyx

724.79 Disorders of coccyx, other

724.8 Other symptoms referable to back

724.9 Other unspecified back disorders

273


846 Sprains and strains of sacroiliac region

846.0 Sprains and strains of sacroiliac region, Lumbosacral (joint) (ligament)

846.1 Sprains and strains of sacroiliac region, Sacroiliac ligament

846.2 Sprains and strains of sacroiliac region, Sacrospinatus (ligament)

846.3 Sprains and strains of sacroiliac region, Sacrotuberous (ligament)

846.8 Sprains and strains of sacroiliac region, Other specified sites of sacroiliac region

846.9 Sprains and strains of sacroiliac region, Unspecified site of sacroiliac region

847 Sprains and strains of other and unspecified parts of back

847.0 Sprain of neck

847.1 Sprain of thoracic

847.2 Sprains and strains of other and unspecified parts of back, Lumbar

847.3 Sprains and strains of other and unspecified parts of back, Sacrum

847.4 Sprains and strains of other and unspecified parts of back, Coccyx

847.9 Sprains and strains of other and unspecified parts of back, Unspecified

1 Codes in bold are used to identify LBP-related claims in the WorkSafeBC claim file. ICD-9 diagnoses are coded to 4 digits in the WorkSafeBC claim file. Claims were included with ICD-9

codes at the 3 and 4 digit level. However, it’s unclear whether 722.0 refers to 722 or 722.0 (neck-related). Similarly, it’s not clear whether 847.0 refers to 847 or 847.0 (also neck-related).

A cross-tab of ICD-9 codes with body part injured suggests that most of 722.0 is coding neck-related injuries and 847.0 is coding either neck or other injuries. The decision was made to

exclude 722.0 and 847.0 from the list of eligible ICD-9 diagnoses in the process of defining the study cohort.

Abbreviations: ICD-9=International Classification of Diseases, 9th Revision

274

Appendix Table B-2 ICD-9 and ICD-10-CA diagnoses used to exclude claimants with LBP-related hospitalizations or serious LBP-related outpatient health care visits and/or hospitalizations within 5 days of injury date

ICD-9

Code1,2 ICD-9 Diagnosis

ICD-10-CA

Code1,2 ICD-10-CA Diagnosis

170 Malignant neoplasm of bone and articular cartilage C40

C41

Malignant neoplasm of bone and articular cartilage of limbs

Malignant neoplasm of bone and articular cartilage of other and

unspecified sites

171 Malignant neoplasm of connective and other soft tissue C49 Malignant neoplasm of other connective and soft tissue

192 Malignant neoplasm of other and unspecified parts of

nervous system

C70

C72

Malignant neoplasm of meninges

Malignant neoplasm of spinal cord, cranial nerves, and other parts of

central nervous system

195 Malignant neoplasm of other and ill-defined sites C76 Malignant neoplasm of other and ill-defined sites

198 Secondary malignant neoplasm of other specified sites C79 Secondary malignant neoplasm of other and unspecified sites

213 Benign neoplasm of bone and articular cartilage D16 Benign neoplasm of bone and articular cartilage

215 Other benign neoplasm of connective and other soft

tissue

D21 Other benign neoplasms of connective and other soft tissue

225 Benign neoplasm of brain and other parts of nervous

system

D32

D33

Benign neoplasm of meninges

Benign neoplasm of brain and other parts of central nervous system

237 Neoplasm of uncertain behavior of endocrine glands

and nervous system

D42

D43

D44

Neoplasm of uncertain or unknown behaviour of meninges

Neoplasm of uncertain or unknown behaviour of brain and central

nervous system

Neoplasm of uncertain or unknown behaviour of endocrine glands

238 Neoplasm of uncertain behaviour of other and

unspecified sites and tissues

D45

D47

D48

Polycythaemia vera

Other neoplasms of uncertain or unknown behaviour of lymphoid,

haematopoietic and related tissue

Neoplasm of uncertain or unknown behaviour of other and

unspecified sites

239 Neoplasms of unspecified nature D48 Neoplasm of uncertain or unknown behaviour of other and

unspecified sites

275

ICD-9


ICD-10-CA


344 Other paralytic syndromes G83 Other paralytic syndromes

353 Nerve root and plexus disorders G54 Nerve root and plexus disorders

720 Ankylosing spondylitis and other inflammatory

spondylopathies

M45

M46

Ankylosing spondylitis

Other inflammatory spondylopathies

721 Spondylosis and allied disorders M47

M48

Spondylosis

Other spondylopathies

7223 Intervertebral disc disorders M50 Cervical disc disorders

M51

M96

Other intervertebral disc disorders

Postprocedural musculoskeletal disorders, not elsewhere classified

7243 Other and unspecified disorders of back M43

M46

M48

M53

M54

Other deforming dorsopathies

Other inflammatory spondylopathies

Other spondylopathies

Other dorsopathies, not elsewhere classified

Dorsalgia

733 Other disorders of bone and cartilage M80

M81

M84

M85

M87

M89

M94

Osteoporosis with pathological fracture

Osteoporosis without pathological fracture

Disorders of continuity of bone

Other disorders of bone density and structure

Osteonecrosis

Other disorders of bone

Other disorders of cartilage

738 Other acquired deformity M95 Other acquired deformities of musculoskeletal system and

connective tissue

805 Fracture of vertebral column without mention of spinal

cord injury

S12

S22

S32

Fracture of neck

Fracture of rib(s), sternum and thoracic spine

Fracture of lumbar spine and pelvis

806 Fracture of vertebral column with spinal cord injury S12 Fracture of neck

276

ICD-9


ICD-10-CA


S14

S22

S24

S32

S34

Injury of nerves and spinal cord at neck level

Fracture of rib(s), sternum and thoracic spine

Injury of nerves and spinal cord at thorax level

Fracture of lumbar spine and pelvis

Injury of nerves and spinal cord at abdomen, lower back, and pelvis

level

839 Other, multiple, and ill-defined dislocations S13

S23

S33

S43

Dislocation, sprain and strain of joints and ligaments at neck level

Dislocation, sprain and strain of joints and ligaments of thorax

Dislocation, sprain and strain of joints and ligaments of lumbar spine

and pelvis

Dislocation, sprain and strain of joints and ligaments of shoulder

girdle

8463 Sprains and strains of sacroiliac region S33 Dislocation, sprain and strain of joints and ligaments of lumbar spine

and pelvis

8473 Sprains and strains of other and unspecified parts of

back

S13

S16

S23

S39

Dislocation, sprain and strain of joints and ligaments at neck level

Injury of muscle and tendon at neck level

Dislocation, sprain and strain of joints and ligaments of thorax

Other and unspecified injuries of abdomen, lower back and pelvis

876 Open wound of back S21

S31

Open wound of thorax

Open wound of abdomen, lower back and pelvis

877 Open wound of buttock S31 Open wound of abdomen, lower back and pelvis

926 Crushing injury of trunk S38

S28

T21

Crushing injury and traumatic amputation of part of abdomen, lower

back and pelvis

Crushing injury of thorax and traumatic amputation of part of thorax

Burn and corrosion of trunk

277

ICD-9


ICD-10-CA


952 Spinal cord injury without evidence of spinal bone

injury

S14

S24

S34



Injury of nerves and spinal cord at abdomen, lower back, pelvis level

953 Injury to nerve roots and spinal plexus S14

S24

S34



Injury of nerves and spinal cord at abdomen, lower back, and pelvis

level 1 Codes are used to identify and exclude claimants with LBP-related hospitalizations or serious LBP-related outpatient healthcare visits within five days of injury date in the Medical

Services Plan (MSP) and Discharge Abstract Database (DAD). ICD-9 codes in MSP data are only considered valid to 3 digits. To maintain consistency with the codes used in the MSP file,

only 3-digit ICD-10-CA codes are also used. 2Both ICD-9 and ICD-10-CA codes are available in the DAD. Corresponding ICD-10-CA codes have been mapped to the appropriate ICD-9 codes. CIHI provided backward conversions for

ICD-10-CA to ICD-9 codes from 2001/2002 to 2007/2008 and these conversions were used to help guide the mapping. 3Only applies to hospitalizations.

Abbreviations: ICD-9=International Classification of Diseases, 9th Revision; ICD-10-CA=International Classification of Diseases, 10th Revision, Canada; LBP=low back pain

278

Appendix Table B-3 Specific prescription drugs included in our study according to drug class MAIN DRUG CLASSES ADJUVANT ANALGESICS

Drug Class Drugs Drug Class Drugs

Opioids Anileridine Methadone (pain only) 1 Antidepressants Amitriptyline2 Maprotiline2 Buprenorphine1 Morphine1 Clomipramine2 Nortriptyline2 Butorphanol1 Oxycodone1 Desipramine2 Protriptyline2 Codeine Pentazocine Desvenlafaxine Trazodone2 Fentanyl1 Propoxyphene Doxepin2 Trimipramine2 Hydromorphone1 Tramadol Duloxetine Venlafaxine Meperidine Imipramine2 Nonsteroidal Anti-Inflammatory Drugs

ASA Meloxicam Sedative Hypnotics Celecoxib Nabumetone Benzodiazepines Alprazolam Lorazepam

Diclofenac Naproxen Bromazepam Midazolam Diflunisal Oxaprozin Chlordiazepoxide Nitrazepam Etodolac Phenylbutazone Clorazepate Oxazepam Fenoprofen Piroxicam Diazepam Temazepam Floctafenine Rofecoxib Flurazepam Triazolam Flurbiprofen Salsalate Ibuprofen Sulindac Z-Drugs Zaleplon Indomethacin Tenoxicam Zopiclone Ketoprofen Tiaprofenic Acid Ketorolac Tolmetin Sodium Lumiracoxib Valdecoxib Mefenamic Acid Skeletal Muscle Relaxants

Baclofen Anticonvulsants Carbamazepine Botox Clonazepam4

Carisoprodol Gabapentin Chlorzoxazone Lamotrigine Cyclobenzaprine Phenytoin Dantrolene Pregabalin Methocarbamol Topiramate Tizanidine Valproic Acid

279

1Considered to be strong opioids in study analyses. 2These antidepressants were only included if daily dose≤100mg/day in an attempt to include only antidepressants acting as adjuvants for pain management, rather than depression

treatment 3Drugs were assigned to drug class categories according to the American Hospital Formulary System (AHFS). While clonazepam is a benzodiazepine, it is classified as an anticonvulsant in

the AHFS. It is also classified as an anticonvulsant in the World Health Organization’s Anatomical Therapeutic Chemical (ATC) classification system.

280

Appendix Table B-4 Ratios used to convert opioid daily dose to a standardized morphine-equivalent dose (MED)

Opioid Equivalence to oral morphine 30mg Morphine-equivalent ratio to convert dose to oral morphine equivalent

Anileridine 75 mg 0.4 Buprenorphine 0.273 mg 110 Codeine 200 mg 0.15 Hydromorphone 7.5 mg 4 Meperidine 300 mg 0.1 Morphine 30 mg 1.0 Oxycodone 20 mg 1.5 Pentazocine 180 mg 0.17 Propoxyphene HCL 130 mg 0.23 Propoxyphene napsylate 200 mg 0.15 Tramadol 300 mg 0.1 Fentanyl1,2 1-60 mg morphine=12 mcg/hr 104.2 60-134 mg morphine=25 mcg/hr 161.7 135-179 mg morphine=37 mcg/hr 176.8 180-224 mg morphine=50 mcg/hr 168.3 225-269 mg morphine=62 mcg/hr 270-314 mg morphine=75 mcg/hr 162.2 315-359 mg morphine=87 mcg/hr 360-404 mg morphine=100 mcg/hr 159.2

Abbreviations: hr=hour; mcg=microgram; mg=milligram 1Includes only fentanyl strengths available in the study data. 2Midpoint of morphine range used to estimate the morphine-equivalent ratio (e.g., 97mg morphine = 25 mcg/hr of fentanyl)

281

Appendix Table B-5 ICD-9 and ICD-10-CA diagnoses used to identify comorbid health conditions in the two years before injury date Comorbidity Reference(s) ICD-91,2 ICD-9 Code Description ICD-10-CA1,2 ICD-10 Code Description

Arthritis / Rheumatism and Other Joint Disorders (Excluding Neck and Back)

(1) 274 Gout M00 Pyogenic arthritis 446 Polyarteritis nodosa and allied

conditions M01 Direct infections of joint in infectious and

parasitic diseases classified elsewhere 710 Diffuse diseases of connective tissue M02 Reactive arthropathies 711 Arthropathy associated with infections M03 Postinfective and reactive arthropathies in

diseases classified elsewhere 712 Crystal arthropathies M05 Seropositive rheumatoid arthritis 713 Arthropathy associated with other

disorders M06 Other rheumatoid arthritis

714 Rheumatoid arthritis and other inflammatory polyarthropathies

M07 Psoriatic and enteropathic arthropathies

715 Osteoarthrosis and allied disorders M10 Gout 716 Other and unspecified arthropathies M11 Other crystal arthropathies 717 Internal derangement of knee M12 Other specific arthopathies 718 Other derangement of joint M13 Other arthritis 719 Other and unspecified disorders of

joint M14 Arthropathies in other diseases classified

elsewhere 725 Polymyalgia rheumatic M15 Polyarthrosis 726 Peripheral enthesopathies and allied

syndromes M16 Coxarthrosis

727 Other disorders of synovium, tendon, and bursa

M17 Gonarthrosis

728 Disorders of muscle, ligament, and fascia

M18 Arthrosis of first carpometacarpal joint

729 Other disorders of soft tissues M19 Other arthrosis M22 Disorders of patella M23 Internal derangement of knee M24 Other specific joint derangements M25 Other joint disorders, not elsewhere classified M30 Polyarteritis nodosa and related conditions M31 Other necrotizing vasculopathies

282

Comorbidity Reference(s) ICD-91,2 ICD-9 Code Description ICD-10-CA1,2 ICD-10 Code Description

M32 Systemic lupus erythematosus M33 Dermatopolymyositis M34 Systemic sclerosis M35 Other systemic involvement of connective

tissue M36 Systemic disorders of connective tissue in

diseases classified elsewhere M60 Myositis M61 Calcification and ossification of muscle M62 Other disorders of muscle M63 Disorders of muscle in diseases classified

elsewhere M65 Synovitis and tenosynovitis M66 Spontaneous rupture of synovium and tendon M67 Other disorders of synovium and tendon M68 Disorders of synovium and tendon in diseases

classified elsewhere M70 Soft tissue disorders related to use, overuse

and pressure M71 Other bursopathies M72 Fibroblastic disorders M73 Soft tissue disorders in diseases classified

elsewhere M75 Shoulder lesions M76 Enthesopathies of lower limb, excluding foot M77 Other enthesopathies M79 Other soft tissue disorders, not elsewhere

classified Back and Neck Pain

(2-4) 720 Ankylosing spondylitis and other inflammatory spondylopathies

M40 Kyphosis and lordosis

721 Spondylosis and allied disorders M41 Scoliosis 722 Intervertebral disc disorders M43 Other deforming dorsopathies

283


723 Other disorders of cervical region M45 Ankylosing spondylitis 724 Other and unspecified disorders of

back M46 Other inflammatory spondylopathies

737 Curvature of spine M47 Spondylosis 738.2

738.4 738.5

Acquired deformity of neck Acquired spondylolisthesis Other acquired deformity of back or spine

M48 Other spondylopathies

739.1 739.2 739.3 739.4

Cervical region Thoracic region Lumbar region Sacral region

M49 Spondylopathies in diseases classified elsewhere

805 Fracture of vertebral column without mention of spinal cord

M50 Cervical disc disorders

839.0 839.1 839.2 839.3 839.4 839.5

Cervical vertebra, closed dislocation Cervical vertebra, open dislocation Thoracic and lumbar vertebra, closed dislocation Thoracic and lumbar vertebra, open dislocation Other vertebra, closed dislocation Other vertebra, open dislocation

M51 Other intervertebral disc disorders

846 Sprains and strains of sacroiliac region M53 Other dorsopathies, not elsewhere classified 847 Sprains and strains of other and

unspecified parts of back M54 Dorsalgia

M95.3 Acquired deformity of neck 905.1 Late effect of fracture of spine and

trunk without mention of spinal cord lesion

M96 Postprocedural musculoskeletal disorders, not elsewhere classified

M99 Biomechanical lesions, not elsewhere classified

S12 Fracture of neck

284


S13 Dislocation, sprain and strain of joints and ligaments at neck level

S16 Injury of muscle and tendon at neck level S22.0 Fracture of thoracic vertebra S22.1 Multiple fractures of thoracic spine S23 Dislocation, sprain and strain of joints and

ligaments of lumbar spine and pelvis S32.0 Fracture of lumbar vertebra S32.1 Fracture of sacrum S32.3 Fracture of coccyx S33 Dislocation, sprain and strain of joints and

ligaments of lumbar spine and pelvis S39 Other and unspecified injuries of abdomen,

lower back, and pelvis T91.1 Sequelae of fracture of spine Neuropathic Pain

(3) 053.1 Herpes zoster with other nervous system complications

B02.2 Zoster with other nervous system complications

250.6 Diabetes with neurological manifestations

E10.4 Type 1 diabetes mellitus with neurological complications

337.2 Reflex sympathetic dystrophy E11.4 Type 2 diabetes mellitus with neurological complications

350 Trigeminal nerve disorders E12.4 Malnutrition-related diabetes mellitus with neurological complications

352.1 Glosspharyngeal neuralgia E13.4 Other specified diabetes mellitus with neurological complications

353 Nerve root and plexus disorders E14.4 Unspecified diabetes mellitus with neurological complications

354 Mononeuritis of upper limb and mononeuritis multiplex

G50 Disorders of trigeminal nerve

355 Mononeuritis of lower limb G52.1 Disorders of glossopharyngeal nerve 357 Inflammatory and toxic neuropathy G53.0 Postzoster neuralgia 733.7 Algoneurodystrophy G54 Nerve root and plexus disorders

285


G55 Nerve root and plexus compressions classified elsewhere

G56 Mononeuropathies of upper limb G57 Mononeuropathies of lower limb G58 Other mononeuropathies G59 Mononeuropathy in diseases classified

elsewhere G61 Inflammatory polyneuropathy G62 Other polyneuropathies G63 Polyneuropathy in diseases classified

elsewhere G90 Disorders of autonomic nervous system G99 Other disorders of nervous system in diseases

classified elsewhere M89.0 Algoneurodystrophy Migraine and Headache Disorders

(5, 6) 307.8 Tension headache G43 Migraine 339 Other headache syndromes

(*introduced in 2008) G44 Other headache syndromes

346 Migraine R51 Headache 784.0 Headache Mood and Anxiety Disorders

(7-9) 296 Episodic mood disorders F30 Manic episode 300 Anxiety, dissociative, and somatoform

disorders F31 Bipolar affective disorder

308 Acute reaction to stress F32 Depressive disorder 309 Adjustment reaction F33 Recurrent depressive disorder 311 Depressive disorder, not elsewhere

classified F34 Persistent mood (affective) disorders

50B Depression/anxiety (*only for MSP) F38 Other mood (affective) disorders F39 Unspecified mood (affective) disorder F40 Phobic anxiety disorders F41 Other anxiety disorders

286


F42 Obsessive-compulsive disorder F43 Reaction to severe stress, and adjustment

disorders F99 Mental disorder, not otherwise specified Intentional Self-Harm or Suicide Attempt

(9-12) E950 Suicide and self-inflicted poisoning by solid or liquid substances

X60 Intentional self-poisoning by and exposure to nonopioid analgesics, antipyretics and antirheumatics

E951 Suicide and self-inflicted poisoning by gases in domestic use

X61 Intentional self-poisoning by and exposure to antiepileptic, sedative-hypnotic, antiparkinsonism and psychotropic drugs, not elsewhere classified

E952 Suicide and self-inflicted poisoning by other gases and vapors

X62 Intentional self-poisoning by and exposure to narcotics and psychodysleptics [hallucinogens], not elsewhere classified

E953 Suicide and self-inflicted poisoning by hanging, strangulation, and suffocation

X63 Intentional self-poisoning by and exposure to other drugs acting on the autonomic nervous system

E954 Suicide and self-inflicted poisoning by submersion [drowning]

X64 Intentional self-poisoning by and exposure to other and unspecified drugs, medicaments and biological substances

E955 Suicide and self-inflicted poisoning by firearms, air guns and explosives

X65 Intentional self-poisoning by and exposure to alcohol

E956 Suicide and self-inflicted poisoning by cutting and piercing instrument

X66 Intentional self-poisoning by and exposure to organic solvents and halogenated hydrocarbons and their vapours

E957 Suicide and self-inflicted poisoning by jumping from high place

X67 Intentional self-poisoning by and exposure to other gases and vapours

E958 Suicide and self-inflicted poisoning by other and unspecified means

X68 Intentional self-poisoning by and exposure to pesticides

E959 Late effects of self-inflicted injury X69 Intentional self-poisoning by and exposure to other and unspecified chemicals and noxious substances

287


E980 Poisoning by solid or liquid substances, undetermined whether accidentally or purposely inflicted

X70 Intentional self-harm by hanging, strangulation and suffocation

E981 Poisoning by gases in domestic use, undetermined whether accidentally or purposely inflicted

X71 Intentional self-harm by drowning and submersion

E982 Poisoning by other gases, undetermined whether accidentally or purposely inflicted

X72 Intentional self-harm by handgun discharge

E983 Hanging, strangulation, or suffocation, undetermined whether accidentally or purposely inflicted

X73 Intentional self-harm by rifle, shotgun and larger firearm discharge

E984 Submersion [drowning], undetermined whether accidentally or purposely inflicted

X74 Intentional self-harm by other and unspecified firearm discharge

E985 Injury by firearms, air guns and explosives, undetermined whether accidentally or purposely inflicted

X75 Intentional self-harm by explosive material

E986 Injury by cutting and piercing instruments, undetermined whether accidentally or purposely inflicted

X76 Intentional self-harm by smoke, fire and flames

E987 Falling from high place, undetermined whether accidentally or purposely inflicted

X77 Intentional self-harm by steam, hot vapours and hot objects

E988 Injury by other and unspecified means, undetermined whether accidentally or purposely inflicted

X78 Intentional self-harm by sharp object

E989 Late effects of injury, undetermined whether accidentally or purposely inflicted

X79 Intentional self-harm by blunt object

X80 Intentional self-harm by jumping from a high place

288


X81 Intentional self-harm by jumping or lying before moving object

X82 Intentional self-harm by crashing of motor vehicle

X83 Intentional self-harm by other specified means

X84 Intentional self-harm by unspecified means Y87 Sequelae of intentional self-harm, assault and

events of undetermined intent Y10 Poisoning by and exposure to nonopioid

analgesics, antipyretics and antirheumatics, undetermined intent

Y11 Poisoning by and exposure to antiepileptic, sedative-hypnotic, antiparkinsonism and psychotropic drugs, undetermined intent

Y12 Poisoning by and exposure to narcotics and psychodysleptics [hallucinogens], not elsewhere classified, undetermined intent

Y13 Poisoning by and exposure to other drugs acting on the autonomic nervous system, undetermined intent

Y14 Poisoning by and exposure to other and unspecified drugs, medicaments and biological substances, undetermined intent

Y15 Poisoning by and exposure to alcohol, undetermined intent

Y16 Poisoning by and exposure to organic solvents and halogenated hydrocarbons and their vapours, undetermined intent

Y17 Poisoning by and exposure to other gases and vapours, undetermined intent

Y18 Poisoning by and exposure to pesticides, undetermined intent

289


Y19 Poisoning by and exposure to other and unspecified chemicals and noxious substances, undetermined intent

Y20 Hanging strangulation and suffocation, undetermined intent

Y21 Drowning and submersion, undetermined intent

Y22 Handgun discharge, undetermined intent Y23 Rifle, shotgun and larger firearm discharge,

undetermined intent Y24 Other and unspecified firearm discharge,

undetermined intent Y25 Contact with explosive material,

undetermined intent Y26 Exposure to smoke, fire and flames,

undetermined intent Y27 Contact with steam, hot vapours and hot

objects, undetermined intent Y28 Contact with sharp object, undetermined

intent Y29 Contact with blunt object, undetermined

intent Y30 Falling, jumping or pushed from a high place,

undetermined intent Y31 Falling, lying or running before or into moving

object, undetermined intent Y32 Crashing of motor vehicle, undetermined

intent Y33 Other specified means, undetermined intent Y34 Unspecified event, undetermined intent Y87 Sequelae of intentional self-harm, assault and

events of undetermined intent

290


Substance Use Disorders

(8, 13) 291 Alcohol-induced mental disorders F10 Mental and behavioural disorders due to use of alcohol

292 Drug-induced mental disorders F11 Mental and behavioural disorders due to use of opioids

303 Alcohol dependence syndrome F12 Mental and behavioural disorders due to use of cannabinoids

304 Drug dependence F13 Mental and behavioural disorders due to use of sedatives or hypnotics

305 Nondependent abuse of drugs F14 Mental and behavioural disorders due to use of cocaine

F15 Mental and behavioural disorders due to use of other stimulants, including caffeine

F16 Mental and behavioural disorders due to use of hallucinogens

F17 Mental and behavioural disorders due to use of tobacco

F18 Mental and behavioural disorders due to use of volatile solvents

F19 Mental and behavioural disorders due to use of other psychoactive substances

F55 Abuse of non-dependence-producing substances

Diabetes (1, 14-17) 250 Diabetes mellitus E10 Type 1 diabetes mellitus E11 Type 2 diabetes mellitus E12 Malnutrition-related diabetes mellitus E13 Other specified diabetes mellitus E14 Unspecified diabetes mellitus Hypertension (18, 19) 401 Essential hypertension I10 Essential (primary) hypertension 402 Hypertensive heart disease I11 Hypertensive heart disease 403 Hypertensive chronic kidney disease I12 Hypertensive renal disease

291


404 Hypertensive heart and chronic kidney disease

I13 Hypertensive heart and renal disease

405 Secondary hypertension I15 Secondary hypertension Heart and Cerebrovascular Disease

(1, 20-26) 393 Chronic rheumatic pericarditis G45 Transient cerebral ischaemic attacks and related syndromes

394 Disease of mitral valve G46 Vascular syndromes of brain in cerebrovascular diseases

395 Diseases of aortic valve I05 Rheumatic mitral valve diseases 396 Disease of mitral and aortic valves I06 Rheumatic aortic valve diseases

397 Disease of other endocardial structures

I07 Rheumatic tricuspid valve diseases

398 Other rheumatic heart disease I08 Multiple valve diseases 410 Acute myocardial infarction I09 Other rheumatic heart disease 411 Other acute and subacute forms of

ischemic heart disease I20 Angina pectoris

412 Old myocardial infarction I21 Acute myocardial infarction 413 Angina pectoris I22 Subsequent myocardial infarction 414 Other forms of chronic ischemic heart

disease I23 Certain current complications following acute

myocardial infarction 416 Chronic pulmonary heart disease I24 Other acute ischaemic heart disease 424 Other disease of endocardium I25 Chronic ischaemic heart disease 425 Cardiomyopathy I27 Other pulmonary heart diseases 426 Conduction disorders I34 Nonrheumatic mitral valve disorders 427 Cardiac dysrhythmias I35 Nonrheumatic aortic valve disorders 428 Heart failure I36 Nonreheumatic tricuspid valve disorders 430 Subarachnoid hemorrhage I37 Pulmonary valve disorders 431 Intracerebral hemorrhage I39 Endocarditis and heart valve disorders in

diseases classified elsewhere 432 Other and unspecified intracranial

hemorrhage I42 Cardiomyopathy

433 Occlusion and stenosis of precerebral arteries

I43 Cardiomyopathy in diseases classified elsewhere

292


434 Occlusion of cerebral arteries I44 Atrioventricular and left bundle-branch block 435 Transient cerebral ischemia I45 Other conduction disorders 436 Acute, but ill-defined, cerebrovascular

disease I46 Cardiac arrest

437 Other and ill-defined cerebrovascular disease

I47 Paroxysmal tachycardia

438 Late effects of cerebrovascular disease I48 Atrial fibrillation and flutter I49 Other cardiac arrhythmias I50 Heart failure I60 Subarachnoid haemorrhage I61 Intracerebral haemorrhage I62 Other nontraumatic intracranial haemorrhage I63 Cerebral infarction I64 Stroke, not specified as haemorrhage or

infarction I65 Occlusion and stenosis of precerebral arteries,

not resulting in cerebral infarction I66 Occlusion and stenosis of cerebral arteries,

not resulting in cerebral infarction I67 Other cerebrovascular diseases I68 Cerebrovascular disorders in diseases

classified elsewhere I69 Sequelae of cerebrovascular disease Chronic Respiratory Disorders

(1, 17, 27-29) 491 Chronic bronchitis J41 Simple and mucopurulent chronic bronchitis 492 Emphysema J42 Unspecified chronic bronchitis 493 Asthma J43 Emphysema 496 Chronic airway obstruction, not

elsewhere classified J44 Other chronic obstructive pulmonary disease

518.8 Other diseases of lung J45 Asthma J46 Status asthmaticus J96 Respiratory failure, not elsewhere classified

293


Thyroid Disorders

(30) 242 Thyrotoxicosis with or without goiter E02 Subclinical iodine-deficiency hypothyroidism 244 Acquired hypothyroidism E03 Other hypothyroidism 245 Thyroiditis E05 Thyrotoxicosis (hyperthyroidism)

E06 Thyroiditis Chronic Kidney Disease

(31) 403 Hypertensive chronic kidney disease I12 Hypertensive renal disease 404 Hypertensive heart and chronic kidney

disease I13 Hypertensive heart and renal disease

581 Nephrotic syndrome N03 Chronic nephritic syndrome 582 Chronic glomerulonephritis N05 Unspecified nephritic syndrome 583 Nephritis and nephropathy, not

specified as acute or chronic N08 Glomerular disorders in diseases classified

elsewhere 585 Chronic kidney disease (CKD) N18 Chronic kidney disease 586 Renal failure, unspecified N19 Unspecified kidney failure 587 Renal sclerosis, unspecified N26 Unspecified contracted kidney 588 Disorders resulting from impaired

renal function N25 Disorders resulting from impaired renal

tubular function Chronic Liver Disease

(32-35) 070 Viral hepatitis B18 Chronic viral hepatitis 571 Chronic liver disease and cirrhosis K70 Alcoholic liver disease 572 Liver abscess and sequelae of chronic

liver disease K71 Toxic liver disease

573 Other disorders of liver K72 Hepatic failure, not elsewhere classified K73 Chronic hepatitis, not elsewhere classified K74 Fibrosis and cirrhosis of liver K75 Other inflammatory liver diseases K76 Other diseases of liver K77 Liver disorders in diseases classified

elsewhere Upper Gastrointestinal Disorders

(36-40) 530 Diseases of esophagus K20 Oesphagitis 531 Gastric ulcer K21 Gastro-oesophageal reflux disease 532 Duodenal ulcer K22 Other diseases of oesophagus

294


533 Peptic ulcer, site unspecified K25 Gastric ulcer 534 Gastrojejunal ulcer K26 Duodenal ulcer

535 Gastritis and duodenitis K27 Peptic ulcer, site unspecified 578 Gastrointestinal hemorrhage K28 Gastrojejunal ulcer K29 Gastritis and duodenitis K92 Other diseases of digestive system Inflammatory Bowel Disorders

(41, 42) 555 Regional enteritis K50 Crohn disease (regional enteritis) 556 Ulcerative colitis K51 Ulcerative colitis

Traumatic Brain Injury

(43-45) 310.2 Postconcussional syndrome F07.2 Postconcussional syndrome 800 Fracture of vault of skull S02 Fracture of skull and facial bones 801 Fracture of base of skull S06 Intracranial injury

802 Fracture of face bones S07 Crushing injury of head 803 Other and unqualified skull fractures T02.0 Fracture involving head with neck 804 Multiple fractures involving skull or

face with other bones T04.0 Crushing injury involving head with neck

850 Concussion T06.0 Injuries of brain and cranial nerves with injuries of nerves and spinal cord at neck level

851 Cerebral laceration and contusion T90 Sequelae of injuries of head 852 Subarachnoid, subdural, and

extradural haemorrhage, following injury

853 Other and unspecified intracranial haemorrhage following injury

854 Intracranial injury of other and unspecified nature

905.0 Late effect of fracture of skull and face bones

907.0 Late effect of intracranial injury without mention of skull fracture

907.1 Late effect of injury to cranial nerve 925.1 Crushing injury of face and scalp

295


959.0 Injury, other and unspecified, head, face and neck

Seizure Disorders

(46-48) 345 Epilepsy and recurrent seizures G40 Epilepsy 780.3 Convulsions G41 Status epilepticus

R56.8 Other and unspecified convulsions Sleep Apnea and Insomnia

(49-51) 307.4 Specific disorders of sleep of nonorganic origin

G47.0 Disorders of initiating and maintaining sleep [insomnias]

327.0 327.2

Organic disorders of initiating and maintaining sleep [organic insomnia] Organic sleep apnea

G47.3 Sleep apnoea

780.5 Sleep disturbances F51.0 Nonorganic insomnia 786.0 Dyspnea and respiratory abnormalities R06.8 Other and unspecified abnormalities of

breathing 1 Codes are used to identify comorbid health conditions in the Medical Services Plan (MSP) and Discharge Abstract Database (DAD). ICD-9 codes in MSP data are only considered valid to 3

digits. To maintain consistency with the codes used in the MSP file, only 3-digit ICD-10-CA codes are also used. In some cases, four digits are used if the three digit term is overly broad and

likely to result in severe misclassification. 2Both ICD-9 and ICD-10-CA codes are available in the DAD. Corresponding ICD-10-CA codes have been mapped to the appropriate ICD-9 codes. CIHI provided backward conversions for

ICD-10-CA to ICD-9 codes from 2001/2002 to 2007/2008 and these conversions were used to help guide the mapping.

Abbreviations: ICD-9=International Classification of Diseases, 9th Revision; ICD-10-CA=International Classification of Diseases, 10th Revision, Canada

296

Comorbidity References

1. Lix LM, Yogendran MS, Shaw SY, Burchill C, Metge C, Bond R. Population-based data sources for chronic disease surveillance. Chronic Dis Can. 2008;29(1):31-8.

2. Noonan VK, Thorogood NP, Fingas M, Batke J, Belanger L, Kwon BK, et al. The validity of administrative data to classify patients with spinal column and cord injuries. J Neurotrauma. 2013;30(3):173-80.

3. Lacasse A, Ware MA, Dorais M, Lanctot H, Choiniere M. Is the Quebec provincial administrative database a valid source for research on chronic non-cancer pain? Pharmacoepidemiol Drug Saf. 2015;24(9):980-90.

4. Sinnott PL, Siroka AM, Shane AC, Trafton JA, Wagner TH. Identifying neck and back pain in administrative data: defining the right cohort. Spine (Phila Pa 1976). 2012;37(10):860-74.

5. Kolodner K, Lipton RB, Lafata JE, Leotta C, Liberman JN, Chee E, et al. Pharmacy and medical claims data identified migraine sufferers with high specificity but modest sensitivity. J Clin Epidemiol. 2004;57(9):962-72.

6. Carlson KF, Taylor BC, Hagel EM, Cutting A, Kerns R, Sayer NA. Headache diagnoses among Iraq and Afghanistan war veterans enrolled in VA: a gender comparison. Headache. 2013;53(10):1573-82.

7. Canadian Institute for Health Information. Hospital Mental Health Services in Canada, 2005-2006. Ottawa, ON: CIHI; 2008. 8. Frayne SM, Miller DR, Sharkansky EJ, Jackson VW, Wang F, Halanych JH, et al. Using administrative data to identify mental illness: what

approach is best? Am J Med Qual. 2010;25(1):42-50. 9. Kisely S, Lin E, Gilbert C, Smith M, Campbell LA, Vasiliadis HM. Use of administrative data for the surveillance of mood and anxiety

disorders. Aust N Z J Psychiatry. 2009;43(12):1118-25. 10. Fransoo R, Martens P, Team TNtK, Prior H, Burchill C, Koseva I, et al. The 2013 RHA Indicators Atlas. Winnipeg, MB: Manitoba Centre for

Health Policy; 2013. 11. Schneeweiss S, Patrick AR, Solomon DH, Mehta J, Dormuth C, Miller M, et al. Variation in the risk of suicide attempts and completed

suicides by antidepressant agent in adults: a propensity score-adjusted analysis of 9 years' data. Arch Gen Psychiatry. 2010;67(5):497-506.

12. Iribarren C, Sidney S, Jacobs DR, Jr., Weisner C. Hospitalization for suicide attempt and completed suicide: epidemiological features in a managed care population. Soc Psychiatry Psychiatr Epidemiol. 2000;35(7):288-96.

13. Kim HM, Smith EG, Stano CM, Ganoczy D, Zivin K, Walters H, et al. Validation of key behaviourally based mental health diagnoses in administrative data: suicide attempt, alcohol abuse, illicit drug abuse and tobacco use. BMC Health Serv Res. 2012;12:18.

14. Southern DA, Roberts B, Edwards A, Dean S, Norton P, Svenson LW, et al. Validity of administrative data claim-based methods for identifying individuals with diabetes at a population level. Can J Public Health. 2010;101(1):61-4.

15. Hux JE, Ivis F, Flintoft V, Bica A. Diabetes in Ontario: determination of prevalence and incidence using a validated administrative data algorithm. Diabetes Care. 2002;25(3):512-6.

297

16. Chen G, Khan N, Walker R, Quan H. Validating ICD coding algorithms for diabetes mellitus from administrative data. Diabetes Res Clin Pract. 2010;89(2):189-95.

17. Borzecki AM, Wong AT, Hickey EC, Ash AS, Berlowitz DR. Identifying hypertension-related comorbidities from administrative data: what's the optimal approach? Am J Med Qual. 2004;19(5):201-6.

18. Tu K, Campbell NR, Chen ZL, Cauch-Dudek KJ, McAlister FA. Accuracy of administrative databases in identifying patients with hypertension. Open Med. 2007;1(1):e18-26.

19. Quan H, Khan N, Hemmelgarn BR, Tu K, Chen G, Campbell N, et al. Validation of a case definition to define hypertension using administrative data. Hypertension. 2009;54(6):1423-8.

20. Tamariz L, Harkins T, Nair V. A systematic review of validated methods for identifying ventricular arrhythmias using administrative and claims data. Pharmacoepidemiol Drug Saf. 2012;21 Suppl 1:148-53.

21. Schultz SE, Rothwell DM, Chen Z, Tu K. Identifying cases of congestive heart failure from administrative data: a validation study using primary care patient records. Chronic Dis Inj Can. 2013;33(3):160-6.

22. Joshy G, Korda RJ, Abhayaratna WP, Soga K, Banks E. Categorising major cardiovascular disease hospitalisations from routinely collected data. Public Health Res Pract. 2015;25(3):e2531532.

23. Tu K, Wang M, Young J, Green D, Ivers NM, Butt D, et al. Validity of administrative data for identifying patients who have had a stroke or transient ischemic attack using EMRALD as a reference standard. Can J Cardiol. 2013;29(11):1388-94.

24. Quan H, Sundararajan V, Halfon P, Fong A, Burnand B, Luthi JC, et al. Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Med Care. 2005;43(11):1130-9.

25. McCormick N, Bhole V, Lacaille D, Avina-Zubieta JA. Validity of Diagnostic Codes for Acute Stroke in Administrative Databases: A Systematic Review. PLoS One. 2015;10(8):e0135834.

26. Tu K, Mitiku T, Lee DS, Guo H, Tu JV. Validation of physician billing and hospitalization data to identify patients with ischemic heart disease using data from the Electronic Medical Record Administrative data Linked Database (EMRALD). Can J Cardiol. 2010;26(7):e225-8.

27. Lacasse Y, Montori VM, Lanthier C, Maltis F. The validity of diagnosing chronic obstructive pulmonary disease from a large administrative database. Can Respir J. 2005;12(5):251-6.

28. Gershon AS, Wang C, Guan J, Vasilevska-Ristovska J, Cicutto L, To T. Identifying patients with physician-diagnosed asthma in health administrative databases. Can Respir J. 2009;16(6):183-8.

29. Gershon AS, Wang C, Guan J, Vasilevska-Ristovska J, Cicutto L, To T. Identifying individuals with physcian diagnosed COPD in health administrative databases. COPD. 2009;6(5):388-94.

30. Marrie RA, Yu BN, Leung S, Elliott L, Warren S, Wolfson C, et al. The incidence and prevalence of thyroid disease do not differ in the multiple sclerosis and general populations: a validation study using administrative data. Neuroepidemiology. 2012;39(2):135-42.

298

31. Vlasschaert ME, Bejaimal SA, Hackam DG, Quinn R, Cuerden MS, Oliver MJ, et al. Validity of administrative database coding for kidney disease: a systematic review. Am J Kidney Dis. 2011;57(1):29-43.

32. Desai S, Peltekian KM. Canadian mortality rates for liver disease: taking a closer look at ICD coding. Can J Public Health. 2004;95(3):198-200.

33. Goldberg D, Lewis J, Halpern S, Weiner M, Lo Re V, 3rd. Validation of three coding algorithms to identify patients with end-stage liver disease in an administrative database. Pharmacoepidemiol Drug Saf. 2012;21(7):765-9.

34. Niu B, Forde KA, Goldberg DS. Coding algorithms for identifying patients with cirrhosis and hepatitis B or C virus using administrative data. Pharmacoepidemiol Drug Saf. 2015;24(1):107-11.

35. Kramer JR, Davila JA, Miller ED, Richardson P, Giordano TP, El-Serag HB. The validity of viral hepatitis and chronic liver disease diagnoses in Veterans Affairs administrative databases. Aliment Pharmacol Ther. 2008;27(3):274-82.

36. Lopushinsky SR, Covarrubia KA, Rabeneck L, Austin PC, Urbach DR. Accuracy of administrative health data for the diagnosis of upper gastrointestinal diseases. Surg Endosc. 2007;21(10):1733-7.

37. Curtis JR, Chen SY, Werther W, John A, Johnson DA. Validation of ICD-9-CM codes to identify gastrointestinal perforation events in administrative claims data among hospitalized rheumatoid arthritis patients. Pharmacoepidemiol Drug Saf. 2011;20(11):1150-8.

38. Pisa F, Castellsague J, Drigo D, Riera-Guardia N, Giangreco M, Rosolen V, et al. Accuracy of International Classification of Diseases, 9th Revision, Clinical Modification codes for upper gastrointestinal complications varied by position and age: a validation study in a cohort of nonsteroidal anti-inflammatory drugs users in Friuli Venezia Giulia, Italy. Pharmacoepidemiol Drug Saf. 2013;22(11):1195-204.

39. Raiford DS, Perez Gutthann S, Garcia Rodriguez LA. Positive predictive value of ICD-9 codes in the identification of cases of complicated peptic ulcer disease in the Saskatchewan hospital automated database. Epidemiology. 1996;7(1):101-4.

40. Castellsague J, Holick CN, Hoffman CC, Gimeno V, Stang MR, Perez-Gutthann S. Risk of upper gastrointestinal complications associated with cyclooxygenase-2 selective and nonselective nonsteroidal antiinflammatory drugs. Pharmacotherapy. 2009;29(12):1397-407.

41. Benchimol EI, Guttmann A, Mack DR, Nguyen GC, Marshall JK, Gregor JC, et al. Validation of international algorithms to identify adults with inflammatory bowel disease in health administrative data from Ontario, Canada. J Clin Epidemiol. 2014;67(8):887-96.

42. Rezaie A, Quan H, Fedorak RN, Panaccione R, Hilsden RJ. Development and validation of an administrative case definition for inflammatory bowel diseases. Can J Gastroenterol. 2012;26(10):711-7.

43. Ng R, Maxwell CJ, Yates EA, Nylen K, Antflick J, Jetté N, et al. Brain Disorders in Ontario: Prevalence, Incidence and Costs from Health Administrative Data. Toronto, ON: Institute for Clinical Evaluative Sciences; 2015.

44. Carroll CP, Cochran JA, Guse CE, Wang MC. Are we underestimating the burden of traumatic brain injury? Surveillance of severe traumatic brain injury using centers for disease control International classification of disease, ninth revision, clinical modification, traumatic brain injury codes. Neurosurgery. 2012;71(6):1064-70; discussion 70.

299

45. Bazarian JJ, Veazie P, Mookerjee S, Lerner EB. Accuracy of mild traumatic brain injury case ascertainment using ICD-9 codes. Acad Emerg Med. 2006;13(1):31-8.

46. Tu K, Wang M, Jaakkimainen RL, Butt D, Ivers NM, Young J, et al. Assessing the validity of using administrative data to identify patients with epilepsy. Epilepsia. 2014;55(2):335-43.

47. Reid AY, St Germaine-Smith C, Liu M, Sadiq S, Quan H, Wiebe S, et al. Development and validation of a case definition for epilepsy for use with administrative health data. Epilepsy Res. 2012;102(3):173-9.

48. Jette N, Reid AY, Quan H, Hill MD, Wiebe S. How accurate is ICD coding for epilepsy? Epilepsia. 2010;51(1):62-9. 49. Hermes E, Rosenheck R. Prevalence, pharmacotherapy and clinical correlates of diagnosed insomnia among Veterans Health

Administration service users nationally. Sleep Med. 2014;15(5):508-14. 50. Anderson LH, Whitebird RR, Schultz J, McEvoy CE, Kreitzer MJ, Gross CR. Healthcare utilization and costs in persons with insomnia in a

managed care population. Am J Manag Care. 2014;20(5):e157-65. 51. McIsaac DI, Gershon A, Wijeysundera D, Bryson GL, Badner N, van Walraven C. Identifying Obstructive Sleep Apnea in Administrative

Data: A Study of Diagnostic Accuracy. Anesthesiology. 2015;123(2):253-63.

300

Appendix Table B-6 CCI and CCP codes used to identify spinal x-rays in the DAD file and fee item codes used to identify spinal x-rays in the MSP file

CCP Code1 Description CCI Code1 Description Fee Item Description

02.11 02.12 02.13 02.19 16.83

X-ray of cervical spine X-ray of thoracic spine X-ray of lumbosacral spine Other x-ray of spine Contrast myelogram

3.AW.10 3.SC.10 3.SE.10 3.SF.10

Xray, spinal cord Xray, spinal vertebrae Xray, intervertebral disc Xray, sacrum and coccyx

8540 8541 8542 8543 8545 8546 8548 8549 19138 19139

X-ray – cervical spine X-ray – thoracic spine X-ray – lumbar spine X-ray – sacrum and coccyx X-ray – sacro-iliac joints Scoliosis film Myelogram and/or posterior fossa positive contrast X-ray – spine additional views (e.g., flexion, bending) WCB chiro spinal series WCB chiro spinal series recheck

1Both CCI and CCP codes are available in the DAD. Corresponding CCI codes have been mapped to the appropriate CCP codes. CIHI provided backward conversions for CCI to CCP codes from 2001/2002 to 2007/2008 and these conversions were used to help guide the mapping. Abbreviations: CCI=Canadian Classification of Health Interventions; CCP=Canadian Classification of Diagnostic, Therapeutic, and Surgical Procedures; CIHI=Canadian Institute for Health Information; DAD=Discharge Abstract Database; MSP=Medical Services Plan

301

Appendix Table B-7 CCI and CCP codes used to identify spinal surgeries in the DAD file

CCP Code1 Description CCI Code1 Description

16.02 Reopening of laminectomy site 16.09 Other exploration and decompression of spinal

canal

1.SA.80 1.SC.80 1.SF.80

Repair, atlas and axis Repair, spinal vertebrae Repair, sacrum and coccyx

16.1 16.2

Division of intraspinal nerve root Chordotomy

1.AW.59 Destruction, spinal cord

16.3

Excision or destruction of lesion of spinal cord and spinal meninges

1.AW.87 1.AX.87

Excision partial, spinal cord Excision partial, spinal canal and meninges

16.43 16.5

Repair of vertebral fracture Freeing of adhesions of spinal cord and nerve roots

1.AW.72 1.SF.73

Release, spinal cord Reduction, sacrum and coccyx

89.78 Other partial ostectomy, other specified site (pelvic bones, phalanges of foot or hand, vertebrae)

1.SC.87 1.SF.87

Excision partial, spinal vertebrae Excision partial, sacrum and coccyx

89.88 Total ostectomy, other specified site (pelvic bones, phalanges of foot or hand, vertebrae)

1.SA.89 1.SC.89 1.SF.89 1.SF.91

Excision total, atlas and axis Excision total, spinal vertebrae Excision total, sacrum and coccyx Excision radical, sacrum and coccyx

90.58 Internal fixation of bone (without fracture

reduction), other specified site (pelvic bones, phalanges of foot or hand, vertebrae)

1.SA.74 1.SC.74 1.SE.53 1.SF.74

Fixation, atlas and axis Fixation, spinal vertebrae Implantation of internal device, intervertebral disc Fixation, sacrum and coccyx

90.68 Removal of internal fixation device, other specified

site (pelvic bones, phalanges of foot or hand, vertebrae)

1.SA.55 1.SC.55 1.SE.55 1.SF.55

Removal of device, atlas and axis Removal of device or appliance, spinal vertebrae Removal of device, intervertebral disc Removal of device, sacrum and coccyx

302

CCP Code1 Description CCI Code1 Description

92.31 Excision (or destruction) of intervertebral disc

1.SE.59 1.SE.87 1.SE.89

Destruction, intervertebral disc Excision partial, intervertebral disc Excision total, intervertebral disc

93.01 93.02 93.03 93.04 93.05 93.06 93.07 93.08 93.09

Atlas-axis spinal fusion Other cervical spinal fusion Dorsal spinal fusion Dorsolumbar spinal fusion with Harrington rod Other dorsolumbar spinal fusion Lumbar spinal fusion Lumbosacral spinal fusion Refusion of spine Other spinal fusion, unspecified site

1.SA.75 1.SC.75 1.SF.75

Fusion, atlas and axis Fusion, spinal vertebrae Fusion, sacrum and coccyx

1Both CCI and CCP codes are available in the DAD. Corresponding CCI codes have been mapped to the appropriate CCP codes. CIHI provided backward conversions for CCI to CCP codes from 2001/2002 to 2007/2008 and these conversions were used to help guide the mapping. Abbreviations: CCI=Canadian Classification of Health Interventions; CCP=Canadian Classification of Diagnostic, Therapeutic, and Surgical Procedures; CIHI=Canadian Institute for Health Information; DAD=Discharge Abstract Database

303

APPENDIX C SUPPLEMENTARY FIGURES FOR MANUSCRIPT 2 (CHAPTER 5)

304

Appendix Figure C-1 Proportion of claimants dispensed their first opioid prescription by week after injury among claimants with at least one dispensed opioid prescription. Each bar represents a 1-week period post-injury.


305

Appendix Figure C-2 Proportion of claimants dispensed their first NSAID prescription by week after injury among claimants with at least one dispensed NSAID prescription. Each bar represents a 1-week period post-injury.


306

Appendix Figure C-3 Proportion of claimants dispensed their first SMR prescription by week after injury among claimants with at least one dispensed SMR prescription. Each bar represents a 1-week period post-injury. Note: Blue=first 8 weeks post-injury / yellow=9-52 weeks post-injury

307

Appendix Figure C-4 Count of dispensed opioid prescriptions in 8-week blocks from 48 weeks pre- to 48 weeks post-injury among claimants with at least one opioid prescription. Note: Blue=pre-injury / yellow=post-injury Abbreviations: Rxs=prescriptions

308

Appendix Figure C-5 Count of dispensed NSAID prescriptions in 8-week blocks from 48 weeks pre- to 48 weeks post-injury among claimants with at least one NSAID prescription. Note: Blue=pre-injury / yellow=post-injury Abbreviations: Rxs=prescriptions

309

Appendix Figure C-6 Count of dispensed SMR prescriptions in 8-week blocks from 48 weeks pre- to 48 weeks post-injury among claimants with at least one SMR prescription. Note: Blue=pre-injury / yellow=post-injury Abbreviations: Rxs=prescriptions

310

Appendix Figure C-7 Median cumulative days supply / claimant in 8-week blocks from 48 weeks pre- to 48 weeks post-injury for claimants with at least one day supplied of opioids and/or SMR with codeine.

311

APPENDIX D SUPPLEMENTARY TABLES FOR MANUSCRIPT 3 (CHAPTER 6)

312

Appendix Table D-1 Multinomial logistic regression analysis of the association between pre-injury supplementary health care and drug class(es) dispensed within 8 weeks of injury among claimants with injury years 1998-2001 (n=34,657) (no drugs as the reference, n=15,405)






SMRs (n=6,008)




SMRs (n=6,008)

Number of physiotherapy visits in the year before injury

0 29,544 (85.2) 1.00** 1.00 1.00 1.00*** 1.00 1.00 1-5 2,682 (7.7) 0.98 (0.90-1.08) 0.91 (0.78-1.06) 0.94 (0.84-1.05) 0.94 (0.85-1.03) 0.79 (0.67-0.92) 0.79 (0.70-0.89) ≥6 2,431 (7.0) 1.21 (1.10-1.34) 1.03 (0.88-1.21) 1.06 (0.95-1.20) 1.05 (0.94-1.17) 0.75 (0.63-0.89) 0.73 (0.63-0.83)

Number of chiropractor visits in the year before injury

0 27,615 (79.7) 1.00*** 1.00 1.00 1.00*** 1.00 1.00 1-5 3,673 (10.6) 0.74 (0.69-0.81) 0.77 (0.68-0.88) 0.74 (0.67-0.82) 0.78 (0.71-0.85) 0.75 (0.65-0.87) 0.71 (0.63-0.79) ≥6 3,369 (9.7) 0.68 (0.62-0.74) 0.82 (0.71-0.93) 0.69 (0.62-0.76) 0.71 (0.65-0.79) 0.75 (0.65-0.87) 0.63 (0.56-0.71)

Number of massage therapy visits in the year before injury

0 32,164 (92.8) 1.00** 1.00 1.00 1.00* 1.00 1.00 1-5 1,437 (4.1) 0.85 (0.75-0.97) 0.97 (0.80-1.18) 0.96 (0.83-1.11) 0.89 (0.78-1.02) 0.96 (0.78-1.19) 0.98 (0.84-1.16) ≥6 1,056 (3.0) 0.81 (0.70-0.94) 0.82 (0.65-1.04) 0.80 (0.67-0.96) 0.83 (0.71-0.97) 0.75 (0.58-0.96) 0.76 (0.63-0.93)

Abbreviations: CI=confidence interval; GP=general practitioner; ICD-9=International Classification of Diseases 9th Revision; NSAIDs=nonsteroidal anti-inflammatory drugs; OR=odds ratio; SMRs=skeletal muscle relaxants


313

Appendix Table D-2 Multilevel multinomial logistic regression analysis of factors associated with drug class(es) at first prescription within 8 weeks of injury among claimants with injury years 1998-2001 (n=18,777) (NSAIDs and/or SMRs only as the reference, n=11,449)




Opioids with NSAIDs and/or SMRs (n=3,387)


Opioids with NSAIDs and/or SMRs (n=3,387)



SMRs (n=3,387)

Number of physiotherapy visits in the year before injury

0 15,964 (85.0) 1.00* 1.00 1.00*** 1.00 1.00** 1.00 1-5 1,424 (7.6) 1.00 (0.87-1.15) 0.91 (0.79-1.06) 0.85 (0.72-1.02) 0.89 (0.75-1.06) 0.86 (0.72-1.03) 0.89 (0.75-1.06) ≥6 1,389 (7.4) 0.87 (0.76-1.00) 0.82 (0.70-0.96) 0.69 (0.57-0.84) 0.79 (0.66-0.95) 0.71 (0.58-0.85) 0.79 (0.66-0.96) Number of chiropractor visits in the year before injury

0 15,410 (82.1) 1.00* 1.00 1.00 1.00 1.00 1.00 1-5 1,784 (9.5) 1.06 (0.93-1.19) 0.96 (0.84-1.10) 1.03 (0.87-1.21) 0.89 (0.76-1.05) 1.01 (0.86-1.19) 0.89 (0.76-1.05) ≥6 1,583 (8.4) 1.20 (1.06-1.36) 0.92 (0.80-1.07) 1.17 (0.98-1.39) 0.85 (0.71-1.01) 1.16 (0.98-1.38) 0.85 (0.72-1.02) Number of massage therapy in the year before injury

0 17,509 (93.2) 1.00 1.00 1.00 1.00 1.00 1.00 1-5 740 (3.9) 1.09 (0.91-1.31) 0.92 (0.75-1.13) 0.99 (0.78-1.25) 1.01 (0.79-1.27) 0.99 (0.78-1.25) 1.01 (0.80-1.28) ≥6 528 (2.8) 1.06 (0.85-1.31) 0.97 (0.77-1.23) 0.80 (0.61-1.06) 1.10 (0.83-1.44) 0.82 (0.62-1.08) 1.10 (0.84-1.45)

Abbreviations=CI=confidence interval; GP=general practitioner; ICD-9=International Classification of Diseases 9th Revision; NSAIDs=nonsteroidal anti-inflammatory drugs; OR=odds ratio; SMRs=skeletal muscle relaxants


314

APPENDIX E SUPPLEMENTARY TABLES FOR MANUSCRIPT 4 (CHAPTER 7)

315

Appendix Table E-1 Results of a zero-inflated negative binomial regression analysis of the association between early drug class(es) dispensed, drug class(es) at first prescription, and strength of opioid received in the first 4 weeks and time on benefits and benefit status after the four week exposure window up to 1 year post-injury

Predictor

Time on Benefits After Four Week Exposure Window On Benefits After Four Week Exposure Window # of days on

benefits after 4 week exposure window, mean (SD), median

Receiving benefits

after 4 week exposure window,

n (%)

IRR (95% CI) OR (95% CI)


Drug class(es) dispensed in first 4 weeks (n=52,723) NSAIDS and/or SMRs only (n=28,533) 29.3 (53.5), 6.0 1.00 1.001 17,263 (60.5) 1.00 1.001 Opioids only (n=7,610) 37.9 (64.4), 10.0 1.22 (1.18-1.27)*** 1.10 (1.06-1.14)*** 4,908 (64.5) 1.18 (1.11-1.25)*** 1.11 (1.05-1.18)*** Opioids with NSAIDS and/or SMRs (n=16,580) 48.6 (71.0), 20.0 1.39 (1.35-1.43)*** 1.26 (1.22-1.29)*** 12,162 (73.4) 1.86 (1.77-1.95)*** 1.67 (1.59-1.75)*** Drug class(es) at first dispense in first 4 weeks (n=52,723) NSAIDS and/or SMRs only (n=31,585) 32.6 (57.2), 8.0 1.00 1.002 19,822 (62.8) 1.00 1.002 Opioids only (n=10,207) 43.6 (68.2), 15.0 1.22 (1.18-1.26)*** 1.11 (1.07-1.14)*** 7,080 (69.4) 1.36 (1.29-1.44)*** 1.26 (1.19-1.33)*** Opioids with NSAIDS and/or SMRs (n=10,931) 41.8 (66.7), 13.0 1.19 (1.15-1.23)*** 1.12 (1.09-1.16)*** 7,431 (68.0) 1.27 (1.20-1.34)*** 1.23 (1.16-1.30)*** Strength of opioids dispensed in first 4 weeks (n=24,190) Weak opioid(s) only (n=21,410) 42.0 (65.9), 15.0 1.00 1.003 14,821 (69.2) 1.00 1.003 Strong opioid(s) only (n=1,674) 61.5 (82.4), 27.0) 1.33 (1.24-1.42)*** 1.27 (1.19-1.35)*** 1,291 (77.1) 1.52 (1.33-1.74)*** 1.40 (1.23-1.61)*** Weak and strong opioids (n=1,106) 84.9 (91.1), 51.0 1.64 (1.52-1.77)*** 1.49 (1.38-1.60)*** 1,958 (86.6) 3.20 (2.57-3.99)*** 2.65 (2.14-3.28)***

Abbreviations=CI=confidence interval; IRR=incidence rate ratio; NSAIDs=nonsteroidal anti-inflammatory drugs; OR=odds ratio; SD=standard deviation; SMRs=skeletal muscle relaxants

316

*Significant at p<0.05 **Significant at p<0.01 ***Significant at p<0.001

1Adjusted for=sex, age, year of injury, income, diagnosis code, sedative hypnotic dispense within first 4 weeks, health authority, spine x-ray within first 4 weeks, cumulative days of opioids

in previous year 2Adjusted for=sex, age, year of injury, income, diagnosis code, cumulative days supplied of opioids in prior year, health authority, sedative hypnotic dispense within first 4 weeks, spinal x-

ray within first 4weeks, pain specialist visit within first 4 weeks 3Adjusted for=sex, age, year of injury, income, diagnosis code

317

Appendix Table E-2 Results of a zero-inflated negative binomial regression analysis of the association between cumulative days supplied and average daily MED in the first 4 weeks and time on benefits and benefit status after four week exposure window up to 1 year post-injury

Predictor

Time on Benefits After Four Week Exposure Window On Benefits After Four Week Exposure Window IRR (95% CI) OR (95% CI)

Unadjusted Adjusted Unadjusted Adjusted Cumulative days supply in first 4 weeks

Opioids (n=24,524) For every 7 day increase 1.17 (1.15-1.19)*** 1.14 (1.12-1.16)1, *** 1.46 (1.41-1.52)*** 1.41 (1.36-1.47)1, ***

NSAIDs (n=37,048) For every 7 day increase 1.09 (1.07-1.10)*** 1.04 (1.03-1.06)2, *** 1.32 (1.29-1.36)*** 1.22 (1.18-1.25)2, ***

SMRs (n=22,448) For every 7 day increase 1.09 (1.07-1.11)*** 1.03 (1.01-1.05)3, * 1.35 (1.30-1.41)*** 1.17 (1.12-1.22)3, ***

Average daily MED in first 4 weeks (n=27,666)

For every 30mg/day increase 1.06 (1.04-1.09)*** 1.05 (1.03-1.07)4, *** 1.02 (0.99-1.05) 1.02 (0.99-1.05)4

Abbreviations=CI=confidence interval; IRR=incidence rate ratio; NSAIDs=nonsteroidal anti-inflammatory drugs; OR=odds ratio; SD=standard deviation; SMRs=skeletal muscle relaxants *Significant at p<0.05 **Significant at p<0.01 ***Significant at p<0.001 1Adjusted for: sex, age, year of injury, income, diagnosis code 2Adjusted for: sex, age, year of injury, income, cumulative days of opioids within 4 weeks, cumulative days of SMRs within 4 weeks, spine x-ray within first 4 weeks 3Adjusted for: sex, age, year of injury, income, cumulative days of opioids within 4 weeks, cumulative days of NSAIDs within 4 weeks, spine x-ray within first 4 weeks 4Adjusted for: sex, age, year of injury, income, diagnosis code, spine x-ray within first 4 weeks, health authority, NSAID dispense within first 4 weeks, pain specialist visit within first 4

weeks, sedative hypnotic dispense within first 4 weeks

318

Appendix Table E-3 Results of a zero-inflated negative binomial regression analysis of the association between early drug class(es) dispensed, drug class at first prescription, and strength of opioid received and time on benefits and benefit status after exposure window up to 1 year post-injury among claimants with injury years 1998-2001

Predictor

Time on Benefits After Exposure Window On Benefits After Exposure Window

# of days on benefits after

exposure window, mean (SD), median

Receiving benefits after

exposure window, n (%)

IRR (95% CI) OR (95% CI)

Adjusted for previous

confounders


confounders + supplementary

health care


confounders



health care

Drug class(es) dispensed in first 8 weeks (n=19,252) NSAIDs and/or SMRs only (n=10,307) 22.5 (52.6), 0.0 1.00 1.001 3,964 (38.5) 1.00 1.001 Opioids only (n=2,937) 26.4 (58.0), 0.0 1.05 (0.97-1.13) 1.05 (0.97-1.13) 1,182 (40.3) 0.97 (0.88-1.06) 0.99 (0.90-1.09) Opioids with NSAIDs and/or SMRs (n=6,008) 41.8 (70.7), 4.0 1.23 (1.16-1.30)*** 1.23 (1.16-1.30)*** 3,213 (53.5) 1.50 (1.39-1.62)*** 1.44 (1.34-1.55)*** Drug class(es) dispensed in first 4 weeks (n=18,217) NSAIDs and/or SMRs only (n=10,051) 32.3 (58.9), 6.0 1.00 1.002 6,069 (60.4) 1.00 1.002 Opioids only (n=2,865) 40.3 (67.0), 12.0 1.09 (1.03-1.16)** 1.10 (1.04-1.17)** 1,856 (64.8) 1.11 (1.01-1.23)* 1.14 (1.03-1.27)* Opioids with NSAIDs and/or SMRs (n=5,301) 51.7 (74.0), 21.0 1.21 (1.16-1.27)*** 1.21 (1.16-1.27)*** 3,874 (73.1) 1.60 (1.47-1.74)*** 1.54 (1.41-1.68)***

319

Predictor






IRR (95% CI) OR (95% CI)


confounders



health care


confounders



health care Drug class(es) dispensed at first prescription in first 8 weeks (n=19,252) NSAIDs and/or SMRs only (n=11,624) 25.9 (56.5), 0.0 1.00 1.003 4,796 (41.3) 1.00 1.003 Opioids only (n=4,047) 33.7 (64.7), 0.0 1.07 (1.00-1.14)* 1.07 (1.01-1.14)* 1,922 (47.5) 1.12 (1.03-1.22)** 1.15 (1.05-1.25)** Opioids with NSAIDs and/or SMRs (n=3,581) 34.6 (65.7), 0.0 1.16 (1.08-1.23)*** 1.16 (1.09-1.24)*** 1,641 (45.8) 1.09 (1.00-1.19)* 1.08 (0.99-1.18) Drug class(es) dispensed at first prescription in first 4 weeks (n=18,217) NSAIDs and/or SMRs only (n=10,984) 35.0 (61.5), 9.0 1.00 1.004 6,854 (62.4) 1.00 1.004 Opioids only (n=3,732) 45.8 (70.5), 16.0 1.11 (1.05-1.17)*** 1.11 (1.07-1.14)*** 2,584 (69.2) 1.22 (1.11-1.34)*** 1.22 (1.11-1.33)*** Opioids with NSAIDs and/or SMRs (n=3,501) 45.4 (70.7), 14.0 1.14 (1.08-1.21)*** 1.12 (1.09-1.16)*** 2,361 (67.4) 1.17 (1.07-1.28)*** 1.18 (1.06-1.28)***

320

Predictor






IRR (95% CI) OR (95% CI)


confounders



health care


confounders



health care Strength of opioids dispensed in first 8 weeks (n=8,945) Weak opioid(s) only (n=8,337) 34.6 (65.0), 0.0 1.00 1.005 3,997 (47.9) 1.00 1.005 Strong opioid(s) only (n=288) 48.0 (74.8), 8.0 1.05 (0.88-1.25) 1.06 (0.89-1.27) 165 (57.3) 1.15 (0.89-1.50) 1.12 (0.85-1.46) Weak and strong opioids (n=320) 83.4 (94.1), 42.5 1.35 (1.16-1.57)*** 1.34 (1.15-1.56)*** 233 (72.8) 1.99 (1.50-2.64)*** 1.97 (1.47-2.64)*** Strength of opioids dispensed in first 4 weeks (n=8,166) Weak opioid(s) only (n=7,636) 45.4 (69.6), 16.0 1.00 1.006 5,291 (69.3) 1.00 1.006 Strong opioid(s) only (n=292) 64.3 (82.8), 31.0) 1.15 (0.99-1.34) 1.14 (0.98-1.33) 229 (78.4) 1.50 (1.08-2.07)* 1.45 (1.04-2.03)* Weak and strong opioids (n=238) 100.0 (99.0), 63.0 1.57 (1.34-1.84)*** 1.56 (1.33-1.82)*** 210 (88.2) 2.88 (1.79-4.63)*** 2.86 (1.72-4.55)***

Abbreviations=CI=confidence interval; IRR=incidence rate ratio; NSAIDs=nonsteroidal anti-inflammatory drugs; OR=odds ratio; SMRs=skeletal muscle relaxants *Significant at p<0.05 **Significant at p<0.01 ***Significant at p<0.001 1Adjusted for: sex, age, year of injury, income, diagnosis code, health authority, pain specialist visit within first 8 weeks, cumulative days of opioids in previous year, sedative hypnotic

dispense within first 8 weeks, spinal x-ray within first 8 weeks, hospital contact within first 8 weeks, back/neck pain in prior 2 years, cumulative days of NSAIDs in previous year, physio visit

in first 8 weeks 2Adjusted for: sex, age, year of injury, income, diagnosis code, sedative hypnotic dispense within first 4 weeks, health authority, spine x-ray within first 4 weeks, cumulative days of opioids

in previous year, cumulative number of physio visits in year prior to injury, physio visit in first 4 weeks. 3Adjusted for: sex, age, year of injury, income, diagnosis code, cumulative days supplied of opioids in prior year, pain specialist visit within first 8 weeks, health authority, sedative hypnotic

dispense within first 8 weeks, spinal x-ray within first 8 weeks 4Adjusted for: sex, age, year of injury, income, diagnosis code, cumulative days supplied of opioids in prior year, health authority, sedative hypnotic dispense within first 4 weeks, spinal x-

ray within first 4weeks, pain specialist visit within first 4 weeks

321

5Adjusted for: sex, age, year of injury, income, diagnosis code, pain specialist visit within first 8 weeks, sedative hypnotic dispense within first 8 weeks, back/neck pain comorbidity within

prior 2 years, cumulative number of chiro visits in year prior to injury, physio visit in first 8 weeks. 6Adjusted for: sex, age, year of injury, income, diagnosis code, cumulative number of physio visits in year prior to injury, cumulative number of chiro visits in year prior to injury, physio

visit in first 4 weeks.

322

Appendix Table E-4 Results of a zero-inflated negative binomial regression analysis of the association between cumulative days supplied and average daily MED and time on benefits and benefit status after exposure window up to 1 year post-injury among claimants with injury years 1998-2001

Predictor


Adjusted for previous confounders

Adjusted for previous confounders +

supplementary health care




Cumulative days supply in first 8 weeks

Opioids (n=9,032) For every 7 day increase 1.10 (1.07-1.12)*** 1.10 (1.07-1.12)1, *** 1.34 (1.30-1.38)*** 1.34 (1.30-1.38)1, ***

NSAIDs (n=13,396) For every 7 day increase 1.04 (1.02-1.06)*** 1.04 (1.02-1.06)2, *** 1.30 (1.27-1.34)*** 1.28 (1.25-1.32)2, ***

SMRs (n=6,957) For every 7 day increase 1.04 (1.01-1.07)** 1.04 (1.01-1.07)3, ** 1.27 (1.21-1.33)*** 1.26 (1.20-1.32)3, **

Cumulative days supply in first 4 weeks

Opioids (n=8.299) For every 7 day increase 1.13 (1.10-1.16)*** 1.13 (1.10-1.16)4, *** 1.36 (1.28-1.45)*** 1.36 (1.28-1.45)4 ***

NSAIDs (n=12,513) For every 7 day increase 1.03 (1.01-1.06)*** 1.03 (1.01-1.06)5, * 1.29 (1.22-1.36)*** 1.25 (1.19-1.32)5, ***

SMRs (n=6,546) For every 7 day increase 1.05 (1.01-1.10)* 1.05 (1.01-1.10)6, * 1.15 (1.07-1.25)*** 1.13 (1.04-1.22)6, **


For every 30mg/day increase 1.03 (0.97-1.07) 1.02 (0.97-1.07)1 1.02 (0.96-1.08) 1.02 (0.96-1.09)7

323

Predictor









For every 30mg/day increase 1.05 (1.00-1.10)* 1.05 (1.00-1.10)2, * 1.01 (0.94-1.09) 1.01 (0.94-1.09)8

Abbreviations=CI=confidence interval; IRR=incidence rate ratio; NSAIDs=nonsteroidal anti-inflammatory drugs; OR=odds ratio; SMRs=skeletal muscle relaxants *Significant at p<0.05 **Significant at p<0.01 ***Significant at p<0.001 1Adjusted for: sex, age, year of injury, income, cumulative days of opioids in previous year 2Adjusted for: sex, age, year of injury, income, cumulative days of opioids within 8 weeks, diagnosis code, cumulative days of SMRs within 8 weeks 3Adjusted for: sex, age, year of injury, income, cumulative days of opioids within 8 weeks, cumulative days of NSAIDs within 8 weeks 4Adjusted for: sex, age, year of injury, income, diagnosis code 5Adjusted for: sex, age, year of injury, income, cumulative days of opioids within 4 weeks, cumulative days of SMRs within 4 weeks, spine x-ray within first 4 weeks 6Adjusted for: sex, age, year of injury, income, cumulative days of opioids within 4 weeks, cumulative days of NSAIDs within 4 weeks, spine x-ray within first 4 weeks, physio visit in first 4

weeks, chiro visit in first 4 weeks. 7Adjusted for: sex, age, year of injury, income, diagnosis code, spine x-ray within first 8 weeks, pain specialist visit within first 8 weeks, health authority, NSAID dispense within first 8

weeks, cumulative number of physio visits in year prior to injury, physio visit in first 8 weeks, massage visit in first 8 weeks. 8Adjusted for: sex, age, year of injury, income, diagnosis code, spine x-ray within first 4 weeks, health authority, NSAID dispense within first 4 weeks, pain specialist visit within first 4

weeks, sedative hypnotic dispense within first 4 weeks, cumulative number of physio visits in year prior to injury, cumulative number of chiro visits in year prior to injury, cumulative

number of massage visits in year prior to injury, physio visit in first 4 weeks, chiro visit in first 4 weeks, massage visit in first 4 weeks

Date post:	24-Feb-2023
Category:	Documents
Upload:	khangminh22
View:	0 times
Download:	0 times