BASELINE PREDCTORS OF PAIN AND DlSABll lM SIXMONTHS FOLLOWING DlSTAL RADIUS FRACTURE

Joy C. MacDemid

Graduate Program in Epidemiology and Biostatistics

Submitted in partial fuifilment of the requirements for the degree of

Doctor of Philosophy

Faculty of Graduate Studies The University of Western Ontario

London. Ontario May 1999

National Library w * I ofCanada BiMimthèque nationale du Canada

Acquisitions and Acquisitions et Bibliographie Services senrices bibliographiques

395 Wellington Street 395. rue Weüïngtan Onawa ON KIAON4 OltawaON KfAON4 canada Canada

The author has granted a non- exclusive licence dowing the National Lîbrary of Canada to reproduce, loan, distribute or seil copies of this thesis in microform, paper or electronic formats.

The author retains ownership of the copyright in this thesis. Neither the thesis nor substantid extracts fiom it may be printed or otherwise reproduced without the author's permission.

L'auteur a accordé une licence non exclusive permettant à la Bibliothèque nationale du Canada de reproduire, prêter, distriiuer ou vendre des copies de cette thèse sous la forme de microfiche/nim, de reproduction sur papier ou sur format électronique.

L'auteur conserve la propriété du droit d'auteur qui protège cette thèse. Ni la thèse ni des extraits substantiels de celle-ci ne doivent être imprimés ou autrement reproduits sans son autorisation.

ABSTRACT

Distal radius fradures (DRF) are a common injury that results in pain and

disability. This thesis is conœmed with detemination of baseline factors that

predict patient-rated pain and disability sixmonths after a distal radius ftadure.

A number of supporting studies were completed to establish reliable methods to

evaluate fiadure severityI physical impairmenls, pain and disability after DRF. A

scale to masure wrist-related pain and disability was developed, resulting in a

1 5-item 'Patient-Rated Wrist Evaluation" (PRWE). High test-retest reliability of

the PRWE was established in patients with DRF or scaphoid fracture (1 CCs>0.90). Construd and aiterion validity were established by comparing the

PRWE to'the SF-36 (general health) and the DASH (Upper extrernity quality of

Me). The PRWE nies most responsive in detecting recovery after DRF

(SRM=2.29), followed by the DASH (SRM=2.01) and the SF-36 (Physical

Surnmary Score SRM =1 .O). Based on its ability to detect recovery and its

correlation with other instruments the PRWE was judged to be a valid indicator

of patient rated pain and disability. A pilot study was undertaken with 50

patients to narrow dom the number of potential patient related factors that might

contribute to outcome. Based on this study, a number of factors were found not

to be predictive of mon th PRWE. These included smoking status, age, sex

and a three-point injury intensity scale. Secondary gain (legal or Workers

Compensation involvement) and education were found to be significant

predidors. A prospective study w s undertaken wïüi 70 patients evaluated at

their baseline visit at the clinic for DRF and s k month after DRF- Baseline

factors recorded included age, se& sacondat'y gain. education level, fracture

"type" (according to A 0 system), prweduction and post-redudion radial

shortening. Patients completed questionnaires and were tested for ROM. grip

strength and dexterïty. Forward stepwise multiple regression was used and

identified that secondary gain (p<O.ûUûl), radial shoftening at injury (p<0.001)

and educaüon (pe0.05) vvere significant predidofs of outcome amunting for

iii

29% of the variance in PRWE score at six months after fracture. The amount of

pain and disability reporteci by six-months mer being treated for a

DRF in a specializeâ hand unit is greater in patients with potentia1 fw seoondary

gain, even Mer controlling for the srnaller eff8C1s of severity of bony injury and

education.

KEYVVORDS: distal radius, ftacture. wrist, outcorne, prospective, pain. disability,

radiographs, predidion, s8condary gain, health statu$. responsiveness,

rel iability , gtip, ROM. questionnaires, dexterity

CO-AUTHORSHIP

The thesis question. the primary OUfcome masure for this study and the

design and analyses for all thesis articles w r e conducteci by me. Co-

investigators m e recruited Men additional testers or raters ~ Ï t h specific

expertise, such as teading radiographs, wre required. Thesis advisors are

included as coauthors for specific chapters for Mich they contn'buted input prior

to submission for publication. The authors and specific roles for each

component of the thesis are Iisteâ below.

1 ) Chapter 1 - Background: Joy MacDermid - sole author

2) Chapter 2: Development of a Scale for Patient Rating of Wfist Pain and

Disability - JoyC MacDemid-saleauthor

3) Chapter 3: Patient Rating of Wrist Pain and Disability: A Reliable and Valid

Measurement Tool

Joy C MacDemid: study design, data colledionlsupervision; data analysis.

wrote manuscript

Tom Turgeon: data collection; reviewed manuscript

Mark Beadfe: data collection; revievined manuscript

Robert S Richards: reviewad manuscript; Co-diredor of Clinical Research Lab

James H Roth: reviewed manuscript; Co-director of Clinical Research Lab

4) Chapter 4: Reliability of Range of Motion Measurement in the Elbow and

Fciream

April Armstrong: tester, CO-wrote paper

Joy C MacDemid: study design, tester, wrote portions of the paper, data

anal ysis

S hrikant Chinchalkar: tester, reviewed manuscript

Rob Stevens: tester; reviewed manuscript

Graham JW King: supplied patient Ji&, tester, manusaipt revisions,

communication with journal

5) Chapter 5: Test-retest Reliability of Static and Endurance Grip Strength Tests

Performed on the Jarnar and NK Devices

Joy C MacDemid: study design, data analysis, wrote article

Tawfeik Aiyafi: tester

Roberts S Richards: review8d manusai* Codrectof of Clinical Research Lab

James H Roth: reviewed manuscript; Co-diredor of Clinical Research Lab

6) C hapter 6: Inter-rater Rel iability of Radiographie Measurement and

Classifications in Distal Radius Fractures

Joy C MacDermid: study design, data analysis, wmte article

Marilena Mautgnani: tester

Kevin Hilldebrant: tester, input on manuscript

Allan Donner: data analysis consultation, reviewed article

Robert S Richards: consultation on cfinical content, reviewed article, Co-diredor

of Research Lab, advisor

James H Roth: reviewed manuscript; Co-diredar of Clinical Research Lab

7) Chapter 7: Responsiveness of the SF-36, Dash, Patient Rated Wrist

Evaluation and Physical Irnpaiments in Patients mer Distal Radius Fracture

Joy C Macilennid: study design, data collection, data analysis, wote article

Allan Donner : advisor, revi- manuscript

Robert S Richards: consultation on clinical content, reviewed article, Codiredor

of Research Lab, advisor

Niwlas Bellamy: advisor, reviewed manuscript

James H Roth: consultation on clinical content, reviewed article, Codirector of Research tab

8) Chapter 8: Baseline Predictors of Pain and Disability Six Months After Distal

Radius Fracture

Joy C MacDennid: study design, data collection (tester and supervisionlttaining

of assistants), data analysis, wote article

Allan Donner : advisor, reviewed manuscript

Robert S Richards: cmsuitation on dinical content, reviewed article, Co-director

of Research Lab, advisor

N kolas Bellamy: advisor, reviewed manuscript

James H Roth: consumm an dinical content, reviewed article, Co-dirador of

Research Lab

9) Chapter 9: Condusions: Joy C Macûermid sale author

1 O ) Appendix I r Tomards Oevelopment of Perfonnanœ-based Wrist Outcorne

Score- Theory paper (wwk in progress) - Joy C MacDemid sole author

vii

ACKNOWLEDGMENTS

I would Iike to acknowledge the following people for their assistanœ in

making this projed possible. My advisor and fellow co-director of the HULC

Clinical Research Lab, Bob Richards, for his support, cornmitment to dinical

research and dinical insight To Jirn Roth, Director of the HULC and Co-diredar

of HULC Clinical Research Lab INho is directly responsible for the excellence in

patient are, teaching and research that takes place within the HULC. Without

his vision to be on the leading 8dg8 of excellence, his leadership in cfeating the

team who make 1 happen, his cornmitment to finding the required resources and

his drive to keep on grmng. the HULC wwld not b8 the wwld rencnm center

that it has becorne-

Thanks to my epidemiology advisors, Allan Donner and Nicholas Bellamy,

for their insight and experienœ. Special thanks to my supervisor Allan Donner

for highlighting the essential statïstical e(emants of the project and helping me

survive graduate sdiool and to the professors in the Department of

Epiderniologyl8iostatistics for conveying the disàpline of epidemiology with

professionalism.

Thanks to the students and research assistants who aided data

collection, in partiwlar. Carol Fazekas whose technical assistance was always

thorough and dependable. The physicians of the HULC, who cared for the

patients in the study and who maintainad a supportive interest in the project and

the staff at the HULC ch ic who assisted with scheduling and directing patients.

Thanks to Catherine McOwan for assisting vvith proofreading.

Funding fiom the Ontario Graduate Society and the Ministry of Health for

Fellowship provided essential support over the course of graduate education. I

appreciate the support fKwn the Pooled Research Trust Fund at St. Joseph's

Health Center that funded the survey wwk that was used to develop &teria for

the Patient Rated Wrist Evaluation.

viii

1 have received valuable feedback from hand therapists that have

adopted the Patient Rated Wnst Evaluatïon for their dinical pradice and

research. I appreciate that Cedaron Medical have induded the scale in their

touch screen sofhmre outcorne modules, thus improving acœss for other

centers. Other "outcomesaiented hand therapists" have embraced the

concept of outcornes researdi and have baen supportive of our comrnon interest

to improve the evaluation and treatment of upper extremity problems.

ft is tradition that w e place our thanks to ouf families at the end of the

acknowledgments. Perhaps this is because it is personal in nature, perhaps a

bad habit - ie wa too infrequently forrnally recognize Mat we hope they always

know - that they are first in prionty. Thanks to my husband ,Steve, who is alweys

foremost in my thoughts. Thanks to Jamey, Katrina and Myles for not only

surviving having a mother who w s a student for th& entire lives, but grawing

to become intelligent, ethical and accomplished individuals and to Kevin for

slmng d o w the thesis Mi le making life more fun and meaningil. I cherish my

father, who vms unabie to be with me until the end of this project, but is still the

guidepost which keeps the path lit and my mother for her continual love and

support. My brothers and sisters and their children are distant in miles but

always close at hean Thanks to my nieces, for teaching me ( a variation of)

this song about my thesis

TABLE OF CONTENTS

.......................................... ACKNOVVLEDGMENTS viii

LIST OF TABLES ............................................... xvi

LIST OF FIGURES ............................................. xviii

LIST OF APPENDICES .......................................... xix

CHAPTER 1: INTRODUCTION ..................................... 1 ................................................. Overview 1

Epidemiology of WrÏst Fracture ................................ 1 Severityoflnjury ........................................... 3 Etiolagy .................................................. 4

....................... Radiologie Classification of lnjury Severity 5 Classification systems .................................. 5 Radiographic measurements ............................ 8

Treatment of Distal Radius Fracture ........................... 10 Issues in Outcorne Measurement in Upper Extremity Otthopaedics ... 1 3 Methods of Rating of Outcorne After Distal Radius Fracture ......... 17

...................... Previous Studies of Predidors of Outcorne 19

...................... Theoretical Aspects of Prognostic Studies 23 .................................. Reliability of Measurement 24

....................... DefinitioMmportance of Baseline factors 30 Justification of Potential Predictors ............................ 30

Age ............................................... 31 Sex ............................................... 32 Secandarygain ...................................... 32

...................................... Smoking status 34 Radial shortening .................................... 36

....................... Aiticular Involvernent/Cornminution 37 Education .......................................... 38

.................... Defïnitionllmportanœ of Mediating Variables 39 Treatment .......................................... 41 Complications ....................................... 42

................................ Thesis Components-Overviw 44 Referenœs ..............................................

CHAPTER 2: DEELOPMENT OF A SCALE FOR PATIENT RATING OF VVRlST PAIN AND DlSABlLlTY ............................... 76 Background .............................................. 76 Overview ................................................ 77

............................... Summary of Findinge of Sunmy 78 ......................................... Theoretical Issues 79

Tooldevelopment ......................................... 81 Reliability testing of the subjective outcorne scale ................. 83 Validity of Subjective Wrist Outwme Scoring .................... 84 Figure 'i . Patient Rated Wrist Evaluation ....................... 87 Sconng .................................................. 89 Conclusions .............................................. 89 Referenœs .............................................. 91 Table 1 . Tat-retest Reliability of the Patient-Rated Wrist Outcome Fom

.................................................. 95 Table 2 . Validîty of the Patient-Rated Wnst Outcome Fom ......... 96

CHAPTER 3: PATIENT RATING OF WRIST PAIN AND DISABILITY A RELIABLE AND VALlD MEASUREMENT TOOL .................. 97 Introduction .............................................. 97 Material and Methods ...................................... 98

The Wrist-Outcorne Expert Survey ....................... 98 Swvey Methds ................................ 98

Development of the Patient-Ratd Wrist Evaluation Questionnaire . . 1 01 Reliability Testing of the PRWE ........................ 102

Reliability Testing Results .................................. 104 Construct Validity .............................. 104

............................... Criterion validity 104 .................................... Validity Testing Resuits 106

Construd validity ............ ,. ...................... 106 Criterion validity .................................... 106

Discussion .............................................. 106 Referenœs ............................................. 111

. Table 1 Respondent Characteristics ......................... 115 Table 2 . Survey Questions and Responses .................... 116 Table 3 . Subjed Characteristics for Reliability Study ............. 121 Table 4 . Reliability Coefficients for The PRWE and its Subscales ... 122 Table 5 . Reliability SF-36 Subscales retested &ter one year ....... 123 Table 6 . Subjeds for Construct Validity Study .................. 124

Table 7 . Validity of the PR= Correlations Be-n Outcome Measures ................................................. 125

CHAPTER 4 . RELIABtLIlY OF RANGE OF MOTlON MEASUREMENT IN THE ELBOW AND FOREARM .................................. 126 Introduction ............................................. 126 Methads ................................................ 127

Subjects .......................................... 127 Instruments ........................................ 127

............................. Tester Preparation 128 .............................. Elbow FlexlOnlExtensian 129

.................................. Foreamsupinat ion 129 Farearm pronation ................................... 129

................................. The Testing Session 129 .................................. Statistical Analysis 131

Resuits ................................................. T31 .............................................. Discussion 132

Conclusions ............................................. 139 ............................................. References 140

........................................ Table 1- Subjects 142 ............................... Table 2 . lntrarater Retiability 144

Table 3 . inter-rater Reliability ............................... 146 Table 4 . Average DïfFerenœs in Scores (summed over raters and

occasions) B e w n Different Instruments Dufing Measurement of ................................... Forearm Rotation 147

................ Figure 1- Instrument Variability in Flexion Scores 148 Figure 2 . Instrument Variability in Extension Scores .............. 149

CHAPTER 5: TEST-RETEST RELlABlLlTY OF STATIC AND ENDURANCE GRIP STRENGTH TESTS PERFORMED ON THE JAMAR AND NK

.............................................. DEViCES -150 Introduction ............................................. 150

................................................ Methods 154 ................................................. Results 155

.............................................. Discussion 156 ............................................. Conclusions 159

Referenœs ............................................. 160 . ........... Table 1 Static Gnp and asdated reliability statistics 166

Table 2 . Cyciic Grip Repetitions and Reliability Statistics ......... 167 Table 3 . Test-retest Reliability uf NK Cornputer-Analysed 10 Second Grip

................................................. 168

xii

Table 4 . Relationship Betwaen Static Grip and Other Variables .... 169

CHAPTER 6: RELIABILITY OF RADlOGRAPHlC MEASURES AND CLASSlFlCATlON IN PATIENTS \MTH DISTAL RADIUS FRACTURES ....................................................... 171

lntrodudion ............................................. 171 Methods ................................................ 176

Materials .......................................... 176 Radiographie Measures .............................. 177 Figure 1 . Analysis of Radial Shortening and Dorsal Angulation

............................................ 178 Fracture Classifications ............................... 17 9 Figure 2 . Fracture Classification Systems ............... 179 Data Analysis ...................................... 184

Results ................................................. 185 Discussion .............................................. 185 Conclusions ............................................. 188 References ............................................. 190

........................................ Table1 . Subjects 195 Table 2 . Inter-rater Reliability of Radiagraphic Measures in Distal Radius

Fractures. ......................................... 196 ......... Table 3 . Inter-rater Reliability of Fracture Classifications 197

CHAPTER 7: RESPONSIENESS OF THE SF.36. DASH. PATIENT AATED WRIST EVALUATION AND PHYSICAL IMPAIRMENTS IN EVALUATING

............ RECOVERY AFTER A DISTAL RADIUS FRACTURE 198 Introduction ............................................. 198 Methods ................................................ 201

Patient Testing .......................... .. ......... 201 Outcame instruments ................................ 201

Questionnaires ................................ 201 Physical Capacity (Impairment) Testing ............. 203

Data Analysis ...................................... 204 Results ................................................. 205

.............................................. Discussion 205 Conciusions ............................................. 212 Referenœs ............................................. 213

............. Table 1 . Responsiveness Reported for Other Scales 217 ............ Table 2 . Subjects with Distal Radius Fractures (n=59) 223

.......... Table 3 . Responsiveness of Outcorne Measures in DRF 224 Figure 1 . Score for AROM ................................. 225

xiii

CHAPTER 8 . BASELINE PREDICTORS OF PAfN AND DISABILITY SIX MONTHS AFTER A DlSTAL RAûlUS FRACTURE ............... 226 Introduction ......-............---..--................-.. 226 Methods ................................................ 235

An Overview of the Prospedve Study Design ............. 235 Definition of Independent Variables ..................... 236 Sample Ske Estimation .............................. 237 Testing Schedule ................................... 239 Recruitment ........................................ 239 Impairment testing ................................... 240

ROM ........................................ 2a GripStnmgth ................................. 241 Dexten ty ..................................... 241 TotalScore ................................... 241

Data Analysis ...................................... 241 Results ............................................... . .242 Discussion .............................................. 243 References ............................................. 252 Table 1 . Subject Charadenstics ............................. 260 Table 2 . Univanate Association Between Predictm and Outcorne ... 261 Table 3 . Stepwise Multiple Linear Regression Results ........... 262

CHAPTER 9: SUMMARY ........................................ 264 OverviewoftheStudy ..................................... 264 Specific Recomrnendations Based on Study Findings ............ 266 Questions Raised by the Present Study ....................... 267 Strengths of the Present Study .............................. 270 Limitations of Present Study ................................ 270 Directions for Future Research .............................. 272 References ............................................. 275

APPENDIX 1: TOWARDS A PERFORMANCE-BASED SCALE FOR WRIST IMPAIRMENT ............................................ 278

............................................. Introduction 278 RangeofMotion ......................................... 279 GripStrength ............................................ 280

.......................................... GripEndurance 281 De*eriv ................................................ 282 Complications ........................................... 284 Structural Criteria ......................................... 284 Summary of Developmental Work ............................ 285 References ............................................. 285

xiv

............................ APPENDIX 2: Results From Pilot Study 289 ............................. . Table 1 Subject Characteristics 289

............ Table 2- Correlation Be- Variables And Outcorne 290 . .... Table 3 Dflerences between subgroups (p values for ANOVA) 291 . .........-.... Table 4 Resuk of 6-rd Stepwïse Regression 292

............................... APPENDIX 3: E'THICS WPROVALS 295

APPENDIX 4- RELEASE TO USE ARTICLES ACCEPTE0 FOR PUBLICATION ....................................................... 300

......................... APPENDR 5: DATA COLLECTION FORMS 304

VITA ........................................................ 318

LIST OF TABLES

CHAPTER 2: DEVELOPMENT OF A SCALE FOR PATIENT RATlNG OF VVRiST PAlN AND DISABlLlW Table 1 . Test-retest Reliability of the Patient-Rated Wrist Outcome Form

.................................................. 95 Table 2 . Validity of aie Patient-Rated Wrist Outcorne Fom ......... 96

CHAPTER 3: PATIENT RATING OF W I S T PAlN AND DISABILITY A RELIABLE AND VALID MEASUREMENT TOOL Table 1 . Respondent Characteristics ......................... 115 Table 2 . Sunrey Questions and R~sponses .................... 116 Table 3 . Subjed Charaderistics for Reliability Study ............. 121 Table 4 . Reliability Coefficients for The PRWE and its Subscales ... 122 Table 5 . Reliability SF-36 Subscales retested after one year ....... 123 Table 6 . Subjeds for Criterion Validity Study ................... 124 Table 7 . Validity of the PRWE: Correlations Between Outcome Measures

................................................. 125

CHAPTER 4 . RELlABlLlTY OF RANGE OF MOTION MEASUREMENT IN THE ELBOW AND F O R M M Table 1 . Static Grip and assaUated reliability statistics ........... 142 Table 2 . Cyclic Grip Repetitions and Reliability Statistico ......... 144 Table 3 . Test-retest Reliabiiity of NK Cornputer-Analysed 10 Second Grip

................................................. 146 Table 4- Relationship Betwen Static Grip and Other Variables .... 147

CHAPTER 5: TEST-RETEST RELIABILITY OF STATIC AND ENDURANCE GRlP STRENGTH TESTS PERFORMED ON THE JAMAR AND NK DEVICES

Table 1 . Static Grip and associated reliability statistics ........... 766 Table 2 . Cyclic Grip Repetitions and Reliability Statistitics ......... 167 Table 3 . Test-retest Reliability of NK Cornputer-Analysed 10 Second Grip

................................................. 1 M Table 4 a Relationship Between Static Grip and Other Variables .... 169

CHAPTER 6: RELIABILITY OF RADIOGRAPHIC MEASURES AND CLASSIFICATION IN PATIENTS WITH DISTAL RADIUS FRACTURES

Table 1 . Subjects ...... .. .............................. 195 Table 2 . Interiater Reliability of Radiographic Measures in Distal Radius

Fractures .......................................... 196 ......... Table 3 . Interiater Reliability of Fracture Classif~tions 197

CHAPTER 7: RESPONSIVENESS OF THE SF.36. DASH. PATIENT RATED WRlST EVALUATION AND PHYSICAL IMPAIRMENTS IN PATIENTS AFTER DCSTAL RADIUS FRACTURE Table 1 . Responsiveness Reporteci for Other Scales ............. 217

. -...-.-...*. Table 2 Subjeds with Distal Radius Fradures (n=59) 224 Table 3- Responsîveness of Outcome Measures in DRF .......... 225

CHAPTER 8- BASELINE PREDICTORS OF PAIN AND DlSABlLlTY SIX MONTHS AFTER A DISTAL RADIUS FRACTURE Tabte 1 . Subjed Characteristics ............................. 260 Table 2 . Univariate Association BetHieen Predictors and Outcorne ... 261 Table 3 . Stepwise Multiple Linear Regression Results ........... 262

APPENDIX 2: Results From Pilot Study ............................ 289 Table 1 . Subject Characteristics ............................. 289

. Table 2 Correlation Between Variables And Outcorne ............ 290 Table 3 MOVA Results for Bi-variate Data .................... 291

. Table 4 Rewlts of Backward Stepwise Regression .............. 292

xvii

LIST OF FIGURES

CHAPTER 1: INTRODUCTION Figure 1. Outcome after Distal Radius Fradure .................. 40

CHAPTER 2: DEVELOPMENT OF A SCALE FOR PATIENT M I N G OF WRlST PAIN AND DISABILITY

Figure 1. Patient Rated Wrist Evaluation ....................... 87

CHAPTER 4. RELIABILITV OF RANGE OF MOTION MEASUREMENT IN THE ELBOW AND FOREARM Figure 1. Instrumemt Variability in Flexion Scores. ............... 148 Figure 2. Instrument Variability in Extension Scores. ......-...... 149

CHAPTER 5: TEST-RETEST RELIABILITY OF STATIC AND ENDURANCE GRlP STRENGTH TESTS PERFORMED ON THE JAMAR AND NK DEVICES Figure 1. Sample Endurance Data From NK 1 O-sec Grip .......... 170

CHAPTER 6: RELIABILITY OF RADlOGRAPHlC MEASURES AND CLASSlF CATION IN PATIENTS WTH DISTAL RADIUS FRACTURES Figure 1. Analysis of Radial Shortening and Dorsal Angulation ....................................................... j78

Figure 2. Fracture Classification Systems ..................... 179

CHAPTER 7: RESPONSJVENESS OF THE SF-36, DASH, PATIENT RATED VVRlST EVALUATION AND PHYSICAL IMPAIRMENTS IN PATIENTS AFTER DISTAL RADIUS FRACTURE

Figure 1 . Score for AROM ................................. 225

LIST OF APPENDICES

APPENDIX 1: TOWARDS A PERFORMANCE-BASED SCALE FOR WRIST IMPAIRMENT ................................................. 278

Introduction ............................................. 278 RangeofMotion ......................................... 279 GripStrength ............................................ 280 GripEnduranœ .......................................... 28f Dexterity ................................................ 282 Complications ........................................... 284 Structural Criteria ......................................... 284 Summary of Developmental Work ............................ 285 References ............................................. 285

APPENDIX 2: Results From Pilot Study ............................ 289 Table 1- Subjea Characteristics ............................. 289 Table 2 . Correlation B e W n Variables And Outcorne ............ 290

.... Table 3 . Diffmnœs behiiueen subgroups (p values for ANOVA) 291 .............. Table 4 . Rewlts of Bachard Stepwïse Regression 292

APPENDIX 3: ETHlCS APPROVALS ............................... 295

APPENDIX 4 . RELEASE TO USE ARTICLES ACCEPTED FOR PUBLICATION 300 .......................................................

APPENDIX 5: DATA COLLECTION FORMS ......................... 304

CHAPTER 1: INTRODUCTION

Overview

The primary research question of this thesis deals factors that predict

pain and disability six months after distal radius fradure (DRF). The pffmary

research question is addressed in the final article (Chapter 8). The first chapter

provides a basic understanding of distal radius M u r e s (DRF) by summarizing

information on the epidemiaIogy, etiology, treatment and outcame.

Methodologic issues relating to the thesis are also addressed. This chapter

also defines the knowledge base that resulted in the development of the present

thesis question. Deficiencies in knowiedge related to measurement of outcomes

w r e addressed ptior to conducting the main thesis prospective study and are

presented as separate artideslchapten. Chapters 2-7 build a foundation for the

primary research question by describing the studies which were conduaed to

develop a reliable and valid measure of wist-related pain and disability and to

evaluate the reliability of secondary outcomes (physical impaiments) and

radiographie rneasures. The final chapter disasses the study findings, their

implications and out1 ines directions for future research.

Epidemiology of Wrist Fracture

Distal radius M u r e s are the most prevalent of al1 fractures (Larsen &

Lauritsen 1993). They ocwr in al1 age groups secondary to trauma (Larsen &

Lauritsen 1993; Lawson et al 1995) but have an increased incidence in

postmenopausal wmen related to osteoporotic changes in the bones (Dias et al

1 987; Lauritzen et al 1993a; Lauritzen et al 1993b; Mallmin 8 Ljunghall 1 994).

Fractures of the distal radius comprise 74.5% of al1 forearm fraaures (Alffram 8

Goran 1962). The incidence of distal radius M u r e has recentl y been

estimated as 27 per 10000 per year (16 for males and 37 for females) (Docri &

Renner 1994). Estimates ftom oher studies indicate an overall rate bebeen 23

and 6711 0,000 (Doczi & Renner 1994). Data from Rochester, Minnesota

suggest that incidence rates may be higher in North Arnerican communities

(Owen et al 1982). This study, b a W on a 30-year population cohort

demonstrated that the incidence of distal radius fracture associated with

moderate or severe trauma among female, Rochester, Minnesota residents rose

from 100 (per 100,000 person years) at ages 35-44 to more than 500 for those

aged 55-75 years. Based on the Minnesofa data. it has been estirnateci that the

lifetime risk of distal radius m u r e for white m e n at age 50 is 1 5%

(Cumm ings et al 1 985; Peltier 1 964). Under-40 years of age, males are 1 -4

tirnes more likely to have a wrist fiacture than fernales. Thereafter, there is a

near linear rise in the incidenœ, with fernales 60-94 years of age being 6.2 times

more Iikely to sustain a wrist fradure Vian males (Singer et al 1998).

A number of factors have been shown to be risk factors for distal radius

andfor hip fracture. These include aga and sex, (Larsen & Lauritsen 1993;

Mallmin 8 Ljunghall f 994; Owen et al 19û2) race, (Farmer et al 1984) nonuse of

estrogen therapy post menopause, (Hutchinson et al 1979; Weiss et al 1980;

Williams et al 1982) thin build, ~ i l l i a m s et al 1982) alcoholism, (Hutchinson et

al 1979) and cigarette smoking (Williams et al 1982). The factors reported to

modify risk of fracture are numerous; however, the primary underlying

mechanisms appear to be bone quality and rïsk of falling.

Low bone mineral density has been show to be a strong predidor of

distal forean fractures (Kelsey et al 1992). Certain risk factors such as

smoking, weight, race and menopause Mect risk of distal radius M u r e through

their effect on bone strength. The rising incidence of distal radius ftacture Mer

menopause is assoc1*ated with Osteopototic bone loss. m i l e osteopurosis is

widely acceptecl as a cause for DRF and it is considered that DRF is an early

sign of osteoporosisl (Mallmin 8 Ljunghall 1994) this relationship has been

questioned by several investigators Hlho reported that mineral content of bone in

patients DRF is neariy the same as in a g ~ t c h e d wntrols kaout M u r e

(Krolner et al 1982; Lamke et al 1978). Furthemore, the rise in incidence in

wrkt fractures in females occuro at age 40, leading others to question the role

of menopause as a primary risk factor (Singer et al 1998).

Other factors associated with distal foream M u r e include: poor viswl

acuity, number of falls reported in a year and frequent walking (Kelsey et al

1992). These fadon, refiect the risk of falling, which is the primary mechanism

of injury for distai radius fradure. A fall from level grouid acawnts for 87% of

injuries in females and 64% af men (Graff & Jupiter 1994). Falls are more likely

to occur in middle aged wmen than in middle aged men which may pactially

account for the increased incidence of fracture in middle aged females. The

dwerential in rates of falling is l a s extreme behrvean the sexes in the elderly

population. In addition, the elderly are le- Iikely to put out an outstretched

hand to break their fall and therefore are more likely to land on their hip

(Femandez 8 Jupiter 1996). That DRF occurs as a result of increased risk of

falling in rniddle aged females, has been supported by wwk which has

demonstrated greater postural instability in these patients (Crilly et al 1987;

Femandez 8 Jupiter 1996)-

Severity of lnjwy

Normal Anatomy

The distal radiocarpal joint and distal radiwlnar joint (DRUJ) fom the

"wnkt", which comprises three separate articulations. The scaphoid fossa and

lunate fossa are hm concave articular surfaces on the distal end of the radius

which are separated by a dorsal-volar ridge which articulates with the lunate and

scaphoid. A separate articulation. the sigmoid notch, is present for the head of

the ulna. The distal articular surface of the radius is aligned to the longitudinal

axis of the radius at 1 1 O of volar tilt (palmar angulation) and 22" of radial

inclination (angulation). Radial length Rom the tip of the styloid to the ulnar

head is 12mm (Szabo 8 Weber 1988). The ulnar side of the wrist is supported

by the triangular fibrocartilage cornplex (TFCC) which articulates Moi both the

lunate and triquetnim. Bony anatomy is evaluated through radiographs. either

with direct measurements or classification of overall type of injury; both are

describeci in more detail belaw.

Etiofogy

Fractures of the distal radius typically ocair as the result of a fall on an

outstretched hand. There is generally a sharp fraaure on the palmar aspect of

ine radial metaphyseal area with those that have a shear or compression

component producing intra-artiwlar fractures that are more unstable. A Colle's

fracture involves the distal metaphysis of the radius, which is dorsally diaplaceâ

and angulated. It occurs within 2cm from the distal end of the radius but may

extend into the radiocarpal or uhocarpal joint Dorsal angulation. dorsal

displacement, radial angulation and radial shortening are expected. There may

be an accompanying m u r e of the ulnar styloid. Smith's fracture are volar

angulated fractures of the distal radius and occur relatively rarely. A Barton's

fracture is a fracture-dislocation in which the rim of the radius is displaced wïth

the hand and the carpus. Components of injury that are thought to be

indications of severity of anatomic disruption include: involvement of the joint

surfaces, the extent of displacement of bone fragments and the number of

fragments (Graham 1997; Jupiter 1991 ; McMurtry 8 Jupiter 1996).

Associated soft tissues are often injured with a fracture. Ligament or

TFCC injuries are common and can be missed dumg initial treatrnent as they

may not be evident Virough radiographie examination (Fischer et al 1996). It

has been demonstrated that chondral lesions occur in 32% of intra-articular

distal radius fractures. TFCC tears occur in 78% of cases and the scapholunate

ligament is at least partially tom in 54% of cases. Soft tissue injuries can

account for impairment in the preseme of welt-healed hgctures (Lindau et al

1997). While soft tissue injuries have an impact on the impairment experienced

by a patient mi a distal radius M u r e , they are not directly asseSSBd by

radiographs nor accounted for in classification scfwrnes.

Stability is related to the type and extent of anatomic disnrption.

lnstability has therapeutic implications as unstable fredures ~ Ï l l require more

sophisticated intewentions to maintain anatornic position. Stable fradures are

usuall y extta-articular with mitd to moderate displacement, such mat *en

reduced. they do not redisplace to their original defomity. Unstable fractures

are more commonly cornminuteci and shortened and have artiwlar fradures that

involve not only the radiocarpal joint but also the distal radiwlnar joint (DRUJ).

Radiolog ic Classification of tnjury Seventy

Classification systems

As stated by Jupiter ' There are few areas of skeletal injury where

eponymic descriptions have enjoyed such longevity as in fracture of the distal

end of the radius" (Femandez 8 Jupiter 1996). Classification systems describe

the severity or type of fracture. Ideally, classification systems should not only

describe the ftacture pattern but also guide treatment and prognosis. A number

of anatornic disruptions due to fracture are typically incurporated into the vanous

levels of a classification systems. Seventy of injury is intuitively thought to be

critical to outcome mer an injury, therefore, identification of a reliable and valid

rneasure of severity of injury is imperative. Classification an* other

radiographic indicators of anatomic position are thought to be important

prognostic factors in a multivariate madel describing outcome, consequently, the

reliability and validity of these measures is fundamental.

A number of classification systems have been reported, arthough none

has been universally adopted (Graff 8 Jupiter 1994). Early classification

systems were bas& on: the fracture line, the direction of displaœment of the

fracture, the extent of articular involvement and involvement of the distal

radioulnar joint (Lidstrom f 959). The anatomic variables selected Vary for

different classification systems. Frykrnan reported one of the first classification

systems and, as such, this system has been used in many clinical studies. The

eight subdasses dflerentiated between radiocarpal and radioulnar involvement;

however, the extent or direction of the initial displacement, dorsal comminution

or radial shortening were not addressed (Frykman 1967). Older (1965)

described a ciassification systern for extra-articular fractures that was based on

measures of dorsal angle, presenœ of dorsal comminution and length of the

radial styloid as measured on radiographs. Lidstrom (1959) based a

ciassification system on the direction of displacement and expanded on articular

involvement More recent classification systems have tended to include both

intra-articular and extra-articular fractures and consider treatment issues

(Femandez 8 Jupiter 1996).

A symposium on DRF resulted in a classification system which considen

joint involvement, displaœment and stability (Cooney et al 1990; Rayhack 1990).

Cooney's "universal classificationn has four basic types type four fiactures

(intraarticular, displaced and unstable) divided into a further four groups based

on their stability after redudion (Cooney et al 1990). This systern is thought to

be suited to guiding treatment. Femandez developed a classification system

based on the mechanisrn of injury and the enticipated M u r e pattern

(Femandez 7982). The classification of th8 A 0 (McMurtry 8 Jupiter 1996;

Muller et al 1990) separates more severe injuries and addresses the extent of

articular involvement This system was designeci to be compatible w*th

cornputer coding. Each bone and segment is given a number (forearrn 1Y2 and

distal segment #3) This system has three types based on the extent of articular

involvement (A, 6, C) with 27 subtypes (-1, 2 , -3) allowing a more detailed

description of the fracture type. Documentation of additional ulnar lesions

wuld produœ 144 possible combinations for DRF (Femandez 8 Jupiter 7996).

Melone divided radiocarpal articular fiadures into five types (Melone 1993).

McMurtry and Jupiter (1996) proposecl that three distinct dimensions were

in need of charactefization to aid in treatment of fractures- These considered

the anatomic type, the local factors and patient consideratians. Their view

addressed the role of both the patient and the injury and how these impact on

treatment and outcome. They considered local factors and patient

charaderistics which may impact on care. Their discussion of 'anatomic typen

is an injury severity scale in the spirit of previous classification systems. Their

classificetion described the fracture according to involvement of the articular

surface and the nurnber af parts ie fragments of bone of suffident size to be

manipulatedlfixateâ and considerd fundionally signifiant Extra-articular

fractures do not involve either the radiocarpal or distal radioulnar joint

lntraartiwlar m u r e s are either hepart, three-part. four-part or fiveor-more

parts,

The present literature does not adequately address Mich classification

systems are either reliable or valid. Good interobserver reliability has been

reported for Older's classification system wed for exîra-articular m u r e s

(Andersen et al 1991), however, this systern has not been used for intradrticular

fractures where outcome has been most unpredictable. Reliability has been

reported as cnly moderate for the Mayo classification and only fair for the

Frykman, Melone, and A 0 classifications (Andersen et al 1996)- Further intra-

observer reliability was only fair to moderate for 3 of 4 observers for these

scales. Caution was recommended in using classification as the sole means for

directing treatrnent and for direct compatison of results in different dinical

studies (Andersen et al ? 996).

Radiographie measurements

Severai diredly measured radiographie parameters have been reported

as indicators of anatomy. These indicatm of severity of injury can be used in

isolation or as wmponents of an injury dassification system. Five

measurements are most offen desaibed: radial inclination, radial length, ulnar

variance, radial tilt and radial shR On the Posteranterior (PA) radiograph the

inclination of the radial plaffonn is the angle between a line drawn fiom the tip of

the radial styloid to the most distal ulnar asped of the lunate faœt and a line

perpendicular to the longitudinal axis of the radius. The average radial

inclination is 22".

Radial length (also temed radial shortening) is measured on a PA

radiograph by detenining the longitudinal difference betiiiieen a Iine

perpendicular to the long axis of the radius drami a the radial styloid and

another tangential to the distal artiwlar surface of the ulna. By this method,

normal radial length is 11 mm. Another indication of the relative radial length

wmpared to that of the ulna is ulnar variance (or radioulnar length). Ulnar

variance can be calculated in a variety of ways and is indicative of the radial

shortening that occurs after a fracture. If a line is drawn perpendicular to the

long axis of the radius at the oigmoid notch; the portion of the ulnar head distal

to that line is rneasured as ulnar variance.

Palrnar (radial) angulation (tilt) is measwed on a lateral radiograph. The

normal palinamrd orientation of the articuler surface is measured by comparing

a Iine drawn through the volar and donal margins of the distal radius with the

perpendicular to the long a i s of the radial shaft The normal volar tilt is

approximately i 1 O. Dorsal tilting of the radius can ocair after a M u r e but

does not occur in the uninjured wist (Mann et al 1992).

Radial shift or radial wuidth is the distance between the longitudinal axis of

the radius and a line d m tangential to the radial styloid. This refleds the

shortening and radial collapse &en seen in comminuted distal radius fractures.

lt is related to radial indination. radial length and ulnar variance (Graham 1 997)

and tends to be used legs fiequently than other indicatm.

Although some authors recommend the cornparison of values with the

uninjured wrist, (Graham 1997) it has been demonstrateci mat aie measures

from one wrist are not a good predicator of values for the contralateral Wst

(Schuind et al 1996). Therefore, radiographic measures tend to be considered

in the context of m a t is considered normal. Adequacy of reduaion is evalwted

by the extent to Hihich normal or corrective anatomy is attained. AcceptaMe

radiographic measures after distal radius fracture have been described as:

Radial shortening <5mm, radial indination 2 1 5 O , sagittal tilt between 1 5 O dorsal

tilt and 20" volar tilt and s2mm of incongruity of intraartiailar surfa- (Graham

1 997).

The relative valuefreliability of various indices has not been clearly

defined and thus a number of different measures are reported in the current

l iterature (Palmer et al 1 982; Warwick et al 1 993). Evaluation of radial length

(shortening) has been suggested to be more prognostic of outcome than other

radiographic measures (Wllar et al 1987) and a preferable method of

standardized measurement has been established (Wawidc et al 1993).

A variety of radiographic parameters and classifications can be used to

quantify fracture severity. Inconsistent reliability in published data and a paucity

of validity data complicate aprior seledion of appropriate variables for statistical

analysis. For this reason, it was decided to focus on hm variables which are

frequentl y mentioned as predictors in the dinical Iiterature ie radial shorten ing

and joint involvement The first variable, radial shortening, has documented

reliability, whereas, the second variable joint involvement does not Joint

involvement, however, has a strong theoretical basis for inclusion. The lack of

substantive evidence on the value of distal radius fradure classifications

systems, the proliferation of systems and la& of comparative data on these

systems leads to the secondary thesis question on the reliability and validity of

these radiographie variables.

Treatment of Distal Radius Fracture

Intervention following distal radius M u r e is directed at restorïng

anatomical alignment of fradured bones, promoting repair of injured structures

and fostering normal fundion of these structures during the heaiing process

(Ark 8 Jupiter 1993; Collins 1993; Jupiter 1992). A number of studies have

proposed various treatment regimens for distal radius fiadure, although there

are few randomizeâ trials in this area (Abbaszadegan et al 1991; af Ekenstam

et al 1 989; AKelrod 1 991 ; Cohen & Frillman 1 997; Home 1991 ; Home et al 1990;

Hutchinson et al 1 995; Rournen et al 1991 ; Shankar & Craxford 1992;

Sommerkamp et al 1994). Fractures require redudion and fixation sufficient to

maintain reduction (Agee 1993; Bass et al 1995; Cooney 1993; Ledingham et al

1 991 ; Mennen & Labuschagne 1993; Raskin 8 Melone 1993; Rayhack 1993).

Fracture fDcation can occur through bracing, (Ferris et al 1989) casting, (Altissimi

et al 1 986; Altissimi et al 1991 ; Chester 1990; Pool 1973; Sarmiento et al 1 980;

Stewart et al 1984; Tang et al 1996) percutaneous pinning (Bennett et al 1989;

Carnonella 8 Stem 1988; Casteleyn et al 1992; Chapman et al 1982; Clanœy

1984; Cole 8 OMetz 1966; De Palma 1952; Greatting 8 Bishop 1993; Green

1975; Habemek et al 1994; Harper 1989; Hochwald et al 1997; Hutchinson et al

1995; Jensen 8 Jensen 1996; Lenoble et al 1995; Lucas 8 Sachtjen 1981 ;

Mil liez et a[ 1 992; Naidu et al 1997; Pany 1997; Rayhack 1 993; Sdieck 1962)

open redudion intemal fixation (ORIF) (Agee et al 1994; Axelrod et al 1988;

Axelrod 8 McMwty 1 990; Bradwy et al 1989; Culp & Osteman 1995; Geissler

8 Femandez 1991 ; Geissler 1995; Gesensway et al 1995; Hastings & Leibovic

1993; Hove et el 1997a; Jupiter et al 1998; Leung et al 1990; Mcûimie et al

1 995; Melone 1986; Rader et al 1995; Rikli 8 Regazzoni 1 996; Trwible et al

1993) or extemal fixetion (Agee 1993; Bartosh & Saldana 1990; Braun 8

Gellman 1994; Cannegieter 8 Juttmann 1997; Clybum 1987; Cooney et al 1979;

Edwards 1991 ; Femander 8 Flury 1994 Frylanan et al 1989; F-an et al

1 993; Goslings et al 1994; Graff 8 Jupiter 1994; Horesh et al 1 991 ; Hove et ai

1997b; Huch et al 1996; Hutchinson et al 1995; Jakim et al i 991 ; Jakob 1994;

Jenkins et al 19ûû; Kaempffe 1996; Lennox et al 1989; Melendez et al 1989;

Mehk et al 29% Merchan et al 1992; Nakata et al 1985; Patterson et al 1997;

Pennig 1993; Raskin 8 Melone 1993; Riis 8 Fruensgaard 1989; Rommens et al

1992; Schrnalhoiz 1990; Schuind et al i989; Seitz 1993; Se@ Jr- et al 7990

Seitz, Jr. et al 1991 ; Seitz et al 1990; Simpson et al 1994; Steffen et al 1994;

Vandersluis et al 1993; Verhaven et al 1991 ; Zanotti 8 Louis 1997) or

combinations of these methods (Bass et al 1995; Leung et al 1989; Putnam 8

Fischer 1997). In extemal fixation a variety of deviœs are available, some which

provide uniplsnar and some multiplanar l igament~t~s (Agee 1993). Bone

cement (Schmalholz 19û9; Schmalholz 1990) or bone grafting has also been

used (Cannegieter 8 Juttmann 1997; Jebson 8 Blair 1996; Leung et al 1989;

McBimie et al 1995)-

Radiographs and other imaging techniques are used to ass8ss the

anatomic positioning before and after fiadure reduction. Arthroscopy can be

used to visualire fradure fragments during reduction to improve alignment of

intra-articular ftactures (Culp 8 Ostman 1995; Geissler 1995; Whipple 1995).

Arthroscopie guided reduction is indicated Men articular ste- greater than

2mm remains after attempted closed reduction, or when a carpal Iigamentous

injury ador distal radioulnar instability is suspected. Arthroscopy aids in

restoration of the articular surface as wll as allows for removal of srnall loose

bodies of bony or cartilaginous matefial. Assocïated ligamentous injuries not

apparent on radiograph may be visualized and treated. Various positions and

durations of immobilization and rehabilitation procedures have been describeâ.

(Christensen et al 1995; Dias et al 19û7) Both supination and dorsifiexion have

been advocated as positions of immobilization (Gupta 1991 ; Sarmiento et al 1980).

The various mades of treatment offer dif5erent ways of achieving anatomie

reduction of M u r e s . Different fracture patterns may require specific types of

fixation to maintain anatomic redudion, aithough &en, ftactures can respond to

a variety of treatment options (Cooney 1993; Doczi & Renner 1994; GrafF &

Jupiter 1994; McMurtry 8 Jupiter 1996). The treatment plan for a patient is

individualized and based on a number of facfofs ïncluding: the fiactufe pattern,

bone quality, cornminution, saft tissue damage. medical conditions, patient age,

mental status and cornpliance (Grafï 8 Jupiter 1994; McMurtry 8 Jupiter 1996).

In general, castinglimmobilization tends to be used for undisplaced or

stable fractures. (Dias et al 1987) K-wireslpinning tends to be used in fractures

wiih no more than fragments, (Clancey 1984) whereas, comminuted,

severely displaced, comminuted or unstable fractures tend to be treated wïth

extemal fixation (Clybum 1987; Cooney et al 1979; Howard et al 1989). Cooney

et al (1993) describe a treatment algorithm which coincides with the universal

classification system. They suggest that undisplaced extra-articular, stable

displaced extra-articular and undisplaced intraarticular fractures should be

treated with cast immobilization. Displaced unstable extra-articular fractures are

treated with perwtaneous pins whereas displaced reduced intra-articular

fractures are treated with extemal fixation /percutaneous pins- Displaced

irreducible intra-articular fractures should be treated with open recluction,

extemal fixation, K-wkes andlor tplates. Complex fractures are treated with

open, arthroscopie, or closed reductions, extemal f-ion, bone grafting,

supplernental k-wires and intercarpal ligament repain (Cooney & Berger 1993).

This algonthm describes the decision making process that surgeons use to

formulate a treatment plan. A number of options for redudion and treatment exist

in al1 but the more simple fractures Mich illustrates that physician preferences

or experienœ remain an influence in treatment seledion. McMurtry and Jupiter

(1 996) emphasize that a repertoire of treaûnent choices provides the

orthopaedic surgeon wïth the capability of matching the patient needs and

fracture type to the tr8afment option most tikely to provide for success. Despite

improvements reporteci in outcorne in comminuted M u r e s , aigre rernains

difficulty in controlling impacted die-punch fragments and difficulty in controlling

the reduction of an unstable distal radioulnar joint (Cooney et al 1979).

Issues in Outcome Measuremnt in Upper Exaernity Orthopaeâics

Outcome measurement has become an important element of dinical

practice in recent years within a health care field dominated by increasing

demand and shrinking resources. There has been pressure frorn govemment,

other funding sources and professional bodies to incorporate standardized

methods of measurement of patient outcomes as part of ongoing evaluation of

clinical practice. To analyse variations in dinical practice and how these

contribute to outwme requires standardized measurement tobls and

standardized measurement protocols. M i l e the field of ortho paedics. l i ke most

other areas of medicine, tries to incorporate this philosophy into standard

medical care there are diffïculties due to lack of instruments, time or data

management capabilities (MacDemid et al 1998; Sanders 1 997). The recent

healthcare trend to incorporate patientieported outcomes ha$ changed the

focus towards these measures and away from physical capacity (impairment)

measures, such as strength, movement and fundional tests (Jette 1995).

Current health outwme evaluation practices Vary considerably between

centres, specifically in relation to: the types of measurement tools utilized, time

periods at which outcomes are reported, consideration to possible modifying

variables, data colledion, utilization and reporting strategies.

The World Heaith Organization defined teminology and concepts around

health in 1980 (World Health Organization Wûû) and is wrrently revising this

work Levels of health descriptors w r e describeci as: impairment, disability and handicap. Impairnent is any loss or abnomlity of psychological. physiological

or anatomical strudure or fundion. Disability is any resfridion of ability to

perfom an adivity in the manner withni the range considered normal for a

human being. Handicap is the disadvantage for a given individual that Iimits or

prevents fulfilment of a role that is normal for that individual and thus by nature

considers age, sex, social and cultural factors. Revisions are expeded to adopt

a more positive teminobgy by renaming disability as =adivity limitation' and

handicap as "participation restncüonn (Anonymous 1997). While previous

orthopaedic Iiterature has foaissed on impairment measures of outcome, it has

k e n recugnized that a shR has ocairred in measurement of outcome to focus

on disability (Jette 1995). Measures of impairnent for upper extremity

orthopaedics fows on either indicators of abnorrnal anatomy ie radiographs or loss in physical capacity ie grip strength, and range of motion of the wrist joints.

Other impairments such as loss of sensation, decreased foream and pinch

strength and loss of dexterity are known to occur after a DRF, but are

infrequently reparted (Arnadio et al 1996).

Recognizing the shift in the health care field towards measurement of

standardized patient -rated outcornes, a number of professional associations

have tried to facilitate operationaliration of this philosophy by assisting Ath

developrnent of databases (UeNet, Mdems) or tools such as the DASH

(Disability of the atm, shoulder and hand - a questionnaire developed by the

American Academy of Orthopaedics). The American Society of Hand Therapists

developed UeNet, a computerizeâ hand therapy charting system which

performs patient documentation, evaluation and billing and can upload this

information into a large nationwide database. These databases do not yet have

sufficient sample size or consistency in mode of data colledion to be useful for

epidemiologic research.

Clinical pradice is moving towacds standardized outcorne data colledion

as a wmpanent of standard medical care. Infmation from outcorne

'databases" serve various functions in the dinical setting. Dun'ng clinical caret

databases are used to monitor individual progress. Program evaluation and/or

outcorne studies use the data to provide documentation of prograrn/freatrnent

effectiveness, Hwever, these data can alsa serve a source of observational

data for epiderniologic research questions. Pradice patterns in ternis of data

colleaed and foflowup proœdures for orthopaedic patients Vary from œntre to

centre, making large scale epidemiolog ic studies diffiailt. S tandardization of

outcome procedures w w Id facil itate cornparison of cl inical results between

centres and through published media. Furthemore, the reduction in

measurement error associated with standardized procedures is essential for

research analyses.

The recent development of the DASH (Disabilities of the am, shoulder

and hand) was intended to facilitate between œntre reporüng of upper extremity

patient reported disability (Hudak et al 1996). This questionnaire focuses on the

upper extremity and the effect d pathology on patient fundion and health

domains. It emphasizes function, as 21 items are specific fundional items, 2 are

pain, 3 relate ta other symptoms (sîiffness, weakness and paraesthesia), 1 to

social role, 1 to work rde, 1 ta sleep and 1 one to perceived capability. The

DASH is considered a region-specific, quality-of-life scale. The DASH was not

available during the initial design of the study but ~ w s added to the data

collection process and thus evaluation of its measurement properties for DRF

became a secondary question.

Another approach to outcame evaluation is to use generic health status

instnirnents as outcome tools ie the SF-36 or Sidviess Impact Profile. The

advantage to generic tools is that they address a variety of health domains and

allow one to compare across a variety of health problems induding mental and

physical disordem. The mnœm in orthopaedics is that aiese tools may not be

sensitive to upper extrenity pathology. The SF-36 is the generic health

instrument used at the Hand and Upper Limô Centre (HUE )( Iiœnsed user.

J. C. MacDemid: ID # 2841 3) (Anonymous 1 994a; Anonymous 1 994b; Ware et

al 1993; Ware et al 1994). The SF-36 has been used to evaluate DRF in f o l l w

up studies (Femandez et al 1997) and some subscales have been shown to be

responsive to DRF recovery (ie Physical role) (Arnadio et al 1996).

Responsiveness is fundamental to the performance of outwme instruments in

clinical evaluation. consequently, it is a d d m as a secondary thesis question.

Issues in Raüng Outcome Afbr Distaî Radira Fmctum

There are a variety of outcornes mer injury and treatment. The definition

of outcome varies according to the problem, purpose of the evaluation and

perspective from which the evaluation is performed. Distal radius fractures

cause immediate pain and loss of fundion. Problem pain. motion loss and

weakness can result fnmi malalignment. soft tissue injuries andlor complications.

Previous reports of outcome after distal radius fradure are plagued with a

variety of problems related to conceptualization of wtcome, experimental design

and la& of precision in measurement of outcome. To date measurement of

outcome has focussed on measurements of impairment such as radiographie

position, grip strength and range of motion. Despite these limitations, it is knom

that distal radius fracture can cause signifiant disability in ability to perfom

activities of daily living, occupation or leisure. The outcome after distal radius

fracture can range from a complete return to nomal fundion to severe long terni

impaiments (Field et al l992a; Steffen et al 1994).

The time to achieve a final outwme after distal radius fracture is

debatable. Short tenn follaiiwp does not allow sufficient time for hiIl recovery of

injured tissue or for rehabilitation to achieve its full result Conversely, long

term complications, sudi as osteoarthritis, may result in a loss of fundion over

time. Field (1 992a) suggested that dinical resuits improve Iittle af&er three

months Mer observing that 8996 of patients had satisf-ory results after three

rnonths and 85% after 10 years. Hwver , when comparing the change in

scores for individual patients over time, 42% improved, 38% remained the same

and 20% deteriorateci by no more than one grade. FurViennom, as this sample

airnprised nonoperative patients, results may not refiect the recovery time

required where more aggressive treatment interventions are used and recovery

from surgical proωures may be required. As most outcorne studies are

retrospective, they encompass a variety of follow-up times. One prospective

study of 90 distal radius fradures which evaluated fadon Mich M8d outcome

(grip strength and ROM ) used three years pst-fracture as the time of outcome

evaluation (Vïllar et al 1987). As these authors did not present data from earlier

time frames, the natural history of recovery was not defined.

This study will use six months as the outcame time of interest. This is not

necessarily the final outcome for some of the patients. Most patientslclinicians

anticipate recovery frorn a mist fracture to approach normal fundion Hiithin six

months. The extent to which this is not possible refiects a burden to society.

Longer term followup may result in higher loss to follow-up, compromising

study validity without providing additional information. It is anticipated that minor

improvements may ouAir over the 6-12 month time frame but that these will not

rnarkedly affect the outcame or conclusions based on outcome evaluations.

Methods of Rating of Outcom After Distal Radius Fmturo

There are several outcome rating scales for distal radius fractures that

have been used in previous case series. Gartland and Wertley (1 951 ) published

a sale which vms modifiecl by Sarmiento (1973). This score is a demerit

scoring system ie patients lose points if they have certain dinical problems.

This scale provides a score based on the pain and disability recorded by the

clinician, ranges of motion, strength, complications, and radiographic

appearance. This scale has not been fomally validated but has been

frequently reported in case series. It has been suggested that the scale may be

"flattering in that in order to have a 'poof result ie a loss of 21 points, a patient

would have to cornplain of pain, stifhess, restn'cted adivity, have a deformed

wrist, loss of dorsifiexion and supination and have severe osteoarthritic changes

in the joint A study nihich used a modifieci scale inwrporating direct

measurement of motion and strength into outobme concludecl that only 56% of

patients had good to excellent results compared to 81 % vvhen evaluated by the

Gertland and Werley system (Bradway et al 1 989). A radiographic outmme

scale wes proposed by Lidstrom in 1959. This scale addresses dorsal

angulation and shortening and ha$ four levels. While useful for describing the

anatomic results, radiographic appearance cannot be considerd to represent

functional performance (Bain et al 1995).

Conœptually, rating of patient fundion rnay be best performed by the

patient as they are the ones who experience the effects the injury. Despite this,

past scales have been based on the clinicïans recording of patient function.

These scales were ~ ~ 8 f u l to allow dinicians to grossly compare their treatrnent

results. Wth the recent increased focus on having standardized patient rated

outcorne scales incorporated into dinical practice and research studies, the

limitations of these scales have been more recognized (Cole et al 1993).

At the time this thesis was k i n g designed there was no tool other than

generic health tools which were appropriate for use distal radius fracture.

There was a need to have a tool which allowed one to capture the patient rating

of pain and disability relatecl to their affeded wn'st. The work to develop a

standardized tool for patient rating of wrist pain and disability was a neœssary,

preliminary step required to complete the prospective study. Despite this, the

development of the scale became a major foais and product of the study

19

(MacDemid 1 996; MacDermid et al 1 998). This resultant outcome tool, the

Patient Rated Wrist Evaluation (PRWE) has been published and is in use at a

number of health centres to evaluate patients M s t pathology (MacDennid

1996). (See induded artides on developrnent, reliability and validity of the

scale). It has b e n modified for use with tennis el- patients by adding

forearm specific items (Overend et al 1999). It has been accepted as an option

in the Dexter Outcornes software which is osed by al1 Modems data colledon

through the AmefÏcan Academy of Orthapaedic Surgeons.

Previous Studierr of Predictors of Outcorne

Fracture redudon and fixation are the immediate goals of intervention in

alf fractures. Appropriate treatment selection and application is essentiaf to

attaining the best possible outcome. Treatment selection is not fully defined for

distal radius fractures. Few randomized trials exist, with case series k i n g the

primary mode of evidence in this area. Case series demonstrate that even with

similar fractures and treatment regimes, outcome can V a r y wïdely. A variety of

non-treatment factors could affect outcome. Some of these factors muid be

present at baseline (Tnimble et al 1994) and others may be complications that

develop during the treatment process (Altissimi et ai 1994; Bromi & Bell 1994;

Kwasny et al 1994).

While certain baseline charaderistics have been previously reported as

predictors of outcome, (Tnrmble et al 1993; Trumble et al 1994; Villar et al 1 987)

these have been studied, most frequently, through univariate analysis of

retrospective studies. Several authors have supported the viewpoint that

restoration of articular congniency is important in the long tem clinical and

rad iographic results Mer distal radius fracture (Trumble et al 1 993; Villar et al

1987). However, controversy exists as to whether anatomy is an important

indicator. A prospective three-year followup study of 90 consecutive distal

radius fractures lwked at Pearson correlations betuwn a varkty of variables

including grip, radiogaphic parameters and range of motion. Shortening of the

radius at one week was the most significant correlate of grip strength.

Shortening also influenceci range of mist flexion and rotation (Villar et al 1987).

This finding was supported by othen, in retrospective reviews (Eelma &

McElfresh 1 983; Lidstrom 1959; Older et al 1965; van der Linden 8 Ericson

1981 ). Adams (1 996) suggested ihat the greatest change in kinematics at the

DRUJ occurs as a resuk of radial shortening, with loss of radial indinaton and

dorsal tilting causing more moderate effects. He also observed that defomity of

the radius alone with an intact TFCC did not permit dislocation of the DRUJ.

The DRUJ is not the only source of impairment after a distal radius fradure. but,

it does comprise a significant articulation of the wrïst and irnpaiments of the

DRUJ resulting from a fracture do affect the overall result (Graham 1 997).

Some authon, have suggested that intra-articular involvement, (Eelma &

McElfresh 1 983) volar tilt, (Rubinovich & Rennie 1 983) injury score based on the

number of fracture fragments, carpal injury and fracture classification (lnimble

et al 1994) are correlated *th outcorne. Femandez obsewed a pathologie

displaœment of the flexion-extension arc with a change in radial tilt greater than

25' (Fernandez 1995). Coss of radial inclination with malalignment has also

been cited as a source of loss of function (Lindau et al 1997).

Residual displacement of 2mm or more in the artiailar surface or

extensive dorsal cornminution has been proposed as an indication for open

reduction intemal f~ation (ORIF) (Cooney et al 1979; Knirk 8 Jupiter 1986). In

reviewing 26 displaced intra-artiailar fractures of the distal aspect of the radius

after a mean of 7.1 years, osteoarthrosis was evident in 76% of the patient%

wrists. A significant association was found be-n residual dispiaœment of

articular fragments at the time of union and osteoarthrosis @=O-74). However,

fundianal status as measured by physician examination was not significantly

conelated with residual step or gap displaœment (r = -0.40). The authors noted

that even patients with amrosis experienced good or excellent fundional results

(Catalane et al 1997). A number of authors have supported the view that

anatornic restoration is the key to fundion (Porter & Stoddey 1987; van der

Linden & Ericson 1981 ). Laboratory studies have al- supparted the concept

that malalignment impairs fundion in the short terni as well as leads to arthrosis

in the long terni (Pogue et al 1990; Short et al 1987).

A preponderance of studies suggest that anatomic fadon play a role in

outcome. Anatomic variables can be measured at basdine to indicate fiacfure

severity or immediately post reduction ta indicate the adequacy of the reduction.

Delayed post redudion radiographs ie followvp radiographs detect pst-

treatment loss of anatomy through M u r e slippage, or non-union. Delayed

anatornic abnmalities are complications. Although complications may a f k t outcome, they are not known to the treating physician et baselnie and therefore

cannot be a component of initial prognostication. The rnajority of studies have

focussed on final anatornic alignment as a predidor of outwme. However,

Stewart et al (1984) wnciuded that severity of initial displaœment had more

effect on functional results than did the acairacy of reduction. Severity of initial

displacement may also be associateci with severity of soft tissue injury. Soft

tissue injury is not diredly measured in DRF injury classifications or bony

displaœment measures. Howver, it may have substantive impact on final

outcome. Confounding this possible relationship, however, is the fad that bone

quality varies across individuals and thus displaœment is not necessarily

indicative of a similar amount of trauma across individuals.

McMurtry and Jupiter (1996) suggest that treatment and outmme are

rnultifaciorial and considerations include fracture type, cornminution, bone

quality, energy of injury, displacement and patient considerations. Patient

considerations specifically mentioned included substance abuse, age, hand

dominance, occupation and lif8style. Few studies have addressed patient

factors as predictots of outcorne,

An et al (1988) dernonstrated that 68% of the patients with traumatic

injuries Hlho deve10ped refiex sympathetic dystrophy (MD) smoked, whereas

only 38% of nonsmokers developed this complication. Although it is generally

accepted that RSD has an adverse Med on outcorne, its diagnosis remains

controversial, (Atkins et al 1989; Atkins et al 1990; Cooney et al 1980; Field et al

1992b; Pollock et al 1993; Steumrt et al 1985; Zemel 1987). Abnomal pain. loss

of motion and loss of fundion are the clinical symptoms that define RSD (Atkins

et al 1990; Giedc 8 Buton 1987; Merchan et al 1992). Whiie some have

suggested that a airee-phase borie scan is diagnostic for RSD (Mackinnon 8

Holder 1984), this not universally accepted. The diagnosis is more frequently

made on the basis of the symptomatology. Pain, loss of motion and loss of

function are natural consequences of a wrist fradure. The point where these

symptoms become sa severe that it is uabnorrnal" is a sumaive decision.

The ladc of agreement on diagnosis of RSD may account for the wicie

variation in reported incidence rates for RSD Mer DRF ie from 0.1 % (Bacorn &

Kurtzke 1953) to 25% (Field et al 1992a). A follow-up of patients who

developed RSD after distal radius fracture demonstrated that 25% of patients

still showed features of the syndrome after ten years (Field et al 1992a). A

study, Hihich examined 125 cases of RSD Mer a varïety of injuries, reported

that 14.5 months after diagnosis 35% were officially disabled, only 30% had

retumed to work and only 38% w e able to do 75% of their premorbid activities

(Subbarao 8 Stillwell 1981). Due to the controversial nature of diagnosis of

RSD, the purpose of this study is not to delineate the incidence of RSD, but it is

expected that some patients in the study sample m'Il develop RSD and that this

will Mect their outcome.

Previous studies that have addressed outcome have tended to focus

primarily in radiographie rneasures- There is an absence of studies Mich

consider patient charaderistics, fracture type and displacement in a

multifactorial mode1 to determine predictors of outcome.

Theoretical Aspects of Prognostic Studim

Prognostic studies typically involve a multivariate analysis of a relevant

pool of variables to detemine which variables are related to an outcorne of

interest, the dependent variable. There are tw basic reaaans for idenWying

which variables are related to an outcome: explanation and predidion.

Explanatory models w-sh to explain an outcome accordin9 to a biologically

plausible hypothesis. The ultimate goal is a better understanding of the

underlying mechanisms involved in occurrence of the dependent variable-

Predictive models are developed to predid an O U ~ C O ~ ~ without concem for the

underlying biologic mechanisms. The primary objective of performing the study

has implications mth respect to variable selection. In an explanatory study,

variables are justified on the basis of previous wrk. In a pfedidion study, any

variable that is thwght to be related to outcome, is a viable choice. A

predidion study emphasizes practical applications. Therefore the goal is to

develop a simple model, from pradically obtainable variables which provide

amrate predidion. Interactions and acairacy of individual parameter estimates

are minimally important as ernphasis lies on variance explaineci- In this thesis,

variables which might predid outcame wwe sooght through literature review and

clinical insight The goal w s to identify variables which dinicians rnight readily

consider during routine clinical pradice, therefore, only simple, pradical

predictors were wnsidered clinically important.

While statistical packages have eased the computational burden of

multivariate analysis, there is inueased potential for abuse of the methodology.

A reasonable analysis includes a limited number of variables that have a

convincing rationale for inclusion either on the basis of a strong scientific

rationale from the literature or pilot dinical data (Norman 8 Streiner 1994).

Analyses where large numbers of potential variables are entered, even with an

appropriate sample size, can be expected to produœ some falsely positive

predictors.

Reliability of Measumment

Reliability is the degree of stability when a rneasurement is repeated. It

can be defined within raters (intrarater reliability), betwwn raters (interiater

rel iability), between occasions (test-retest reliability) or betwwn instruments

(inter-instrument reliability). Reliability is an important component of clinical

rneasurement, yet, many routine patient assessrnent tachniques have not been

examined in ternis of their reliability. Because it iiiras the intent of Uiis study to

use standardized dinical measurement techniques, a number of acœpted

cl inical ptactices needeâ to be evaluated for aieif reliability.

Methodology for ass~ssing reliability can be cornplex and controvenial.

There is not unifom agreement on the type of statistics that should be used to

evaluate reliability or on the standards which measurements should reach for

various applications. This is partially due to the faa that reliability requirements

Vary according to the specified application and partially due to different

perceptions of reliability.

As stated by Fleiss (1986a) ' The most elegant design of a clinical study

not overcome the damage caused by unreliable or imprecise measurement".

The consequences of unreliability include attenuated correlations, increased

sample size requirements and biased sample setedion (if selection is based on

a measurement). Fleiss illustrateci that sample correlations may be seriously

underestimated and/or found to be insignificant, where a tnie correlation exists,

due to the effects of unreliable measurement- Attenuation of obsewed

correlations is a partiwlar problem with respect to the primary research question

on prediction of outcorne. Prediaion studies rely on correlations (R2) to

detemine the dinical importance of identifieci predidors Furthemore,

secondary thesis questions address the validity of the PRWE Validity studies

are also based on observed correlations. Consequently, the ability to evaluate

validity is also thwarted Men a reliable measurement protocol has not been

estabiished. Th-, reliability is fundamental to the intemal validity of thesis

research questions-

Reliability is also important frorn a dinical perspective. Measurements

performed on patients on difterent occasions or by difFwent dinicians should

provide the same (or very similar) results, unless there has b e n a change in the

clinical phenornena k i n g measured- Unstable measurements make it more

diffiwlt to distinguish tnie clniical change from measurement error.

It has been suggestted mat health status measures should be tested for

stability over a 1 -2 wek period (Streiner & Norman 1 991 ). Retest situations

which allow patients to cornplete responses within 1-2 days are subject to recall

bias. Longer intervals may be subjed to nonspecific changes in patient status.

However. Strafford suggests tbat assessment of long tenn stability is afso

beneficial (Stratford et al 1 996). When evaluating reliability, it is important to

retest over a pend when tnre clinical change is not occurring-

Clinical measures include true cardinal data where an unlimited number

of responses are possible, very ordered ordinal data Hihere a large number of

well ordered categories exist, less uordered" ordinal (ie f ~ r l l e s s well defined

categofies exist) data and nominal data - where no order exists b e W n

categories. Different statistical methods have been proposed for assessment of

reliability such as: the Pearson correlation coefficient (r), the in tradass

correlation coefficient (ICC) and kappa (k). Selection of statistics for assessing

reliability may be basai on the type of data or on other considerations.

It has been pointed out that the Pearson productinoment correlation has

been used exîensively to assess reliability, partiwlarly in the quality of life arena

(Deyo et al i99l). The problem with this approach is that measures may be

systematically different and yet highly (even perfectly) correlatecl (a perfect

straight Iine relationship, but slope t 1 and interœpt #O). Thus the potential

exists for false conclusions on reliability to be made M e n the Pearson r is used.

However. in practical ternis, Me predaninant source of error is usually random

error and under these cirwmstanœs. the Pearson r and the ICC m'Il be very

close (Streiner 8 Norman 4 991 ). As an ïnterciass correlation COBfficient, the

Pearson produd moment correlation is appropriate for comparing the

association be-n measurements of dinerent Wasses", Were aie

relations hip be-n va riables is of interest.

Where the consistency of cardinal data is an issue, the appropriate

correlation is an intracla- correlation coefficient (ICC). This statistic assesses

not only the strength of the correlation, but also whether the dope and iiterœpt

of the regression Iine between the im measures Vary from those expected with

replicate measures ie slope=l and intefœpt =O (Deyo et al 1991 ; Streiner &

Norman 1991 ). Simply, the ICC is the proportion of total variance accounted for

by the vatiability amongst persons. Where most of the variability L accounted

for due to subject variability and little ia due to intrasubject or examiner

variability, then a measurement will attain a high ICC. One disadvantage of this

rnethod, is that Mer8 the sarnple is homogeneous, the ICC tends to be low

(Streiner & Norman 1991).

Shrout and Fleiss (1979) describe calailation of three different foms of

the intraclass correlation coefficients and associateci confidence intervals based

on how raters are chosen and assigned to subjects. In Type-1 each subjeds is

assessed by a set of different raten randomly chosen from a pool of raters. In

the Type-2 CC, the same raters assess each patient and are these raters are

randomly selected from a pool of raters. This research design should be

generalizable to raters from the same pool of raters, testing similar patients. In a

Type3 ICC, each rater assesses each subject, but the raten, are the only raters

of interest (ie fd). Each of the thme m a l s can be expssed in two foms

depending on whether isolateci or poaled raters' scores will be utilized. The type

of I CC selected for a given reliability study is based on the required application.

The analytical differences betwwn foms of the ICC relate to the number

of components of variance which are calculated and whether ratez's are fixed or

random. In a Type 1 ICC, variance is calarlated as either between or within

subjects (Shrout & Fleiss 1979). In Type 2 and Type 3 ICCs, based on twpwiay

repeated measure ANOVAs, variance estimates are also attributed to raters. In

a Type 2 ICC . the raters are randorn, vrtiereas, in the Type 3 they are fixed.

Other foms of ICC have been desaibed and the appropriate seledion of

research design and fom of ICC has been discussed (Bartko 1976; Eliaszk et

al 1994; Leung et al 1990; Muller 8 Buttner 1995; Vargha 1998). The ICC

described by Shout and Fleiss remeins Vie most *dely utilized in fhe dinical

Iiterature-

The strength cianiied for the ICC is iYs ability to express reliability in a

single number (Deyo et al 1 991 ). The weakness of the ICC is that it states

reliability as a single dimensionless number. It is not framed in a quantitative

method, readily interpreted in ternis of the amount of dinical emr to be

anticipateci (Fleiss 1 986b; Streiner & Norman 1 991 ). A variety of methods have

been desaibed to deal wïth problem. The standard error of measurement (SEM)

is one (Streiner & N m a n 1991). The SEM is based on the standard deviation

and the reliability (CC) and is expreosed in the same units as the base

measurement. Bland and Altman (1988) described a technique vrtiere

differences between successive measures are plotted against the average of the

measures. This allaws a visual inspection of how measures Vary over the

spectrum of clinical measurement, the average difference or bias between the

two measures and the 95% confidence intenral for the average difrenœ. This

method is particularly useful where bias may exist. such as when different

instruments or techniques are us& to derive a clinical masure. This msthod

cm, however, be labourious where multiple measurements are being evaluated.

The ICC is widely accepted as an appropriate statistic for evaluation of

cardinal data. However, many clinical measures such as outcorne scales,

severity of injury sales and exposure variables are ordinal in natwe.

Furtherrnore, other clinical measures are nominal ie hand dominance,

occupation. diagnosis. or injured extremity. The choice of statistic for these

measures is less obvious-

Kappa was originally designed for measuring inter-rater agreement

between two observers classifying subjeds into nominal (nonordered) categoiies

(Cohen 1990). It is supedor to using rate of agreement because it takes into

consideration the M8ds of chance agreement. Its original application has been

extended, so that it can be used with more than hiK, raters and with ordinal data,

The Mect of ordering is taken into consideration by weighting the deviafion frorn

exact agreement Maclure and Willet (1 987) argue that arbitrary, ordinal

categories Kappa is so arbitrary, that it is meaningless. Furthemore, with

naturally ordered data. the intradass correlation coefficient is superïor. A

weighted kappa allows one to Wight" disagreements, consequently,

standardization of wighting is required to minimize the arbitrary nature of

resuits. However. Men standard wights are selected, the kappa becornes

equivalent to the intradass correlation coefficient Kappa tends to be very

affected by the number of categories whereas, the ICC is less aff8Cfed by the

number of categories. A weighted kappa becames exadly equal to the ICC

when the wighting is quadratic (ie the square of the deviation of the pair of

observations from exact agreement) and the marginals are equal. Macclure and

Willet (1 987) conclude that kappa is best resenred for use with dichotomous

data. This wuld make the ICC the statistic of choice for evaluation of reliability

except for the situation where hm observers rate nonordered data. Despite

this, kappa remains popular when evaluating ordinal data and the benchmark

established for acceptable reliability of radiographie classification systems has

been stated as a kappa of k20.55 (Sanders 1997). Streiner and Norman (1 991 )

state that because Pearson correlations are usually fairly close to the intraclass

correlations and kappa tends to be almost equivalent to the intradass

correlation coefficient, the choiœ of statistics should be didated by ease of

calw lation. The majority of studies evaluating the reliability of radiographic

classifications continue ta use kappa as their evaluative measure.

Response bias can be anticipateci as a potential problem in radiographic

evaluations by different raters (Ker 1991 ). Where cases are ambiguous. raters

are more likely to select responses consistent w-th their own pemnal

experience. This bias will negate the "equal marginal$' requirement for the ICC

to exactly equal the quadratically w i g M kappa. Consequently, same

discrepancy between the weigMed kappa and the ICC may exist under

conditions where response bias exists.

Base rates are the relative propoftïon of a specific classification or

disease in a sample. It is k n m that high base rates affect the kappa obtained.

It has been suggested that base rates should be reported with kappa,

particulariy where benchmarks are being used (Ker 1991).

A number of benchmarks for appropriate levels of reliability have been put

forward. Landis and Koch (1977) describe a five level system (9.20 poor, 0.21-

0.40 fair, 0.41 -0.60 moderate, 0.61 0.80 substantial and >0.80 almost perfect).

Fleiss (1 986a) descri bes three categones (20.40 poor, 0.40-0.75 fairlmoderate

and 0.75 excellent). The Journal of Orthopaedic Trauma ~ M l l only consider

articles for publication if a ks0.55 has been established for the radiographic

classification reported. Minimum standards of 0.94 and 0.85 have been

suggested if tests are to be used to make decisians about individuals (Streiner 8

Norman 1991).

Given a couple of clinical realities, the need to establish benchmarks is

not compelling. Firstly, clinical measurements may need to continue even in the

face of substandard reliability. Radiographie classifications are a prime

example. The radiologie evaluation uf M u r e seventy is fundamental to

M u r e treatrnent The majority of studies have demonstrated that these

systems have poor ta moderate agreement b0-n or within raten. Despite

mis, they floufish in authorïtative texts on M u r e management. Authors

comrnonly present their favourite system without discussing issues of reliability.

Secondly, kappa will Vary eccording to base rates (Ker 1991) and readers will

have diffiailty evaluating the eff8ct of base rates on resultant kappas.

Frequently, reliability statistics are used to compare alternative dinical

measurement strategies. Which ciassification system is most reliable? What

effect does training have on reliability? M i c h measurement tool provides the

highest reliability? These questions are answered by cornparing reliability

statistics obtained in one ciraimstance (or scale) those obtained in the

same sample in an alternative ciraunstance (or scale). Benchmarks are not

required in these studies. Furthennore, as few measures are abandoned on the

basis of a substandard reliability score, benchmarks are rarely used as such.

Definitionflmporturce of Baseline factors

Baseline factors include information that may be accessible to treating

clinicians a baseline clinic visits. This information can play a role in the

dinicians decision making andlor prognostication. If baseline factors are

important predictors of outcome, this w l d assist physicians in providing a more

definite prognosis to their patients at these initial consultations. In addition,

sig nificant baseline predictors cauld be wntrol led in future trials which evaluate

the efkacy of various interventions or efFect of various complications on

outcome. Modifiable baseline factors that relate to outcome may be a foais of

intervention strategies. For example, if specific postiedudion anatomic

variables are predictive of outcome, this may guide physicians on how to

evaluate the adequacy of their initial treatment interventions,

Justification of Potential Predictom

Through literature review, dinical insight and discussion with other

clinicians a variety of baseline factors were identified that could a d as potential

predictors of outcome. As patient-rated outcome has not been previously

reported some of these potential factors are factors reported to be related to

other outcornes such as anatornic position on radiographs or dinician rated

outcornes. Not al1 variables considerd as potential predictors are disarssed

within the wntext of this proposal, as a number were eliminated based on

pradimlity of measurement or lack of substantive support for their inclusion.

Variables of particular importance. from either a theoretical or practical

viewpaint, are discussed. A number of the variables belaw were elhinateci as

predictors based on pilot work (See Appendix Tm).

1) Age While age is a signifiant risk factor for occurrence of a distal radius

fracture, (LaurÏtzen et al 1993a; Lauritzen et al 1993b; Mallmin & Ljunghall 1994)

it is not dear whether it has a significant efF8d on outcome Mer a fracture has

been sustaîned. Osteoporosis is an agedependent disease Wich is known to

increase the risk of fiadure, but may also act on outcome due to its Med on

bone quality. Reduced healing capacity, due to age, may alsb have a negative

effect on outcame. Converoely, patients with tower levels of physical demand,

such as elderly retired patients, rnay experienœ less pain and restriction than

younger patients who typically hava higher physical demands in their occupation

and recreation. Given that age rnay decrease both bone quality and physical

demands, it is possible that factors may negate each other with no net age

effect being observed. Another possible confounder to the patential age Mec3 is

that less force is required to cause a fradure in older individuals. Thetefore,

younger patients may actually have more severe soft tissue injury associated

*th a fracture than an elderly person with the same extent of fradure- This

suggests the possi bility of an age-sex interaction- Although some interactions

do not make intuitive sense to clinidans, the age-sex interadion is a cummon

dinical occurrence. For example, cl inicians are already ainiiare of the tendency

for older females to have a higher risk of fracture than young males and that the

mechanism of injwy is different be-n these groups.

2) Sex

Distat radius fractures are more common in Hiornen, (Dias et at 7987;

Lauritzen et al 199%; Lauritzen et al 1993b; Mallmin 8 Ljunghall 1994) again

related to bone strength issues. This is particularly tnie in post-menopausal

women where osteoporosis is highly prevalent in fradured wrists (Dias et al

1987). While sex is known to be a risk factor for M u r e occurrence, its effect

on outcorne is not known. Issues of bone quality and physical demand rnay

differ based on $ex

3) Sewndary gain - compensation status

From a dinical standpoint, a WSlB (Workers Safety lnsuranœ Board) or

legal daim is thought to have a negative impact on patient perception of pain

and disability. Patients on Workars Compensation Benefits or ~ Ï t h legal daims

related ta their fracture may be expeded to have a poorer outwme due to

issues with secondary gain. That is, patients who are being compensateci for

pain and disability either by the Worker's Compensation Board or the legal

system may be inclined to report higher levels of pain and disability. Thus,

patient ratings of outcorne could be less favourable in these patients. Another

possible explanation for differential outwme is that patients injured at wwk may

have higher occupational demands and restoration of a higher level of physical

function may be required in order to achieve satisfactory results.

The terni "secondary gain" is often used by clinicians. Freud defines

secondary gain as "interpersonal or sacial advantage attained by the patient as

a consequence of hislher il lness" (Freud 1 959). Secondary gain al lows a patient

to attain help, vutiich migM othewhe be unavailable, f ' m oieir environment as

a result of his illness behaviour. While clinicians &en consider secondary gain

as an issue when inappropriate illness behaviours are abservecl, the definition of

seoondary gain is not specifically defïned. Sscandary gain actually

encompasses al1 psychosocial rewards; however, monetary rewards are those

most frequently considered by clhicians in the pairVorthopaedics arena (8etter

et al 1979). Monetary rewards are more straightforward to measure than

psychosocial reumrds-

While there is considerable anecdotal and retrospective research

reporting that compensation has an M6Ct on patient behaviour and many dinical

studies report WSiB cases separately, there is little evidence on how

compensation affects recovery from an injury. While Better et al (1 979)

demonstrated that beneficiaries of disability pensions m e less Iikely to be

rehabilitated, th8y did not control for severity of illness. In a multivariate analysis

of predictors of health status in patients with carpal tunnel syndrome, being on

WSlB had a negative impact on overall health and pain perceptions

(MacDermid et al 1994). The majonty of wwk in this area has b e n done on

chronic pain, in particular Iow back pain due to the high prevalence of this

problem and the resultant compensation costs. For example, between 1960 and

1980 the incidence in low back pain has remained relatively constant Mi le the

rate of compensation has increased 2680% (Fordyce 1985). In contrast, lung

disease, the fastest growing disorder saw a 4W0h increase for the same time

frame. A meta-analysis (37 suitable studies derived from a total of 157)

detemined that compensation is related to increased reports of pain and

decreased treatment efficacy in the chronic pain population (Rohling et al

.1995). The effect size of compensation was 0.60 ie compensation patients

could be expeded to report a pain score 60% of one standard deviation higher

than non-campensated patients.

The secondary gains evaluated in this thesis be restrïcted to

possible secondary gain issues, ie Worker Compensation or legal daims related

to the injury. These two were selected as they are thought to act most strongly

as secondary gain variables. Due to low prevalence these two variables were

wmbined into one variable termed secondary compensation. It is notmrthy

that sorne subjects may Main financial gain from their injury by other means, ie

disability pay, reduction in daycare costs etc.

4) Smoking status

Smoking is known to increase risk of distal foreann M u r e , (Kelsey et al

1992) which may be related to a negative M8Cf of smoking on bone quality.

Post-menopausal -men who smdce lose more cortical bone and have more

spinal osteoporosis than do non-smokers (Daftari et al 1994). Weaker bone

stock may. not only increase fisk of fradure, but also have an affect on treatment

and outcorne-

Smoking has been show to have an effect on fracture healing in clinical

studies. Nicotine inhibits the revascularization of bone grafts and may be

responsible for a high pseudoarthrosis rate (Cobb et al 1994). Patients with

tibial shaft or mandibular fradures are at greater risk of delayed or nonunion if

they are cigarette smokers (Haug 8 Schwimmer 1994; Kyro et al 1993). The

effect of smoking on fracture healing has been investigated by using lowdose

ultrasound to modify healing. Healing time for tibial fractures was reduoed by

41 % in smokers and 26% in non-smokers and by 51 O h and 34% respectively for

DRF (Cook et al 1997).

There are other substantive reasons why smoking rnay M8Cf outcome

after a fracture. A number of studies have demonstrated that smoking impairs

the systemic microcirculation (Richardson 1987; Sarin et al 1974). This is

illustrated in the fact that circulation is significantly impaired in replanted digits in

smokers. Patients Mth below knee amputations have more complications and a

poorer outcome if they are smoken (Pohjolainen & Aiaranta 1991). There is a

higher rate of faifure of skin fiap and aiteriovenous shunts in patients who smoke

(Ray 1972). The biologic mechanism for the Mect on microcirculation had been

hypothesized to be an increase in platelet aggregation (Davis & Davis 1979).

Cigarette smoking is kn- to have a negative effW on wwnd healing (Sherwin

& Castwi-rth 1990). Smoking has an effect on bone metabolism, which rot only

underlies an increased risk of fradure. but may also Mect outcorne after a

fiacture. Tobacco use also has a negative effect on osteoblastic function (Ross

& Wu 1 996).

Smoking could also affect outcome by increasing the risk of

complications. I has a b been docurnented that smokhg inmeases the fisk of

reflex sympathetic dystrophy (RSD) (An et al 1988). RSD is k n m to have a

negative impact on outcome (Atkins et al 1989; Stewart et al 1985). Risk of

postoperative infection is higher in srnokm (Capen et al 1996). Nonunion is

also more frequent in smokers (Cobb et al 1994; Lau et al 1 Sû9; Schmitz et al

1995). Dupuytren's contrachire has been reported as a complication of distal

radius fracture and the rïsk of Dupuytren's disease requiring an operation is

higher in smokers (odds ratio 2.8) (Burge et al 7997).

While there is evidence of a biologic basis whereby smoking could have

a negative effect on fracture outcome, the overall impact on outcorne is more

wntroversial. White Hanley and Levy (1 983) stated that smoking had a

negative impact on outcome after lumbosacral fusion for spondylolisthesis. the

association was not statistically significant. Previous studies have tended to be

retrospective and have not controlled for socioeconomic status or other possible

confounders (An et al 1988; Kw~~atkowskÏ et al 1996).

There is considerable evidenœ to support a possible negative effect of

smoking on outcorne Mer a fraaure. There are various biologic rnechanisms

that cuuld be expected to impact on fracture healing. Whether smoking affects patient rated pain and disability, partiwlarly in the context of controlling other

variables has not been previously addressed.

5) Radial shortening

Radial shortening rdiects loss of radial length due to fracture and is the

distance frorn distal radial to distal ulnar surfaces (Altissimi 8 Mancini 1988;

Warwick et al 1993). Biomechanid studies have demonstrated that

progressive shortening of the radius causes changes in muscle function through

changes in moment arms and muscle Iengths. Shortenhg of only 2.5mm caused

statistically significant changes in mist flexw and extensor muscle biomechanics

(Tang et al 1 997). Previous univariate ctinical studies have also doarmented

that radial shortening is an important predidor of outcome (Trumble et al 1993;

Villar et al 1987). In fact, it is the single most commonly reported predictor of

outcome in the wrrent literature on fracture outcorne- Previous clinical studies

have not inaxporated patient-rateâ outcornes and m r e ftequently conducted

through univariate analyses of retrospective data. Therefore, the extent to Mich

radial shortening predicts patient rated outcome is unclear. Furthemore, the

extent to wttich original disniption and subsequent reduction aff6ct outcome has

not been definitely addressed. Nor has the relative role of anatomidinjury

factors and patient characteristics. A multivariate analyses that indudes both

patient and injury characteristics would provide insight as to the relative

contribution of these factors in detemiining six month outcome aiter DRF.

Radial shortening represents collapse of the radius as a result of M u r e

or disease. Radial shortening measured on different occasions represents

different aspects of a DRF. Initial (pre-reduction) radial shortening represents

severity of injury. Post-reduction radial shortening represents the extent of loss

of anatomic alignment Mer treatment Villar (1 9û7) observed that loss of radial

length one week pst-fracture was most highiy correlatecl (Pearson r) wïth three-

year grip strength.

In a DRF, anatomic disniption occurs in various planes and some

methods of measuflng radial shortening may be affecteci by disruption of

surrounding fandmarks. It has been suggested that this problem can be

minimized by using the centte-point of the distal artiw lar surfaœ of the radius

or the capitate vertex as a landmark. because these landmarks are less affectecl

by ulnar and palmar tilt (Bilic et al 1995). However, the method descrïbed by

Wamridc (Wamidc et al 1993) continues to be more commonly used in

evaluating radial shortening.

6) Articular lnvolvement/Comminution

It has been suggested th& intra-articular involvement is a negative factor

in terms of m u r e outcome (Eelma 8 McEMesh 1983). Residual displacement

of the articular surfaœ (greater than 2 mm has been suggested as an indication

for open reduction internai fixation (Cooney et al 1979). Biomechanical cadaver

studies have dernonstrated that increased dorsal cornminut ion/ loss of dorsal

bone results in instability (Bouman et al 1994; Bronstein et al 1997; Waters et al

1 997). Transgression of the midline in more extensive cornminution results in

biomechanical vectors that tend to make the fracture unstable, It is known that

the long terni outcome where there is pronounced incongniity of the joint sucface

is arthritis (Catalarto et al 1997). These fnxlings suggest that amminution or

the extent of joint involvement may be a predictor of outcome.

Cornminution, or fhgmentation of fractures, tends to ocarr in intra-

articular fractures. Presenœ of comminution is incorporated into a number of

classification systems. McMurtry ciassifieci ftactures as non-articular, 2, 3, 4 or

5- part intra-articular fractures (McMurtry 8 Jupiter 1996). The five or more parts

represents those fractures that are highly cornminutecl. As the f m s of this

classification system is extent of cornminution, it was considered a potential

variable for multivatiate analysis to represent this aspect of severity of injury.

The un iversal system classifies fractures as extra-art icular nondis placed, extra-

articular displaced, artiwlar nondisplaced and intra9rtiwlar displaced and thus

also places priority on the extent of joint involvement Because the reliability

and validity of DRF classifcatio systems had not ben delineated in the

literature, the choice of a fiadure classification variable was dependent on

reliability analyses perfmed as part of this thesis. Combining variables for

initial displacement, adequacy of reduction (Residual displacement) and joint

involvement was Viought ta incorporate a more global picture of the fiadute

status than has been perfomied in previous studies.

7) Education

Education rnay be retated to physical demands required for occupation,

cornpliance with rehabilitation. overall health behaviour or options for finding

alternative occupations. Education can be associateci with positive health

behavioun such as better nutrition and physical fitness. For these reasons,

inueased education may be related to better patient raid outcornes. No

evidence was found in the orthopaedic Iiterature on the role of education in

outcorne after orthopaedic injury. Despite this dinicians are familiar with

patients who have difficulty retuming to gainfil employment once an injury Ihits

their physical capacity- This ocarrs most frequently where there are high

physical demands to return to work such as with manual labour andlor where the

patient does not have the education to consider a wider range of ernployment

opportunities. Lower education has b e n associateci with a poorer outcorne

after total hip arthroplasty (MacWilliam et al 1996).

Education is a frequently studied sacioeconomic indicator. In a study of

33,962 Finnish subjeds evaluated for self-rated and examined health, high

education was a consistent predidor of good health (Heistaro et al 1996). This

is consistent with other studies wtiich have examined the efbct of education on

health perceptions (Gudex et ai 1996). Education has not only been associated

with a differential level of chronic disease but also a more rapid progression of

existing health problems (van der Meer & Mackenbach 1998). Education affects

risk parîially through its Mect on health behaviours but through additional

undelineated sources as well. For example, Mi le it is cmmonly accepted that

smoking is a cause of obstructive lung disease (COLD), it is also knomi that the

prevalenœ of smoking decmases as education level inmeases (Bakke et al

1995). Despite this, education level is a risk factor for COLD independent of

smoking and occupational airbome exposure (Bakke et al 1995). Utility scores

in rheumatoid arthritis patients are parüally explaineci by education even after

cantrolling for pain, disability. disease adivity, mood and side Mects (Hurst et al

1 997).

It is recognized that baseline variables are not the only fadors M i c h

affect outme. See Figure 1. Mediation of outcome potential ocairs as

variables oaxir along the path to recovery that alter the capacity for recovery.

Modification of initial outcorne potential commonly occurs through treatment

interventions and complications that o a x ~ after treatment is initiated. While

these variables are deemed ta be important, investigation of their role is not within the scope of this thesis.

Treatment

Treatment of M u r e s begins Men the patient presents for their initial

medical Gare and continues throughout the treatment process as the patient is

evaluated at each subsequent visit Medical treatment indudes: M u r e

reduction, fracture fixation and ongoing management of associated injuries and

complications. A variety of treatment options are wed for distal radius fractures

and were revievved earlier in this background chapter. Some techniques tend to

be used for specitic types of fiactutes. but physician pteférmce remains an

important component of treatment seladion. Physician skill is also considered

as a component of treatment While physician skill and preferenœ typically Vary

widely in large geographic areas, M i n gmup pradices such as the HULC this

effect is minimized through gmup consensus exercises such as ocairs during

teaching or patient care rounds.

Treatment also indudes rehabilitative services such as physiotherapy.

Beneficial effects of physiotherapy after DRF where desaibed as early as 191 9

where it was reported that 83% had 'good healing" and 16% 'bad healing" if

physiotherapy was initiated within tw months. Delayed after two months the

proportion of 'bad healing" increased to 43% (Steinmann 1993). Unfortunately,

80 years later little scientific evidence is available on the role of physiotherapy in

management of DRF- Evidenœ from UeNem suggests that DRF requires

intensive use of hand therapy services (Casanova 1998). Others have reported

that although a physiotherapy program provides patient satisfaction, no

impmvements in fùnctional stafus are obsewed M e n compared to a home

program (Oskarssun et al 1997). From out pilot data we know that the majority

of patients with a distal radius M u r e at Our centre m'Il receive physical therapy

(80%) either wîthin the facility, other local clinics or in their home communities.

A cornmon assumption made in prospective studies wïthin teaching

centres is that 'appropriate caren was deliverad. It is difficult to ascertain the

relative effects of treatment variations as they Vary by indication,

contraindication and Hihefe/how rehabilitative care is undertaken. Further, there

is Iittie evidenœ to delineate definite treatment guidelines (Cooney 8 Berger

1993). The greater the variation in outmme due to treatment variations the

greater the loss of predictive power in modelling outcome M e n these variables

are not wntmlfed-

Cornplications

Complications are an inherertt risk in distal radius fractures- Some ocair

sewndary to the injury, others develop over the course of trestment and

rehabilitation, Mile still others are long terni complications . The rate of

complications has been reportecl to be as high as 31% (Cooney et al 1980).

Whether considered a treatment failure or complication, malalignment may be

considered the most cornmon complication of D M . Another common

complication is compromise of median nerve function (Frykman 1967; Lynch 8

Lipscomb 1963). Other complications include malpositioning, arthrïtis, loss of

motion, tendon rupture and refiex sympathetic dystrophy, pseudo-Dupuytren's

nodules, cubital tunnel syndrome, piso-tn'quetral pain and exacerbation of

existing capometacarpal osteoarthritis of the thumb (McMurtry & Jupiter 1996)-

complications can be due to the nature of the injury or the nature of the

treatment required. For example, extemal fucators and percutaneous pins

(Hochwald et al 1997) are k n m to have an associated complication rate. In

the case of K-wires, iatrogenic nerve injury during percutaneous pinning is a

known complication Mich can be minimized through use of open dissection and

visualization (Vandersluis et al 1993). Beonise neither radiographs. MRI nor

arthrography are entirely reliable for diagnosis of ligamentous injury, the extent

of these with DRF has only recently been delineated. Ligarnentous injuries

ocair in up to 9û% of cases and chondral lesions are fwnd in 32% of DRFs

(Lindau et al 1997). No relationship between fracture pattern and associated

ligamentous injury has been detemineci. Furthemore, the eff8d of ligamentous

injuries an treatment requirements and outcome has yet to be determined-

Complications undaibtedly afféct outcome, but the quantification of

corn pl ications is- dîffÏcult as a methocf for grading the of severity of individual

complications has not been estabikhed. For example, in a randomized trial of

plaster casting versus extemal fixation a higher complication rate w s reporteci

for extemal fixation (nine pin track infedions, and six radial nerve injuries) as

cornpared to patients niho w r e casfed foilowing DRF (fvuo RSDs). The authors

noted that fine of the radial nerve injuries resolved wÏthin six weeks and ocairred early in the trial (ie wwe on the leaming wrve) and that al1 iwedions were

superficial and easily reoolved. Conversely, the h o RSDs were associated with

extreme morbidity and required extensive physiotherapy. Thus, corn pl ication

rates do not necessarily refiect Me impact of complications on management of

DRF.

In a review of 565 patients Cooney et al (1 980) reporteci the incidence of

various complications after Colle's fradure, ie 177 serious complications in 128

patients. There were 45 cases of compression neuropathy, 36 cases of

arthrosis, 30 cases of malalignment after loss of reduction, 12 unrecognized

associated injuries, 13 complications related to fixation, 4 cases of Volkmann's

ischemic contracture, 9 cases of arthrofibrosis and 20 cases of RSD.

Comrninuted displaced intraartiwlar fiadures had higher complication rates.

Post-traumatic arairitis is a long terni complication, thought to occur Men

articular incongniency rernains after M u r e healing. The incidence of post-

traumatic arthritis at mean followup of 4.8 years was 414 for patients with a

residual s t e m greater than 2 mm and 3/12 when the incongniency was less

than 2mm (Bradway et al 1989).

Complications a d as madiating variables to alter the potenaal outcome for

a given patient Certain complications ocair in the short terni, whereas others

occur over the long tem. Thus, there is a time dependency for specific

complications. Field et al (1992a) reported that 31 % of patients had an

unsatisfactory resuit after Colie's M u r e Men reviewed after ten yeam. Of the

patients with RSD, 50% had unsatisfadory results and 28% of patients with

osteoarthtitis had unsatisfadory results. Other variables which may mediate

o u t m e are injurÏes ancilor medicar problems that devekp over the course of

treatment, h o s e occurrence is unrelated to the fracture but that cause

additional disability of the upper extremity- m i l e complications are expected in

P certain proportion of patients, olese are events that oaxir after baseline and

as such cannot be taken into consideration for initial pmgnostication or used in design of research studies. Fu- research in this field is needed to delineate

the extent to Hihich specific complications mediate outcome. Hawever, these

questions are beyond the scope of this thesis.

While it is recognizeâ that these mediating factors w*ll Mact outcome, the

extent to Hihich baseline factors act as predictors without mntrolling for these

mediating variable is also an important issue. For this reason, information on

treatment and complications be mllected, but will not be the foais of mis

thesis.

Thesis Components~erview

The ptimary research question of this thesis is concemed identifying

baseline factors which are predidive of patient-rated pain and disability six-

months after a distal radius fracture. Fundamental to the task of answering this

question was the creation of an outcome measure to assess wrist pain and

disability. This wwk is surnmarized in the first hiiio articles: 1 ) 'Development of a

scale for patient rating of mist pain and disability', published in the Journal of

Hand Therapy in 1996 and 2) 'Patient rating of wrist pain and disability: a

reliable and valid rneasurement tooln, published in the Journal of Orthopaedic

Trauma. The first article provided an ovewiew of the pracess and decision-

making in development of the prirnary outaxne measure, the PRWE The

second article summarked the survey wwk which formulateci the content and

structural criteria f'or the P M , describeci short terni and long terni reliability

and presented criterion and constnict validity evaluations.

The ned stage of foundational wwk was establishing reliable protocols

for measunng secondary outcame measures, ie physical impairment measures.

The conceptual process for developing a global wrkt impairment measure is

outlined in detail in Appendix 1. Literafure deficiencies neœssitated reliability

wwk on several aspects of measuring wrist impairments.

Although a reliable protocol for assessing grip strength was established

~ Ï t h specific positioning using the Jamar hydraulic grip device, this thesis used

a straïn-gauge computerized gflp deviœ (NK digit grip). Grip enduranœ was

considered as a potential physical impairment measure. Thus one article

addresses the reliability of grip measures wïth the NK device. In addition, a

reliable protoool for measuring forearm rotation was not pubiished in curent

literature. Therefore a study mas conduded to establish a reliable method for

measuring forearm rotation with the NK computerized goniorneter. The reliability

of these measures of physical impairment are presented in the foll0wiOWing

articles: Chapter 4) 'Reliability of Range of Motion Measwement in the Elbow

and Foream" - published in The Journal of Shoulder and Elbow Surgery and

Chapter 5) "Testietest Reliability of Static and Endurance Grip Measures

Perfomed on the Jamar and NK Devices".

Once viable outanne measures were identified, evaluation of potential

predictors through the literature ~ i a s undertaken. This literature review was

summarized in this chapter. It vms possible to narrow d m radiographie

measurements solely on the basis of evidence Rom the Iiterature, resulting in the

selection of hm radiogcaphic variables. Hmver, aven after literature review a

variety of patient factors remained as potential predidors of patient outcorne. A

pilot study was wnductad to reduce the number of patient variables to those

most likely to have an impact on patient ratecf outcume and to eveluate the study

processes. The results af the pilot study are reportecl in Appendix 2. The pilot

evaluation of potential predidon established two key patient-reCated variables

(secondary gain and educaüon) and the literature suggested two radiographie

variables (radial shortening and joint involvement) Hihich wamnted further

investigation in a whort study.

While literature review indicated that radiographic variables are predidive

of outcome after DRF, the reIiability of these masures had been variable.

Furthemore, the issue of reliability and validity of the various dassificatïon

systems appeared to be a pmblem. This resulted in the fifth article: (Chapter 6)

"Reliabil ity of Radiographie Measures and Classification in Patients Distal

Radius Fractures?

Over the course of the study, an upper extremity disability scale (the

DASH) was published. This new development resulted in a supplementary thesis

question to address the relative responsiveness of available outcome measures

for evaluating dinical recovery after DRF: 'Responsiveness of Patient Rating

Scales and Ph ysical I mpaiment Measures in Measuring Recovery After Distal

Radius Fracture",

The final artide address the main question 'Baseline Predidors of

Patient Rated Pain and Disability Six Months M e r Distal Radius Fracturen. The

final chaptet summarizes the thesis findings and I imitations.

In summary, a series of papers ware wrÏtten to doaiment a reliable and

valid primary outcome masure specific to the wrist wuhich allows patients to rate

their patient-ratd pain and disability. Secondary outcornes of physical capacity

were established that m e reliable and valid. Reliability of radiographic

rneasures and classifications of DRF m e investigated. Potential predicton

w r e identifid through literature review and pilot study. A prospective study

was used to determine the relative importance of badine indicaton on six-

rnonth patient rateci pain and disability. Supplementary thesis questions

addressed the relative responsiveness of difkrent outcornes measues-

Anonymous: Scoring Exercise for the SF-36 Health Survey. (2nd) Medical

Outcornes Trust, Boston- 1994a-

Anonymous: Sf36 Heafth Survey A ScMng Manual for English-Language

Adaptations: AustraliaMew Zealand, Canada, United Kingdom. Medical

Outcornes Trust, Boston, Mass. 1994b.

Abbaszadegan H, Adolphson Pt Dalen N, Jonsson W. Sjoberg HE, Kalen S:

Bone mineral loss after Colles' fracture- Plaster cast and extemal fucafion

equivalent Acta Orthop Scand 62: 156-1 58,1991.

Adams BD, Samani JE, Holley KA: Triangular fibrocartilage injury: a labofatory

model. J Hand Surg Am 21 : 189-1 93,1996.

af Ekenstam F, Jakobssm OPt Wadin K Repair of the triangular ligament in

Colles' fracture. No M8d in a prosmve randomized study. Acta

Orfhop Scand BO: 393396,1989-

Agee JM: Distal radius fradures. Multiplanar ligamentotaxis. Hand Clh 9: 577-

585,1993-

Agee JM, Szabo RM, Chidgey LK, King FC, Kerfoot C: Treatment of comminuted

distal radius fiadures: an approach based on pathomechanics.

Odhopedks 17: 1 1 1 M 122,1994.

Amam PA & Goran CHB: Epidemiology of fractures of the foream. J Bone Joinf

Surg Am 44A: 10% 14,1962.

Altissimi M, Antenucd R, Fiacca C, Mancini GB: Long terni results of

consenrative treatment af m u r e s of the distal radius. Clin Olihop Zûû:

202-21 O, 1 986,

Altissimi M 8 Mancini GB: Surgical release of the rnedian nerve under local

anesthesia for carpal tunnel syndrome. J Hand Surg 13: 395396.1988.

Altissimi M. Mancini GB, Azzara A, CiMoIoni E: Early and late displacernent of

fractures of the distal radius. The predidion of instability. !nt Orthop 18:

63-65,1994.

Altissimi M, Mancini GB, Ciwoloni Er Pucci G: Comminuted articular fractures of

the distal radius. Results of consenrative treattnent ftal J Orthop

Traumatd 17: 1 1 7-1 23, 1 991 -

Amadio PC, Silverstein MD, Ilstnip DM, Schleck CD. Jensen LM: Outcorne after

Colles fracture: the relative responsiveness of three questionnaires and

physical examination measures. J Hand Suig Br 21: 781 -787, 1996.

An HS, Hawthorne KB, Jackson VVT: Reflex sympathetic dystrophy and cigarette

smoking. J Hanâ Surg Am 13A: 45&Eû, -1 988.

Andersen DJ, Blair WF, Steyers CM, Jr., Adams BD, el Khourï GY. Brandser EA-

Classification of distal radius fractures: an anaiysis of interobserver

relia bil ity and intraobserver reproducibility. J Hand Suig Am 21 : 574582,

1996-

Andersen GR. Rasmussen JB, Dahl B, Solgaard S: Oldefs classification of

Colles' fiadures. Good intraobserver and interobserver repraducibility in

1 85 cases. Acta OIfhop ScaMj 62: 463464, 1991.

Anonymous: Canada's role in the revision of the international cfassIficiation of

irnpaimnts, disabilties and handicaps - ICIDH (WHO. 1 980)

( W o t k ~ h ~ ~ ) . 1-t5, 1997.

Ark J 8 Jupiter JB: The rationale for precise management of distal radius

fractures. O m p Clin North A m 24: 205-21 0,1993.

Atkins RM, DudomJrth TT Kanis JA: Algodystrophy followïng colles' fracture. J

Hand Surg Brt4B: 161464, lm.

Atkins RM, D u W r a i T, Kenis JA: Features of algodystrophy after Colles'

m u r e . J 6one Jant Surg Br 72: 1 05-1 1 O. 1 990.

Axelrod TT Paley D, Green J, McMurtry RY: Limited open reduction of the lunate

facet in comminuted intraarticular fractures of the distal radius- J Hand

SU^ 13A: 372-377,1988.

Axelrod TS: A prospective randornked trial of extemal fu<ation and plaster cast

immobilization in the treatment of distal radius fractures [letter. comment]

[see cornments]. J Orthop Trauma 5: 1 14-1 1 5,1991-

Axelrod TS 8 McMurty RY: Open reduction and intemal faation of comminuted,

intra-artiwlar M u r e s of the distal radius. J Hand Surg Am 1SA: 1-1 1,

1990-

Bacom TM 8 Kurtzke JF: Colles fracture. A study of two thousand cases from

the New York State Workmen's Compensation Board. J &ne Joint Surg

35: 643-65& 1953.

Bain GI, Roth JH, Pugh D. MacDermid JC: Matched hemiresedion interposition

arthroplasty of the distal radioulnar joint J Hand Surg Am ZOA: 944950,

1995-

Bakke PS, Hanoa R, Gulsvik A- Educational level and obstnictïve lung disease

given smoking habits and ocaipationa1 airborne exposure: a Nonmgian

wmmunitystudy. Am J Enmiôdl41: 1080-1088,1995-

Bartko J J: On vanow intraclass correlation coefficients. Psychd Bull 83: 762-

765,1976-

Bartosh RA & Saldana MJ: Intraarticular fractures of the distal radius: a

cadaveric study to determine if ligamentotaxis restores radiopalmar tilt J

HandSurgAm 15: 1821,1990.

Bass RL, Blair WF, Hubbard PP: Results of cdmbined internat and extemal

fixation for the treatment of severe A M 3 fractures of the distal radius- J

Hand Surg Am 20: 373-381 . 1 995.

Bennett GL, Leeson MC, Smith BS: lntramedullary fucation of unstable distal

radius fradures- A method of fixation allawnig early motion. Orthop Rev

18: 210-216,1989-

Better SR, Fine PR, Simison Dl Doss GH, Walls RT, McLauglin DE: Disability

benefits as disincentives to rehabilitaüon. Milbank Q 57: 41 2427. 1 979.

Bilic R, Ruzic L, Zdravkovic VI Boljevic 2, Kovanjanic J: Reliability of different

methods of determination of radial shortening. J Hand Surg Br 2ûB: 97-

101, 1995-

Bland JM 8 Altman DJ: Statistical methods for assessing agreement between

Bouman HW, Messer E, Sennwald G: Measurement of ulnar translation and

carpal height. J Hand Surg Br 19: 325329, 1 994.

Bradway JK, Amadio PC, Cooney WP: Open reducüon and intemal flxation of

displaced, comminuted intra-articuiar fractures af the distal end of the

radius. J Bone Jant Surg Am 71 : 839-847,1989.

Braun RM 8 Gellman H: Dorsal pin placement and extemal fu<ation for correction

of dorsal tilt in fractures af the distal radius. J Hand Surg Am 19: 653-

655,1994-

Bronstein AJ, Tmmble TE, Tencer AF: The eff8Cfs of distat radius fiadure

malalignment on foreann rotation: a cadaveric study. J Hand Surg Am

22: 258-262,1997.

Brown JN & Be11 MJ: Distal radial osteotomy for malunion of wrist fractures in

young patients. J Hand Surg Br 19: 589-593, 1 994.

Burge P, Hoy G, Regan P, Milne R: Smoking, alcohol and the risk of Duputren's

contracture. J Bone Joint Surg Br 796: 2-21 O. 1997-

Cannegieter DM 8 Juttmann JW Canœllous grafting and extemal fuation for

unstable Colles' fractures. J &me Joint Surg Br 79: 428432, 1997.

Capen DA, Caldemne RR. Green A: Perïoperative nsk fadors for wwnd

infections Mer lower back fusions. Orthop Clin North Am 2 2 83-86,

1996.

Canozzella J & Stem PJ: Treatment of comminuted distal radius fractures with

pins and plaster. Hand Ciin 4: 391 -397, 1988.

Casanova J: UeNet Data Trends. 1998-

Casteleyn PP, Handelberg F, Haef@ens P: Bidegradable rods versus Kirschnet

wire fixation of @st Mures. A randomized trial. J Bone Joint Surg Br

74: 858-861, 1992.

Catalano LW, Cole RJ, Gelbemian RH: Displaced intra-articular fractures of the

dîstal aspect of the radius- Long t m results in young adults after open

reductiort and intemal fixation- J Borie J a n t Surg Am 79A: 12904 302,

1997,

Chapman DR, Bennett JB. Bryan WJ, Tullos HS: Complications of distal radius

fractures: pins and piaster treatment, J Hand Surg 7: SB51 2, 982-

Chester MH: Segmental manif@staüon of refiex sympathetic dystrophy syndrome

limited to one finger. Anesthesriobgy 73: 558561,1990.

Christensen OM, Christiansen TG, Krashenbnikoff M, Hansen FF: Length of

immobilisation after fradures of the distal radius. lnt Orthop 19: 26-29,

1995-

Clancey GJ: Percutaneous ffirschner-wire fixation of Colles' fractures. J Bone

Joint Surg 6aA: 10081014, 19û4.

Clybum TA: Dynamic extemal fixation for comminuted intraarticular fraaures of

the distal end of the radius. J &ne Joint Sum Am 6SA: 248-254, 1987.

Cobb TK, Gabtielsen TA Campbell DC, et al: Cigarette smoking and nonunion

after ankle arthrodesis. Foot AnMe 15: 64-66, 1994.

Cohen J: A coefficient of agreement for nominal scales. Educ Psych Meas 20:

37-46, 1990,

Cohen MS 8 Frillman T: Distal radius fractures: a prospective randomized

cornparison offibreglass tape with QuickCast Injury 28: 305-309, 1997.

Cole 6, Finch E, Gowland C: Physical Outcome Measures. Canadian

Physiotherapy Association, Toronto. 1993.

Cole JM & Obletz BE: Comminuted fractures of the distal end of the radius

treated by skeletal fixation in plaster cast An end-result study of thirty - three cases. J &ne Joint Surg Am a: 931-945,1966.

Collins DG: Management and rehabilitation of distal radius Mures - Odhop

Clin North Am 24: 365-378, 1993,

Cook SD, Ryaby JP, McCabe J. Frey JJ, Hedunan JD, Krïstiansen TK

Acceleration of tibia and distal radius fracture healing in patients MO

smoke. CIin Offhop lS8-2O?.l99T.

Cooney WP: Fractures of the distal radius- A modem treatment-based

classÏfication. O~hop Ciin North Am U: 21 1-216,1993.

Cooney WP, Agee JM, Hastings H: Symposium: Management of intrasrtiwlar

fractures af the dista1 end of the radius. Contemp Orthop 21 : 71 -1 04,

1990.

Cooney WP & Berger M Treatrnent of cornplex fractures of the distal radius.

Cornbined use of intemal and extemal fixation and arlliroscopic reduction.

Hand Clin 9: 6û3=612,1993.

Caoney WP, Dobyns JH, Linscheid RL: Complications of Colles' fradures. J

&ne ~oinf ~k ~ r n 6 ~ . 61 3-61 9,1980.

Cooney WP, Linscheid RL, Dobyns JH: Extemal pin focation for unstable Colles'

fractures. J Bone Joint Surg Am 61A: 840-845, 1979.

Crilly RG, Delaquemïere-Richardson LD, Roth JH: Postural instability and Colles'

fraaure. Age Ageing 16: 133-1 38.1987.

Culp RW & Osteman AL: Arthroscopie redudion and intemal fixation of distal

radius fractures. Orthop Clin Norlh Am 26: 739-748,1995.

Cummings SR, Kelsey JL, Nevitt MC, 0'- KJ: Epidemiology of osteoporosis

and osteoporotic fradures. € m m & / Rev 7: 1 7û-208, 1985-

Daftari TY Whitesides TE, Heller JG, et al: Nicotine on the revawlarization of

bone gr* Spine 19: 904-91 1.1994.

Davis JW & Davis RF: Aaite Mect of tobacco cigarette smoking on the platelet

aggregate ratio. Am J Med Sa 278: 139-1 43, 1 979.

De Palma AF: Comrninuted fiactures of the distai end of the radius treated by

ulnar pinning. J&m JointSutg Am 34: 651462,1952.

Deyo RA, Diehr P, Patrick DL: Reproducibility and responsiveness of health

status measures. Statistics and strategies for evaluation. Contmi Clin

Trials 1 2: l4ZS-l58SI 1991.

Dias JJ, Wray CC, Jones JM, Gregg PJ: The value of eady mobilzation in the

treatrnent of Colles' fractures. J Bone Joint Surg Br 696: 463467, 1987.

Doai J & Renner A- Epidemiology of distal radius M u r e s in Budapest. A

retrospective study of 2,241 cases in 1989. Ada Otttnop Scand 65: 432-

433,1994.

Edwards GSJ: Intra-articular fractures of the distal part of the radius treated with

the maIl A 0 extemal fmor. J Bone Joint Surg Am 73: 1241-1250,

1991-

Eelma J 8 M c E M EC: Colles fractures in young adults. Minn Med 487490,

1983-

Eliasziw M, Young SL, Woodbury MA, Fryday-Field K Statistical methodology

for the concurrent assessrnent of interrater and intramter reliability: Using

goniornetnc measuments as an example. Phys Ther 74: 777-788,1994.

Famer ME, White LR, Brody JA, et al: Race and sex dîffwences in hip fiactuce

incidence- Am J Public Nealth 74: 1 3744 380,1984L

Femandez DL: Comon of post Waumatic mist def;onnity in adults by

osteotomy, bone graffing and internai fixation. J Bone Jdnf Surg Am

WA: 11644 178,1982.

Femandez DL: Cotfedl-ve osteotorny for extra-articular malunion of the distal

radius. In Safftar P and Cooney WP: Fractures of the Distal Radius Marün

Dunitz, London- f995-

Femandez DL 8 Flury MC: History, evolution and biornechanics of extemal

fixation of the mist joint Injury 25: sd2-sd6, 1 994.

Fernandez DL & Jupiter JB: Epidemiology, Mechanisrn, Classification Fractures

of the Distal End of the Radius. A Pradical Approach to Management

Springer, NY. 1996.

Fernandez JJ, Gruen GS, Hemdon JH: Outcome of distal radius fractures using

the short form 36 health survey. Clin Orthop 3641,1997.

Ferris BD, Thomas NP, Dewar ME, Simpson DA: Braœ treatment of Colles'

fradure. Acfa Orthop Scand BO: 6345,1989.

Field J, Warwick D, Bannister GC: Features of algodystrophy ten years after

Collees' Fradwe- J Hand Surg Br 1 TB: 31 R320.1992b-

Field J, Warwick Dl Bannister GC, Gibson AG: Long-tenn prognosis of displaced

Col les' Fracture: a 1 O-year prospective review. Injury 23: 529532. 1 992a.

Fischer M, Denzler Cl Senniiirald G: [Carpal Iigament lesions associateci with

fresh distal radius Mures: arthroscopie study of 54 cases]. Swiss Surg

2: 269-272, 1996.

57

Fleiss JL: Reliability of Measurement ln Fleiss Jl- The Design and Analysis of

Clinical Expenments John Wiley and Son. Toronto. 1 W6bb

Fleiss JL: The Design and Analysis of Clinical Expetirnents. John Wiley and

Sons, Toronto, l986a-

Fordyce WE: Badc pain, compensation. and public policy. In Rosen JC and

Solomon LJ: Revention in Health Psyaiology University Press of New

England, Hanovershl H- 1985.

Freud S: lntroductory Lectures on Psychoanalysis. Hogarth Press, London.

1959-

Frykman G: Fracture of the distal radius inciuding sequelae - shoulder hand

finger syndrome, disturbance in the radioulnar joint and impairment of

newe function. A dinical and experimental study- Acta Otthop Scand

i08: 1-1 55, 1967.

Frykrnan GK, Peckham RH, Willard Saha S: Extemal f~ators for treatment of

unstable wist ftactures. A biomechanical, design feature, and cost

cornparison. Hand C h 9: 555565,1993-

Frykman GK Tooma GS, Boyko K, Hendemn R: Cornparison of eleven extemal

fixaton for treatment of unstabfe wrïst fractures. J Hand Surg Am 14:

247-254,1989,

GaNand JJ & Weriey CW Evaluation of healed Colles' fractures. J Bone Joint

Surg Am 33A: 895907,1961.

Geissler WB: Arthroscopically assisted reduaion of intra-articular fractures of

the distal radius, Hand Clin 1 1 : 1 9-29, 1 995-

Geissler WB & Femandez DL: Percutaneous and Iimited open reduction of the

artiailar surface of îhe dishi radius. J Ortf,op Trauma 1: 255-264, 1991.

Gesensway D, Pufnarn MD, Mante PL, LewÎs JL: Design and biomechanics of a

plate fw the distal radius. J Hand Surg Am 20: 1 021 -1 027.1 995.

Gieck J 8 Buxton BP: Refiex symapthetic dystrophy. AthkoC Training 22: 120-

125,1987,

Goslings JC, Tepic S. Broekhuizen AH, Jakob RP, Perren SM: Three-

dimensional dynamic A0 extemal fixation of distal radial fiactures-a

preliminary report Injury 25 Suppi 4 S-085, 1994.

Graff S & Jupiter J: Fracture of the distai radius: classification of treatment and

indications for extemal fixation. lnjury ZS Suppl4: S-û 1 4, 1 994.

Graham TJ: Surgical conedion of malunited fractures of the distal radius. J Am

Acad Otthop Suq 5: 270-281,1997.

Greatting MD & Bishop AT: lntrafocal (Kapandji) pinning of unstable m u r e s of

the distal radius. Odhop Clin Notfh Am 2& 301 -307, 1 993.

Green DP: Pins and plaster treatment of comminuted s e s of the distai end

of the radius. J Bone Joint Surg 57A: 30431 0, 1975.

Gudex C, Dolan P, Kind P. Williams A: Health state valuations from the general

public using the visual analogue scale. Qual Ldk Res 5: 521 -531, 1 996.

Gupta A- The treatment of Colles' fracture. Immobilisation *th the wrist

dorsiflexed. J Bone M n f Surg Br 73: 31 2-39 5, 1 991.

59

Habemek H, Weinstabl R, Fialka C, Scrirnid L Unstable distal radius fiadures

treated by rnodified Kirschnef wire pinning: anatMnic considerations,

technique, and results. J Trauma 58: 8388. 1994.

Hanley EN 8 Levy JA- Surgical hreatment of isthmic lumbrosacral

spondyfolisthesis. Analysis of variables influmcing results. Spine 14: 48-

50, 1989.

Harper MC: A modifieci pins-amlplaster technique inmrporating early motion for

fractures of the distal radius. J Orfhop Tmuma 3: 300305, 1989.

Hastings H2 8 Leibovic SJ: Indications and techniques of open reduction.

lntemal fixation of distal radius fractures. Orthop Clin Nom Am U: 309

326,1993-

Haug RH 8 Schwimmer A: Fibrous union of the mandible: A review of 27

patients. J Oral Maxillofc Sutg 52: 832839,1994.

Heistaro S, Vartiainen E, Puska P: Trends in seff-rated health in Finland 1972-

1992, Prev Med 25: 625432,1996.

Hochwald NL, Levine R, Tometta P3: The nsks of Kirschnef wÏre placement in

the distal radius: a cornparison of techniques. J Hand Surg Am 22: 580-

584,1997-

Horesh 2, Volpin G, Hoerer D, Stein H: The surgical treatment of severe

comminuted intraarticular fiadures of the distal radius with the small A 0

extemal fmtion device. A prospective three-and-onehalf-year follovwp

study. Clin Orthop l4%l!53, 1991.

Home G: A prospedive randomizd trial of extemal fo<ation and plaster cast

immobilkatCon in the treatment of distal radius fiacture. J Ofthop Trauma

5: 246-246, 1991.

Home JG, Devane P, Purdie Gr A prospedive randamized trial of extemal

fixation and plaster cast imrnobilization in the treatment of distal radial

fractures. J Omhop Trauma 4: 3034, 2 9 9 0 -

Hove LM, Fumes O, Nilsen PT, Oulie HE, Solheim Er Molster AO: Closed

reduction and extemal f d l o n of unstable ftactures of the distal radius,

Scand J Plast Reconsir Hand Surg 31 : I5WM,l997b.

Hove LM, Nilsen PT, Fumes O, Oulie HE. Solheim El Molster AO: Open

redudion and intemal fucation of displaced intraarticular fiactures of the

distal radius. 31 patients followed for 3-7 years. Acta Orthop Scand6û:

59-63, 1 997a.

Howard PW, Stewart HD, Hind RE, Burke FD: &temal fixation or plaster for

severely displaced comminuted Colles' fractures? A prospective study of

anatomical and fwidional results. J Bone Joint Surg Br 71 : 68-73, 1989.

Huch K, Hunerbein M. Meeder PJ: Extemal fixation of intra-artiwlar fracture of

the distar radius in young and old adults. Arch Offhop Trauma Surg 115:

3842,1996-

Hudak PL, Amadio PC, Bombardier C, The Upper Extremity Collaborative

Group: Development of an upper extremity outcorne measure: the Dash

(disabilities of the am, shoulder and head (sic)). Am J Ind Med 29: 602-

908,1996.

Hurst NP, Kind P, Ruta Dr Hunter M, Stubbings A: Measuring health-related

quality of life in rheumatoid arthntis: validity. responsiveness and

reliability of EuroQol (EOSD). Br. J RheumafoI. 36: 551 -559,1997-

Hutchinson DT. Strenz GO, Cautilli RA: Pins and plaster vs exfemal fixation in

the treatment of unstable distal radial fractures. A randomized prospective

study. J Hand Surg Br al: 365372,1995.

Hutchinson TA, Polansky SM, Feinstein AR: P o s t . s a i oastragens pmted

against fractures of hip and distal radius: a caseantrol study. Lamet 2:

705-709,1979-

Jakim 1, Pieterse HS, Sweet MB: Extemal fiication for intra-artiwlar fractures of

the distal radius. J Bone Joint Surg Br 73: 302-306.1991.

Jakob RP: Development of the small A 0 fuatof to the aiment set Injury 25

Suppl4: S-M6,1994.

Jebson PJ 8 Blair WF: A surgical protocol for combined intemal and extemal

fixation of cornplex intraarticular distal radius fractures using dorsal and

volar approaches- lowa Omop J 16: 3548,1996-

Jenkins NH, Jones DG, Mintowt Czyz WJ: Extemal fixation and recovery of

fundion following fiadures of the distal radius in ywng adults. lnjury 19:

235-238, 7988.

Jensen CH 8 Jensen CM: Biodegradable pins versus Kirsdiner wkes in hand

surgery. J Hand Surg Br 21 : 507-51 O, 1996.

Jette AM: Outcornes research: shifang the dominant research paradigm in

Physical Therapy. Phys -75: 965-970, 1995.

62

Jupiter JB: Curent concepts review Fractures ofthe distal end of the radius. J

Bone J'nt Suq Am 73A: 461469,1991.

Jupiter JB: Fractures of the distal radius. Surg Annu 24 Pt 1 : 1 43-1 60. 1 992.

Jupiter JB, Femandez DL, Toh CL, Fellman T, Ring O: Operative treatrnent of

volar intra-articular fractures of the distai end of the radius- J Bone Jor'nt

Surg Am 78: 181 74828.1996.

Kaempffe FA: Extemal fixation for distal radius fractures: adverse M e s of

excess distraction. Am J Orthop 25: 205-209, 1 996-

Kelsey JL, B r w e r WS, Seeley DG. Nevitt MC, Cummings SR: Risk fadors for

fractures of the distal foreami and proximal humenis. The Study of

Osteoporotic Fractures Research Group. Am. J Eplilemrol. 135: 477489,

1992.

Ker M: Issues in the use of kappa. lnvest Rad 26: 7883,1991.

Knirk JL 8 Jupiter JB: Intra-articular fradures of the distel end of the radius in

young adults. J Bone Joint Surg Am 68A: 647459,1986-

Krolner 8, Tondervold E, Toft B: Bone mass of the axial and appendicular

skeleton in wmen wiai Colles' fradures: Its relation to phsyical advity.

Clin Physb12: 147-1 5ï. 1 982.

Kwasny O, Fuchs M. Schabus R: Opening wedge osteotomy for malunion of the

distal radius *th neuropathy. 13 cases followed for 6 (1 -1 1 ) years. Acta

Orthop Scand 61: 207-208, 1 994.

Kwiatkowski TC. Hanley EN, Jr., Ramp WK: Cigarette smoking and its

orthopedic consequemes. Am. J Otthop 25: 590597, 1996.

Kyro A, Usenius JP, Aamio M. Kunnamo 1, Avikainen V: Are srnokers a risk

group for delayed healing of tibia1 shaft M u r e s ? Ann Chir Gynaecd 82:

254-262, 1 993-

Lamke B. Sjoberg HE, Sylven M: Bone mineral content in m i e n wïth Colles'

fracture: Eff- of calcium supplementation. Acta Ofthop Scand 49: 143-

146,1978-

Landis JR & Koch GG: The measurement of observer agreement for categm-cal

data. Bbrnetfüs 33: 1 59-1 74, 1 977-

Larsen CF & Lauritsen J: Epidemiology of amte M s t trauma. Int J Epiie/emiùi

22: 91 1 -91 6,1993,

Lau GC, Luck JV, Marshall GJ, Griffith G: The effect of cigarette smoking on

fracture healing: an animal model. Clin Res 37: 1 32A, 1 989.

Lauritzen JB, Schwarz P, Lund B, McNair P. Transbol 1: Changing incidenœ and

residual IKetime rïsk of cornmon osteoporosis-related fractures.

Osfeoporos Int 3: I27-l32.1993b.

Lauritzen JB, Schwarz P, McNair P, Lund B, Transbol 1: Radial and humeral

fractures as predicton of subsequent hip, radial or humeral fractures in

women, and their seasonal vanation. Osteopoms lnt 3: 1 33-1 37, 1 993a.

Lawson GM, Hajducka C, McQueen MM: Sports fiactures of oie distal radius-

epidemiology and outcorne. lnjury 26: 33-36, 1995.

Ledingham WM, wch R, Goring CC, Mathieson AB, Wardlaw D: On imrnediate

fundional bracing of Colles' fradure. Injury 22: 1 97-201 , 1 991.

Lennox JO, Page 642, Mandell RM: Use of the Clybum Extemal Fkator in

fractures of the distal radius. J Tiauma 29: 326-331, 1989.

64

Lenable E, Dumontier CE Goutallier D, Apoil A: Frachwe of the distal radius. A

prospective mmpaiwn behiueen trans-styioid and KBpandji fiwtions. J

Bone Joint Surg Br 77: 562467,1995-

Leung KS, Shen WY, Leung PC, Kinninmonth AW. Chang JC, Chan GP:

Ligamentotaxis and bone grafting for comminuted m u r e s of the distal

radius. J Bone J M t Suw Br 71 : 838842,1989.

Leung KS. Tsang HY Chiu KH: An effective treatment of cominuted fractures of

the distal radius. J Hand Surg Am 1 SA: 1 1 -1 7, 1 990.

Lidstrorn A: Fractures of the distal end of the radius: A clinical and statistical

study of end results. Acta Orfhop Scand Supp141: 1959.

Lindau Tl Amer M, Hagberg L: Intraartiwlar lesions in distal radius fardures of

the radius in young adults. A descriptive arthroscopie study in 50 patients.

J Hand Suq Br 226: 638643.1997-

Lucas GL 8 Sachtjen KM: An analysis of hand function in patients with Colles

fractures treated by rush rad fuatïon. Clin Orthop 155: 1 72-1 79, 1 981.

Lynch AC & Lipscomb PR: The carpal tunnel syndrome and Colles' fradures.

JAMA 185: 363-366, i 963.

MacDennid JC: Development of a scale for patient rating of &st pain and

disability. J Hand Ther9: 178-1 83, 1 996-

MacDemid JC, Roth JH, Kohls J: Predicüon of eledrophysiologic, pain and

health status variables in patients with carpal tunnel syndrome.

Physiofher Can 46 (suppi): 1 39-1 39, 1994.

MacDennid JC, Turgeon T, Beadle M, Richards RS, Roth JH: Patient rating of

wrist pain and disabiliw A reliable and valid measurernent tool. J Orthop

Trauma t 2: 577-586, 5398-

Mackinnon SE 8 Holder LE: The use of thmephase radionulide bone scanning

in the diagnosis of refiex sympathetic dystrophy- J Hand Suq Am SA:

556-563, 1984.

Maclure M 8 Willet WC: Misinterpretation and misuse of the kappa staüsticC J

Epid 'l26: 1 6f -1 68,1987.

MacWilliam CH, Yood MU, Vemer JJ, McCarthy BD, Wafd RE: Patient-related

risk fadon that predict poor outcorne after total hip replacement Heaith

Serv Res 31 : 623-638,1996.

Mallrnin H & Ljunghall S: Distal radius fraaure is an early sign of general

osteoporosis: bone mass measurements in a population-based study.

Osteopo~os Int 4: 357-361.1994-

Mann FA, Wilson Al, Gilula LA: Radiographic evaluation of the wrïst: what does

the hand surgeon want to k n M Radblogy 184: IS24,l992.

McBimie J, Court B r m CM, McQueen MM: Eariy open reduction and bone

grafting for unstable fractures of the dista( radius. J &ne Joint Surg Br

77: 571-575,1995.

McMurtry RY 8 Jupiter JB: Fractures of the distal radius. In Bromier BD, Jupiter

JB, Levine AM and Trafton PG: Skeletal Trauma WB, Saunders Co,

Philadelphia. 1 996.

Melendez EM, Mehne DK, Posner MA: Treatment of unstable Colles' fradures

with a new radius mini- fo<ator- J Hand Surg Am 14: 80781 1, 1989.

66

Melik N, Tognïnalli D, Biegger P: Retroswve analysis of ainsecutively treated

distal radius fradures vith the extemal fmator- Unfalkhiiurg 97: 64!564û,

1994-

Melone CP: Open treatment for displaceci artîailar fractures of the distal radius.

Clin Orthop 202: 103-1 1 1, 1 986.

Melone CP: Distal radius fradures: Patterns of articular fragmentation. Orfhop

CIh North ARP W: 239-253,1993.

Mennen U & Labuschagne FJ: An alternative method of treating comminuted

intra-articular fiadutes of the distal radius. S Afr J Surg 31 : 81 -84, 1 993.

Merchan EC, Breton AF, Galinda E, Peinado JF, Beltran J: Plaster cast versus

Clybum extemal fucation for fractures of the distal radius in patients under

45 years of age. Oiolop Rev 21 : 1203-1 209,1992.

Milliez PY, Dallaserra M. Oefives T, el Ayoubî L, Thomine JM: Effect of early

mobil ization follownig Kapandji's method of intrafocal wiring in M u r e s of

the distal end of the radius. Results of a prospective study of 60 cases.

Inf OIfhop 16: 3943,1992.

Muller ME, Nazarian S, Koch P. Schatzker J: The compreshensive dasification

of long bones. Springer-Verlag, New York 1990.

Mu1 ler R & Buttner P: A critical discussion of intraclass correlation coefficients-

Sfat Med 13: 2465-2476,1995-

Naidu SH, Capo JT, Moulton M, Ciccone W2, Radin A: Percutaneous pînning of

distal radius fiacturesr a biomechanid study. J Hand Surg Am 22: 252-

257,1997.

67

Nakata RY. Chând Y, Matiko JD: Extemal fixators for w-st Mures: A

biomechanicai and dinical study. J Hand Surg Am 10A 845-851, 9û5.

Norman GR 8 Streiner DL: Biostatistics. The Bare Essentials. (f ) Mosby, St

Louis, MO- 1994.

Older TM, Stabler EV, Cassebaum WH: Colles fracture: evaluation and setedion

of therapy- J Trauma 5: 469476, 1965.

Oskarsswi GV, Hjall A, Aaser P: Physiotherapy: an overestimated fador in after-

treatment of M u r e s in the distal radius? Ard, Olthop Trauma Sutg 116:

373-375, 1997.

Overend T, Wuorï J. Kramer JF. MacDemid JC: Test-retest reliability d a

foream rating sale. J Hand mer 1999.

Owen RA, Melton W. Johnson KA. Ilstnip DM. Riggs BL: Incidence of Colles

fracture in a North Arnerican comrnunity. Am J Publk Heaith 72: 607,

1982,

Palmer AK, Glisson RR. Werner FW: Ulnar variance determination. J Hand

SU^ 376-379.1982-

Pany BR: Colles' M u r e : efficacy of pins and plaster. Am J Orthop 26: 45-50,

1997.

Patterson RM, Nicodemus CL. Viegas SF, Elder W, Rosenblatt J: Normal wrist

kirtematics and the analysis of the effect of various dynamic extemal

fixators for treatment of distal radius fractures. Hand Clin 13: 129-1 41,

1 997,

Peltier LF: Fractures of the distal end of the radius- An historical accourt Clin

OIfhop 187: 18-22. t 984.

Pennig DW: Dynarnic extema1 W i o n of distal radius fradures- Hand Clin 9:

587-602, 1993.

Pogue DJ, Vegas SF. Paftemn füül, Peterson PD. Jenkins DK, S m TD, et al r

Effects of distal radius fiadure malunion on Hirist joint mechanics. J Hand

Surg Am f 5: 721 -727.199O.

Pohjolainen T 8 Alararita H: Predictive factors of fundional ability &ter lower-

limb amputation. Ann Ch& Gyn 80: 3639,Iggl.

Pollock FE, Koman LA, Smith BP, Poehling GG: Patterns of microvasailar

response assoc-ated wïth reflex sympathetic dystfophy of the hand and

W s t J Hand Surg Am 18A: 847852. 1 993.

Pool C: Colles M u r e : A prospective study of treatment J Bone Joint Sug Br

55B: 540, 1973.

Porter M & S t d e y 1: Fractures of the distal end of the radius- Intemediate and

end results in relation to radiologie parameters. Clin Odhop 220: 241 - 251,1987.

Putnam MD 8L Fischer MD: Treatrnent of unstable distal radius fractures:

methods and camparison of exîernal distraction and ORIF venus extemal

distraction-ORlÇ neutralization. J Hand Sutg Am 22: 238251, 1997.

Rader CP, Rauber C, Rehm KE, Koebke J: Interna1 fmtion of the distal radius- A

comparative. experïmental study. A m Olorop Trauma Surg 114: 340-

343,1995-

Raskin KB & Melone CPJ: Unstable artiwlar fractures of the distal radius.

Comparative techniques of Iigamentotaxis- Otfhop Clin Nom Am 24:

275-286, 1993.

69

Ray R: Vaailarizaion of bone grafts and implants. Clin Orthop 87: 4348,1972.

Rayhack J: Symposium on distal radius fractures. Edited by William Cooney.

Contemp Orthop 21: 75,1990.

Rayhack JM: The history and evolution of percutartems pinning of displaced

distal radius fiadures. Oifhop Ciin Nom Am 24: 287-300,1993.

Richardson D: Effeds of tobaaxi inhalation on capillary b l d flow in human

skin- Arch GMmn HeaIttr 42: 1 9-25,1987.

Riis J & Fmensgaard S: Treatrnent of unstable Colles' m u r e s by extemal

fixation. J Hand Surg Br 148: 1451 48, 1989.

Rikli DA 8 Regazzmi P: Fractures of the distal end of the radius treated by

intemal fmtion and early fundion. A preliminary report of 20 cases. J

Bone Joint Surg Br 78: 58&592,1996.

Rohling ML, Binder LM, LangminrichsenRohling J: Money matters: A meta-

analytic review of the association be-n financial compensation and the

experienœ of chronic pain. HeaRh Psych 14: 537447,1995.

Rommens PM, Vanderschot PM, Broos PL: Extemal fmtion of the distal end of

the radius: the sarne tedinique for different patient populations? Arch

Orfhop Trauma Surg 11 1 : 165-1 70,1992.

Ross CE 8 Wu CL: Education. age, and the cumulative advantage in health. J

HeaRh Soc &hav 37: f 04-1 20,1996.

Roumen RM, Hesp WL, Bruggink ED: Unstable Colles' fractures in elderly

patients- A randomised trial af extemai fixation for redisplacernent J

Bone Joint Sutg Br73B: 30731 1,1991.

Rubinovich RM & Rennie W- Colles' fiadure: end tesults in relation ta

radiologie parameters. Can J Surg 26: 361 -363. 1983.

Sanders RW: The problem with apples and oranges. J Otthop Trauma 11 : 465-

466, 1997.

Safin CL, Austin JCt Nickel WO: Effects of smoking on digital bload-flow

velocity. JAMA 229: 1327-1 828,1974.

Sarmiento A, Prati GW, Beny NC, Sinclair WF: Colles' fractures: Functional

bracing in supination. J Bone Joinf Surg Am 57A: 31 1-31 7, 1973.

Sarmiento A, Zagorskï JBt Sinclair WF: Fundional bracing of Colles' fradures: A

prospective study of imrnobilzaüon in supination vs pronation. Clin

0 d h 0 ~ 146: 175-7 83,1980.

Scheck M: Long terni followup of treatrnent of corninutad fractures of the distal

end of the radius by transfiation with Kinhner wires and cast J &ne

Joint Surg Am 44A: 337351,1962.

Schmalholz Ar Bone cernent for redislocafed Coiles' fracture. A prospective

comparÏson with doseci treatment. Acta Oroiop Scand BO: 21 2-21 7,

1989,

Schmalholz A: Extemal skeletal fmation versus cernent fixation in the treatment

of redislocated Colles' M u r e . Clin Orthop 236-241, 1990.

Schmitz M, Finnegan Mt Champine J: The effed of smoking on the clinical

healing of tibia l shaft fradures 1 995-

Schuind F, Alemtadeh S, Stallenberg B. Bumy F: Does the normal contralateral

h s t provide the best reference for X-ray film measurements of the

pathologie mist? J Hand Surg Am 21 : 2430,1996.

Schuind F, Donkedcke M, Raspuin C. Bumy FI Extemal fixation offradures of

the distal radius: a study of 225 cases. J Hand Sutg Am 14: 40rl-407,

1989.

Seitz WHJ: Extemal fmtion of distal radius fiadures- Indications and technical

principles. Otfhop Clin Norfh Am 255-264, 1993.

Seitz WH. Jr., Froimson Ai. Brooks DB, Postak PD, Parker RD, Laporte JM. et a[

: Biomechanical analysis of pin placement and pin site for extemal

fixation of distal radius fradures- ClIn ORhop 207-21 2, 1 990-

Seitz WH, Jr-, Froimson Al, Leb RB: Reduction of treatment-related

complications in the extemal fmtion of corn plex distal radius fradures.

Oehop Rev 2û: 169-1 77,1991.

Seitz WR, Putnarn MD. Di& HM: Limited open surgical appmach for extemal

fixation of distal radius fradures. J Hand Sutg Am 1 SA: 288293, 1 990.

Shankar NS & Craxford AD: Comminuted Colles' fractures: a prospedive trial of

management. J R Col1 Surg Edinb 37: î99-202,W92.

SheMn MA 8 CasM-rth CM: Detrimental effeds of cigarette smoking on lower

extremity m n d healing. J Foof Surg 29: 84-87, 1 990.

Short WH, Palmer AK, Werner FW, Murphy DJ: A biomachanical study of distal

radius fractures. J Hanâ Surg Am 12A: 529-534, 1987.

Shrout PE 8 Fleiss JL: lntraclass correlations: Uses in assessing rater reliability

Psycho/ Bull 86: 420-428, 1979.

Simpson NS, Wilkinson R, Barbenel JC, Kinninmonth AW: Extemal fixation of

the distal radius. A biomechanical study. JHand Surg Br19: 1-192,

1994.

72

Singer BR, Mdauchlan GJ, Robinson CM, Christie J: Epidemiology of fractures

in 1 5000 adufts. J 8one Jant Surg Br BOB: 243-24û.1998.

Sommerkamp TG, Seeman M, SiIliman J, Jones A, Pattemon S, Walker J, et al :

Dynamic extemal fixation of unstable RadUres af the distal part of the

radius. A prospective, randomized cornparison with static extemal fixation.

J Bone Joint Surg Am 76A: 1 149-1 161,1994.

Steffen T, Eugster T, Jakob RPr Twelve years followrip of fractures of the distal

radius treated with the A 0 extemal fixator. lnjury 25 Suppl4: S-W.

1 994.

Steinmann Fr Lehrbuch der funktionellen Behandlung der Kiiochenbniche und

Gelenkverletzungen. Verlag Ferd. v. Enke Stuggart, 191 9. In Pennig DW:

Hand Clin 1993.

Stewart HD, lnnes AR, Burke F D: Fundionai cast-bracing for Colles' fractures: a

cornparison betwaen cast-bracing and conventional plaster casts. J Bone

Joint Surg Br 668: 749-753, 1 984.

Stewart HO, lnnes AR, Burke FD: The hand complications of Colles' fractures J

Hand Surg Br 108: 103-106.1985-

Strafford PW, Binkley JM, Riddle DL: Health status measures: stategies and

analytic methods for assessing change scores. Phys Ther 76: 1 109-

1 123,1996-

Streiner DL 8 N o m n GR: Reliability. Health Measurement Scales. A Practical

Guide to their Development and Use O>dord University Press, Odord.

l99l.

Subbarao J 8 StilWll GK Refiex Sympathetic Dystmphy Syndrome of the

Upper Extmmity: Analysis of Total Outwme of Management of 125

Cases. Arch Phys Med Rehabii 72: 549-!j54,1981.

Szabo RM & Weber SC: Comminuted intraarticîuar fiadures of the distal radius-

Clin Orthop 230: 39-48,1988.

Tang JB, Ryu J, Kish V, Wearden S: Effect of radial shortenhg on muscle length

and moment amis of aie HInst fiexors and extenm. J Orfhop Res IR

324-330, 1997-

Tang JB, Shi D, Gu YQ, Zhang QG: Can cast irnmabifization successfülly treat scapholunate dissociation associated with distal radius fractures? J Hand

Surg Am 21 : 583-590, 1996.

Tnimble TE, Schmitt SR. Vedder NB: lntemal fixation of pilon M u r e s of the

distal radius. Yak JBWMedBb: 179491. 1993.

Tmmble TE, Schmitt SR. Vedder NB: Fadors Mdng functional outmrne of

dispiaced intra- artiwlar distar radius fractures. J Hand Surg Am 19: 325-

340, 4994.

van der Linden W 8 Erïcson R: Colles fracture: how should its displacement be

measured and how should 1 be Îrnmobilized? J Bone Joint Surg Am =A:

1285-1 288,1981.

van der Meer JBW 8 Mackenbach JPr Course of heaRh sMus among

chronically il1 penons: differentials according to levei of education. J Clin

EpidemrOl51: 1 71 -1 79,1998.

Vandersluis R, Richards RS, Roth JH: Use of the extemal fixation apparatus for

peraitaneous insertion uf pins in the distal onethird of the radius: an

anatomic study. Can J Surg 36: SI?-519, 1 993.

Vargha P: A critical discussion af inbaclass correlation coefficients (Letter ta the

Editor). Stat Med 16: 821 -823, 1998.

Verhaven E Haentjens P, Casteleyn PP. De Boeck H, Opdecam P: Nonunion of

a cornminutecf intra-articutar mst fiadure &er extemal fixation- A case

report and review of the literature. Arch Orthop Tmma Surg 110: 1 1&

120,1991.

Villar RN, Manh O, Rushton N, Greatorex R A Three years after Collesr fracture:

A prospective review- J Bone Joinf Surg 698: 63543û, 1987.

Ware JE, Snow KSI Kosinski M. Gandek BI SF-36 Health Sunrey Manual and

lnterpretation Guide. The Health Institute, New England Medical Center,

Boston- 1993-

Ware JE, Kosinski M. Keller SD: SF-36 Physical and Mental Health Summary

Scaies: A User's Manuai. (3rd) The Health Institute, New England Medical

Center, Boston, Mass- 1994

WaMnck D, Prothero Dl FieM J, Bannister G: Radiological measurement of

radial shortening in Colles' Fracture. J Hand Surg Br 186: 5042, 1 993.

Waters PM, Min- CM, Hipp JA, Snyder BD: Noninvasive measurement of

distal radius instability. J Hand Surg Am 22: 572379, 1997.

Weiss NS, Ure CL, Ballard JH, et al: Decreased risk of fractures of the hip and

lower foream with postmenopausal use of estrogens. N Engl J M M 3û3:

ll95-ll98,l980.

Whippie TL: The role of arthroscopy in the treatrnent of intreartkular rn-st

fractures, Hand Clin 11 : 1 3-1 8, 1 995.

Williams AR, Weiss NS, Ure CL: Eff- ofweighc smoking, and estrogen use on

the risk of hip and forearm fradures in postrnenopusal m e n . Obstef

Gynecof 6& 695699,1982-

World Health Organization: International Classification of Impairments,

Disabilities and Handicaps- A Manual of Classification Relating to the

Consequenœs of Disease. 1980.

Zanotti RM & Louis OS: Intra-articular ftactures of the distal end of the radius

treated with an adjustable fmor system. J Hand Surg Am 22: 4 2 M ,

1997.

Zemel NP: The prevention and treatment of complications from ftactures of the

dista[ radius and ulna. Hand Clin 3: 1 -1 1, 1 987-

CHAPTER 2: DEVELOPMENT OF A SCALE FOR PAtlENT RATlNG OF

WRIST PAiN AND DISABIUlY'

Background

Wrist fractures are the rnost prevalent of all fractures. They ocuir in al1

age groups secondary to trauma but have an increased incÏdenœ in

postmenopausal wmen refated to osteoporotic changes in the banes (Habaubi

& Hudson 1991 ). Measurement of outcome, detenninatian of prognostic factors

and cornparison of dïerent treatment methods has been hampered by the la&

of outcorne tools (Cole et al 1 993; Nash 1992).

A number of scoring systems have been developed for Hlnst m u r e

outcome (Gartland & Werley 1951 ; Sarmiento et al 1973). These have been

pnmariiy designed by dinicians based on their persona1 experienœ with wrist

fracture evaluation and treatment They rely on radiographie measures and

clinician/observer evaluation of functional deficits, VVhile these scales have

been usehl in encouraging difrent authors to report dinical results in a simifar

manner, they have several methodologic and dinical flaws. First, they tend to

fows on the clinician's interpretation of the patient's loss of function, rather than

asking the patient how much dificulty they experience- In reœnt years there

has been a refocusing of health evaluation Mich ernphasizes the patient

evaluation of their fundional ability. Secondarily, these scales are not

A version of this chapter has been published in the Journal of Hand Therapy 9:17&183,1996

standardized as they do not indude details on measurement procedures,

reliability or validity (Anonymous 1992; Cole et al 1993).

This paper outlines the progress on development of a wrist outcome

measurement tooL The study has evolved to the stage where a patient rated

wrist score is now ready for general use. A perfmancebad impairnient

rating requires further development to find mathods that poclrey a valid globa1

indicator of wrist fundion and can be performed reliably in an average hand

clinic. During the research and development of these bols a number of

decisions had to made based on results of an expert survey, Iiteratwe review,

pilot testing , consensus exercises and reliability and validity testing. Choices

were based on the need to aeate a tool INtiich was both scientifically sound as

wll as practical in the clinical situation.

This projed to date has taken three years and has resulted in

background studies that were used to develop or test the measure as it

developed (Armstrong et al 199û; MaCDennid et al 1998; MacDemid 1999;

MacDermid et al 1999). Based on ouf experience, a number of global outcome

tools, such as the SF-36 (Jenkinson et al 1994; McHorney et al 1994; Ware, Jr-

8 Sherboume 1992) or the Sickness Impact Profile,(Anderson et al 1993;

Greidanus et al 1994; Katz et al 7 992) are not optimal for upper extremity

practice as they are too lengthy and nonspecificC Our goal was to develop a tool

that would be useful for clinical pradice and research to evaluate wrist

impairment and disability.

Ovenriew

A survey of the International Wrist Investigators (IWI) was perfomed to

establish criteria on wrrent use of outcame measures, opinions regarding

appropriate structure and content of a wrist outcome tool (MatSemid et al

1998). This gmup was seleded as it is the only gmup of dinicians M o are

offkially recognized as specialists in HBst treatment It is an international group

of physicians/researdiers wtio have a high level of interesflexperience in

management of the M s L The group members have academiclclinicaifresearch

components to their pradice and membership is determined by cornmittee

review,

This survey established the content validïîy of the newwrkt scales and

helped define a format that would facilitate its clinical usefuiness- The

developmental stage where specific items for the scale m e generated was

corn pleted after performing a Viorough literature review, interviewhg and pilot

testing patients, reviMng other disabiiity scales, and consensus exerckes with

upper extremity therapists and surgeons. Consensus exercises m e perfomed

by repeateâ suwey of clinicians and patients on content and structure of

developed items or foms. A long Iist ~ w s complied of al1 possible items and

Vien condensed to select the best items. These specific items were compiled

into pilot measures that fit the criteria established in the expert suwey for

structure of a clinically useful outcome tool. The scales were then pilot tested,

revised and reviewd by expert clinicians until they wre satisfactorïly refïned.

Pilot testing was performed by using the tool on a series of patients and

intewiewnig thern about the tool (camprehension, content and ability to

complete) - Statistical analyses w r e used to compare the outcome scale results

with other clinical indicators (Bain et al 1995; MacDermid et al 1995). Only then

was official reliability and validity testing commenced.

Summary of Findings of Survey

The findïngs of the mist outcome survey (Macüenid et al 1998) indicate

that the IWI supports outcome measwement as an essential component of

clinical practice. Hcnmverr, the constraints of clinical pradice dictate Viat a

measure accepted into common dinical use must be simple, not require sWal

equipment and the time allotteci to outcame masurement canmt be more than 1 5 minutes for either patient or dinician-

The survey indicates that clinicians regarded the most important

subjective indicators to be pain, funcüonal ability and satisfaction. Objective

measures of outcome agreed upon by the majority of clinicians induded

functional tests, grïp strength and range of motion. Functional tests were rated

most highiy 84% of respondents suggesting that #ey should be

incorporated Ïnto an outcorne measure. Interestingly only 13% of respandents

currently used fiindional tests to evaluate the outcome after mist fradure.

Theoretical Issues

A number of theoretical issues had to be addressed in the design of the

scales. First, ttiere was the issue O€ how uutwme shoufd be measuredl

classifiad. Historïcally, questions directed to patients w r e temed "subjectiven

responses and measurements taken on patients were termed "objective. "

However, sincerity of effort can be questioned when the physical measurements

taken on patients require their cooperaüon (Hildreth et al 1989; Niebuhr 8

Marion 1987; Saunders 8 Bohannon 1991 ; Smith et al 1989; Stokes 1983).

Thus, physical capacitylirnpaiment rneasures are not necessarily objective.

Tests that require patient moperationleftort are termed by this author as

performance tests. This reflects the fa& that they are based on patient

performance of the test Mich rnay be limited by physical impaiments,

understanding of tests procedures and patient woperation.

Previous wrk at OUT centre lwking at "objective" measures of patient

function including radiographs and fundional performance tests correlateci

pwrly with patient rating of pain, function and satisfaction (Bain et al 1 995;

MacDemiid et al 1995). Many clinicians may have exprienced the

phenornenon where patients exhibit function beyond that expected based on

their physical status, Mi le others report far more disability than expeded based

on their level of physical impairment Therapists need to understand if th8ir

intervention made the changes in joinVnewe/musde system(s) that were

anticipated, as these indicate whether the treatment results comply with its

theoretical basis- More importantIy, patient fundion Mich incorporates

integrated use d these systems is also an element that needs to be measured to

detemine the eff8ds of a treatment, Clinicians need to understand M a t

physical changes result Rom an intervention and how this physical status

translates into disabilÏty and handicap. The research data and dinical

experience suggest that separate performance based and patient ratnig

outcorne tools are reqoired. It was also felt that a combination of these two

levels of rating (subjective and perfomance based) into one scale wu ld

diminish problem solving on the separate sources of success or failure in

evaluation of treatment intewentions,

Dividing the 'patient rate6 venus patient 'performance" components of

outcome fulfilled our theoretical framewrk - Hovuever, a second theoretical

issue arose in developing a scale vvhich w o l d indicate "overall h s t outcorneu.

Different functional tasks can represent widely varying importance to different

individuals. For example, lifting a 10 Ib weight may exceed the functional

requirement of an elderly arthritic person, but be far belawthat required by a

heavy manual labourer- To Say that the same numeric score on this task

represents a similar outcome behmen these hm patients might grossly

misrepresent the real handicap experienced by these two patients.

The Worid Healai Organization classifies health deficits into impaiments,

disabilities and handicaps (Worid Health Organization 1980). Impaimnts

represent limitations in specific structures ie loss of nerve condudion velocity,

loss of joint motion or loss of muscle contractile ability. Disabilities are the loss

of the ability to perfom a specific task such as lifting a 10 Ib weight or rising fmm

a chair. Handicaps are the loss of the ability to perforrn 1 one's roles, such as:

self care, family or occupational commitments or -al responsibilities. m i l e

minimization of handicap is the ulümate goal of surgery and therapy, the achial

handicap experienced by two hdividuals Mth the same impaiments can vary

greatl y de pending on a variety of sacial, psychologica 1. environmental, cultural

and physiologie factors (Cole et al 1993; World Heaith Organization 1980). This

issue was addressed in the patient rated fom. The pain scale can be thought

of as measure of impairment Fundionai abiliw has hiuo subscales: 'specific

activities* where patients rate their ability to do specific wrïst related aaivities

(disa bility) and a 'usualw scale where they rate their ability as cornpareci to their

nomal role (handicap). While the majority of authors will prefar to have a single

numeric score (which m'Il combine measures of impairment, handicap and

disability) a separation of these elernents of outcorne is possible based on the

design of the form.

Tool development

Literature r e v i d to assist with tooi construction included al1 articles

which Iooked at physical requirements to perform a variety of specific functional

activities and those that used standard patient questionnaires to evaluate

outcorne, pain or disability (Roach et al 1991 ).

The ~WU basic subjective efements of outcome are pain and functional

ability. Questions generated in the Iiterature review m e pilot tested through

interviews of Hirist-injured patients and hand surgeons to afflve at the four "best"

questions of pain intensity and one on pain frequency. The final questions wre

selected to represent a spectrum of Wst pain scores with some questions that

may be responded to by those ~ Ï t h mild intemittant pain (latter items of pain

subscale) and one that would be smred highly only wïth those whose pain was

so severe that it is present even at rest (first item).

From the Iiterature a Iist of tasks specifc to wrist physical capability was

corn piled. These items were combined wîth items generated fmm patient

interviews as ones that patients found to be fundional limitations mer wrist

injury. This list was reviewed by the study investigators and the surgeons at ouf

centre to eliminate any that w e unclear or ifrelevant to the wrist Items that

mntained a gender bias or wre perfomed by less than 80% of respoodents

were also eliminated. Pilot testing and consensus exercises with ouf dinicians

wre used ta shorten this list to those included in the final version of the

questionnaire. The final Iist was designed to sample a range of physical

irnpainnents and thus included tasks Wich required different ranges or muscle

strength capabilities.

The specific functional questions m e selected to represent 1 ) heavy

and IigM adivity 2) a spednuii of Wst movements 3) fundional tasks that cwld

be perfomed with either hand or bimanwlly and 4) tasks that wwld be

perfomed frequently by a majotity of individuals.

Activities of daily living were selected for the wrist scale because of the

need to target specific wist-related activities that the majority of respondents

would experience, on a frequent basis. lt was recognized that these may

represent the lower end of physical demands and not be sensitive to milder

irnpairments. That is, these partiwtar physical disabilities, Mi le specific to wrist

pathology, may not represent higher lever abilities necessary for many patients.

Therefore, the ipecific activities" subscale daes not r e m handicap because it

does not cover the spectnrm of fundional tasks that are relevant to the individual

patient For exarnpie, handicap in ocaipationa1 or leisure activities is not

reflected in these items. membre, a section where the person compares his or her functioning to his or her "usual' ptemorbid fundioning was added to the

scafe for eacb of the four major domains: self care, household duties, occupation

and leisure.

Our pilot study indicated mat patients (partiwlarly elderly patients) were

uncornfortable with the visual analogue rating scales. The 0-10 scale was found

to be more accepted by patients. This level of cornfort rnay refiect the f i uency

which rating of a scale of 0-1 O is used in daily Iife and dinical situations.

Therefore, an 1 1point scale was incorporated in al1 questions- It was also

decided to keep the diradion af scale constant for simplicity in patient

understanding and dinician (Jaeschke et al 1930).

The questionnaire was desQned to have five pain items and 10 function

items. This was the minimum number that mxild allow a spearum of

impaimentldisability/handicap to be sampled, Mile still permitting a simple

scoring system. The questionnaire provides several scoring options. The total

wist subjective score could be made into a scale which rates pain and fundional

limitations equally by adding the fundion score (divided by two) and the pain

score together to give a total out of 100. Pain and fundion could be separated

by gïving ratings out of 100 for each. If there was desire to break domi

subjective scores into im paimnts (pain), disabil ities (specific hndional tasks)

or handicap (usual funaion) then scores out of 100% could be tabulated for the

three separate subscales.

Reliability tesang of the subjective outcome seale

Reliability testing was perfomied on two series of patients. Tvvo patient

groups were selected. One group, M ich is Mill in progress. is being used to

evaluate the long terni stability of the scales. A group of patients treated for

nonunion of the scaphoid has been evaluated several years after their surgery.

Because their level of impairment is relatively stable, the reliabif ity on test-retest

with a prolonged interval could be evaluated- They were tested in the HVLC

Clinicaf Research lab during a followp study and are now k i n g resurveyed by

mail approximately one year later. A group of patients with M s t fractures Hiho

were more than six months post injury was also evaluated. They completed the

subjective questionnaire once at our clinic and viiere also provided another

questionnaire Mich ums to be completed 2-7 days later and returned by mail.

The intraclass cwrelation coefficients (1,l) for each item, specific subscales and

the overall saxe for the mist fracture group are presented in Table 1 (Shrout &

FIeiss 1 979)-

Validity of Subjective Wrïst Outcoene ScMng

Validity or ''tnieness" is a more difficult concept to measure. For this

reason, various types of validity have been describecl based on the rnethods

used to detemine howwell a tool measures the trait for whkh it ~ i a s designed.

Content validity is the extent to Mich a tool appears to measure what is

was intended to measure. The foundation for content validity of the mist scale

was laid by surveying the Intemational Wnst lnvestigators for the general

framework of Hihat type of items wwe thought to refiect the important elements of

outcorne. Specific items were generated from the literature and patient

interviews desaibing Mich functional tasks were most demanding on the wrïst

Wording of questions was improved through repeated pilot testing of draft foms.

The pilot versions were sent to the surgeons at the HULC program, the chair of

the IWI subwmmittee working on development of a wrist outwme measure and

fellow clinicat researchers for comments. The final format is based on

amalgamaiion of al1 of this input.

Constnid validity is the extent to which a tool behaves as anticipated

based on the underlying theoretical wnstructs of the tool. Wrist fi-adure is an

acute injury and it is anticipated that wîth normal healing subjective scores would

improve. That is. a time dependency was expected (with a certain unknm

plateau ocairring once healing and rehabilitation was complete) (Field et al

1992). Secondly. it was assumed that more severe m u r e s would tend to have

more residual cornplaints once fradure healing m s complete (Aro 8 Koivunen

1991 ; Szabo 1993). That is, scores wwld be higher in patients moi more

severe fractures. Data on t h e and fradure severity wwe calculateci and

indicated moderate correlations (~0.450.70) with scores improving as time from

fracture increased and decreasing as fracture severïty inaeased, supporting the

underlying constnrct that the scales measured mist pain and disability-

Criterion validity involves compafhg one scale with another- Typically

one wishes to compare a newiy devetoped tw l with a gold standard toal.

However, in this case there was no other tool vvtrich mas designed to measure

wist pain and disability. Therefore, it was necessary to make the most

appropriate cornparisons possible. These were thougM to be cornparison of

scores be-n 1) subscaIes O€ the subjective quesüannaire 2) the wbjedive

and perfonnancebased scales and 3) subjective and physical and pain

subscales of the SF-36 (Jenkinson et al 1994; McHomey et al 1993; McHomey

et al 1 994; Streiner 8 N m a n 1992; Ware, Jr- 8 Sherboume 1992).

mile subscales of the Wst scale and the SF-36 do not cover the same

domain (wrist versus overall health), it was expedeâ that the correlation

between the pain subscales would be relatively high (~4 .79) . This was

expeded because their wrÏst fracture vms the primary source of pain for many of

the subjeds during the evaluation penod. As expected, correlations betwen

d*Merent scales (ie function versus pain) were significant, but not as strong as

between scales covering the same domain. As the SF-36 is a general heakh

status instrument, physical disabilities and pain aimplaints need not be related

to the wkt. Therefore, the associations between the Sf-36 and the wrist scales

were anticipated to be moderate-

Finally, the relationship between subjective scores and performance

subsmes was moderate, as anlicipated based on our premise that perceived

outcome is related to physical capacity, but mediated by a variety of

psychosocial and environmental factors. Correlations between subecales of the

patient rated wrist evaluation and performance baseâ indicators of wrist

impairment and subjective health outcorne subscales f m the SF-36 are

presented in Table 2. These indicate that moderate correlations were found

betwen patient ratings of pain and function with measures of grip strength, mist

AROM (a composite score of al1 six mist movements) and the checken subtest

of the Jebsons (Jebson et al 1969). Correlations tended to be higher for speafc

function tasks than for pain or usual role. One wwfd exped pain or usual role to

be less strongly related to specific fundional task ability than more direct

measures ofwrï& impairnient These correlations supporteci aie conclusion that

the wrkt scaie (patient-rated W s t eva~uation) wms capturing the important

dimensions of Mst pain and disability. Because of the la& of another wist

scale as a axnpamtor, ~ i e must assume that the uinst scale w s more specific to

wrist pain and disability than the SF-36 Mich addresses overall health and al1

body parts.

Figure 1. Paient R a t d Wrist Evaluation

Name Date

The questions below wrll heîp us undemtand how much dltWWty you have bad with your wnst in the p s t m k . You Mil be describing yourav- wrist sympfums overfhe nmf -on a scele of G1O.Pîease provide an answer for A U questions. If p u dîd not pelbrm an acdivity, please ESTiMATE the pain or difkulty you would ex-

1. PAIN

Rate the average amount of pain in your owr the past w e k by circing the number that besf desaites your min on a seale from û4O.A zero (O) means that you dîd not have any pain and a ten (IO) means that you had the worst pain you haw evwexper&nceà or that p u couM not do the actNity because of pain-

Sample scale -, 0 1 2 3 4 5 6 7 8 9 1 0 No Pain Worst

Ever

RATE YOUR PAIN:

At rest 1 0 1 2 3 4 5 6 7 8 9 1 0

When doing a ta& w i ~ a 0 1 2 3 4 5 6 7 8 9 1 0 repeated wrist movement

When lifting a heavy object 1 0 1 2 3 4 5 6 7 8 9 1 0

How ofien do you have pain? 0 1 2 3 4 5 6 7 8 9 1 0 Never Aiways

Please tum the page. . . . . . . . . .

A. SPECIFIC ACTlVlTlES Rate the amount of d M f c u ~ you expeanced perfoming each of the

items listed bebw , ovw the past ~ e e k . by Crrding the number that describes your difficulty on a SC& of 0-1O.A zem (O) means you did not expeti6ncle any diffcuffy and a f.n (IO) rneans it was so diffiicuit you wete unabîè to dD it a all.

Sample scale * 0 1 2 3 4 5 6 7 8 9 1 0 No Unable

Tum a door knob using my 0 1 2 3 4 5 6 7 6 9 1 0 affected hand

Cut meat using a knife in my I 0 1 2 3 4 5 6 7 8 9 7 0 affected hand

Fasten buttons on my shirt I o 1 2 3 4 5 6 7 8 9 1 0

Use myaff8Cfed hand to push up O 1 2 3 4 5 6 7 8 9 10 from a chair 1 Carrya10Ibobjectinmyaffected O 1 2 3 4 5 6 7 8 9 10 hand 1 Use bathrwm tissue my I 0 1 2 3 4 5 6 7 8 9 1 0 affected hand

6- USUAL ACTIVITIES Rate the anrounf of dHVku@ you expe*nced perfonning your usual

activilies in each of the amas Iiisted beîow, over the pat week, by t i ~ l i n g the number that besf describes yourditkuRy on a scaie of O-10.8~ 'usual acfivities: we mean the acfJv&s you peribnned belbm you starfed having a problem wth your wtist. A zem (O) means that you d i not expenenCe any diffculty and a fen (10) means it was so diffii=utt you were unable fo & any of

1. Personal care adivities 0 1 2 3 4 5 6 7 8 9 1 0 (dressing, washing)

2. Household wk (cleaning. I O 1 2 3 4 5 6 7 8 9 1 0 maintenance)

3. Work (your job or usual I O 1 2 3 4 5 6 7 8 9 1 0 everyday wwk)

Scon'ng

The scale provides the followïng options in scoring:

1) Total wrist score: This score can be computed as the average of the pain and

the disabilitylhandicap score. That is the total pain score (out of 50) is added to

the disabilitylhandicap saire (score out of 100 divided by 2). This provides a

swre out of 100. A higher score is associated with more pain and disability and

thus a poorer outcorne-

The three separate sedions of the scale can be scored separately, as

percentages, to provide a pain score, a disability score and a handicap score if

one wishes to follow the recommendation of the World Health Organization on

the classification of heaith.

Finally, one could separate pain and fundion into separate scales

(percentages) - Dependhg on the clinical purpose, one migM select to use wbscales.

For example, in a clinical situation Mere one is primarily treating pain as a

problem, it might be beneficial to separate pain and function so that one could

determine if the intervention was impacting on pain. The reliability coefficients of

the subscales are al1 excellent (Table l), vuhich justEfi8~ the use of subscales

where they suit the clinicai goals. Howver, the simplest approach is to use the

total wist score (Which equally weights pain and fundion) as an indicator of

overall stanis. Our data is in accordance with statistical theory that the reliability

of overall swles tends to be higher than specific items or subscales.

Conclusions

The final patient rated scale has taken three yean to develop and test A

performance based impairment scale is now in development. A number of

complications. that relate to developing valid and reliable methods, have

delayed its introduction. The patient rated scale is not the only outwme scale

that can be used in evaluation dwst patients. Howsver, it is the only one, of

which we are aware, that provides a brie#, yet reliable and valid assessrnent of

Wst related pain and disabiiity/irnpaimient. We are currently wing this sale in

a variety of research and dinical applications and invite others to do so as wll.

Anonymous: Clinical Assessrnent Recommendations. Ediüon 2. American

Society of Hand Therapists, Chicago- 1992.

Anderson RT, Aaronson NK, Wil kin DI Critical review of the international

assessments of health-telated quality of life. Qua! Mis Res 2: 369-395,

1993,

Armstrong A, Macûemid JC, Chinchalkat SI Stevens RS. King G: Reliability of

range of motion measurements in the el& and foream. J ShouM Ubow

Surg 7: 573580.1 998.

Aro HT 8 Koivunen T: Minor axial shortening of the radius affects outcorne of

Colles' M u r e treatment J Hand Surg Am 16: 39239û, 1 991.

Bain GI, Roth JH, Pugh D, MacDermid JC: Matched hemiresection niterposition

arthroplasty of the distal radioulnar joint J Hand Surg Am 2OA: 944-950.

i 995.

Cole B. Finch E, Gowland C: Physical Outcorne Measures. Canadian

Ph ysiotherapy Association, Toconto. 1 993.

Field J, Wacw~kk DI Bannister GC, Gibson AG: Long-tenn prognosis of displaced

Colles' fracture: a 1 O-year prospective review. lnjury 23: 529-532, 1 992.

Gartland JJ & Werley CW: Evaluation of haaled Colles' fractures. J Bone Joint

Surg Am 33A: 895-907,1951.

Greidanus J, Groenier KH, Schuling J: [Fundional status of the elderly

measured the Sickness Impact ProfileI. TZjfdJcn~ Gemnfol-Genatr-

25: 1 1-1 6, 1994-

Haboubi NY & Hudson PR: Factors associateci with Collesr fracture in the

elderly. Gemntology 37: 335338, 1991.

Hildreth DH, Breidenbach WC, Lister GD, fiodges AD: Detection of submaximal

effort by the use of the rapid exchange grip. J Hémd Suq Am 14A: 742-

745,1989-

Jaexhke R, Singer J, Guyatt GH: A cornparison of sevenpoint and visual

analogue scales. Data h-om a randomized trial. Contml Clin TMs 11 : 43-

51, 1990.

Jebson RH, Taylor N. Tneschmann RB, Trotter MJ, Howwrd W An objective and

standardized test of hand fundion. Arch Phys Med Rehabil50: 31 1 -31 9,

1969.

Jenkinson Cl Wright 1, Coulter A: Criterion validity and reliability of the SF-36 in

a population sample. Qua[ Loé Res 3: 7-1 2.1994.

Katz JN, Larson MG, Phillips CB, Fosse1 AH, Liang MH: Comparative

measurement sensitivity of short and longer health status instruments.

Med Cam 30: 917-925,1992-

MacDermid JC: Development of a Performance Based Impairment Scale for the

Wrist. Unpublished wrk, 1999.

MacDermid JC, Alyafi T. Richards RS: Test-retest reliability of staîic and

endurance grip strength tests perfomied on the Jamar and NK

Unpublished work, 1999.

MacDerrnid JC, Bain GI, Rarnpersaud R. Roth JH: Factors related to satisfaaion

in review of patients mth hemirection arthroplasty or silastic thumb

arthroplasty. J Jand Ther 8: 5050, 1995.

MacDermid JC, Turgeon T, Beadle M, Richards RS, Roth JH: Patient rating of

HRst pain and disability: A reliable and valid measurement twl. J OIfhop

Trauma 12: 577406,1998-

McHorney CA, Ware JE, Jr., Lu JF, Sherboume CD: The MOS -item Short-

Forrn Health Survey (SF-36): III. Tests of data quality, scaling

assumpttons, and reliability across diverse patient groups- Med Care 32:

4046,1994.

McHorney CA, Ware JE, Jr., Raczek AE: The MOS 36-Item Short-Fom Health

Survey (SF-36): II. Psychometric and dinical tests of validity in measuring

physical and mental health constnrcts- Med Care 31: 247-263, 1993.

Nash DB: The state of the outmmeslguidelines rnovement M ~ o n s N, lmaging

Emnomics 1992-

Niebuhr BR & Marion R: Detecting sincerity of Mort Men rneasuring grip

strength. Am J Phys Med 66: 16-24,1987-

Roach KE, BudimanMak E, Songsiridej N, Lertratanakul Y: Development of a

shoulder pain and disability index Atth Care Res 4: 143449, 1991.

Sarmiento A, Pratt GW, Berry NC, Sinclair WF: Colles' fraaures: Functional

bracing in supination. J Bone Joint Surg Am 57A: 31 1 -31 7. 1973.

Saunders NE 8 Bohannon RW Can feigned maximal efforts be distinguished

from maximal Morts? J Hum Musck ~ r f o t m 1 : 1 6-24, 1991.

Shrout PE 8 Fleiss JL: intraclass conelations: Uses in assessing rater reliability.

Psychd Bull 86: 42W28.1979.

Smith GA, Nelson RC, Sa- SJ, Sacloff AM: Assessing sincerity of effort in

maximal grïp strength tests. Am J Phys Med Rehabii 6û: 7380,1989-

Stokes HM: The serïously uninjured hand - Weakness ofgnp. J Occup Ther 25:

683484,1983-

Streiner DL 8 Norman GR: Health Measurement scales. A pradical guide to

their development and use. Oxt~ord University Press, Oaord. 1992.

Szabo RM: Extra-artiailar fractures of the distal radius. Offhop Clin North Am

24: 229-237, 1993,

Ware JE, Jr. 8 Sherboume CD: The MOS 36-item short-fom health survey (SF-

36). 1. Conceptual framewwk and item seledion. Med Care 30: 473-483,

1 992.

World Health Organizatïon: International Classification of fmpairments,

Disabilities and Handicaps. A Manual of Classification Relating to the

Consequences of Disease, 1980-

Table 1. Test-retest Rdiaôility of the Paüent-Rated W-t Outcome Fomi

(n=38 patients with wrid frcicarw)

1 Subscales 1 ICC

1 Pain 0-89

Specific Fundion 0.89

Usual Function 0-90

, Total Score 0-93

Table 2. Validity of the Patient-Rated Wrirt Outcome Fom (ni63 patients with wrist fractures)

Total Grip C heckers Physical Role B d l y Overall

AROM Strength Subtest Function Limitation Pain L

Pain -0,35 4 4 4 -0. 54 4,58 -0,62 -0.79 -0. 74

Specifio Funotion -0.52 -0,74 -0.65 4.66 4 4 4 -0.63 -0.72

Usual Function -Ot41 -0. 50 -0,44 -0.69 -0.M -0,74 -0,86

Overall Score -0.59 -0.60 -0.77 4.80 Pearson correlations ôetwwn different test scores

CHAPTER 3: PAT)ENT RATlNG OF HlRlST PAIN AND MSABIUIW A

REUABLE AND VAUD MEASUREMENT TWL2

Introduction

Outcorne measurement ha$ reœived an increased focus in the current

health care climate (Tarlov 1996). Some reasons for this include the tising cost

of health care Ath littie documentation of the cause, variation of practice

amongst centres and individuals and deficiencies and dismepancies in the

current Iiterature (Bradham 1994). While outcome measurernent has becorne a

priority issue in health care, considerable diversity of opinion exists with respect

to the definition of outcorne measurernent and which measures will best define it,

In general, the trend in patient evafuation has k e n to decrease emphasis on

measures of impairment and to Ricrease the foais on disability and handicap

(Tarlov 1996; World Health Organization 1980).

The wrïst is one of the most common sources of orthopaedic disorders.

Distal radius fractures are the single most common fracture (Alffrarn 8 Goran

1962; Doczi 8 Renner 1994; Owen et al 1982). A number of validatecl health

measures currently exist that focus on general health or quality of Me, such as

the SF-36 (Ware, Jr. 8 Sherboume 1 992) or Sickness Impact Profile (de Bmin et

al 1992). While these general health measures appear to be adequate for a

variety of conditions, this may be less Vue for viinst injury, as few SF-36 items

are directly relevant to the hand or wist Impact on overall health is an

important component of evaluating health intervention- Howver, where

A version ofthis chapter has b e n published in the Journal of Orthopaedic Trauma l2(8) 5iï-587, 1 998 .

treatment is direded to the wist joint, additional measures that reRed wrÏ~t-

related pain and disability are imperative.

The la& of a standardized tool for patient evaluation of Hinst symptoms

resulted in the decision to develop a tool 4th hm basic principles in rnind: 1 )

instruments must be construded using methodology far standardization (Cole et

al 1993; ffirsfuier & Guyatt 3985; Streiner & Norman 1992) and 2) instruments

must be viable in aie dinic environment A stmctured survey of the International

Wrist Investigators was conducteci to define 4 ) aiment usage of outcome

measures by clinicians 2) content items which experts considered essential for

inclusion in a outcorne measure and 3) structural charaderistics required

by clinicians who might incorporate a resuitant scale into their ciinical practice.

The survey las to address issues in content validity and ciinical applicability of

the resultant measure. Based on the survey iiifomation and following research

methodology for instrument development, an instrument was designed and its

reliability and validity ware tested. The steps involved in designing and testing

the Patient Rated Wnst Evaluation (PRWE) were: 1 ) using the survey

information to direct the structure and content of the tool2) Item Generation 3)

Item ReductionlSelection 4) Questionnaire Construction: refining itemsl scoring

system 5) Pilot testing 6) Evaluation: Reliabilitylvalidity testing.

Material and Methods The Wrist-Outcorne Expert Survey

Survey Methods

The general purpose of the survey was to define current practice and

opinions with respect to outcome measurement after wrist injury. The swvey

was designed to address current outcMe measurernent pradices, opinions on

appropriate content for a wrist o u t m e measure, structural guidelines for a tool

that wwld be usehl in the clinic and barriers to use of outcome measures- The

draft survey was evafuated for mors in design or mxding by a panel of

professional therapists and epidemiologists. A pretest swvey distributeci to

seven hand surgeons not contained in the study sarnplhg frame determineci that

the questionnaire had to be very brief to be acceptable to respondents.

Pretestirtg resulted in streamlining. removing open-ended questions and

shortening of the survey so that mean campletion time was four minutes.

The final questionnaire was mailed to al1 adive members of the

International Wrïst lnvestigators (M) (n=151). This grwp wnsists of

physicians with a demonstrateci interest in the investigation and Weafment of

wrist pathology. Acœptanœ into the IWI is based on expwïenœ in clinical

practice, research and teaching related to the wrï~t as evaluated by a panel of

the IWI. An initial response rate of 55% was adiieved. A second mailing

resulted in only 1 7 additional c m pleted responses.

Survey results were entered in SPSS sofhiirare which was used to provide

sumrnary desaiptive data on response patterns. Consensus opinions were used

to devise a new measurernent t w l that wauld suit clinical pradice and yet meet

scientific standard on measurernent,

Expert Survey Results

The final response rate observed was 66% over a pend of 6 months

during 1995. See Table 1. Four surveys wete tetumed unanswered, for various

reasons, and were eliminated from the summary of results. It has been

suggested that confidence in survey results is lawered Men response rates are

less than 60% (Salant 8 Dillman 1994). The present study sample induded

experts in thirteen dmerent countries from a group of surgeons M o are

approached for a Hnde variety of professional activities- Surgeons responding to

the survey were experienœd dinicians involved in a number of wrist-related

academic adivities in keeping wïth their membership in the IW and because of

their professional adivities are frequently surveyed. For these reasons the

response rate of 66% was considerd hvourable-

The responses to individual suwey questions are listed in Table 2. While

physicians did not have a standardized tool to measure wrkt pain and disability,

responses regarding current pradice indicated that there are certain items which

are commonly used by surgeons to evaluate these parameters. Pain, wrk

capability and limitation in movement wwe most frequently used by dinicians as

considerations in subjedive evaluation of oufcome. The majority of respondants

also routinely perfomed objective measurements: range of motion, radiographs

and grip strength. Only a small minorïty of respondents fek that amently

available outcome tools were satisfactory for clinical(15Jb) or research (1 0%)

needs. A minonty of respondents were familiar wîth general health suweys

(question 12) and of these, few thought that these tools wu ld be suitable for

evaluation of wrist ftactures. The primary reasons for lack of suitability were that

the foms are too long and not specific to the W s t The majority of respondents

were reœptive to incorporating a new instrument into Vieir pradice.

When asked to rank which measures w r e most important for inclusion as

outcome indicators (question IO), grip strength, range of motion and patient

evaluation of their function were the three most highly rated. Respondents felt

that direct measurement of strength, movement and fundional ability was

important both as an indicator of outcorne (question 10) and as aimponents of

an objective scale (question 13). Radiographie appearanœ was not seen as an

important indicator of outcorne and was not suggested for inclusion in outcome

measures (question 13). Subjective domains that were ranked rnost highly by

respondents w r e pain, limitation in activities of daily living and work (question

14).

The structural diaradeflstics that survey respondents considered

important in development of a wrist outcorne measure are Iisted in Table 2

(question 9). Those rated as most important wre requirements of the surgeon's

time, complexity of the measure and the papenmrk required fw its completion.

Incorporating outcome measures into ciinic pracüce (question 15) was thougM to

be hampered by the pressure to see numerous patients. lack of data

management capabilities and insufficient support staff- Time commitment for

both the patients and the surgeon was seen as an important consideration in

whether a tool vms suitable fot dinical practice (questions 16 and 17). The

single most cornmon recommendation made by respondents was to keep

measures simple and "user -fnendly.m As patient ratïngs and performance

indicators of outcome tend to be poorly related, (Bain et al 1995) it was decided

that these two constnias should be measured separately. The presant study

developed a brief (15-item) measure of mist pain and disability questionnaire

(Figure 2) guided by the opinions obtained in the expert survey-

Development of the Patient-Rated Wrist Evaluation Questionnaire

Developrnent of a questionnaire for patients to rate w-st pain and

disability involved the following steps: 1) using information from the expert

survey to decide on structure and content of the tool2) item generation 3) item

reductionlselection 4) questionnaire construction: refining items/scoring systern

5) pilot testing fom and 6) reliability and validity asseosment.

Once the expert survey identifid the domains of pain and fundion duting

activities of daily living as priorities, a master list of specific mist-relatecl

problems was generated by: 1 ) patient and expert interviews 2) evidence on

wrist fundion derïved Born the biomechanical literature and 3) pain and function

items from other questionnaires. This master list was reduced through expert

and patient review and pilot testing. Pain items ware reduced to cover a full

spectrum of seventy and address both intensity and frequency. A pain subscale

was designed to be sensitive to mild pain ie only ocwrs wïth adivity as well as

more sevete pain such as occurs at rest.

Functional items were reduceâ by a proœss Hkiereby items had to satisfy

several aiteria ie items that wre: 1) mrnonly perfonned with ei#er hand 2)

perfomed by a majorïty of patients and 3) could be understood by most

respondents. The specific fundion items were designed to assess parameters

of the spectnirn of wrïst motions and strength as used in specific daily activities.

The specific items most sensitive to di f rent M s t pathologies were detemined

by pilot testing the questions on small groups of patÏents. As wll, patient

specific role limitation vas a d d r e d in fow queations asking about their ability

to perform their usual activity in the domains of self care, work role, home life

and recreation.

The questionnaire format was designed to balance the clinical

requirement that fomn be simple and brÏef Hiith the methodologic requirement to

portray a spectrum of mst disability. ltems m e limited to five pain questions

and ten function questions to permit a simple scoring system. A 0-10 scale was

selected because of its patient acceptance. simplicity in scoring and

responsiveness to change. The %est" items for the subscales were seledeci

through statistical evaluations perfomed duflng pilot testing and expert

consensus. ltems had to be acceptable on the basis of having a high response

rate with patients. refiect an important dimension of the subscale and be affected

by wrïst pathology. The scale uws designed so that a total score (out of 100)

can be corn puted by equally weighting the pain score (sum of 5 items; w r s t

score is 50) and the disability score (sum of the 1 O items; divided by two).

Separating the pain and disability subscales may be appropriate in certain

applications. The final instrument is Figure 1.

Reliability Testing of the PRWE

Preliminary reliability analysis was conduded during pilot testing and has

been previously reported (MacDemid 1996). In this study, the reliability of the

final version of the PRWE is addressecl. Testt-retest reliability anabsis of the

PRWE was designed to masure instrument stability over diffmnt aaiities and

pathologies. The diagnostic groups selected wre: q) subaaite distal radius

fractures still involved in physiotherapy, 2) distal radius fradures Hiho had

completed their medical and physiotherapy tr8atrnent and 3) patients who were

an average of six years post treatment for nonunion of a scaphoid fradure.

Patients initially completed the PRWE as a component of their routine dinical

evaluation and wre asked to participate in a swey Hihereby they wmpleted a

second PRWE at a later date. The two Wst m u r e groups represent active

and immediate post-treatrnent phases, wtiere shofi-temi retest reliability was

assessed,

Patients m e given a second PRWE Mich was completed between iw

and seven days later and retumed by mail to the senior author. Patients in

group three (patients post nonunion scaphoid) were contacted, one year after

participating in a long term outcome study, to complete a second PRWE and

thus provide an estimate of the long term stability of the tool. Response rates

were high for the short temi distal radius m u r e patients (groups one and two)

(>9O0h). Despite the prolonged interval between occasions for group three, the

retum rate for the scaphoid group was also cansidered high (76%).

Reliability coefficients ( CC's (type 23)) wsre calculated using SPSS

(Norusis 1990) according to Shrout and Fleiss (Shrout & Fleiss 1979). ICC's

were compared €0 a subjective rating scale e0.40 poor, 0.40-0.75 fair and ~ 0 . 7 5

excellent (Fleiss 1 986). Because, one year is a prolonged retest interval and

little comparative data exists on long terni stability of instruments, the reliability

of subscales of the SF-36 was also calculated in group three. Subscale scores

for the SF-36 and PRWE wwe calculated as describecl by their authors

(MacDemid 1996; Ware et al 1993; Ware et al 1994)

Reliability Testing Results

Subject characteristics are presented in TaMe 3. Reliability results are

presented in Table 4. Excellent reliability was demonstrated by the total PRWE

score for all three patient groups (ICCs >0.90). Even with a one year time lapse

betwen completions of the PRWE. the 1CC was still0.91- The pain subscale

demonstrated excellent reliability in al1 thme groups (ICCs2 0.90). Fundion

subscales had excellent reliability in the distal radius m u r e groups, (ICCst

0.85) but only moderate reliability over the long terni in scaphoid non-union

(ICCsr 0.61). The only SF36 subscale to demonstrate excellent test-retest

reliability more than one year was the physical firnction scale (ICC=0.89 ).

(Table 5)

Validity Testing of The PRWE

Constnrct Validity

Construct validity assesses the extent to which PRWE scores correspond

to theoretical constructs. Rie constnict evaluated was that patients with wrist

fractures shouM demonstrate less pain and disability mer tirne- Patients Hiho

had sustained a wist fracture ( ~ 1 0 1 ) wmpleted the PRWE at baseline, 2, 3

and 6 months after their fracture. Repeated rneasures ANOVA with post-hoc

Student-Neuman-Keuls tests was used to determine if a statistically signifïcant

improvement occurred over time (Norkin 8 White 1995). This analysis was

perfomed for both the SF-36 summary score and an impairment score (not

measured at baseline).

Criterion validity

Criterion validity requires cornpanson with a criterion comparator at a

single point in tirne to detemine the extent to vrtiich two instruments provide

sirnilar results. Where gold standards exist they are preferable; more typically

new instruments are compared to more established instruments. Tw groups of

patients (wrïst fracture and resolved scaphoid non-unions) wre used as

subjects in this wmponent of the study. The 35 patients with scaphoid fractures

reported on in the reliability wmponent of this study and 101 patients with vurist

fractures wwe enrolled in this component of the study. The critenon

comparators were the SF36 and an impairment saxe based on a physical

assessment.

Impairment testing was performed by tesüng wrkt ROM (fiexion,

extension, pronation, supination, radial and ulnar deviation), (LaStayo &

Wheeler 1994) grÏp strength ( M a t h i ~ t z et al la) and dexterity testing

(Jebson checkers subtest) (Jebson et al 1969)- ROM was compared to

normative values, grip and dexterity were expressed as a percentage of the

uninjured side ~Mth an adjustment for dominance. (See Appendix 1. A

composite objective score ~ n 3 5 derivecl by weighting: 30 points ROM. 40 points

grip strength and 15 points dexterity* This score ptovided an overall rating of

measured M s t physical impairment

The SF-36 is a general health questionnaire that provides scores for eight

difberent domains of health. as well as summary scores for physical and mental

health- While the physical subscales are not well referenced to the upper

extremity, they provide a standardized health rneasure that serves as an

altemate aiterion of patient reported physical status. The physical summary

score is a composite score weighting indicators of physical impairment and

bodily pain from the SF-36 into a single number which is referenced to US

population scores. As such, it is an indicator of overall physical health. The

Physical Cornponent Summary score and the subscales Physical Fundion and

Bodily Pain w r e expeded to be moderately related to PRWE scores. The

Mental Summary score covers psychosocial dornains of health and as such was

expected to be less related to PRWE scores. The association be-n the

difïerent scores was mea~ured by Pearson product moment wrrelations.

Validity Testing Results

Constnict validity

A statistically significant improvement in the PRWE score ocairteci over

time (p4.000?). This was paralielled by improvements in the impairnent score

and the SF-36 Physical Surnmary Scores (p4.0001). The amount of

improvement was 74% as measured by the PRWE and 16% as measured by the

SF-36 Physical Surnmary Score Fable 6).

Criterion validity

Correlations between domains of the PRWE. the measured impairment

s a l e and the SF-36 are in Table 7. Correlations were highest between the

PRWE pain s a l e and the SF-36 bodily pain scale. Correlation betwen PRWE

scores and physical summary scores were higher than with mental health

surnmary scores. Moderate correlations w r e observed between direct1 y

measured impairment (grïp, AROM, dexterity) and patient reported pain and

d isabil ity. All correlations were statistically significant (p<O.Ol) These

correlations confomed to our expctations as to the strength of association

between these different measures and thus supported the validity of the PRWE

in rneasurïng wrïst specific pain and disability.

Discussion

Outcorne measures are used to assess severity and document clinical

change over time. Stanâardited measurement tools are designed for a spe«fic

purpose in a certain population and have detailed instructions on administration,

scoring interpretation, reliability and validity(Cole et al 1993). Traditionally, wrïst

treatrnents have been evaluated with measures of impairment: abnomaiities of

physiological or anatomical structure (World Health Organization 1980) such as

radiûgraphic appearame, ROM or grïp strength. Measures of impairment include

measures which are tnily objective, ie not conttalled by patient cooperation,

such as radiographs. and tests uuhiai are aff6ded by patient cooperation such

as grip strength, ROM and fundional tests. The survey of expert dinicians

reveals that these continue to fom the primary evaluative tools for monitoring

care of wrist fractures-

A number of scales have been used to sumrnarize these impairment

measures into global ratings ( B r a h y et al 1989; Gartland 8 Weriey 1951 ;

Sarmiento et al 1973)- These scales have tended to rely on radiographic

measures anaor clïnician opinion. While these scales have hel ped dinicians

communicate with respect to overall outcome after Hnist fracture, none have

reported reliability or vafidity and thus cannot be considerd as standardized

scales-

Traditional measures of impainnent may be valuable in reporting the

physiologie or anatomic results of treatment but do not represent functional

results (Bain et al 1995). This finding was supportad in this study where direcüy

measured impaiments were only moderately correlated with patient reported

pain and disability. The tool developed in the present study provides a

standardized tool to measure -st pain and disability as reported by the patient.

However, rather than re placing tradit ional measures, these patient-rated

outcornes tend to be added on to wrrent evaluations. Thus, the challenge as

dictated by expert survey respondents was to find methods which do not

encroach on valuable patient contact tirne- The survey confimed that measures

of health status rneasures such as the SF-36 have not been readily accepteci

into upper extremity clinics. The major drawback reportecl was that they are not

specific to upper extremity pathology and have not been validated for this use.

FuNiemore, clinicians find them too lengthy for clinical usage. Some œntras

and databases use the SF-36 as a genetic health instrument but supplement it

with more specific measures-

There wre same inconsistem-es betw8en current pradiœs and rnethods

required for outcome measurnent Fable 2)- For example, radiographs ware

used by 97% of surgeons dunng finaf evaluation of a Mst frsdure. yet, X-ray

appearance was considered the most important objective measure by only 6% of

surgeons and was remmmended for indusion in an outcome measure by none

of the respondents. This is thought to refiect the fact that radiographs were

wnsidered essential to document anatomic alignment (impairment), but, may not

be predictive of disability(6ain et al f 995). SimilarIy, fundional adivity tests

were considered the most important objective measure by 59% of the

respondents, yet, standardized functional tests were reportedly perfomed by

only 13% of the sample.

Surgeons expresseci concerns about the general use of the measum,

whether or not it could be both flexible (applicable to a number of situations,

operative and nonoperative) and efficient Another issue was the amount of

education requked fw members of the health care team who might be expected

to administer and interpret this measure. Surgeons recognized that tracking

outcome information requires data management capability and reported this as a

limitation in incorporatïng outcome measwes into their clinical practice. The

PRWE provides a quick visual representation of the level of a patients pain and

disability which is useful in the clinic. A simple score can be computed out of 1 00

and used for outcorne studies or dinical trials-

The PRWE allouis patients to rate their status from 0-10. This type of

scale was selected as pilot testing demonstrated that visual analogue

scales(VAS) were not wll accepteci, partiwlarfy by elderly patents.

Furthemore, a 7 point likert sale has been reported to be as respnsive as a

VAS scale (Jaeschke et al 1990). Rating on a scale from 0-1 0 is a common

clinical pracüce. Responsiveness is better with more respanse options, with 7-10

choices k i ng considered reasonable. Information on the effect of the number of

points on scale reliability have demonstrated that converthg distributions wÏth

known correlaüons, a small loss in reliabilïty oca i r s for 7-10 categofles, Hihile

the use of hre categoties reduces reliability by about 12 percent and tw

categorïes reduœ it by 35%. Convenely, t m many categodes may also have

an adverse efFect on reliabifity by making choices too difficult for the

respondents (Streiner 8 Norman 1992). The high reliability coefficients

demonstrated in this study suggest that the measurement approach seleded

fulfills both measurement th- aiteria as wll as those stated by expert

clinicians in that l is btief and simple to score. Even with a one year time lapse

between completion of the PWRE exœllent lCCs ware obtained, indicating that

the patient responses on pain were highly consistent aven over an extended

period of tirne, during which racall bias wuld be definitdy eliminated.

Fundional performance items demonstrated relatively high reliability in

the fracture groups, but were low in long temi evaluation. This may be because

these questions wwe perceiveci less consistently or that patients changed

funetional perfonnanœ over the course of a year in the absence of changes in

pain or joint status. A further factor contributing to the lower reliability

coefficients obsewed on long terni followup was the la& of fundional problems

reported at this late time frame. As sample variability contributes to calculation of

the ICC, (Shrout 8 Fleiss 1979) homogeneow samples tend to appear less

reliabie.

The SF-36 includes a variety of health domains Ath questions whose

scales range f m 2-5 points. A number of health issues other than the status of

the M s t could Med their SF-36 scores. The consistencies reported on the

PRWE suggest that change in mist pain and disability w s not a contributing

factor to change in health status. The l m r reliability of most SF-36 subscales

may refiect, in part, the fa& mat it has fewer options per scale. Another possible

explanation is that the dimensions of health other than pain and physical

limitation are less stable-

The PRWE vms design& to complement traditional impairment and

radiographie measures. The correlations behiiieen measured impainnent and

patient reported pain and disability w r e moderater indicating that fadors other

than strength, movernent and dexterity contribute to -ent perceived disability.

The PRWE w s designed to conforni to the prioriües set by clinicians

using pain, adivities of daily IKe and wwk as indicators of subjective outcome. The PRWE equaily rates pain and fundional problerns- While pain is one

possible reason for Iimitab'on in acüvity, a separate pain scale with equal weight

was deemed as necessary by expert consensus- The importance that expert

clinicians placed on a pain scale was supporteci by the strength of association

between the pain sale and overall physical health as reported on the SF-36.

Other measurement options for wist fradures include the SF-36 wtiich is

a generic qualiv of life instrument (Ware, Jr. (L Sherboume 1992) and the DASH

(Hudak et al 1996) wuhich is a region specific quality of Iife instrument. The DASH

contains 30 questions 21 of which relate to upper extremity funetional activities

and 2 of which relate to upper extremity pain- Some variation in outcorne c m be

anticipated based on the specificity and measurernent properties of Viese

scales-

A comprehensive evaluatïon of outcome after treatment requires reliable

measures of impairment, disability and quality of life indiCators. Clinical pradice

demands may require thaï decisions be made to select outcome toals which

provide useful information with respect to outcorne and are feasible for routine

administration in the clinic. The PRWE was designed to provide dinicians a

simple , brief, reliable and valid tool to assess mist related pain and disability.

Alfkarn PA 8 Goran CHB: Epidemiology of M u r e s of the foream. J Bone Joint Surg Am 44 . : 1 051 1 4, ? 962.

Bain Gl, Roth JH, Pugh O, MacOmid Je: Matched hemiresedion interposition

arthroplasty of the distal radioulnarpint J Hand Surg Am 20A: 944-950,

1995-

Bradham DD: Outcornes research in orthopedics: History, perspectives,

concepts and future- Arthmscopy: me Journal of A t t h m p y and

Related Surgery 10: 493501. 1994.

Bradway JK, Arnadio PC, Cooney WP: Open reduction and intemal f~at ion of

displaced, comminuted intra-articular M u r e s of the distal end of the

radius. J Bone Joinf Surg Am 71: 839847, 1989.

Cole B, Finch E, Gawland C: Physical Outcorne Measures. Canadian

Physiotherapy Association. Toronto. 1 993.

de Bruin AF, de Witte LP, Stevens F, Diederiks JP: Sickness Impact Profile: the

state of the art of a generic fundional status measure. Soc Sci Med 35:

1 003-1 01 4, 1 992-

Doczi J & Renner A: Epidemiology of distal radius ftactures in Budapest. A

retrospective study of 2,241 cases in 1989. Acta Orthop Scand 65: 432-

433,1994.

Fleiss JL: The Design and Analysis of Clinical Expefïments. John Wiley and

Sons, Toronto. 1 W.

112

GaNand JJ 8 Weriey CW: Evaluation of healed Colles' fradures- J Bone Joint

Sug Am 33A: 895-9û71 1 951.

Hudak PL, Amadio PCl Bombardier Cl The Upper Mramity Collaborative

Group: Development of an upper extremity outcome measure: the Dash

(disabilities of aie am, shoufder and head (sic)). Am J lnd Med 29: 60%

908,1996.

Jaeschke R, Singer J, Guyatt GH: A cornparisan of severt-point and visual

analogue scales. Data from a randomized trial. Confrol Clin T M s i 1 : 43-

51,1990.

Jebson RH, Taylor N, Trieschmann RB, Trotter MJ, Howard LA; An objective and

standardizeâ test of hand function. Arch Phys Med RehabiI 50: 31 1-31 9,

1969.

Kirshner 6 8 Guyatt G: A methodological framewrk for assessing health

indices. Journal of Chtank Diseases 38: 2746,1985.

LaStayo PC 8 Wheeler DL: Reliability of passive wrist fiexion and extension

goniometnc measurements: a multicenter study. Phys Ther 74: 162-1 74,

1 994.

MacDermid JC: Development of a scale for patient rating of wrist pain and

disability. J Hand Therg: 17û-183, 1996.

Mathiowetz V, Weber K, Volland G, Kashman N: Reliabilty and validity of grip

and pinch strength evaluations. J Hand Sug Am SA: 222-226, 1984.

Norkin C 8 White DJ: Measurement of Joint Motion: A Guide to Goniometry. FA

Davis, Philadelphia. 1 995.

Nonisis MJ: SPSSPC+. SPSS Inc.Chicago,lllinois1990.

113

O\iuen RA, Melton LJ, Johnson Kq Ilstcup DM, Riggs BL: Incidence of Colles

M u r e in a North American community. Am J Pubk Healtf, 72: 607.

4 982.

Salant P & Dillman DA How to Conduct Your Own Survey. (1) John Wiley 8

Sons, New Yocù, NY. 1994.

Sarmiento A, Pratt GW, Berry NC, Sindair WFr Colles' fractures: Fundional

bracing in supination. J Bone Joint Surg Am 57A: 31 1-31 7,1973.

Shrout PE 8 Fleiss JL: lntraclass correlations: Uses in assessing rater reliability-

Psychol Bull 86: 420428,1979-

Streiner DL 8 Norman GR: Health Measurement scales. A practical guide to

their development and use. Odord University Press, O~ord. 1992.

Tarfov AR: The era of outcornes in health care. An ovewiew of the outcames

field. Medcaf Outwmes T i SourciePages 6-7,1996.

Ware JE, Snow KS. Kosinski M, Gandek B: SF-36 Health Sunrey Manual and

lnterpretation Guide. The Health Institute. New England Medical Center,

Boston. 1993.

Ware JE, Kosinski M, Keller SD: SF-36 Physical and Mental Health Summary

Scales: A User's Manual. (3rd) The Health Institute, New England Medical

Center, Boston, Mass. 1994-

Ware JE, Jr. 8 Sherboume CD: The MOS 36-item short-fom health survey (SF-

36). 1- Conceptual framewwk and item selection. Med Cam 30: 473483,

1 992-

World Health Organization: International Classification of Irnpaiments,

Disabilities and Handicaps- A Manual of Classification Relating to the

Consequences of Disease. 1980.

Table 1. Respondent Characten'stics (n=tW)

Trait -

University Affîl Ïaîion

Il lnvolved in Reseatch Studies . . - . -

11 lnvolved in Training Residenîs

Mean Years Sinœ Graduation From

Medicaf School --

Funding of Medical Practice

I Number of fradures treated per year

24 yean

Range 746 yeam

Private lnsurance 57%

Govemment SalaryRlMO 41 %

Private lnsurance 16%

Govemment Fee for service 9%

Worker's Compensation 3%

Other 8%

USA 60% Japan 12%

France 6%

Switzerland 5%

Spain, UY Netherlands, Finland

ZOh each

Canada, Gennany, Sweden,

Denmark, lndia l%each

Mean 54

27% treatad 0-25 per year

39% treated beîwen 2550 per

year

Table 2. Suwey Questions and Responses

1 . Which of the follow-ng syrnptoms do you routinely ask your patients with mist fractures about, in evaluation of the final result?

2. Which of the followïng objecüve measures do you routinely perfom, in evaluation of the patient's final status after a wrist fracture?

3- Do you have a standard time for Followup for al( wrïst fradure patients?

4. Which of the following personnel routinely participate in final evaluation of outcmme after Colles m u r e at your facility?

5. Is there an outcome measure (s) for mist fractures that satisfies your clinical needs? If 'yes" Iist

- 1 9

Pain 99% lnability to umrk 97%

Limitation in movement 92% lnability to perfwm recreaüonal activities

86% Overall satisfadion 77%

Weakness 10% Numbnesdparaeshesia 5%

Range of motion 99% Radiographs 97% Grip strengfh 85%

Sensary testing 40% Pain evaluation 40%

Patient satisfaction 21 % Standardized fundional tests 13%

Global outcome measures 6%

Yes 25% Most commonly at 1 year No:62%

Surgeon 83% Resident 54%

Ocaipational Therapist 28% Physiotherapist 9%

Hand Therapist 45% Nurse 6%

Fellows 5% Research Assistant 5%

No 62% Yes 15%

Don't know 20% Reported tmls in use:

Mayo 2.3 % O'Brien 2_3%

7 others by 1 % each

-on

6. Is there an outcome measure(s) satisfactory for researcti purposes?

7. Do you feel that research projeds and clinicat practiœ can use the same outcome measures?

8. Would you consider a new or different o u t m e measure in your cl inical practiœ

9. How important are the following characteristics to you in ternis of whether you would incorporate an outcome measure into your clinical pracu'ce?

Requirements af my time Requirements of patients time: The complexity of the measure: Papemrk required: Requirement of support personnel Whether it is useâ by other surgeons:

10. How important is it to include the following indicators in an outcome measure?

The surgeon's opinion Objedive measurements of movement Objedive measurements of strength Objective measurements of hand Function Patient evaluation of their Fundional ability Pain scales Patient satisfaction with treatrnent

No 3% Yes 10%

Don't know 13%

No 20% Yes 58%

Don't know 14%

No 1 jOh Yes 81 %

Don't know 2%

lmportanœ ranked as 'very", "somewtiat" or %or

Importance Somwhat

Not 20°h

0%

OOh

1%

1% 8% 2%

1 1 - Are you familiar with general health questionnaires sudi as the Sickness lmpad Profile or the Medical Outcorne Survey?

12. Do you think that hese general health questionnaires muid be usefùl in evatuating wrist fracture treatment?

13- If a mean M s t fradure outcame was based on objective measures- rank the following in ternis of importance

14- If a subjective outcorne scale was derived, rank the importance of the following items

, RF 1

No 80% Yes 20%

No 20% Yes 40%

Don't know 40%

Problerns wifh general heaith farrns: Tm long 33%

Not specific ?O mist 33%

Mean ranks (1 most important, 5 least important)

Fundional activity tests. 1 -8 Strength: 2.2

Range of motion: 2.7 Sensation: 3-7

X-ray appearance: 4.1

Mean ranks (1 most important, 5 least important)

Pain: 1-8 Limitation in ADL: 2- 1 Limitation in wric: 2-7

Limitation in rnovement: 3-9 Limitation in leisure 4- 1

Question

1 5- How important are the following barriers to nicorporaüng an outcame measure inta your clinical practice?

Pressure to see numerous patients Changing staff makes 1 dficult to maintain consistent clinlc procedures lnsuffcient support personnel lnsufkient opportunity to learn outcorne measure options Lacking of training of support personnel La& of data management capability

16. What is the maximum amount of time that patients should be expected to take to complete an outcome measure?

17- How much time should a treating physician be expected to commit to outcome evaluation at a final recheck visit of a Colles fracture patient?

1 8. How important are the following barriers to establishing an outcome measure that would be utilized by the majorïty of major centres treatïng Wnst fractures?

Difficulty attainïng agreement betwen clinicians Lack of communication betwieen centres Different needs by different clinicians Differences in practice at difFerent centres

lmmrtance

Somewhat Not

39% 46% 15%

Less than 15 minutes 54% 15-30 minutes 37%

more than 30 minutes 5%

5 minutes 8% 10 minutes 21% 1 5 minutes 41 O h

greater than 15 minutes 19% Mean 1 5 minutes

Immrtanœ Sornewhat Not

Percentages less than 5% are not teporteci; where numben add up to more than 1W%, it is because multiple responses were permitteci; d e n less Wan 100%, it is due to nonrespanse on that item

Table 3. Subject Characteristics for Reliability Study

Wtist Fracture Wrist Fracture in Treatment Post-treatment

W3(11. 2) 44.7 (1 0.2)

Months Post lnjury 4.5 (1. 3) 9.0 (5. 8)

Sex 1 1OM 18F 1 12M 24F

Average PWRE Scores*

Pain (150) 21 (13) 16 (13)

Overall Score UI 001 :M=male; F= fernate);' higher scores indicate more pain anc

Scaphoid Fracture

Treated for

disability

Table 4. Reliability Coefficients for The PRWE and its Subscales

Evaluaüon Item Wflst Fracture in Treatment

Wrist Fracture Post-treatment

Scaphoid Fracture Treated for Non-union

ICC ! 95% CI

Pain Scale

Specifîc Aciivit ies

- --

Usual Activities

Total Score relation menicient; SI= confidence inter

Table 5. Reliability SF46 Subscales mtesteâ afbr one year

Evaluation Item Scaphoid Fracture Treated for Non-union

II I icc 1 95% CI

[ ~ o d i l ~ Pain 1 0.67 1 0.474.80

i~eneral Health 1 0.50 1 0.24-0.70

~entsl Health 1 0.19 1 -0.11-0.46

(1 CC=l ntraclass &relation coefficient; CI= confidence

Table 6. Subjects f6r Constmd Validity Study

-

Months post Ïnjury

Sex

PRVVE Score'

Impairment score2

SF-36 Physical Sumrnary Score

SF-36 Mental Sumrnary - score3

Distal Radius

I 1 months

not 162(13)* 183(11)' tested

1. Score out of 100: &no pain or disability 2. Impairment score out of 85: 30 points AROM, 40 points gnp strength. 15 points dexterity 3. SF-36 Summary Scores are nonnalizd to the US population for a mean of 50 and standard deviation of 1 O ' ANOVA (p4.0000) and posthoc tests (pe0.05) for al1 pairim'se cornparisons statistically different

126

CHAPTER 4. REUABIUiY OF RANGE OF MOTlON MEASUREMENT IN THE

ELBOW AND FOREARM.

Introduction

Reliability, definecl as the consistency of a measurernent, becornes an

extremely important clinical issue when measurements of joint range are shared

among health care pruf8ssionaIs~ This is paftiwlarly true Men dechions regarding

a patient's physical, financial, social, and psychological well-being are based on

these measurements,

Measurements of elbow and forearm range of motion using standard

goniometer methads have b e n unreliable, in some circumstances, in Our clinical

practice. Range of motion measurements are frequently exchanged between

different dinicians. For example, residents report assessrnent findings and

physiotherapists provide progess reports to orthopaedic consultant surgeons as

a routine pradice in the clinic. These reports are valuable to the consultant, if they

provide the same information that he/she muld have obtained, if heishe had

perfomed the measurements. While a number of studies have evaluated the

i ntratester and intertester reliabil ity of el bow and foream goniometry, few have

used a standardized clinical rnethod on a group of patients with restricted elbow

and foream motion (Fish & Wingate 1985; Hellebrandt et al 1949; Rothstein et al

1 983).

The standard mechanical goniometer is the most frequently used

measurement tool to objectivelyquantify joint range of motion. This is probably due

to the fact that it is simple to use, inexpensive, portable, and durabie (Boone et al

1978; Fish & Wingate 1985; Hewitt 1928). A number of newer devices have been

A version of this chapter has been published in The Journal of Shoulder and Elbow Surgery 7:573-5ûû, 1998. Permission to reproduœ this text has been granted by Mosby-Year Book Inc.

developed to measure elbow and foreann motion, however, the comparative

reliability of these instruments to the standad goniometer has not been reported.

In this study we wmpared a universal goniometerviiith a computerized goniometer

and a mechanical rotation measuring device-

Three elements that may contribute to emx during ROM rneasurements

include the tester, the patient, and the instrument. The intent of this study was to

examine the extent to which these three elements impact on reliability. Specifically,

this study examinsd in(rat8ster reliability (stability oves tirne), intertester reliability

(stability between examinen) and interdevice reliability (stability between different

instruments).

Methods

Subjects

Sample size calculations indicated that thirty-ight patients were necessary

to detect greater than moderate reliability (ICC>0.60 (Landis & Koch 1977)) (a=

0.05 and 8-0.20; nuIl hypothesis ICC-0.60) (Donner & Eliasziw 19û7). Thirty-eight

former patients of one of the authon (GK) wre rewited by telephone by one of

the principal investigators who had no prior contact with the patients. The

volunteers had undergone a surgical procedure for an injury to the elbow, foream.

or M s t and were not wrrently reœiving physiotherapy as their range of motion had

plateaued (Table 1). Nineteen males and nineteen females agreed to participate

with an average age of 44 years (range 14 to 72 years). The average length of

tirne from their rnost recent surgical procedure was 17.5 months (range 7 to 81

months). (Table 1)

Instruments

The universal standard goniometer is a 180 degree protractor with one axis that joins h m amis (Smith and Nephew Rolyan Inc. , PO. Box 555. Menomonee

Falls, WI, 53051). One 25cm (from ais) a n îs stationary and the other 25 cm

(tom a i s ) ami is movable amund the axis or futmm of the protractor. Three

standard goniometers were available to the testen. This was done to replicate the

dinical situation Were a number of standard goniometen, are typically available in

the clinic. Testers were ftee to select any available standard goniometer to use

M e n perforrning their measurernents.

The NK Hand Assessmmt Laboratocy joint motion module (NK Biotechnical

Engineering Company, Mînneapoiis, MN) provides active and passive range of

motion data for all joints. The sensors are based on variable rotational resistanœ

corelated to the measurementof angularity. Thîrty cm (fmm goniometer amis

wre used in this study. Data collection was perfomied using standard computer

software and a foot so the examiner's hands were free for goniometrîc

alignments. For purpases of this study an independent assistant reairded

computer generated results so that testen were blinded to the ROM results.

The Super Pro Goniometer (Creative Medical Designs tnc., Tampa, FL)

measures solely supination and pronation of the foream. It consists of a central

vertical shaft (25 an from axis) the measuring cornpanent attached. Projeding

from the shaft is a free-rnoving, gravity dependant needle Hihidi indicates the

rotational amplitude on a fixed horizontal semicircular scale. The clinician places

the central shaf't across the ventral or dorsal aspect of the patients foreann and

measures the foream rotation from the measuring camponent.

Tester Preparation

Five testers ranging in professional experience measured each subject. The

testers included an orthopaedicsurgeon, tm experimced hand therapists, and two

inexperienced physical therapists. Pilot studies indicated that variations in IOcafing

landmarks was a signifiant saurce of test error. Therefore, the testers were given

specific directions that defined the extrmity position and the goniometnc alignment

for each motion of elbow flexion, elbow extension, foream pronation, and foream

supination. The followïng landmarks wem discussed and agreed upon among the

testers _

Elbow FlexionlExtension - a lateral placement wgs used Ath the goniometnc axh

aligned with the lateral epitmdyle of the humerus, the stationary a m was parallel

to the longitudinal axÏs of the humerus. pointhg t~~lllilrds the tip of the aaamion

process, and the movable a m was parallel to the longitudinal axis of the foream,

pointing towards the styloid process of the radius.

Forearm supination - the goniorneter was aligned on the media1 aspect of the distal

foream. one a m was perpendiwlar to the flwr, and the other moved to the Iimit

of supination by plachg it along the proximal wrist crease so that the goniometer

a m was in contact with the middle half of the Hinst

Forearm pronation - the goniometer was aligned on the lateral aspect of the distal

foream, one a m was perpendiarlar to the loor, and the other moved to the Iimit

of pronation by placing it along the mist just distal to the ulnar styloid so that the

goniometer a m ums in contact with the middle haif of the wrist

The testers discussed avoidance of changes M posture and use of 'Yrick"

movements. The teminology used in the present study has k e n describeci by the

Amencan Academy of Orthopaedic Surgeons (Amencan Academy of Orthopaedic

Surgeons 1 988).

The Testing Session

Seven to eight subjects were reauited for each testing session and w e

divided into three groups of to three subjeds. Each group was then assigned

to a testing room. The testers rotated in a randomized order to each room and

measured the elbwand/orforeami motion of the involved extremity of each subject

with the deviœ situated in that room. Subjeds m e rotated to three separate

rooms to be tested on each of the three devices-

The subjedo were positioned in a standardized manner. For

fiexionledension measwements the subjed sat Ri an amless chair adjusted so that

helshe sat in a cornfortable erect position with the trunk supported and the thighs

horizontal and feet flat on the ffoorffoor The shoulder, a m and tnink wre maintained

in a neutral position throughout testÏng. For wpination/pmnation measurements,

in addition to the above position, the elbow was flexed to ninety degrees, *th the

foream in midpootion (zero degrees) and the unist at neuhal.

The subjeds were asked to gently move their a m through their maximum

active range of motion. The subjeas w e instnicted, and ffequently reminded, to

make movements of maximal amplitude. Testers observeci the subjects for gross

changes in posture and for 'Ynck" movements. The tester measured the active

foream supinatian/pronation and elbaw flexion/extmsion using standardized

technique as previously describeci-

After the subjects were tested Ath each instrument by each of the testers in

the first round of measurements, a second round m s perforrned in the same

rnanner. A new randomized order was pmvided to testers for the second round of

test in^. Throughout the procedure, the subies had no knavvledge of their results

and efforts were made to maintain their interest during the 45 minute testing

periods. Each subjed was tested by five testers on each of three instruments

for a total of 30 measurements per subject per movement Individual data sheets

were recordeci for each measurernent made and placed irnmediately in an envelope

which was given directly to the data analysis personne(. Data w s racorded by the

testers for the universal goniorneter and Super Pro device and by an independent

study assistant for the NK goniometnc measurements.

Statistical Analysis

lntradass correlation COBffiaents (ICCs) w r e calailated for intratester and

intertester reliability using the Statistical Package for Soaal Sciences (version #4,

444N. Michigan Ave. , Chicago. Illinois, 6061 1) (Nomsis 1990) and methods

descrïbed for ICC(2,l) (Shrout 8 Fleiss t 979). lntradass correlation coefficients

were considered law if less than 0.40, moderate if greater than 0.40 and less than

0.75, and high ifgreater than 0.75. In addition, the average difkrenœ between the

tester's measurernents and the upper Iimit of 95% confidence interval of that

difference vms calculatecl to provide an estimate of the quantity of the intratester

and intertester measurement error. Instrument variability w s detennined by

cornparhg the measurements obtained by each examiner wing each of the three

different tools. The average amount of variation be-n measurernents obtained

by the same examiner Mile using the three different instruments indudes both the

amount of error related to the instrument and intratester variation in performing the

measurement The average ciifference of the rawmeasurements was calculated as

an indication as to whether there was any bias in the use of any instrument ie. did

one tool provide higher scores than anothef? Graphic representation of instrument

error was performed as desmbed by Bland and Altman (1 986).

Results

The intraclass condation coefficients for intratester reliability ware high for

active el bow flexiodextension and pronationlsupination with al1 three instruments

with 57/60 ICC exceeding 0.89 (Table 2). The average difference between trials

for each tester, the group average difference and the upper 95% confidence

intervals of these differences indicate similar amounts of intratester errot,

regardless of tester experience or tool used. On average. intratester eror did not

exceed five degrees for any of the measures, hawever, the 95% confidence

intervals for individual testers indicated mat emr wuld be as high as 22 degrees.

lntratester emir HWS higher for measures of foream rotation than for measures of

fiexion/extension. The vanability in measurements for each of the fnre examiners

was similar regardrass of the level of training. Despite the faa that NKgoniometric

rneasures were blinded and universal goniometrïc meawres were not, intratester

measurement~ssimilarforbothdevices~ Theintraclasscorrelation

coefficients for intertester reliability for fiexiorVext8nsion measurements were

moderate for the universal goniorneter and high for the NU goniameter ([CC's = 0.5û-û.87) (Tabie 3). Higher absolute enw in the universai goniometric

measurements is indicated by the higher average difference between testers as

compared to the NK computerïzeâ goniorneter. For supination and pronation al1

three instruments shawed high Ïntraclass correlation coefficients. (Table 3)

Analysis of the data suggests that the bias in the three instruments vms

relative1 y small but instrument relateâ measurement emr was substantial. The

absolute difference be-n successive rneasures performed with diffetent

instruments summed overall Ïndicates the arnount of measurement error (Figures

1,2 and Table 4 ). These data msistently show that instrument error is relatively

unimportant compared to measurement error. The 95% confidence intervals for

mean instrument difference indicate that measures taken with dwerent instruments

can Vary wïthin a wide range (ie over 30 degrees for most measures).

Discussion

Our results cleariy show that intratester reliability was high, regardless of

midi instrument ums used. This flnding agrees wïth past literature, confinning that

measurements taken by the same individual are reliaMe over time (Rothstein et al

1983). The average diKerenœ M i n the tester's measurements never exœeded

five degrees (range from 2.4 to 4.9 degrees) with eioier of the three devlces for any

joint angle measured. The upper limit of 95% confidence interval ranged fmrn 5.2

to 6.6 degrees for flexion/extension and 7.7 to 14. 3 degrees for

pronationlsupinatim when compaflng the Viree deviœs. The upper limit of 95%

confidence intewal of differenœ using the universal goniometer w r e 5.9 degrees

for flexion, 6.6 degrees for extension. 7.7 degrees for pronation, and 7.7 degrees

for supination. From this. one could extrapolate that, on average. motion changes

greater Vian 6 degrees for fiexion, 7 degrees for extension. and 8 degrees for

pronation or supination may be meaningful Men using the universal goniometer.

that is, unf ikely to be a result of vanation ~Mthin a tester's rneasurements. M e n

measurements are performed by the same tester with the same instrument

The 95% confidenœ intervals for measurement variation were lmwhen

rneasuring flexion and extension than when rneasuring supination and pronation.

This would suggest thatthe emcomponent of flexion or extension measurernents

about the elbow w r e 1-r than measurements of foream rotation- The

confidence intervals for the Super Pro goniometer ware particufarly high suggesting

that it may not be as precise an instrument,

It was noted that intratester variations w r e similar between testers with

different levels of expeflence. This muid suggest that level of experienœ does not

affect intratester reliability Men standardized methods are useâ to measure range

of motion. The critical issue in obtaining reliable results in this study may have

been the standardized meaiods useû by al1 testers.

Intertester reliabif ity for el bowflexion and extension was higher using the NK

cornputerized goniometer than when ushg the universal goniometer- The average

dïfferenœ between results reported by different testers and the 95% confidence

Iimit of this average were also less for the NK goniometer Men compared to the

universal goniometer. For example, our data illustrates that changes greater than

4 degrees for flexion and 6 degrees for extension. using the NK goniorneter, wwld

be meaningful, whereas 10 and 10 degrees respedively would be required if the

universal goniometer was used. The smaller value for the 95% confidence Iimit of

measurement e m between d i imn t testers, implies that the measurements made

with the computerized device were more repeatable-

An unexpeded finding wgs that reliability for elbow range was inconsistent

between the universaI goniometer and the NK goniometec A number of potential

causes can contrÏbute to rneasurement enor, although not al1 can be related to the

discrepancy behnieen the universal and computefïzed goniorneters. Error may be

attributed to variation in the tester. üie patient. andlor the instrument Identification

of bony landmarks, positioning and stabikaüon of the joint, and correct application

of the instrument have been suggested as sources oferror in the past Hellebrandt

et ai (1 949) noted that structural characteristics at certain joints make it inherently

difficult to apply the goniometer. Sinœ the same standardized technique was used

for both the NK and universal goniometer. it is unlikely that this could account for

the discrepancy in results.

Other components to be considered include the issue of individual tester

biases in reading and recording measurements This, obviowly. could be a source

of error for the universal goniometer, sinœ oie testers ware not blinded Men using

this instrument Wile they were for NK goniometer measurements. It is possible

that testers unknowingly bring into play a subjective bias t m r d s a certain

measurement Straford et a1 (1984) have been able to show that testers do tend

to read values that end with a particular digit Testers biases do not necessarily

contribute to error when the same tester repeats the measurement (intratester

reliability), however, they can play a signifcant role in explaining variations betwean

different testers. In cases vutiere test results are obtained by the same tester, a bias

may be repeated resulting in the tester agreeing with himsefflherself more

consistently. On the other hand, this wu ld tend to create less agreement between

testers. While we cannot define the adual reasons behind rneasurement ewors,

our results suggest that tester bias may be one factor in intertester measwement

error M e n using the universal goniometer to measure range of motion at the elbow.

Patient related sources of efror are another possible swrce af e m ~ but it is

unlikely that patient e m selectively varied betwaen the two instruments. The

subjects were not pmvided *th any incentives The environment was similar to the

clinic in that variables such as temperature and humidity were not controlfed.

These factors could have afFéct8d the subjeas musde strength and tissue

distensibility, resulting in differenœs in measured range between tests. Physical

characteristics could also amtribute to differences in measurements. The presenœ

of ederna and pain can Vary over time aven over the 45 minutes required to

complete the repeatecf measurements. As well the repeated movements

themselves could alter joint resistance and pain and aius change the range of

motion of Mich the subject was able to actively peHom. In addition, one cannot

be confident that the subjects put forth their best effort consistently with aach tester-

However, these sources of patient relateci mors are more likely to contribute ta the

total amount of random emr reported and are not likely to have aeated any

systematic differences beheen testers or instruments.

One source of error related to the use of the universal goniometer was

thought to be related to variation betwwn different models of the goniometer.

Three dïfFerent universal goniometers were utilizeci dunng the testing sessions as

these goniometers w r e typicall y used interchangeabiy in the ci inic. Differences

rnay have existed be-n these three devices such as in their degree markings

and stiffness. An examination of nine such universal goniometers demonstrateci

marking difrences of less than Mm degrees supporting that this played only a

minor role, if any, in amunting for the discrepancy in measurements. Different

instruments have dissimilar levels of stiftiess with HIftich the amis rotate around the

axis- DifFerent testers may use 'stif€' devices differently than 'loose' devices. They

may exert more or less force in positioning the goniometer amis to align the

anaiornical landmarks- This interaction behiifeen tester and individual universal

goniometers may have conûÏbuted to less reliable results wïth standard

gon iometers.

Our results differ from those demonstrated by Rothstein et al (1983). Their

study demonstrated a high intertester reliability (CC = 0.85 - 0.97) fw passive

elbow flexion and extension range of motion measurements using a standard

goniometer. THFBnty-four patients with upper extremity pathology were tested; the

examiners were blinded and the technique for measuring wms not standardized. A

selection bias for subjects in this study may have contributeci to high reliability. The

testing therapist identifid patients vrtio they felt ware "suitable" for the study during

a regularly schedoled treatment session. Seledion of particular may have

reduced the patient related source of enor in the measurement proces.

Furthemore, as the movements were passive, the results are not difectly

comparable to those of active motion. Adive movements tend to be used more

frequently in the clinic and Vieoretically are more directly related to fundionai

a bil ity. The testers in Rothstein's study were blinded and reported results with high

intertester reliability. This is consistent with our study, where vue also found high

intertester reliability M e n the blinded testers used the eledrogoniometer, whereas

we only had moderate intertester reliability Men unblinded testers us& the

universal goniometer,

Godwin et al (1 992) found that significant differenœs existed among testers

M e n measuring active elbow range of motion, a standard goniometer, on

Wnty-three healthy female individuals. This was a partiwlarly surprising finding

since the landmarks for measuring were actually marked on the skin of the subjeds-

The faa that this was done, makes this study dissknilar Rom the clinical situation.

In addition, healthy subjects wwe examïned which does not replicate the clinical

situation-

Statistical methods may also have contributed to the difkrent conclusions

obtained in recent studies. The ICC is a ratio of intersubject varianœ divided by

total variance (intrasubject + intersubject variance). Intrasubject varianœ refiects

error and intersubject varianœ reflects population variabilit. Thus the high

intersubject variation that tends to occurs in patient groups favours obtaining a high

ICC. Lmer vanation is antidpated in nomals (partiailariy for charaderistics which

plateau such as for ROM). The abgolute ernx for individual measurements taken

on groups of patients or 'nomals" may be similar, although the inter-subject

variation is more variable. For this teasan, one must be caotious about making

decisions on reliability of measurements based on [CC's alone. We campensated

for this statistical property by presenüng both the CC, the average diMarences, and

95% confidence intervals be- measurements to provide a mare cornpiete

representatim af measurement e m ~ . For example, aithough intertester [CC's for

elbaw flexÏon and supination, as measured by the NK goniorneter, are both very

high, the absolute value of the rneasurement errot refleded in the average

difference behNeen testers is twiœ as large for supination as it is for flexion.

High intertester reliability COBfficients were obsewed for pronation and

supination using all three deviœs. Gravity andlor the perpendicular axis to the flwr

were utilized to establish the position of the stationary a m of the goniometen.

Therefore, only one landmark really had to be identifieci by the testers for placement

of the movable am. Furtheme, this land marking required the tester to place the

goniometer directly on top of the landmark rather than aligning wïth a distant

landmark as occurs in lateral goniometer placements. Identification of only one

landmark may contribute to aie high reliability found between the testers in this

study, since landmark detemination has been implicated as a possible source of

error that could reduce reliability (Hellebrandt et al 1949). It should be noted,

however, that the average difference beîwen testers' measurements ranged from

6.6 to 8.5 degrees for pronation and 6.2 to 123 degrees for supination. The upper

limit of 95% confidence intewal of differenœ ranged ftom 7 2 to 10.0 degrees for

pronation and 7.1 and 14.5 degrees for supination. The upper limit of 95%

confidence intewal of differenœ was 9.5 degrees for pronation and 1 1.2 degrees

for supination Men using the universal goniometer. This would suggest that range

of motion changes greater than 10 degrees for pronation and 11 degrees for

supination are rneaningfbl Men mmpaflng measurements between testers using

the universal goniometer-

Reliability of a masurement device cannot ensure valida* H~~lll~ver* if low

interdevice variability c m be established, then this can alsa be thought of as

concurrent validity for the measurement devices- If similar resuits are obtained with

different instruments, this supports the likelihaod that the measurements are valid

indicaiors of the trait being measured; because it is more unlikely that three invalid

instruments viiould al1 measure range of motion the same -yY Figures 1 and 2 and

Table 4 demonstrate that al1 three instruments producd similar results. This

supports the use of goniometry as an indicator of joint movernent capacity-

In summary, it is evident that a number of factors need to be considered

M e n measurïng joint range of motion. Standardized techniques for rneasuring

range of motion of the elbowand foream shouid allow even inexperienced testers

to produce consistent resufts. It is neœssary for dinicians to describe the

instruments and test procedures used when reporting range of motion results in

clinical studies and to try to maintain cansistency in these aspects within their

clinical practice. Measurement error is srnall when the same tester repeats a

measurement Ath the same device and is larger when either the instrument or

tester is different-

Conclusions

1. Inbatester reliabilitywas high for measurements ofelbowand foream movement

with three differenttypes of goniometn'cdeviœs. Range of motion changes greater

than 6 degrees f i flexion, 7 degrees for extension, and 8 degrees for pronation or

supination Hlere thought to be meaninghil Men measurements were taken by the

same observer using the universal gonioîneter.

2. Level of experienœ did not seem to affect reliability results Men individuals

were provided with details on landmarks to be incorparateci in their measuflng

technique-

3. lntertester reliability was high for flexion and extension when using the NK

goniometer but moderate when using the universal goniometer. The NK

goniometer being a mmputerized deviœ may have more inherent consistency and

loww examiner bias due ta blinded recording of results. Range of motion changes

greater than 4 degrees for fiexion and 6 degrees for extension w e thought to be

meaningful between testers M e n using the NU goniometer compared to 1 O and 10

degrees respectively if the universal goniometer was used.

4. lntertester reliability was high for pronation and supination using al1 three

devices. Range of motion changes greater than 10 degrees for pronation and 1 1

degrees for supination were thougM to be meaningful between dietent testers

when using the universal goniometer. The average differenœ between testers and

the 95% confidence intervals w r e higher when compared to ffexionlextension

results. This suggests that absolute values of errorfor measunng rotation aboutthe

foream is greater than that seen for measunng fiexion or extension about the

elbow.

5. Instrument bias is a minor problem in measunng range of motion, however inter-

instrument measurement errors can be large and should be avoided.

Amencan Academy of Orthapaedic Surgeons: Joint Motion: Method of

MeasurÏng and Recording. Churchhill Livingstone, New York 1988.

Bland JM & Altman DJ: Statistical methods for assessing agreement betwesn

two methods of dinical measurement L a m f q: 307310,1986.

Boone DC. Azen SP, Lin CM, Spence D, Baron C. Lee Lr Reliability of

goniometnc measurements. Phys mer Sû: 1355-1 390.1978-

Donner A & Eliasziw M: Sample size requirements for reliability studies. Stat

Med 6: 441448,1987.

Fish DR 8 Wingate ML: Sources of goniometric e m at the elbow Phys Ther

65: i 666-1 670,1985-

Goodwin J, Clark C. Deakes J, Burdon D, Lawrenca C: Clinical methods of

goniometry: a comparative study. Disabii Rehabil14: 10-1 5, 1992.

Hellebrandt FA, Duvall EN. Moore ML: The rneasurement of joint motion: Part III

Reliability of goniometry. Phys Ther Rev 29: 302307, 1 949.

Hewitt D: The range of actîve motion of the wrïst in wmen. J Bone Joint Surg

26: 775-787, 1928.

Landis JR 8 Koch GG: The measurement of observer agreement for categorical

data- BIomefniCs 33: 159-1 74,1977.

Norusis MJ: SPSSIPC+. SPSS Inc. Chicago. Illinois, 1990.

Rothstein JM, Miller PJ, ROeffger RF: Goniometnc Reliability in a Clinical

Setthgr Elbow and Kiwe Measurements. Pnys -63: 161 1-1 61 5,

1 983.

Shrout PE 8 Fleiss JL: lntradass correlations: Uses in assessing rater reliability.

Psychof Bull û6: 420-428.1979-

Strafford P, Agostino V, Brazeau C, Gowitzke B: Reliability of joint angle

measurement- a discussion of methodology issues. Physribrnr Can 36:

5-9, 1984-

Table 1. Subjects

1 58 M ORlF distal ulna f gaff 2 65 F ORIF R radius and ulna 3 66 M ankylosis takedown R elbow 4 30 M ORlF L uha, hardware out 5 40 F ORlF 1 elbow 6 40 M ORlF L distal radius I 46 F bercm b in, exfk L disfa1 radius

10 66 M open biopsy L etbow 11 49 M ORlF R elbow 12 14 M ORlFLelbow 13 24 M ankylosis take down L elbaw 14 32 F L elbow capsuledomy, hardware

1 15 1 35 1 M 1 ulnarnerve transposition,

1 debridement R elbow 16 48 1 M ORlF Rdistal humerus

1 17 1 69 1 F 1 ORIF, ex-fuc L distal radius 1 RSD 18 47 F L distal radius osteotomy I graft 19 31 F R elbow debridernent 20 55 F ORlF L distal radius I graft 21 51 F R radial head excision 22 68 F ORIF, ex-fix, L distal radius 23 49 M ORiF Lolecranon graft 24 40 M resection of L radial head 1

1 1 1 1 debridement L

- -

25 69 F ORlF L 8 R distal radius, L distal

uina 1 graft 26 29 F excision R radial head. prosthesis 27 s/ M ORlF R radius& ulna

'

28 46 F ORlF L distal radius, damach

proœdure 29 66 F OR1 F L distal radius. ex-fx I sraft

30 31 32 33 34 35

ORIF= Open reduction intemal fmion; L=M; R=rÏght

36 37 38

24 50 59 44 26 75

64 65 61

M F M M M F

F M F

ORIF Lradius L radial head implant excision of extosis, L elbow ORlF R distal radius, ex-fix I gr* ORlF R scaphoid I gr& ORlF right distal radius, ex-fix I

23 7 24 21 21 27

gr* ' revision L radial head implant ORlF L proximal ulna ORlF R distal radius

14 17 7 i

Table 2. Intramter Reliability

ation W w t d t r El 5-0 12-0 0-96

Table 3. Inter-rater Reliability

Ubow Flexion

Pronaüon

C k r i u ~ ~ d

Gœakma%~

AIK -

Supina- tion

1

2

1 2

UMmrsd Oovriarinliov

MK mmhm&r

Su- Pt0

miimsa

GonimMaf

NK Ganiome($r

Super Pt0 Gmi-

7-7 1 0.9 I I 1 G01)i- I

9-5 8.3

7.9 8.2

6-2

6-5

3.2 3.5

2 i 6-7 I

0.83 I

0.86 ,

I

0.98 I

0.86

I

1 1 8.4

1

2

1 2

9.8 f 0-72

2

1 2

1

7-3

6.8 7 2

8-5

6-9

9-4

6-3 6.2

-

8-9

9.5

3.9 4-1

0-62

0-58 ,

0-95 I

0-94 .

7.8

112

7.2

1 2

0-9

0.93

0-91

7.5 14-5

6-4 12.3

0-86 l

0.92

7-1 r 0.91

Table 4. Average Diffiiuces in Scores (summd ova irtm and oecwions) Between DInerent Insarments During M n m e n t of Po- Ratrtion

Cl= confidence interval

Figure 2. Instrument Variability in Extension Scores

-20 -10 O 10 20 30 40 50 60

Average Extension ROM Score

Figure 2. Instrument Varïability in Extension Swres. The sdid line represents the average difference between masures taken müi the two goniometers (Universal - NK).The dashed line represents the range of measurement enor with 95% confidence.

CHAPTER 5: TEST-RETEST REUABIUTY OF STATiC AND ENDURANCE

GRfP STRENGTH TESTS PERFORMED ON THE JAMAR AND NK DEVICES

Introduction

Outcorne measures which are reliable. valid and responsive are critical in

health evaluation when measuring impairnent disability and handicap. Grip

strength is one of the most commonly assessed measures of impairment It has

been used as an indicator of overall health, aging effeds (Pentland et al 1995)

pre-placement evaluation in irtdustry, (Pattemon 1965) athletic capability,

disa bility ratings (Stokes 1 983), diagnosis (Czitrom 8 Lister 1 98û) as well as to

monitor response to medical intementions (Forbes 8 Hopper 1990; Sheehan et

al 1983; Heyward et al 1975). Standard methods for statichsometric grip

strength testing have been dem*bed (Fess 1 992) which are reliable for both

nonnals (Bilodeau et al 1994; MacDemiid et al 1994; Balogun et al 1990;

Montipetit et al 1965; Strafford et al 1989; Coleman et al 1996; Harth & Vetter

1994; Hamilton et al 1994; Niebuhr et a1 1994; Hamilton et al f 992; Wadswortti

et al 1992; Trossman et al 1990) and patient groups (MacDemid et al 1994).

Methods have been established fo detemine if patients' efforts are sincere in

providing maximum gr@ efforts (Nieburh 8 Marion 1987; Gilbert 8 Knowlton

1 983; Stokes et al 1995; Smith et al 1989; Nieburh 8 Marion 1987). Normative

values have been established (Mathiawetz et al 1 985; Balogun et al 1 990;

Desrosiers et al 1995; Gilbertson 8 Barber Lomax 1994) and the recovery of gr@

strength &ter interventions is commonly reported in a number of outcome

studies.

SMic gr@ strength scores represent the maximum isometric grip

capability of a patient. This is assumed to be related to the fundional status of

the patient Endurance is the ability to sustain adivity b resist fatigue.

Theoretically, it is refated to fundion as it represents the capability to pefiorm

repetitive adivity. For this reason, grip endurance may represent a more direct

measure of fundional ability than static grip skmgth. If this were the case, it

would be a useful outcome measure for dinical evaluation a€ patients Ath upper

extremity pathology. The comrnon usage of static grip strength samw as dinical

indicators may reRed the fact that equipment and standardized protocols are

readil y available for mis type of testing-

There has been some interest in grip endurance measurement and the

relationship between static and dynamic musde adivity (Mundale 1970; Nmiga

1 975; Byrd 1982; Wolf et a1 1996; Robertson et al 1 996; Mundale 1970). Grip

enduranœ has ben related to wwk capaaty (Woif et al 1996)- However a

standardized definition of how enduranœ should be rneasured or a reliable

protowl that could be adopted by dinicians has not been desaibed.

Endurance can be msasured in a variety of ways. Contractions can be

isotonie, isometric or isokinetic or combinations of these three basic types.

Contraction times, percentage workload, rest intervals between contractions,

total test time and endurance indicators can Vary. For example, Hiwk done

(force X distance) and power (force per unit of tirne) and wwk duration (time

work is maintained) hava been discussed as criterion measutes (MacDermid 8

Fish 1991 ; Wolf et al 1996; Robertson et al 1996). Furthemore, the metabolic

pathways that Iimit endurance can Vary depending on the type of muscle adïvity

that is perfomed. High intansity contractions perforrned over a short durath

may fatigue due to metabolic inhibition of contradion, whereas during longer

duration exercise performed at a relatively lm perœntage of maximum capacity,

there may be a long-lasting nonmetabolic component that a d s at the level of

excitationcontradion coupling (Baker et al 1993). It has been demonstrated

that at contraction leveis greater than 10% of maximum capability, blood flow is

insufficient to maintain homeostasis (Sjogaard et al 1988).

Woif et al (1996) looked at the loss of isometn'c gtip strengoi followïng

isotonic wwkloads of 75%, 50% and 25% of maximum isotonic capacity- They

found that the best predidor of recavery tirne ~ w s the extent of deficit in grip

strength f o l l ~ n g the activity. They found no significant relationship behiiieen

grip strength and duration dwork, although total work done was related ta

isometric strength at the 75% wwkload (19% of the variance in wwk done

explained by isometric grip strmgth) and 50% workload (1096 varience

explained). This is explained by the fact that wrk done incorporates the force

(of individual contractions) and the distance (detennined by the number of

repetitions). These authors did not report the reliability of the enduranœ test

protocol.

Robertson et al (1996) evaluated endurance using the Strength

Decrement Index defined by Clarke (1955). This index is expressed as [(initial

strength - final strength)IInitial strength] X 100%. A standardized test procedure

in ternis of pacing, contraction cycles and test duration is important for this type

of endurance pmtocol. Robertson et al (1 996) described an isometric test

protocol perfomed on a modified power grip attachment conneded to a digital

rnonitor and an analogue recorder. Subjects altemately gripped for 1.25 seconds

and relaxed for 1.25 seconds for 24 cycles (total test time one minute). They

reported high correlations be-n test and retest scores for the initial strength

(trials 1-3) and the final strength (trials 22-24) that were used to cornpute the

Strength Decrement Index They also demonstrated that the Strength Deaement

Index was significantly different b e m n injured patients and uninjured subjects.

Grïp endurance was defined as the time that a sustained force of 70% of

maximum grip couid be maintaineci, in a study evaluating the efféct of sustained

rock climbing on grip strength and enduranœ (Watts et al 1996). Again, the

reliability of the endurance protocol was not reported.

The exercise physiology literature tends to report studies of muscle

enduranœ fmrn the perspective of improving air understanding of muscle

physiology. Clinical authors tend to select endurance test protocols that simulate

work situations (Kennedy 8 Bhernbhani 1991). predict work capacity/recovery

(Wolf et al 1996; Robertson et al 1996) or to test metabolic pathrays that may

be affected by disease (Cheung et al 1991)- Thus a varïety of endurance

protocols have been used to assess muscle endurance-

Despite interest in fatigue patterns for isometric strength in exercise

physiology literature (Kroll1968; KFoll1971) and research on grip endwanœ

specialized equipment (Robertson et al 19961, measurement of gn'p

endurance has not becorne a commonly used dinical procedure. This may be

partially due to la& of defined test protocols with readily available dinical t a s -

Isometric grip devices such as the Jamar dynamometer have been used in the

clinic primaflly to detect maximum grip strength. It is technically possible to use

these devices to test wwk duration or a Strength Decrement Index H-ver,

the reliability of this application has not been investigated.

The ment commercial availability of cornputer-based hand evaluation

systems Mich incorporate endurance measurement protods has enhanced

interest in enduranœ as an outcorne measure- Endurance can be measured

over a sustained gr@ Moi real time assement of grip farce throughout the

contraction. The reliability of these test protocols has nut been established.

Furthemore, while the reliability of static measures taken with the Jamar device

has been well established (MaCDennid et al 1994; Harkonen et al 1993; Niebuhr

et al 1 994; Hamilton et al 1 992) the reliability of static measures performed on

the newer camputerized devices has not yet been reported. There is a need to

determine if grip endurance can be measured reliably Mth currentfy avaiiable

equipment and if static measures taken with the recently introduœd NKHAS qip

device provide reliability similar to that attained with the more traditional Jamar

device.

Conversely, the need to rneasure grip endurance muid be obviated if

grip endurance was highly related to static gnp strength measures ie no

additional infornation mnrld be provided if the hiiib measures are highly

correlated. Therefore, the purposes of this study were ta: detemine the

reliability of measurement pmtocols on the traditional deviœ (Jamar) and the

more recently developed device (NK). The measurement protocols tested were:

1 ) a static isometrk grip and 2) three grip endurance protocols. A further

purpose was to determina the relationship be-n grip enduranœ and static

grïp strength-

Methods

Forty subjeds any pathology in the upper extrmity (80% Right-

hand dominant, 52% female, average age 30, Standard deviation=IO ) were

tested for static grïp and enduranœ protocols on two occasions within one week

The testing was perfomied using two devices: the Jamar hydraulic hand

dynamometer and the N-K hand assessrnent system (NKHAS) digigrïp device5

which measures grip strength using a straingauge systern.

Static grip strength w s measured according to the measurement protacol

recommended by the Amencan Society of Hand Therapists (ASHT) (Fess 1992)

which recommends standard positioning and averaging of three repetitions.

Testing w s perfomed using the second handle position on both measurement

devices.

Endurance protocols were of basic types. The 5- second-onl5-

second- ofF grip protocol was tested using both the Jamar and the NKHAS. This

test involved performing a fivesewnd gtip immediately fol lowed by a 5-second

- - - - - - - pp

4

Jamar, TEC, 60 Page Road, ClÏfton, New Jersey

'NK Biotechnical Engineerhg Company, 10850 Otd County Road, Minneapolis. MN

rest period. This grip sequence nias repeated until subjeds wwe unable to reach

a speafic percentage of their static grip strength. Two specific percentage -off

points wre evaluatedr 80% and 70%. The aiterion measure for this test ~ w s

the number of repetitions the subject m e able to perfonn before falling below

this percentage. When wing the Jamar to paifom the test a stopwatch w s

used for timing. With the NKHAç the timing was pertdnned by the cornputer-

The second endurance protocol wes a computerized evaluation af fague

during a l Gsewnd grip pfimmd only on the NKHAS- Measures of rise time,

response t h , peak focce. fatigue and capacify m e cornputecl by the NKHAS.

(See Figure 1 )

Subjects were testeâ within a 14day period. Order of hand tested

and instn~ment-used w r e randomized for each test session,

Descriptive statistics cansisting of means and standard deviations were calwlated using SPSS statistical sofhnwre(Norwis 1990). Test-retest reliability

of each of the criterion rneasures MS evaluated by wmputing intraclass

correlation coefficients (ICCs) Type (2.1) and their 95% confidence intervals

(S hrout 8 Fleiss 1 979). The follkng subjective classification of reliability

coefficients was used: c 0.40 poor, 0.40-0-75 fair, >0.75 excellent (Fleiss

1986a). Pearson correlations m e used to determine the association between

grip strength and endurance measurements.

Results

All subjects completed the testing protocd. Some subjads reported mild

discornfort at the handle site of contact wïth repetitive gripping- This w s

reported more frequently the Jamar device. The beeper for the NK Hand

evaiuation systern w s too quiet for some subjeds to hear clearty- Thus. the

examiner had to state '&art" at the sound of the beep to ensure that the trial

began when required.

Reliability statistics indicated that static grip reliability is excellent (Fleiss

1986a) as measured with eeither the NKHAS of the Jamar (Table 1 ).

Rie repetitive gripping protocol results are in Table 2. Reliability was

poor to fair for this protocol. Using 80% as a cuioff resulted in lower reliability

with 3/4 of CC's king P r . Reliability staüstics tended to be better for the 70%

cutoff and were higher for the NKHAS man for the Jamar, There was wide

variability on the number of repetitions performed betHRen different subjects,

even though cutufF points were calwlated as a parcentage of the subject's own

static capability.

The 1 O-second grip protocol on the NKHAS demonstrated that peak force or capacity measures had excellent reliability. Measures of rise time, response

tirne and fatigue had poor reliability (see Table 3).

Measures of static grip strength wre strongly as-ated with peak force

or overall capacity during a 1 @second contraction, but not related to fatigue

measures (see Table 4).

Discussion

Static grip strength measufes have been demonstrated to be highly

reliable in a variety of circumstances (MacDennid et al 1994; Balogun et al 1990;

Montipetit et al 1965; Stratford et al 1989; Coleman et al 1 996; Harth 8 Vetter

1994; Hamilton et al 1994; Niebuhr et al 1994; Hamilton et al 1992; Wadsworth

et al 1992; Trossman et al 1990). AIViough grip strengih has been measured

with a variety of devices (Balogun et al 1990; Fike 8 Rousseau 1986; Anderson

et al 1990; Balogun et al 1991 ; Heyward et al 1975; Coleman et al 1996;

Stephens et al 1996; Hamilton et al 1992), the most common clinical practiœ

pattern is for the modified sphygmomanometer to be used for arthritic patients

and the Jamar to be used for patients *th other hand pathologies. This study

adds to previous research in that it demonstrates excellent reliability for static

grip using a wmputerized grip device (NKHAS). The grip device and the

measunng systems (strain-gauge venus hydrwlic) are dR8tent between the

Jamar and NK grip device. These diff'ences did not have any appreciable

eRed on reliability of scores obtained. Future studies should determine

whether the scores are similar betwwn the twb instn~ments at ail five handle

positions and thw determine Wmther commonly used grip strength noms,

estabfished on the Jamar, are suitable for the NK device as mil. If not, new normative studies on the NK ûevice are requimd-

While endurance measures had a strong theoretical reiatianship to

physical fundion, this study did not dernonstrate a reliable protocol for

measurernent of grip endurance. Not only wre the test-retest reliability results

indicative of poar reliability on a day to day basis, the variability behiuaen normal

subjeds was quite high indicating that it might be dimcult to establish normative

values within a range narrow enough to be clinically rneaningful. This indicates

that grip endurance, at present, shoutd not be considerad a useful impairment

rneasure.

This study evaluated two diirent types of endurance protocols, criterion

rneasures and measurernent devices. The cyclic o n 4 protoc01 which evaluated

the duration for which a minimum level of force could be maintained, may

simulate certain patterns of adivity ie wk on an assembly line. The 1 &second

grip for Mich both wwk capcity and fatigue rate were evaluated is a

measurernent protocol for short-terni fatigue created by cornputer based hand

evaluation systerns. The relative metits of different endurance protocols can be

debated. However, neither the number of individual grip repetitions produced

above the predetermineû cutofF point nor the rate of fatigue calwlated by the

NK from a 1 O-second gnp were reliable in a retest situation. Thus, at present,

grip endurance cannot be relied upon to provide useful information on patient

status or respanse to treatment.

Single static grip rmtions have b e n sh- to be reliable Wen a rest

is provided be- repetitions (Macûmid et al 1994). The observation that

fatigue measures, Hihich use isolated gnp repetition SCOT= as cutoff points are

not reliable suggests that cyclic repetitive gripping may exhibit more inter-tnial

variation than static testing. Assessrnant of fatigue may require a sampling

frame composed of more data points to provide stable results. This approach

would require development of a gripfatigue protocol with established

parameters for timing of repetitions that wwtd be feasible ta perfonn mlli both

patients and comparative subjeas without pathology. The wwk done using a

one minute enduranœ test a custmized gripping deviœ found the average

of first three or the last three grips to have high retest reliability (Robertson et al

1996). This suggest that the Strength Deaement Index approach to evaluatnig

endurance may have merit- Investigation of this protoc01 using the gr@ deviœ

commonl y found in the dinic is advisable.

Capacity measures from the NK device were reliable suggesting that it

may be also possible to devise more stable endurance measures by relating

early 'capacitÿ to late 'capacity". For example, wwk done over the first three

seconds of a 1 O-second grip trial wuld be compareci to wwk done in the final

three seconds of gnpping- Studies to document reliability of the any newly

devised enduranœ protocol muid be required.

One modifiable factor on the NK System 1 O-second grip tests was the

poor audibility of the start 'beep". If this problem w r e correcteci, riw times

might become more consistent

Endurance measures represent ability to sustain a repetitive grip force.

While theory suggests that endurance is an important component of ability to

maintain sustained actkty, such as is required to perforrn tasks of daily living or

occupational activities, the role of grip enduranœ as a predidor of wwk

performance has not been clearly established. There has been some suggestion

that power (static strength) and endurance are inversely related (Chatterjee 8

Chowdhuri 1991 ; Barnes l98O), at least in fernales (Nmiga 1 975), although

others have fwnd no correlation between enduranœ and isametfÏc grip (Nwuga

1975; Robertson et al 1996). This study also fbmd that rate of fatigue cannot be

predicted from static grip strBngfh, Hmver , attenuated correlations are a

known problem wïth unreliabfe measurement (Fleiss 19û6b). Thus, it is unclear

whether the lad< of correlation Mumen gflp strength and endurance found in

this study and others refiects a tnie nidependence of these measures or is a

result of measurement em Although the theoretical concept that grip

endurance capacity relates to occupational upper extremity use remains

unproven, Iittle pqress can be anticipated in this regard without a reliable

measurement protocoi. Future studies faaissing on altemative endurance

protocols are required to establish a test th& will remain stable over time (in the

absence of real change) and be able to deted (me) dinical change. Definition

of the relationship be-n static and endurance grip strength scores to overall

hand and occupational funaion and responsiveness of these rneasures to

ciinical intervention wiil provide insight into the utility of these measures.

Conclusions

Static grip can be measured reliably with either the NKWS or the Jamar.

Neither the nurnber of repetitive grip cycles perfomied at a perœntage of

maximum grip strength nor gr@ fatigue rates during a 10 second sustained gfip

were reliable as measured in this partiwlar study. The rate of fatigue of grip

strength mnnot be predided frorn static grip strength scores.

Anderson PA, Chanoski CE, ûevan DL, McMahm BL, Whelan EP: Nonnative

study of grïp and wrïst flexion strmgth employing a BTE Work Stimulatorr

J Hand Surg Am 1SA: 420-425,1990.

Baker DJ, Kostw KG, Miller RG, Weiner MW Slow force recovery mer long

duration exercises: metabolic and activation factors Ri muscle fatigue. J

Appl P h w 71: 2294-2300,1993-

Balogun JA, Adenlola KA, Akinloye KA: Gnp strength normative data for the

Harpenden dynamometer. J Otfhop Sporfs Phys Ther 14: 155-160, 1991.

Balogun JA, Akomolafe CT, Amusa LO: Repmducibility and cfiterion-related

validity of the modified sphygmmanometer for isometric testing of grip

strength. Physrother Can 42: 2-295,1990.

Bames WS: The relationship betwiaen maximum isokinetic strength and

isokinetic endurance. Res Quarf 5 t : 7 14-71 7,1980.

Bilodeau M, Arsenault AB, Gravef D. Bourbonnais D: EMG power spedrum of

el bow extensors: a reliability study- Dedromyagr Clin NeumphysriaI34:

149-1 58, 1994-

Byrd R: Effed of maximal grip strength and initial gnp strength on contration time

and on areas under force - time cuwes during isometnc contractions.

E~onornrcS 25: 387492,1982.

Chatterjee S 8 Chowdhuri BJ: Cornpanson of grip strength and isomerïc

endurance between aie rQht and let% hands of men and their relationship

with age and other physical parameters. J Hum Ergd (Tokyo) a: 41 50,

1991 -

Cheung PY Arnold JMO, Marsh G, Thompsan RT, Driedger AMJ, MacDermïd

JC: Training improves metabolism Ri skeletal muscle in heart failwe. J

Am CollCatdid17: 159a. 1991-

Clarke HHI Shay CT, Matthews DK Strength deuement index A new test uf

mwailar fatique. Amh Pnys Med Rehabii 36: 376, 1955.

Coleman EF, Renfio RR, Cetinok EM, Fess EE, Shaar CJ, Dunipace KR-

Reliability of the manual dynamic mode of the Baltimore Therapeutic

Equipment Work Simulator. J Hand Ther 9: 223-237.1996-

Czitrom AA & Lister GD: Measurement of grip strength in the diagnosis of wrist

pain. J Hand Surg Am 13A: 16-1 9,1986.

Desrosiers J, Bravo G. Hebert R, Dutil E: Normative data for grip strength of

elderly men and women. Am J Oocup Ther49: 637444, 1995.

Fess EE: Grip strength. In Casanova JS: Clinical Assessrnent Recommendations

(2nd) Arnerican Society uf Hand Therapists, Chicago. 1992.

Fike ML 8 Rousseau E: Measurement af adult hand strength: A cornparison of

two instruments. Oocup Ther J Res 2: 4349, 1986.

Fleiss JL: Reliabil ity of Measurement In Fleiss JL: The Design and Analysis of

Clinical Expeflments John Wiley and Son, Toronto. 1986b.

Fleiss JL: The Design and Analysis of Clinical Experiments. John Wiley and

Sons, Toronto, l986a-

Forbes A & Hopper Di The Med of counterforce bracing an grip stfength in

tennis players wïth painful elbows. Aust J PhjmWbr36: 259266, 1990.

Gilbert JC 8 Knowiton RG: Simple Method to Determine Sincereity of Effort

During a Maxim1 Isometnc Test of Grip Stfength. Am J Phys Med 62:

t 35-t 44,1983,

Gilbertson L & Barber Lomax SI Power and pinch gnp strength recorded using

the hand-held JAMAR dynamometer and B+L hydraulic phch gauge:

British normative data for adults. Br J Ocwp W S 7 : 48348. 1994.

Hamilton A, Balnave R, Adams R: Grip strength testing reliability- J Hand Ther

7: 163-i70,1994-

Hamilton GF, Mcûonald Cl Chenier TC: Measwement of g ip strength: validity

and reliability of the sphygmomanometer and Jamar grÏp dynamometer. J

Orfhop S ~ M S Phys Ther 16: 21 5-21 9,1992.

Harkonen R, Harju R, Alaranta H: Acairacy of the Jamar dynamometer. J Hand

Ther 6: 259-262,1993.

Harth A 8 Vetter WR: Grip and pinch strength arnong selected adult

occupational gmups. Occup Ther lnt 1 : 13-28, f 994.

Heyward V, McKeo~n1 B, Geeseman R: Cornparison of the Stoelting hand grip

dynamometer and Iinear voltage difrential transformer for measuring

maximal grip strength. Res Quart 46: 262-266,1975.

Kennedy LE & 8hambhani YN: The Balthore Thefapeutic Equipment Work

Simulator. Reliability and Validity at Three Work Intensities. Arch Phys

Med Rehabil72: 51 1-51 6, 1991.

163

Kroll W: lsometric fatique curves under vaned intertrial rearperation periods.

Res Quart 39: 1064 15,1968-

Kroll W: Isometric fatique patterns in fernale subjects. Res Quart 42: 286-298,

1971-

MacDemid J 8 Fish Ar Reliability of Hnist fiexor and extensor strength and

isotonic endurance perfomied on the Baltimore Therapeutic Equipment

Device. PhysMber Can 43: 16.1991.

MacDerrnid JC, Kramer JF. Woodbury MG, McFarlane RM, Roth JH: Interrater

reliability of pinch and grip strength measurements in patients with

cumulative trauma disordem. J Hand Ther 7: 1 0-1 4, 1994.

Mathiowetz V. Kashman N, Volland G, Weber K, D o w M. Rogers S: Grip and

pinch strength: Normative data for adults. Amh Phys M d Rehabü 88: 69-

74.1985.

Montipetit RR, Montoye HJ. Laeding L: Grip Strength of School Children,

Saginaw. Michigan: 1899 and 1964. Res Quarf 38: 231 -240.1965.

Mundale MO: The Relationship of Intermittent lsometric Exercise to Fatigue of

Hand Gn'p. Arch Phys M M Rehabil M: 532439.1970.

Niebuhr BR, Marion R, Fike ML: Reliability of grip strength assessrnent *th the

computerized Jamar dynamometer. Occup The? J Res 14: 3-1 8, 1994.

Niebum BR & Marion R: Detedïng sinœrïty of effort when measuring grip

strength. Am J Phys MW 66: 16-24.1 987.

Nonisis MJ: SPSS/FC+. SPSS Inc.Chicago,lllinois1990.

Nwuga VC: Grip Strength and Grip Endurance in Physical Therapy Students.

Arch Phys kled Rehabjl58: 296-299, 1975.

Patterson HM: Grip Measurements as a Part of the Pm-Placement Evaluation.

Ind Med Surg July: 555-557,1965.

Pentland WE, Vandenmort AA, T m y LT: Age-related changes in upper Iimb

isakinetic and grip strsngth. Physhfher Thedry Prad 11 : 1651 73,1995.

Robertson LD, Mullinax CM, Brodowicz GR, SwafFord A K Muscular fatique

pattemïng in power grip assessment J Occup Rehabil6: 71 -85,1996.

Sheehan NJ, Sheldon F, Marks D: Grip Strength and Torquometry in the

Assessrnent of Hand Function in Patients Wï# Rheumatoid Arthritis, B H

J Rheum 22: 1581 64,1983-

Shrout PE 8 Fleiss JL: lntraclass corrdations: Uses in assessing rater reliability.

Psychol Bull 86: 420-428, 1979-

Sjogaard G, Savard G, Juel C: Musde blood fiow during isometric adivity and its

relation to muscle fatigue. Eur J Appl Physrol57: 327-335, 1988.

Smith GA Nelson RC, Sadoff SJ, Sadoff AM: Assessing sinœnty of effort in

maximal gn'p strength tests. Am J Phys Med Rehabil 68: 7380, 1989.

Stephens JL, Pratt N, Michlovb S: The reliability and validity of aie Tekdyne

hand dynamometer: part II. J Hand -9: 1 &26, 1 996.

Stokes HM: The seriowly uninjured hand - Weakness of grip. J ûccup Ther 25:

683484,1983-

Stokes HM, Landrieu KW, Domangue B, ffinen S: Identification of ImfVort

patients through dynamometry- J Hand Surg Am 20A: 1047-1056,1995.

Strafford PW, Norman GR, Mdntosh JM: Generalizabitity of Grip Strength

Measurements in Patients With Tennis Elbaw. Phys -6B: 276-281.

1 989,

Trossman PB, Suleski KB. Li PW: Ta-retest reliability and day-to-day variability

of an isometric grip streng!h test using the work sirnulator- ûccup Ther J

Res IO: 266279.1990-

Wadsworth CT, Nielsen DH. Corcoran DS, Phillips CE, Sannes TL: Interrater

relia bility of hand-held dynamometry: M i s of rater gender, body weight,

and gnp strength. J Orthop S m Phys Ther 16: 74-81,1992.

Watts Pl Newbury V. Sulentic J: Acute changes in handgrip strength, endurance,

and blood lactate with sustained sport rodc dimbing. J Sports Med Phys

Fitness 36: 255-260,1996-

Wolf LD, Matheson LN, Ford DD, Kwak AL: Relationships among grip strength,

worû capacity, and recovery. J Occup Rehabil6: 57-70, 1996.

Table 1. Static Grip and associated reliability stab*sücs

Right 1 35-4 (1 2)

Left 131.4(10)

Right 1 38.5 (1 2)

Left 1 36.5 (1 O)

Variation Belmen

j ICC

1 (95%

Confidence

interval)

Table 2. Cyclic Grip RepetitiCons and Reliability SWstics --

Absolute Variation Betwen Days (?Q

icc (95%

Confidence

interval)

Jamar

80°h

Mean

(#) Day 1

Right . .

Left

Mean(%) Day2

Jamar

70°!

Range

Right

Left

Right

Left

NK 70%

Right

Lei?

Table 3. Test- Reliability of NK Cornputer-Anam 10

Second Grip

- -

Mean

Day t

Range -

Absolute Variation Be- Days

- ---

icc (95%

Confidence

intetvar)

Mean

Day 2

Measure

Peak

Force (kg)

1 Right

Response

TÎme

(sec)

1 Right

1 Left

Rise Time

(sec)

Right I r Fatigue

( W s W Right C

l Left

169

Table 4. Relationship Betniieen SWc Grip and Other Variables

Static Grïp

Jamar

Right

Left

NK

Right

Left

Peak

Force

0-79"

0-67"

0-86"

0-70-

Capacïty

0-78"

0-63"

0-79"

0.68"

809(Test

4-28

4-29

70% Test

-0-1 1

4-04

Fatigue

0-04

0-00

1

0-05

0-28 b

L 4-16

- 0-07

4-09

0-04

Figure 1. Sample endurance data from NK 10-sec grip

Static peaks from stmc gnp trial are in box at the center of the f~ure; resuk from f 0-second sustained gfip are ta right of raw data curves,

CHAPTER 6: RELlABlUTV OF RADOGRAPHIC MEASURES AND

CLASSIFICATDN IN PATlENtS HnTH M S T ' RAMUS FRACTURES

Introduction

Ideally, a radiographic fiadure ciassification or measurement should be a

standardized measure Hidi pemits dinicians to d e ~ ~ b e a M u r e for

effective communication with others, suggest a treatment plan and assist with

prediding the results of treatment Standardization of measurement requires

that fraaure classifications are reliable, vafid and that normative values have

been established mth desa-bed standardized methods (Cole et al 1993).

Relia bility, or consistency in measureme* underlies the validity of classification

systems, their ability to predict treatment requirementsl outcornes and the value

of published normative values. The need for reliable radiographic measures in

scientific reporting has been recognized. For example, the Journal of

Orthopaedic Trauma requires that any work submitted by prospective authors

use radiographic classifications mth a ddoaimented reliability of at least

kappa=0.55 (Sanders 1997).

A variety of anatomic defomities may contribute to classification systems

for distal radius ftactures (DRF). The presenœ of articular involvement, the

amount of cornminution, the degree of displaœment in various planes and

fracture stability are a few of the major concepts involved in M u r e

management and thus represented in certain classification systems. Neither

the reliability nor the relative importance af specific anatomic defomities has

been well defined. Thus. a variety of classification systems continue to be used

in clinical literature. Classification systerns allow clinicians to categonze injuries

into a limited number of descriptive levels. The ranking of these sales is

usually ordinal ie inaeased severïty is indicated by increased IeveI on the

classification systern.

Disruption of anatamic structure due to a fiadute can be analyzed diredly

from the radiograph thmugh measurements like radial shortening, radial

inclination and dorsal angulation. These measurements are compared to

normative values or to the contralateral side to a m s the extent of bony injury.

The validity of using the contralateral side as a predidor of preinjury anatomy

has been question& (Schuind et al $996). FwthemKKe, the use of addional

radiographs, taken on the uninjured extrernity has disachantages because it

increases coss and exposes uninjured tissue to radiation, Therefore, the use

of radiographs taken only on the injured mist remains a anmon dinical

pradice. Direct measurement of malalignment in speafic planes provides

quantitative methods for desaibing fiadures. Angles or distances rneasured

from a radiograph are on a cantinuous scale. The problern with a lirnited number

of response options that ocairs ordinal scales is no longer an issue.

However, two dimensional radiographic measures of three-dimensional injuries

are restncted in their ability to fully represent the severity of the injury.

Furthemore. these two-dimensional measures are not independent of each

other, nor is the relationship beniiiieen different measures well understood.

Although classification systems and radiographic measures have some

limitations, they remain one of the most commonly used and important clinical

tools utilized to evaluate bony injuries. For this reason, reliability is a crucial

clinical issue. A number of publications have addressed reliability of

radiographic measures and clessifÏcations systems. Reliability has been

examined for the ankle (Nielsen et al 1 990; Rasmussen et al 4 993; Thomsen et

ai 1 991 ) acromiai morphology. (Bright et al 1997) hip arthroplasty interfaœ

evaluation (McCaskie et al 1996) and M u r e (Kreder et al 1996b).

Moderate to poor agreement has been daaimenteci in classification of. Gustilo

and Anderson open fractures (Hom 8 Retag 1993), the Neer system for proximal

humerus fiactues (Bnen et al 1995; Sidot et al 1993; Siebenrock & Gerber

1993). the A0 system (Johnstone et ai 1993) and the Garden system for femoral

neck M u r e s (Thomsan et al f 996).

Swiontkuwki et al (1 997) observeci that the reported reliability for most

classification systems is in the moderate range and attribute this to forcing

continuous variables into dichotomous cafegories. Radiographie quality is not

thought to be the primary source of disagreements (Martin et al 1997). Despite

questions regarding the reliability af classification systems. the use of the

AO/OTA system has been recommended by Vie pnmary organùation for

musculoskeletal Ïnjury in the wwld and mechanisms for change have ben

established (Sw-ontkowski et al 1997). However, recogniting that increasing the

number of levels of ciassaication systems tends to reduœ reliability, it has been

suggested that authors Iimit themselves to communicating about the most

general "w of fracture-

The A 0 classification systern was evaluated for inter-rater reliability by 36

assessors with a variety of expen'enœ levels ( M e r et al 1996b). The A 0

classification system 'types" are either extra-articular, partial articular or

complete articular- Groups and subgroups provide finer gradations. Raters

m e eight attending surgeons, nine fellows, ffieen residents and four non-

clinicians. lllustrating the lad< of "standard practiœ," only 3/36 of the raters

regularly used the A 0 systern- Fifieen used the F m a n scale, six the

Universal scale and 12 did not regularly use any scale, although 81% of the

clinicians wre familiar with the A0 system. A 30 minute review of the A 0

system was used for training raters prïor to data collection. When lwking at the

three fracture types rated by attending surgeons, an unweighted kappa (k) of

0.68 was obsenred. With finer gradations the kappas feil to 0.48 for groups and

0.33 for the subgroups. Kappas obtained for fellows or attending surgeons

were similar, with residents' kappas k i n g lower and non-clinicians obtaining the

least reliable scores. These authors also noted that the A 0 system addresses

joint involvement but not displacement and suggest that displaced and non-

displaced be added to the system-

Andersen et al (1 991 ) studied the reliability of four different classification

systerns, when two attending hand surgeons and two radiologists evaluated 55

sets of DRF radiographs. lntra and interder reliability was fair to moderate.

Only when the A 0 system was reduced to the (3) types of Ractures las

substantial agreement adiieved The authors suggest that it is inappropriate to

compare results bstweerr studies or guide treatment on the basis of

classification, given the poor observeci reliability in emenced dinicians and

radiolog ists.

Authors have cansistently dernonstrateci that disagreements ocair at

each successive level of tiered systems end thus agreement can be pbor Men

one extends classification past broad categories. The relatively unimpressive

reliability reported in the majority of radiographic classification studies may

reflect the difficulty in taking data that are continuous and translating them into

mutually exclusive categories. For example, disagreements on whether a

fracture is infra-artiwlar or not may occur more frequently Men the M u r e line

is close to the joint surface. disagreements on involvernent on the ulnar surface

when it is rninimally invofved and on stepff where the step is minimal. Once

data are categorized. small differences between these measures becorne

equivalent to large measurement errors.

An alternate method of expressing radiographic findings is through direct

measurements; therefore, the reliability of these rnethods is equally important

Measures of dorsal angulation, radial inclination and radial shortening are

commonly reported in the literature. Bilic et al (1 995) point out that there is a

possibility of incorrect detemination of radial shorteriing when measuing

according to the palmar or dorsal edge of the ulnar notch Hhen a marked

change in palmar tilt oaxrrs. This is because the edge c# the ulnar notch

rotates in a dorsal direction and offen migrates in a distal direction such that

bone loss is not apparent Some authors use radial shortening without stating

whether the palmar or dorsal edge of the ulnar nota was used as a landmark,

despite the fad that these hiuo methods give dietent results (Bilic et al 1995).

Furhermore, it is common dinical pracEÏœ for physicïans to eyeball radiographic

deformities rather than to diredly masure and document on the chart

parameters such as dorsal angulaüon and radiai shortening.

Kreder et al (1 996a) conductecf a reliability study lmkïng at distal radius

fracture films in order to standardize measurement procedures prior to

conducting a rnulticentre trial so that reliability and Iimits of errw wuld be

defined. Instructions w r e modified after the raters met to discuss

standardization to make instructions more specific. They utilized six films from

healed fractures because their raten reported fatigue and loss of concentration

after this number of films. Step, gap. radial angle, radial length, radial shift,

ulnar variance, palmar tilt and dorsal shift were rneasured according to

standardized instructions. Raters used high quality protractors and mlen as

w I I as fine tipped markers to draw lines and angles. Repeat measurements

were taken by the same raters 2-4 weeks later, Substantial inter-rater

consistency was observed for ulnar variance (ICC=û.82). paimar tilt (ICC=0.?4),

radial shift (ICC=0.67), radial length (ICC=0.44 over al1 raten, but wes 0.89 in

attending physicians). Lower reliability in radial length measures was ambutecl

to dwerences in landmarking ie some used the distal ulnar articular surface

Mile others used the distal medial radial articular surface. Some radiographie

measures demonstrated poor-fair reliability ie dorsal shift (ICC=0.42), radial

angle and step deformity (ICCs: 0.27438). Higher agreement was achieved

for intrarater agreement arnongst senior raters. although consistency between

fellows and attendhg surgeons who reportedly routïnely used these measures

was no higher than raters M o did not routhely use them. lntrarater reliability

was similar to inter-rater reliability ie high for ulnar variance (ICC=0.85) and low

for gap and step deformity (ICC=0.22). The tolerance limits for error were

cornputed to be: palmar tilt 1 Sol ulnar variance h m , radial shift 5mm. radial

length 10mm, dorsal shift 8mm. gap 3mm, and step 3rnm. It rnay be that acute

radiographs are more reliably measured than healed fiactures as fracture fines

have not yet been abaaired by the heaiing pmcess (Kreder et al 1996a).

It has been suggested that kappa values (k) of at hast 0.55 should be

dernonstrated before any dassificatbn scale is used to âescribe clinicai results

in peer -reviewed fiterature (Sanders 1997). According to F leiss a kappa of

>0.75 (Fleiss 19û6) is excellent Ooiers suggest thaï k i0.60 is subsfantiat and

kw0.80 is alrnost perf8Ct (landis 8 Koch 1977). RegardIess of the benchmark,

joumals are becoming increasingly reludant to published dinical results

stratified on the basis of classification systerns Mich have not been

substantiated (Burstein 1993; Sanders 1997).

Fracture classifications are popular, partially, because they provide

"names" for various fracture patterns, Hlhich aid in the communication role of

classification. Fracture classifications have been problemaic in ternis of

reliability and universal acceptanœ, consequently, the ability of these

classification scales to serve as communication tools has been restricted.

Furthemore, the other purposes of d e ~ ~ b i n g an injury ie to guide treatrnent and

aid in prognostication may be less well served by dassifications than by direct

measurements. For these reasons, w were motivated to evaluate fracture

classification systems and measurements in radiographie assessrnent of DRF to

detemine aie relative interiater reliability. A secondary question was to

evaluate the affect on rater reliability of using a volar surface to assess radial

length.

Methds

Materials

A total of 120 radiographs of distal radius fradures (64 preiedudion and

post-redaon films) Hlere rated by hand surgery fellow. The hand fellows

were in their Iast year oftraining and had used radiographie dassifications

previously. One rater consistently useâ the McMurtry system and the other

rater had not consistently used any one systm- 80th reporteci that although

they had perfmed the radiogaphic measures previously, the actual

measurements wwe not routinely measured/rBCOCCCed by either themselves, the

consultant or the radiolog ist The observers reviewd the classification systems

together and kept a description of the systems with Viem during review of

radiographs. F o m l training or disaission of resulfs during the sfudy w s not

permitted- It was our intent to document the reliability of these measures as

perfomed in routine clinical pradiœ.

Radiographie Measures

Measurements of radial shortening from the distal ulnar to distal radial

surfaces was measured as desctïbed by Wamidc (1993). A measurement of

radial shortening wing the volar surfaœ was also perfomed to detemine the

effect of using the volar surfaœ as a landmark in termination of radial

shortening. Dorsal angulation was measured Rom lateral radiographs as the

angle fomed by the line perpendicular to the long axis and the artiwlar surface

indicated by a line joining the volar and dorsal margins of the radial scafaœ (van

der Linden 8 Ericson i 981).

Figure 1. Andpis of Raâial Shortening and Dormi Angulation

a=radial tilt

b=radial shortening (loss of radioulnat length

c=dorsA angulation (loss of palmar üIt due to fradure)

179

Fracture Cfassifications

Fractures wwe classifiecf accordin9 to the f o l l ~ * n g classification systems: A 0

(Muller et al 1990). Universal (Cooney 1993), Mayo (Graff 8 Jupiter 1994;

Raskin 8 Melone 1993). McMutry (McMurtry & Jupiter 1996), Olâer (Older et al

1 965). Frykman (Frykman 1 967). The fracture classification scales are listed

below-

Figure 2. Fracture Classification Systems

A Univemal (Rayhack 1990; Rayhack 1990)

Classification of ftacture

II. Non-articular, dispiaceâ

III. Articular, nondisplaced

W. Articular, displaced

A Reducible, stable

B. Reducible, unstable

C- lrreducible

Treaanent prwfennce

Cast immobilization

Cast immobilization

Percutaneous pins

Open redudiorVextemal focation

Cast imrnobil izat ion

Percutaneous pins

CIosed reduction

Percutaneous pins (K-wires)

Closed reduction

Extemal fmtion (I percutamus

pins)

Open redudion/extemal fixation +

Percutaneaus pins

1 percutaneow pins)

Classifiaon of fracture

D. Cornplex

B. Older (Older et al 1965)

Tmaüwni pmhmnce

Open red~~odexternal fixation

Plate fixation + bone graft (&

Type I. Nondisplaced

1. Loss of some volar angulation and up to 5" of dorsal angulation

2. No signifïcant shortening 2mm or more above the distal radius

Type II. Dispiaced

1. Loss of volar angulation or dorsal displaœment of distal fragment

2. Shortening usually not below the distal ulna by occasionally up to 3 mm

below it

3, Minimal cornminution of the distal radius

Type III. Dispiaced with cornminution of the dorsal radius

1. Comminution of the distal radius

2. Shortening usually below the distal ulna

3. Cornminution of the distal fragment usually not marked and &en

charactenzed by large pieces

Type N Displaced with severe cornminuQian of the radiai head

1. Marked cornminution of the dorsal radius

2. Comminution of the distal radial fragment shattered

3. Shortening usually 28mm belw the distal ulna

4, Poor volar cortex in some cases

C. Frykman CIassitication (Frykman 1967)

Type 1. Extaarticular

Type 2.

Type 3.

Type 4.

Type 5.

Type 6-

Type 7.

Type 8.

Type 1 with fracture of distsl ulna

Radiocarpal joint involved

Type 3 wiai haawe of distal ulna

Distal radioulnar joint involveci

Type 5 with fiadure of dïstal ulna

Radiocarpal and radioulnar joints bah involved

Type 7 with fracture of the distal ulna

D. Mayo (nitra-articuiar fracture) (Raskin & Melone i 993)

Type 1.

Extra-articular radiocarpal joint intraartiwlar in radioulnar joint

Type II.

A displaced intraarticular fradure of radioscaphoid joint involving a significant

portion of the artiailar surface of the distal radius (more than a radial styloid

fracture). Associateci dorsal angulation and shortening necessary components

of the fracture-

Type III.

A displaced intraartiwlar fiadure of radiolunate joint Mich ofteri presents as a

'die-punchn fracture of the lunate fossa. A fiacture comment displaced into

the distal radioulnar joint is cornmon.

Type W.

A displaced intraarticular fractura involving both the radioscaphoid joint surfaces

and usually involving the sigmoid fossa of the radioulnar joint- This fracture is

usually comrninuted.

E. McMurby (McMurtry 8 Jupiter 1996)

1. Twogart

The opposite portion af the radioulnar joint remains intact

2. Three-part

1 82

The lunate and scaphoid facets of the disW radius separate from each other and

from the proximal portion of the radius

3. Four-part

Similar to three-part exœpt the lunate facet is furüier divided into volar and

dorsal fragments

4. Five-part

lncludes a wide variety of wmminuted fragments

F. A 0 (Muller et al 1990)

Types (A* B and C) are divided into groups (1,2 and 3) and each group is further

subdivided into subgmups as Iisted b e l w

Type A- Extraamcular

Extraartiailar fractures involving neither the radiocarpal nor the radioulnar joint

A, lsolated fiactures of the ulnar metaphysis

- 1 ulnar styloid

- 2 simple metaphyseal fradure

- 3 wmminuted metaphyseal m u r e

4 Extraartiailar fracture of the radius, simple but impacted. (These are

subdivided into subgroups depending on the orientation of the m u r e line and

the direction of the displaœment of the distal fragment)

. 1 nondisplaced

. 2 dorsal displaœment (Colles-Pouteau)

3 volar displacement (Smith-Goyrand)

4 Extraartiwlar M u r e of the radius, multi-fragment Ath varying degrees

of metaphyseal impaction

- 1 impacted (shortening)

- 2 wïth metaphyseal amiminution

- 3 with metaphyseal-diaphyseal comminution

B. Simple Articular

Fracture aff8Cfs a mon of the articular surfiace, but the continuity of the

metaphysis and epiphysis is intact

Cuneiform articular fiacture of the distal radius

- 1 radiaf styloid

- 2 comminuted radial styloid

- 3 ulnar tuedge" fiadure

Dorsai margin M u r e (Barton's fradure)

. 1 simple

- 2 assaciated radial styloid fragment

- 3 associateci with radio-carpal dom[ dislocation

Volar margin m u r e (reversed Barton's. Letenneur)

- 1 volar radial fragment (sigmoid notch intact)

- 2 volar fragment (afkting sigmoid notch)

- 3 comminuted volar fragment

C Complex articular

A fracture that affi8Cts bath the joint surface and the metaphyseal area.

C, Articular fracture of the radius simple-articular component (2 fragments

and no metaphyseal cornminution)

. 1 Colles fiacture Meding the radiwlnar joint

- 2 Colles M u r e with dorswlnar artiwlar fragment

- 3 T-fracture in the sagittal plane

C, Artiwlar fracture of the radius wïth simple articuiar cornpanent and

meta ph yseal comminution

- 1 Colles fracture Meeting the radiwlnar joint

- 2 T-sagittal fiadure

- 3 T-fiontal fiadure

C, Cornminuted artiailar fracture of the radiw

- 1 without metaphyseaI comminution (-art fradure)

- 2 with metaphyseal cornminution

- 3 with rnetaphyso-diaphyseal comminution

Data Analysis

l nter-rater reliabil ity of radiographie measures w s assessed using hm

different reliability stati*stics. lntraclass Correlation CoetKcients (ICCs) are

commonly used to quanMy rater reliability in cardinal data ie true numerical data

such as angles. number of mm of displaœment (Shrout 8 Fleiss 1979). Kappa is

most comrnonly used to evaluate reliability of nominal (nonordered) data or

ordinal (ordered) data with a iimited number of CZIfegones such as Fracture

classification systems. Debate exists over the appropriate statistic to be used

for ordinal data, especially as the number of categories increases. It haî been

documented that kappa and ICC are equivalent when the marginals are equal

(Maclure & Willet 1987). Because both cardinal and ordinal data were used and

because the wrrent benchmark for dassification systems has been desmhd as

a minimum kappa, both statistics were calw lated. A weighted kappa was

calculated for classification scales (Cohen 1968). Qwdratic weighting was

selected because this wigMnig has been desaibed as giving results equivalent

to to the ICC (Maclure 8 Willet 1987). The ICC type (2.1) was chosen as the

statistic of choice because hm randornly selected raters rated al1 films fitting the

model describecl by Shrout and Fleiss (1979). The 95% confidence intervals

were calculated for al1 ICCS as describeci by Shrout and Fieiss (Shrout 8 Fleiss

1 979).

Results

Subject charaderistics are presented in Table 1 . Inter-rater reliability for

radiographic masures m s higher for pre-reducüon films Uian for post~edudion

films (Table 2). Radial shortening. radial inclination and dorsal angulation wre

al i substantially reliable (ie b O . 6 0 ) (Landis 8 Koch 977) nihm measured on

pre-reduction films. Radial inclination was substantially reliable on pst-

redudion films. Use of a volar landmark of the radial surface resulted in poor

reliability.

Exact agreement (unweighted diagonal agreement) ocwrred in 3-96

of cases where fractures were cfassified. A high preponderance of intra-

artiwlar fractures o c a i r r d in the sample leading to imbalanœ between category utilkation. Of the classification systems only the Older system(k=0.73)

reached substantial reliability ie k>0.60 or the minimum acceptable standard

accepted by the Journal of Orthopaedic Trauma ie b0.55 . Although the Older

classification system had the highest reliability, it i t s designed for use

extra-articular fractures and has not previously been utilized to desaibe nitra-

art icular fractures. In review of the disagreements on m u r e classification, it

was observed that disagreements tended to ocair when the decision was

'borderlinen ie when the joint was minimally involved, a small ulnar head

fracture, minimal extension into ulnar wrist etc-

Discussion

This study illustrates that the reliability of routine radiographic

classification of DRF, be-n hand surgery fellows using a variety of systems,

is below accepted standards. The level of reliability observed refiects that ta be

expected under aiment clinical situations. Reliability may be enhanced

further training and practice suggeating that more of this should be incorporateci

into the routine training of residents and fellows. Notably, the la& of substantial

agreement on rediographic ciassifications found in this study is in agreement

with the majonty af studies in this area, including studies conduded with raters

having different experïenœ levels. using different scales and different joints

(Brien et al 1995; Bright et al 1997; Hom 8 Rettig 1993; Johnstone et al 1993;

Kreder et al 3996b; Nielsen et al 1990; Rasmussen et al 1993; Sidor et al 1993;

Siebenrodc 8 Gerber 1993; Thomsen et al 1991 ).

The only classification to achieve substantial reliability was the Older

system. The Oider dassifrcation system was designed for use with extra-

articular M u r e s but can be applied to any distal radius m u r e as it addresses

displaœment (shortening) and comminution. A previous study of 185 extra-

artiwlar M u r e s demonstrated "high" nitrarater (k=0.75) and interrater (k0.69)

agreement (Andersen et al 1991 ). Currently, the Older system is not used for

intra-articular fiadures. Additional evidenœ on its reliability and validity, in 0 t h

studies, may be required to convinœ dinicians to change their current pracüce

patterns. The Older system addresses radial length, dorsal angulation and

comminution and this study illustrateci that dorsal angle and radial shortening

measurements have high reliability on pre-fedudion films (ICCs>0.77).

Therefore, consideration must be given to the relative benefits and

disadvantages of an ordinal scale incorporating these masures versus use of

the measures themselves,

The Universal system proposed by Rayhadc (1 980) addresses artiwlar

involvernent and stability. It has been suggested that ils simplicity and relation

to treatment indications are appealing (Jupiter 8 Femandez 1997). In the current

study, observed agreement was 65%. although the chance carrected agreement

was only k=0.36. Fellows provided favourable feedback on the use of this

system and liked both its structure and relation to dinical practice. For this

reason, efforts to enhanœ its reliability are wrranted.

The Mayo classification system was designed specifically for intra-

articular fractures and addresses displacement and exient of joint involvement

The McMurtry system for intraartiwlar fradures classifies M u r e according to

the nurnber of parts- Neither of these systems demonstrateci substantial

reliability with the two fellows used in the present study- Because neither rater

used any of these classification system routinely, it is entirely possible that

further training or experience would result in higher reliability for any of the

classification systems evaluated in this study. Ho~liever, our chaR audits

indicate that it is not common dinical practice for physicïans to record either a

classification or a direct measutment of M u r e severity- If thess measures

are "eyeballeb in clinical practice and measured for research purposes, then it

is not surprising that rater reliability is problematic-

This study illustrated that direct measurement exhibited much higher

inter-rater reliability than did classïfÏcations. This observation anmers our

primary research question. The implications of this are that radiographic

measures can pmvide a reliable measure to be used in the chic. Perhaps,

rather than conœntrating on new and improved classifications, wa shouM focus

on incorporating these direct measures into routine clinical pradice and definhg

levels at which specific interventions are required- Radiographie measurements

taken from initial films relate to severity of injury and are also used to assess the

quality of reductions. Although pst-reduaion measures were less reliable, this

may partially refiect the increase in hornogeneity of the study sample as patients

are restored to a more unifom anatomic position after m u r e redudion-

Conversely, casts, healed M u r e Iines etc. may obscure visualization of

fractures in the post-reduction situation and result an increase in error Men

reading films.

The most commonly reporteci radiographic measures are dorsal

angulation and shortening. Van der Linden et al (1 981 ) anaiysed a variety of

radiographic measures and related covariation amongst dorsal angle, dorsal

shift, radial angle, shortening and radial shift in 202 displaced fiadures. They

found that the two measurements of dorsal displacement were highly correlated

ie dorsal angulation and dorsal shïft. They recommended use of dorsal angle

as it was easier to measure. The two measures of compression, shortening and

decrease of radial angle, w r e also highly wmelated. In fact, dorsal angulation

and compression measures ware also highly correlated. Others have also

suggested that up to 50% of measured shortening is related to rotation and

changeci position of the radial styloid (Bilic et al 1995). Measures which are

predidive of outcome that can be simply and reliably obtained should be utilized.

Because there are a variety of measures dm-bed in the literature, prosnostic

studies should be conduded to detemine the relative value of these measures.

Radial shortening can be measured in a variety of methods. Warwick et

ai (1993) demonstrated that measurement friam the articular surfaces of the

radius and ulna was more relateci to outcome than other methods of measuring

radial shortening- However, the articular surface of the radius can be

problematic as it presents a palmar and dorsal edge and is Mected by ulnar and

palmar tilt (Bilic et al 1995). Some authors have suggested alternative methods

of rneasuring radial shortening to address this problem (Bilic et al 1995). This

study illustrated that use of a volar articular s u r f a as a landmark resulted in

inferïor reliability.

Conclusions

Under the conditions where fellows do not routinely use a single

classification systern or routinely record direct measures of bony displaœment

from the radiograph poor interiater reliability exists for al1 but the Older

classification system, whereas, direct measurement of bony displacement has

substantial to almost perfed reliability in pre-reduction films. Post-redudion

measurements are less reliable than preredudion measurements. More

comprehensive training and evaluation of skill may need ta be incorporated into

training prograrns to enhance radiographic classification reliability between

cf inicians. Hmver, g iven the rel iability of direct radiographic measures, efforts

i 89

might be better direded into incorporafing these masures into dinical decision-

making and research communications-

Andersen GR, Rasmussen JB, Dahl 8, Solgaard S: Oldef s classification of

Colles' fradures. Good intraobsenrer and interobserver reproâucibility in 185 cases- Acta OIfhop Scand 62: 463464,1991-

aiic R, Ruzic L. Zdravkovic V, Boljevic 2, Kovanjanic J: Reliability of dïfFerent

methods of detemination of radial shortening. J Hand Surg &?OB: 97-

101, 1995-

Brien H, Noftall F, Mamaster S, Cumrnings T, Landells C, RocWood P: Neefs

classification system: a critical appraisal. J Tmuma 38: 257-260, 1995.

Bright AS, Torpey B. Magid D, Codd Tl McFarland EG: Reliability of radiographic

evaluation for acromial morphology. Skektal RadW 26: 71 8-721, 1997-

Burstein AH: Fracture classification systems: do they wwk and are they useful?

Editorial. J Bone J o n t Suq Am 7 5 k 17434 744, 1 993.

Cohen J: Weighteâ kappa: nominal scale agreement provision for scaled

disagreement or partial credit Psychd Bull 70: 21 3-220, 1968,

Cole B, Finch El Gowland C: Physical Outcorne Measures. Canadian

Physiotherapy Association, Toronto. 1 993.

Cooney WP: Fractures of the distal radius. A modem treatment-based

classification. Orttrop Clin Norfh Am U: 21 1 -21 6, 1 993.

Fleiss JL: The Design and Anaiysis of Clinicai Experiments- John Wiley and

Sons, Toronto. f986.

Frykman G: Frachm of the distal radius including sequelae - shoulder hand

finger syndrome, distu- in the radioulnar joint and impairnent of

nerve fundion- A dinical and experïmental study. Acta Odhop Scand

108: 1-155,l-7.

Graff S & Jupiter J: Fracture of the distal radius: classification of treatment and

indications for extemal fixaüon. Injury 25 Suppl4: S O I 4, 1 994.

Hom BD & Rettig ME: interobserver reliability in the Gustilo and Andersen

ciassificiation of open fradures. J Orthop Trauma 7: 357360, 1 993.

Johnstone DJ, Radford WJP. Parnell EJ: InterObSeiver variation wing the

AOiASlF classification of long bones fiadures. Injury U: 163-165. 1993.

Jupiter JB 8 Femandez DL: Comparative classification for fradures of the distal

end of the radius. J Hand Suw Am 22: 563-571.1997.

Kreder HJ, Hanel DP, McKee M. Jupiter J, McGillivary G, Swi~ntkowski MF: X-

ray film measurements for healed distal radius fradures. J Hand Sutg Am

21 : 31-39,l-a-

Kreder HJ, Hanel DP, McKee M, Jupiter JB, McGillivary G, Swiontkowski MF:

Consistency of A 0 M u r e classification for the distal radius. J Bone Joint Surg Br 788: 726-731,1996b-

Landis JR 8 Koch GG: The measurement of observer agreement for categorical

data Biomettiès 33: 159-1 74, 1977.

Maclure M 8 Willet WC: Misinterpretation and misuse of the kappa statistic. J

Epid 126: 161 -16û. 1 987.

Martin JS, Marsh JL, Bonar SK, Decoster TA, Found EM. Brandser €A:

Assessrnent of the AOIASIF M u r e dassifïcation for the distai tibia. J

Orthop Trauma I f : 477=4û3,1997.

McCaskie AW, Brcnnm AR, Thompson JR, Gregg PJ: Radiologid evaluatïon of

the interfaces after cementeci total hip repiacement Interobserver and

intraobserver agreement J Bone Joint Surg Br78: 191 -1 94, 1996-

McMurtry RY 8 Jupiter J E Fracfures of the distal radius- In Browner BD, Jupiter

JB, Levine AM and Trafton PG: Skeletal Trauma- W.B. Saunders Co,

Phi fadelphia. 1996.

Muller ME, Nazarian S. Koch P, Schatzker J: The compreshensive dasification

of long bones. Springer-Verlag, New York 1990.

Nielsen JO, Dons-Jensen H, Sorenson HT: Lauge-Hansen dassiflcation of

malleorlar fradures: precision improved by instrudion. Acta Orfhop

Scand 64: 93-694, 1990.

Older TM, Stabler EV, Cassebaum WH: Colles ftaaure: evaluation and seledion

of therapy. J Trauma 5: 469476. 1965.

Raskin KB & Melone CP Jr Unstable articular M u r e s of the distal radius-

Comparative techniques of Iigamentotaxis. Otfhop C h NOMI Am 24:

275-286, 1993,

Rasmussen S, Madsen PV, Bennicke K Observer variation in the Lauge-

Hansen classificaion of ankle fratures: precision improved by instruction.

Ada Orthop Scanâ 64: 693494,1993-

Rayhack J: Symposium on distal radius fradures. Edited by William Cooney.

Contemp Orthop 21 : 75,1990-

Sanders RWr The problem wÏth apples and oranges. Editorial. J Otthop Trauma

Il: 466, 1997-

Schuind F, Aiemzadeh S. Stallenberg B. Burny FI ûoes the normal contralateral

wrist provide the best reference for X-ray film measurements of the

pathologie wsf? J Hand Surg Am 21: 2430,1996-

Shrout PE & Fleiss JL: lntradass wnelations: Uses in assessing rater reliability.

PsychoI Bull û& 42W28,1979.

Sidor ML. Zuckemian JD. Lyon Tl Koval Y Cuomcs F, Schoenberg NI The Neer

classification system for proximal humeral fractures. An assessrnent of

interobserver reliability and intraobserver reproducibility. J Borie Joint Surg Am 75: 7 7454 750, 1993.

Siebenrock KA & Gerber C: The reproducibility of classification of fractures of

the proximal end of the humerw. J Bone Joint Surg Am 75: 1751 -1 755.

1993,

Swïontkowski MF, Sands AK, Agel J, Diab M. Sdiwappach JR, Kreder HJ:

Interobserver variation in the AOlOTA fradure dassifcation system for

pilon fradures: is there a pmblem? J Olthop Trauma 11 : 467470, 1 997.

Thornpson WD: Effect Modification and the Limits of Biological Inference from

Epidemiologic Data. J Clin Epi&?miol44: 221-232, 1 991 .

Thornsen NO, Overgaard S. Olsen LH. Hansen Hl Nielsen ST: Observer

variation in the radiographic dassificiation of ankle fractures. J Bone

Joint Surg Br 738: 676478.1 991 .

Thomsen NOB, Jensen CM, Sicovgaard N, Pedersen NS. Pallesen P. Soe-

Nielsen NH, et al : Observer variation in the radiographic classificiation of

fractures of the nedc of the femur using Garden's systern. lnt J Oiihop

20: 326-329,1996-

van der Linden W & Ericson R: Colles fracture: how should its displacement be

measured and how should it be immobilized? J Bone Joint Surg Am 63A:

1 285-1 288,1981-

Wamidc D, Prothero O, Field J, Bannister G: Radio(ogica( rneasurernent of

radial shortening in Colles' Fracture. J Hancf Surg Br 186: 50-52, 1993.

Table 1. Subjects

C hariacteristics Mean (Standard deviationl

Gnp at t year 89% of unaffected side

AROM at 1 year (6 motions) 88% of unaffected side

Jebson's Checkers Subtest 93% of unaffected side

Tester #1 Measures Pre-reduction Post-reduction

Radial Inclination 16.5 (7.6) 20.4 (4.4) - - - - - - - - - - -

Dorsal Angulation 14.3 (20.5) -3.3 (6.0)

Radial Shortening- - 1 -3 (3.1) -0.84 (1 -9)

Radial Shortening,,,, -2.3 (2.5) -2.1 (2.4)

Tester #2 Measures Pre-reduction Post-reduction

Radial Inclination 16.5 (7.7) 20.1 (5.2)

Dorsal Angulation 14.1 (19.5) -0.09 (8.8)

Radial Shortening, 0.34 (3.5) -1 -36 (2.4)

Radial Shortening, -1 -8 (2.5) -1 -36 (2.3)

Table 2. Inter-rater Reliability of Radiographic Measures in Distal

Radius Fractures (N=67). Intraclass Correlations (and 95%

Confidence Intenrals).

Measure Initial Post-Reduction

Radial Inclination

Dorsal Angulation 0.92 0.59 (O,ûû-û.9!5) (0-41-0.72)

Radial Shortening- 0.77 O -44

(OB1 4-85) (0 -26.0 -59)

Radial Shorteningv.- 0.22 0-27

(0,024 -42) (O -07-0 -45)

1 97

Table 3. lntewater Reliability of Fracture Classifications (95%

Confidence Intewab)

Classification Po 1CC (H% coawmce Iirdsnral)

Older 56 0-73 0-71

(0624.84) (O. 594-80)

Frykman 47 0.35 0-44

(0.1 1-0.57) (0.15-0.52)

Un iversal 65 0-36 0-41

(O. 1 54.57) (0.21 -0.58)

Subgroup 0-33

(0.17-0.51) Po= observed diagonal agreement (uninieighted); ICC=lntraclass correlation

coefficient

1 98

CHAPTER T: RESPONSNENESS OF THE SF-36, DASH, PATIENT RATED

W S T EVALUATION AND PHYSlCAL IMPAIRMENTS IN EVALUAmNG

RECOVERY AFTER A MSTAL RAMUS FRACTüRE

Introduction

A m u r e of the distel radius (DRF) is a wrnmon injury (Larsen 8

Launtsen 1993). There are a wide variety of treatment options for DRF and

these Vary according to a number of patient factors, injury charadenstics and

clinician preferences (McMurby & Jupiter 1996). Until recently, the majority of

studies which evaluated outcome after treatment used either observer-based

scales that have not ben tested for reliabilityhcalidity (Sarmiento et al 1980) or

measures of physical capability (ROM and grip strength). With the development

of a nurnber of patient-rating scales which fouis on general healwquality of life

(Ware et al 1993). upper extremity quality of life (Hudak et al 1 996) or Wst pain

and disability (MacDemiid 1996; MacDemid et al 1998) it has been possible to

evaluate the impact of injury and treatment fiom the patienfs perspective.

Although in the past evaluation of recovery from DRF was primarily based on

radiographie results and measured impainnents of range of motion and grip

strength, patient evaluations are k ing increasingly incorporated into outcorne

evaluation. Reœnt case series reports have started to report generic health

outcornes such as the SF-36 physical and mental component summary scales

(Femandez et al 1997).

Responsiveness is the ability to detect important dinical changes and

may be cunsidered one fom of validity (Kirshner 8 Guyatt 1985). Various

indices have been defined to describe instrument responsiveneso (Norman et al

1997; Strafford et al 1996; Wright & Young i 997a). It is k n m that different

indices can provide a different ranking of instruments (Wright & Young 1997a)-

Responsiveness indices relate oie observeâ dinical change to the background

noise in a number ofmays. The Standardized Response Mean (SRM) is

calailateci as the average change score (initial avaluaüon to f o l l w p ) divided

by the standard deviaüon of aie change scores (Liang et al 1990; Stratford et al

1996). Another commonly used responsiveness index is eff&t sue (average

change divided by standard deviation of initial scores) (Kaat et al 1989;

Strafford et al 1996). The SRM is thought to be a superior index of

responsiveness because t is not ïnfîuenced by sample size, (Stratfixd et al

1996) whereas, eff- size is more related to magnitude of clinical change ( W s

et al 1989)-

Responsiveness is a cn-tical measurernent property for evaluative t a s ;

despite this, there is Iittle evidenœ on the relative responsiveness of wrrently

used outcorne measures for DRF. Arnadio et al (1996) studied the relative

responsiveness of three questionnaires and physical capacity measures in 21

consecutive patients treated for Colles M u r e . Patients w r e tested when their

immobilization device was removed and again three months later. The physical

examination included static grip and pinch testing, sensation, dexterity and ROM

testing as perfomied on the NK Hand Evaluation system6. These measures

were expresseci as absolute scores from the affectecl extremity. The SF-36, a

modified Arthritis Impact Measurement scale (AIMSZ) and a modifieci Brigham

and Women's carpal tunnel instrument were completed on bath occasions. The

responsiveness of these instruments in evaluating three month recovery after a

wrkt fracture is Iisted in Table 1 . The highest responsiveness was found for

several function-related subscales of the questionnaires and motion and

dexterity testing. This study concludeci that questionnaires provide a responsive

method to detect clinical change.

NK Biotechnical Engineering Co, Minneapolis, MN

While this paper provided valuable evidence on evaluation of DRF, a

number of issues required Mher development Firstly. as behiiieen

variability is the background noise which limits the ability to detect dinical

change, Morts to reduœ this may enhanœ responsiveness. Certain physical

capability measures, such as grip strength and dexterity, are subjed to high

levels of inter-subject vanability W i n a nomal population. This was illustrated

by '~rnadio et al (1996) Hiho reported standard deviations exceeding mean

values for gr& pinch and sensation absolute scores. Grip sfrength is commonly

used as an outcome measure after upper extremity injury and it is routine clinical

practiœ to express afF&ted hand strength as a perœntage of the unMeded

hand strength. By expressing data as a perœntage of the altemate side,

impairrnents are made more obvious and each person ads as his or her own

control reducing the inter-subject variability uf these physical capacity measures.

Previous studies of responsiveness have used raw grip strength scores (Amadio

et al 1996; Katz et al 1994). It is useful to k n w vvhether the clinical practice of

using ratios does provide a more responsive outcome rneasure-

A further issue is that upper extremity measures such as the DASH

(Disability of the Am, Shoulder and Hand) (Hudak et al 1996) and the P M E

(Patient Rated Wrist Evaluation) (MacDermid t 996; MacDemid et al 1999) are

only recently published. Arnadio et al (1 996) modified a carpal tunnel instrument

and an arthritis instrument to obtain musculoskeletal scales. It is appropriate to

detemine whether recently introduced instruments designed for upper extremity

regional evaluation provide more favourable responsiveness than traditional or

generic measures.

Finaliy generic scales themselves have continued to evolve. The SF-36

was traditionally scored using eight individual subscales scores. While these

scales continue to have value for certain applications, a single global physical

score has advantages for data analysis and reporting of outcornes. Wth the

development of the Physical (and Mental) Component Surnmary Scores (Ware

et al 1 994, this option is now available for SF-36 users. It is now relevant to

compare the responsiveness of aie global sumrnary score of physical health ie

the Physical Component Summary Score to the Sf-36 subscales most relevant to

DRF (Physical Fundion, Physical Rok, Bodily Pain). The purpose of this study

is to increase o u understanding of these measurement issues by evaluating the

responsiveness of physical capacity measures (gnp, motion and dexterity), and

three patient rated instruments (SF-36. DASH and PRWE) .

M e t t i o d s

Patient Testing

Patients with wrist ftactures (n=59 for whom al1 questionnaires were

completed at al1 thtee visits) wwe tested on three occasions. Primary

emergency care took place either in local or outlying hospitals. The badine

orthopaedic chic visit and definitive treatment ocairred wïthin the first 1 O days

after fracture. Patients wre assessed again at three and six months &ter their

fracture. At al1 visits the patients completed the SF-36, Ware et al f994) the

DASH, (Hudak et al 1996) and the PRWE (MacDemid 1996). Immobilization

devices were removed prior to the 3month visit and thus physical capacity

rneasures w r e incorporateci at that tirne, Patient characteristics are presented

in Tabte 2.

Outcome Instniments

Questionnaires

SF-36

The SF-36 is a widely used general healwquality of life index with eight

subscales for domains of health and hm summary scores (Ware et al 1993;

Ware et al 1994). The SF-36 is an evaluative measure recommended by the

Amencan Academy of Orthopaedics, and is wmmonly used ta evaluate patients

wi-th joint disordem. It also is used to evaluate a wïde variety of physical and

mental pathologies and a large database of nomiaove data is available through

the Medical Outcomes Trust (Ware et al 1993; Ware et al 1994)

The scale has eight suOscales which pofûay various domains of

healthlquality of Iife: Physical fundion, Physical Role, Bodily Pain, Vitality,

General Health Perception, Emotional Role, Mental Health and Social Fundion.

These subscales are scorecl out of a maximum score of 100 (higher is better).

The Physical and Mental Components Surnmary Scores represent the

main components of health. These scores are calailated in a three-step

process. First the eigM subscales are sco~ed and standardized using means

and standard deviations from the US. population. Second they are aggregated

using weights ( fador score coefficients) from the US. population. Then

aggregate summary scores are standardized using a Iinear T-score

transformation to have a rnean of 50 and a standard deviation of 10 in the U S

general population (Ware et al 1994).

The physical domains of health and M i l y pain are rnost conceptually

relevant to DRF and their responsiveness in DRF assessrnent has been

previously reporteci (Arnadio et al 1 996). Based on these reasons, these

subscaies (Physical Function, Physical Role, Bodily Pain) and the Physical

Component Summary (PCS) were evaluated as outcome measures of interest

DASH

The DASH (Disability, Am, Shoulder and Hand) is a %item upper

extremity scale which measures function, syrnptoms and quality of Iife issues

related to upper extremity pathology. The ma@rity of questions relate to

fundional activities requirïng use of the upper extremity; the remainder include:

two items specific to pain, one addresses social life and one wwk; three relate to

other symptoms, one sleeping and one perceived capability- lt was scored as a

percentage out of 100 (O= no problems), as recommended by authors (Hudak et

al 1 996; Upper Extremity Collaborative Grwp 1 996).

203

PRWE

The (PRWE) Patient Rated Wnst Evaluation is a 15-item questionnaire

which equally rates Mst-related pain and disability in frndional activities.

Fundional items include six specific tasks and four questions where the patient

rates their ability to perform hi$ or hef wual lave1 of fundon in the domains of

self mre, wwk, howehold work and tecreation. A score out of 1 ûû (O= no pain

or disability) is calculated (RilacDemiid 1996; MacOermïd et al 1998).

Physical Capacity (Impairment) Testing

Static Grip Strength

Grip strength ~ w s measured with NK Hand Assessrnent Systern7- Grip

strength (average of three repetitions) was measured accordhg to standardized

methodology (Fess 1992). A grip strength score was computed as the ratio of

affecteci/ u n a f f ' side wïth an adjustment for dominanœ, (1 -07 used to adjust

ratio to "remove" effect of dominanœ) based on previous author's normative

data (Mathiowetz et al 1985; Mathiowetz et al 1984). Where no affected hand

was available, normative values wwe used (Mathiowtz et al 1985). A total of

40 points in the composite physical impairnent score was allotted to gflp

strength.

Range of Motion

Range of motion (ROM) was tested using computerized goniometry from

the NK Hand Evaluation system. The motions tested wre: radial and ulnar

deviation, flexion, extension, pronation and supination. A standard dorsal

NU Biotechnical Engineering Co, Minneapolis, MN

technique m s used for wrist ffexioniextension (LaStayo 8 Wheeler 1994). A

standard dorsal alignment along the third metacarpal and the foream, with the

wrïst as a fulaum, was used for mist deviations (NoMn 8 White 1995).

Forearm rotation mas measured using a perpendiailar axis and either the

proximal M-st crease or just proximal to the ulnar head as landmarks for

placement of the moving am of the goniorneter (Armstrong et al 1998). ROM

capability was dassifled Rito a 3îLpOint score based on the system presented in

Figure 1.

Dextenty

A brief dexterïty test was seleded to refkct foss of ability to manipulate

objects with the radial side af the hand. This might ocair as a resuit of a median

nerve insult, joint stiffness or other fradure complications. Dextenty was

assessed by the Jebson's Hand Function4eckers subtest (average of three

repetitions) and cornpared the uneffected side Mth an adjustment for

dominance (1 -07) (Jebson et al 1969). A total score out of 15 was computed for

this ratio. See Figurel.

Physical Impairment Score

The total score for physical capacitylimpairment (out of 85 points) was the

sum of these three components; a higher score represented better physical

capability. See Figure 1.

Data Analysis

The standard- response mean and M8Cf sizes w r e calculateci

between basefine and each of the followup visits (Strafford et al 1996).

Results

Over bath the Q.3 and 0-ô month time periods, the PRWE m s more

responsive than the DASH or th8 SF-36- (Table 3) The DASH was more

responsive than the SF36 over these same time fiames. This trend was

observed for bath the SRMs and sire- A larger dinical efféct was

observed during the first three months of treatment than during the second thme

months of treatrnent as noted by the smaller eff&ct sizes and SRMs obsaved.

Differences be- questionnaires were not remarkabie during aie second

interval. Howver, the physical impairment score, vuhich could not be measured

at baseline, was more respansive than the questionnaires over the 3 4 month

time period.

Or the components of the physical impairment score, grip was most

responsive, followed by ROM Of the subscales of the PRWE and SF-36,

fundion subscales were more responsive over the 0 3 month peflod, whereas

pain and fundion were similar over the 3 4 month pend.

Discussion

The data in this study are compared to SRMs obtained by other

investigators for DRF (Amadio et al 1996) and other conditions in Table 1. The

six-month SRMs obtaïned for the SF-36 w r e similaf to those obtained six

months after Hip Arthroplasty, (Wright 8 Young 1997b) or for Physical Role

after DRF (Amadio et al 1996) but higher than that obtained with &ers treated

for Musculoskeletal condition's (Beaton et al 1997) or a variety of other

musculoskeletal problems (Martin et al 1997). This can be attributed to the fact

that DRF are acute injuries and may expeflence a more unifonn response to

intervention and a higher amount of clhical change than more chronic

musculoskeletal conditions. The joint specific scale (PRWE) demonstrated

slightly higher SRMs than reported for other disease or joint speafic scales

including those for the shoulder, (L'lnsalata et al 1997) carpal tunnel syndrome

(Katz et al 1994) or arthritis (Hust et al 1997).

Amadio et al (1996) reporteci that the Bodily Pain and Physical Role

subscales of the SF-36 showed statistically signifiant improvement after DRF;

whereas, the Physical Fundion subscale did not We fwnd a similar SRM for

Physical Role ( 0.81 vs 0.95) but higher SRMs for Bodily Pain and Physical

Fundion,

Amadio et al (1996) reportecl SRMs for grip strength, but higher

SRMs for dextetity and ROM than the prasent study. These differenœs may be

partially explained by diftèrenœs in masurement procedures. Both previous

studies which have looked at grip strenm responsiveness Mer treatment for

upper extremity pathology have used raw strength scores (Amadio et al 1996;

Katz et al 1994). We normalized our strength masures based on the person's

uninjured extremity, thus eliminating the high variability commonly seen in raw

strength scores- Because these ratios are commonly used clinically and provide

more favourable responsivefless, w believe they are a superior way to report

strength scores when construding scales.

Our ROM measure was a graded composite score (Table 3) for ROM

measures in six movements measured with the same computerked goniorneter

as used by Amadio et al (1 996). They reported an SRM of 1 -08, whereas we

found it to be 0.84. While they did not state how ROM was measured, it appears

that they measured wrist motions and cornputed a cumulative raw score. This

suggests that graded scales for ROM may be less responsive than the raw

measurements, This is consistent with loss of infornon that can occur when

data is collapsed into categories. If wa surmise that nomalkation of grip

strength scores enhanced responsiveness, then the same prïnciple may hold

true for ROM. as well. Although, normal ROM is less variable than normal

strength, Mure studies should investigate this possibility.

Amadio et al (1 996) reported that dexterity had a higher SRM than

observed in the present study. The checkers suMe* af the Jebson's Hand

Function test was used, in the present study, as a Mef screening for

pronounceâ loss of dexterlty as may aca ir 4th median nenre injury. Each

repetition takes 3-8 seconds to perforrn. The NK dexterity test is a more detailed

test contained in the NK computetized hand evaluation system. It involves three

subtests (large, medium and small objeds) and takes approximately 30 minutes

to test (Turgeon et al 1998). The NK dexterïty test is not widely available in

clinics and perhaps for this reason has been infiequentIy reporteci as an

outcome measure. Hawever, it is apparent that this more detailed test is more

responsive in evaluation of clinicaf recovery after treatment for DRF.

While Amadio et al (1996) reported that questionnaires tended to have

greater SRMs, this was tnie for a minority of the subscales of the instruments

they evaluateâ. Thus a vatiety of questions, particulariy in the AlMS were

addressing domains not a f f ' e d by DRF. To rninitnize the patient burden

associated with outcome measurement, shorter tools which more directly

address the problems of the patient with DRF should be utilized. In otherwwds,

long instruments with few scales that are responsiveness to change in a specific

condition are not appropriate for use in that patient population- Hmmver, clinical

relevance can not be decided on the basis of responsiveness atone. Patient

rating scales provide a more comprehensive view of the patient status than

isolated impairment measures such as grip strength ratios. Similady, composite

physical impairment scores such as the Sf-3ô Physical Component Summary

score or the total impairment score provide more information than the separate

components. Thus. where responsiveness is considered acceptable, these

measures may be selected because they provide a more global representation

of outcome. Generic instruments may be selected as outcome measures,

although they are k n m ta be less responsive than specific measures. because

they reflect a more global pmpedive and alkw one to compare difFerent

pathologies and their relative impact on overall h e a k

A nurnber of dinical issues impact on the evaluation of responsiveness.

The appropnate time frame over which respansivemss should be avaluated can

Vary, dependhg on the pathology. the aaiity of the condition and the time

course of the antiupated tmatrnent, This can make comparÏson of difbrent

studies problematic. FurMermore, different instruments may be more msponsive

at different aspects of r6cOvefy- For example, an instrument Mich -ses on

pain may be more responsive in the acute phase of fradure healing, but may

not adequately address issues such as inability to retum to normal adivities. In

Viis study, the fundion subscales &the PRWE were most responsive during the

0-3 month the period. This may r d e d the definition of the 'top end" &the

scale, In the function subscale of the PRWE the 'tKlr& one can be is 'unable to

perfom" an adkity. Worst on the pain scale is defined as "worst (pain) ever."

Patients with acute ftactures that are adequately reduced and immobilized rnay

experience mildlrnoderate pain but still be unable to parfarm fundional

activities, With removal of immobilization their function should increase- In fa&,

clinicians using the PRWE for ongoing patient management report that both

pain and function tend to imrease immediately post-immobilization. This is not due to a la& of responsiveness in the pain scale. Conversely, the scale is

reflecting a cornmon clinical phenomena ie immobilization minimizes fracture

pain, whereas initiation of rehabilitation cm result in a temporary increase in

pain.

Another factor ta be cansidered is whether the treatment is efficacious-

The magnitude of the treatment effect influences the responsiveness statistics

obtained. This leads to differences betwean studies baseci on the patient

sample studied. Differences in responsiveness indices, based on patient

sample, are evident in the lowr responsivaness teporteci for the generic

rneasures in wrkers, (Beaton et al 1997) man reporteci in the present study

consisting of patients *th acute injuries. This treatment magnitude effect ums

further illustrated within the present study by the 1owr respomivaness indices

during the 3-6 maith time intewal as cornpared to the baselhe-3 month interval.

Because tmatment M8dl*veness does a f k t respansiveness indices,

some authors saparate patients who responded to treatmnt Rom those wuho did

not Data may be presented for the entire clinical group andlor only those M o

responded to treatrnent (Beaton et al 1997; Hurst et al 1997). Norman et al

(1 997) point out that any post hoc separation of a group of patients. into those

who report a positive treatment e f f B C f and those who do not, result in a

responsiveness ratio that is unrelated to the responshness COBfficient That is

becawe any group of patients wïll fom a distribution from which high and law

change saxes can be partitioned, ueating a ratio that is always greater Vian

zero, even Men there is no detectabie average change in the patient cohort.

One additional problem associated wiai these retrospective evaluations of

responsiveness cornparing improved unchanged patients is that they are

based on global visual analogue scale measures of change which themselves

have not been validated (Norman et al 1997). For this reason, dividing a cohort

of patients based on their apparent response to treatrnent appean il1 advised. In

the case of distal radius fractures, the influence of patients M o did not

experience any positive treatrnent effect is rninimized. Even a suboptimal

treatment is more effective than no treatment, in the case of bony fraaures, thus,

with few exceptions, patients should e w e n œ some improvement even with

inadequate redudion, suboptimal bone healing andlar complications.

As emphasized by Strafford, (1996) a number of study designs can be

used to assess responsiveness. Most incorporate the concept of distinguishing

scale performance Men patients are stable versus that obtained when a

meaningful diniml change has taken place. This concept is particularly

important in conditions such as low back pain (Stratbrâ et al 1996) or workers

with musculoskeletal disease (Beaton et al 1997) vvhere natural history may

mntribute to recovery in the absence of treatment, HRiere treatments may not be

effective for all patients and Mere secondafy gain may minimize reported

clinical effeds. Strafford emphasized that a research design which look$ at

short stabüity of a masure may underestimate the magnitude of random

variability that occufs over üie longer teim in patients whose health status is

tmly stable. Therefore, it is important to determine that consistent scores are

obtained M e n the patient does not change, eiaier over the short terni or the

long term and that real dinical change is reff8deô in a change in scores in

exœss of the M8Cf contributeci by background noise. We have previously

demonstrated that the Patient Rateâ Wnst Evaluation (PRVVE) provides

consistent scores wben patients are retested withi a 2-week pend and also

when the retest ocarrred one year later in patients stable pathology

(MacDermid et al 1938). The present study demonstrates mat the PRWE is also

responsive to real clinical change oaximng over three or six months.

The joint-specifk PRWE was m r e responsive than the DASH wuhich is

specific to the upper extremity and the PCS which is a general physical health

index- Differenœs behiiieen the scales rnay be partially attn-but& to the fact that

the PRWE questions more ciearly target the mist and thus are more able to

detect symptoms related to vwïst pathology. Another difference between the

scales was Uiat they priofitize symptoms differently, for example 2130 questions

on the DASH are pain questions, one-half of the PRWE is based on a pain score

and the bodily pain scale from the SF-36 foms a weighted component of the

PCS. The PRWE does not describe the overail impact of pathology on general

healWquality of Ife, but is more responsive than these generic measures.

Information may be maximized by using bai instruments together.

Although the PRWE had a higher SRM than the DASH, the difkrence

between these nAio measures was small in comparison to the difference beiween

either the DASH or the P M and the SF-36. Furthemiore, the DASH SRM

compared favourably to responsiveness reporteci for other outcorne measures in

other oroiopaedic conditions (See Table 1). For mis reasan, the DASH was

considerd to have favwrable responsiveness, which supports its use as an

evaluative tool in upper extremity clinical ptadice including DRFs. While

administrative efficiency issues may didate that a limited number of measures

are used in the clinic, for clinical trials, seledion of Vie most responsive

instrument *Il rninimize sample sire requirements. Thus, decision making on

outcorne measure selection is influenced not only by methodologic factors, but

also amrding to the underfying regsons for viihich the instrument is use&

This study demonstrated that physical capacity rneasures which are

based on aff8ded/unaffeded ratios are more responsive than three patient

rating scales in measuring reawery Mer DRF dunng the 3 4 month recovefy

period. Grip and ROM, which have b e n traditional outcome measures after DRF, were most responsive. While these measures were more responsive than

patient rating scales, this does not suggest that they should be used in isolation.

They do not necassarily correspond to kindional capability and therefore rnay

not reflect outcornes that are of priotity to patients.

Responsiveness is one element of validity. Other elements wïll Mëct the

decision to select an outcome tool. Instruments must have content validity That

is, clinicians must accept that they measure important elements of the

pathologies for which they are to be used. Treatment Mects can be positive,

provide no benefit or even result in negative health e f fds- The extent to which

these effects are addressed and the relative weighting of symptoms vvi-Il affect

the responsiveness observed. For example, one instrument couid ask questions

that directly measure a common complication which rnay be missed by other

scales. If examining respansiveness statistics in isolation, one might conclude

that this instrument was insensitive to clinical change &en in fa& it is was

measuring negative clinical e f f ' s not observed by other instruments. For this

reason, responsiveness must be considerd as one element of validity, with

other types of validity sudi as face, mnstmd and aiterion validity k i n g

additional factors used Men selecting an instrument for a specific application.

Conclusions

This study demonstrateci that both patient rating scafes (PRWE, DASH

and SF-36) and physical impairment measures (grip, ROM and dexterity) are

responsive in evaluation of recovery after DRF. The more pathology-specific

scales (PRWE and DASH) ware more responsive than the SF-36 subscales or

Physical Component Summary Score The size of the dinical eff- observed

was larger over the first three months than for the next consecutive three

rnonths. Physical impairment measures tendeci to be more responsive than

patient rating scales during the period Hhere both wuld be measured ie over the

Arnadio PC. Silver~tein MD, Ilstnip DM, Schleck CD, Jensen LM: O u t m e &er

Colles fiadure: the relative reswnsiveness of three questionnaires and

physical examination measures. J Hand Surg Br 21 : 781 -787,1996-

Armstrong A, Madiemiid JC, Chinchalkar S. Stevens RS, King G: Reiiability of

range of motion measurements in the elbow and foream. J Should Elbow

Surg 7: 573-580, 1998.

Beaton DE, Hogg Johnson S. Bombardier CI Evaluating changes in health

status: reliability and responsiveness of five generic health status

measures in &ers with rnusculoskeletal disordem. J Clin Ep&mbl5ô:

79-93, 1997-

Beurskens AT, de Vet HC, Koke AJ: Responsiveness of functional status in low

back pain: a cornparisan of different instruments- Pain 65: 71-76, 1996-

Femandez JJ, Gmen GS, Herndon JH: Outcorne of distal radius M u r e s using

the short fom 36 health survey. Clin Oithop 36-41, 1997-

Fess EE: Grip strength. In Casanova JS: Clinical Assessrnent Recommendations

(2nd) Amencan Society of Hand Therapists. Chicago. 1 992.

Hudak PL, Arnadio PC, Bombardier Cr The Upper Extrernity Collaborative

Group: Development of an upper extremity outcorne measure: the Dash

(disabilities of the am, shoulder and head (sic)). Am J lnd Med 29: W2-

908,1996-

Hurst NP, Kind P, Ruta D. Hunter M, Stubbings Ar Measuring health-related

quality of life in rheumatoid arüntis: validity, respansive~ws and

reliability of EuroQol (EQ-SD). Br. J Rhematd. 551 -559, 1997.

Jebson RH, Taylor N, Trieschmami RB, Tmtter MJ, Howard Lk- An objective and

standardized test af hand fundion. Amh Phys Med Rehabiil 50: 31 1-31 9,

1 969.

Katz JN, Gelberniart RH, Wngtn Lew RA, Liang MH: Responsïveness of

self-reportad and abmve masures of disease severity in carpal tunnel

syndrome. Med Care 32: 1 127-1 133,1994,

mis LE, Anderson JJ, Meenan RF: E f f i sites for interpreting changes in

health status. Med Cam 2?: SV&Sl89, 1989-

Kirshner B 8 Guyatt G: A methodological fiamewwk for assessing health

indices- Journal of Chtonïc Diaases 38: 27-36,1985.

Clnsalata JC, Warren RF, Cohen SB, Altchek DW, Petwson MG: A self-

administered questionnaire for assessment of symptoms and function of

the shoulder. J Bone Joint Surg Am 79: 73&748,1997.

Larsen CF & Lauritsen J: Epidemiology of acute wrïst trauma. Inf J Epriemii

22: 91 1-916, 1993-

LaStayo PC 8 Wheeler DL- Reliability of passive Hirist fiexion and extension

goniometnc measurements: a multicenter study. Phys Ther 74: 1 62-1 74,

1994.

Liang MH, Fossel AH, Larson MG: Cornparisans offive health status instruments

for orthopedic evaiuathn. Med Care 28: 632442,1990.

MacDemid JC: Development of a scale for patient rating of W s t pain and

disability. J Hand Therg: 178-1 83, 1996.

MacDemid JC, Turgeon T. Beadle Ml Richards RS, Roth JH: Patient rating of

M-st pain and disability A diable and valid measurement tool. J Olaiop

Trauma 12: 577-586. 1998-

Martin DP, Engelberg R, Agel J, Swiontkowski M F Cornparison of the

Muswloskeletal Fundion Assessrnent questionnaire with the Short F m

36, the Westem Ontario and McMaster Universities Osteoarthritis Index,

and the Siclmess Impact Profile health-stahis measures. J Borie J o n t Surg Am 79: 1323-1 335,1997.

Mathiowatz V, Kashman NI Volland G, Weber K, Oowe M, Rogers S: Grip and

pinch strength: Normative data for adults. Arch Phys Med Rehabil66: 69-

74, 1985-

Mathiowetz V, Weber K, Volland G, Kashman N: Reliabilty and validity of gflp

and pinch strength evaluations. J Hand Surg Am SA: 222-226.1984.

McMurtry RY 8 Jupiter JB: Fractures of the distal radius. In Browner BD, Jupiter

JB, Levine AM and Traf€on PG: Skeletal Trauma W. B, Saunders Co,

Philadelphia. 1996.

Nodün C & White DJ: Measurement of Joint Motion: A Guide to Goniometry. FA

Davis, Philadelphia f 995.

Norman GR, StrMord P, Regefv G: Methodological problems in the

retrospedive computation of responsiveness to change: The lesson of

Cronbach. J Clin E m m a 50: 869879,1997.

Sarmiento A, Zagorski JB, Sindair WF: Fundional bracing of Colles' fiadures: A

prospective study of immobiIzation in supination vs pronation. CM

Orthop 146: t75-183.1980.

Stratford PW, Bïnkley JM, Riddle DL- Health status masures: stategies and

analytic methods for assessing change scores. Phys Ther76: 1 1 OS

1 123,1996,

Turgeon T, MacOermid JC, Roth JH: Test-retest reliability of the NK dextefky

test, J Hand 7Ber 1998-

Upper Extremity Collaborative Group: Measurhg disability and symptoms of the

upper Iimb: a validation study of the DASH questionnaire- Arth Rheum

39: SI 12,1996-

Ware JE, Snow KS, Kosinski M. Gandek 6: SF-36 H ealth Survey Manual and

Interpretation Guide. The Health Institule, New England Medical Center,

Boston. 1993.

Ware JE. Kosinski M, Keller SD: SF-36 Physical and Mental Health Summary

Scales: A User's Manual. (3rd) The Health Institute, New England Medical

Center, Boston, Mass. 1994.

Wright JG & Young NL: A cornparison of different indices of responsiveness. J

Clin Epidembi 50: 239-246,1997a.

Wright JG 8 Young NL: The patient-specific index: asking patients Mat they

want J Bone Joint Surg Am 79: 974983, 1997b-

Patient Popuhtkn

(n=21) Colles fradures; closed redudion evaluated three months after immobilization (Amadio et al W96)

Physical Function

Physical Role

Mentai Health

General Health

Social Function

Emational Role

8odily Pain

Bfigham and Women'r

tunnel instrument)

symptom sevm-ty - - -

Functional Status

Phydcal Capacity M e a s u m 8

Sensation

Dexterity

Motion

Walking and Bending

n=74 patient participathg in a randomized trial on carpal tunnel release who wre 80% satisfied

1 ~ Ï t h their surgery (Katz et al 1994)

Questionnoir, SRM

Hand and Finger 1-16 Fundion

Ami Fundion 10.86

- -

Social Adivity

Support from farnily 1 0.1 8

Arthritis pain 1 0.39

- -

Level of tension

Satisfaction

Health Perceptions

Arthritis Impact

Health Status Components

Ph ysical 1 1.07

- -

Symptom

Social Interaction 1 0.18

Role 1 -0.66

CTS Symptom severity 2.02 scale I CTS Function scale 1 0.96

(II=) Muswloskeletal problems (42% lower extremity. 27% upper extremity, 4% both, 12% repetitive motion disorder, 14% arthritis);(n=21) (Martin et al 1997)

Sensation (SWMF llreshold)

Sensation 0-59 Z-point discrimination

Sf36 Physical Funcîion 0-37 Efîiciemy (MFA

MFA total 10.74 14-06

MFA mobility 1 0.45 1-1-55

MFA housewrk 1 0.56 12-32

MFA leisure 1 0.56 1 2-39

MFA emotional 1 0.64 1 3.11

Sf36 Physical Role 1 0.45 1 MFA total 1 0.74 12-72

Sf-36 badily pain 10-34 1 MFA housewrk

MFA total 1 0-74 1 4.76

MFA emotional 1 0.64 1 3.64

0-56

SF-36 social fwiaim ( 0.32 1

1-56

MFA total 10.74 15.16

MFA family 1 0.26 1 0.62

SRM

(n=113) Hip Arüiroplasty patients; 6 monüis after surgery (Wright 8 Young

Patient Spedic

Hams Hip Score

Pain

Physical Fundion

Pain

N= 46-59 rheurnatoid arthritis patients treated for 3 months (Hurst et al

EQ-5D VAS

EQ-SD Utility

Joint swelling

Joint tender

Disease activity (patient )

ESR

HAQ

n 3 0 patients who had satisfactory results after surgery; pre-op versus one year postop scores (Clnsalata et al 1 997)

L ' i d a t a et al shoulder sale

--

Index of Responsiveness

Global assessrnent

ûther

Pain - -- .

Daily adivities

45 wrkers Hiho had a positive treatment eff&ct and 79 wfio wuld have been expectd to have a treatment Med mer rehabi l itative intervention; patients cornpleted questionnaires at admission and 3-6 weeks later (Beaton et al 1997)

Effbct Size

Sf-36 Pain

Sf-36 Physical funaion

NHP Overall

NHP Pain

NHP Physical fundion

Duke Overall

Duke Pain

Duke Physical function

OHS Overall

OHS Pain

OHS Physical function

SIP Overali

SIP Phvsical function

Patknt Popuktion

n= 38 patients mth low back pain particïpatïng in an randomized trait on the effkacy of continuos motosized traction separated into improved versus noni'mproved; 'Effect size" reporfed HWS change scotdstandard deviation of change score (Beurskens et al 1 996)

J

Roland (0-24)

Roland (standardized to 2-02 scale of 100) I Main cornpliant

Pain last w w k 1 1.58

Table 2. Subjeds with Distal Radius Fractures (ri==)

-

Age

Sex Distribution

lnjured Hand

Dominance

Treatment

Complications

22 males; 37 females

23% cast only; 23% open reductim, Int-fix and ex-fiix; 18% arthroscopic redudion and Irit-fuc and cast; 1 1 % arthroscopic reductim and Int-fDc and ex&; 11 % open reduction and Int* and ex-fix and iliac bone crest gr* 5% closed reduction and int-fix, 5% open redudion and [nt-fix and cast and iliac bone crest graft; 2% ciosed redudion and int-fix and ext-fm

70% none; 12% mild ie resolved without treatment; 1 8% significant complications requiring treatment (ie CTS, newe or tendon injury)

Table 3. Respoiwivenems of Outcome Measures in DRF

PRWE

Pain

Specific Function

Usual Function

TOTAL PRWE

DASH

SF-36

Physical Function

Physical Role

Bodily Pain

Physical Sum.

Im~ainnents

ROM

Grip

Dexterity

TOTAL SCORE

SRM

Figure 1. Score fw AROM

Points

Max Score

O

1

2

3

4

5

6 fi

Extemion 1 Flexion Ulnar Deviation

Radial 1 Pronation Supinatio Deviaüon 1 n

1 points 1 Finger Flexion Score (maximum 4 points)

More than one finger wïth extreme loss of movement, ie-flexion contractures60" andlor extension lags of >4S0

One finger moi extrema loss of movement , ie-flexion contracture >60" and for extension lags of 245'

One or more fïngers with moderate loss of movement, ie-fiexion contradure >30-60° and/or extension lags of 30-45"

One or more fingen with slight los$ of movement, ie.up to 30" in flexion or extension

I 4 1 FUN tight fist with no ~ag Score according to scale above- (When ROM is rot available Rom neutral, subtract W the number of degrees away from neutral ftom available ROM, ie-Extension from 20-40 = 20 degrees-10 degree penalty for distance away from neutral, AROM score given for 10 degrees is 1 point.) TOTAL AROM score = Sum of the above /30-

CHAPTER 8. Ra-cEUNE PREETORS OF PAN AND OlSABlUiY SIX

MONTHS AFTER A MSTAL RAMUS FRACTURE

Introduction

Dista[ radius ftactures (DRF) are common injuries occumng more

frequently than any other fracture (Larsen 8 Lauritsen 1993). While some

fractures are treated with simple cast immobilïzation and normal function is

restored, it has been recognked that fundional limitations are more prevalent

than originally assumed, partiwiarly in intra-articular fractures (Knirk 8 Jupiter

i 986). Much of the fiterature has f-sed on radiographie results and physical

impairment in grip and range of motion-(ROM) as outwme measwes.

Randomized controlled trials have addressed several issues with respect to DRF

including the position of immobilization and the eff8Cfs of extemal f&on as

compared to cast immobilization (Dias et al 1987; Home et al i 990; Howard et al

1 989; Rodriquez-Merchan 1997; Roumen et al 1991 ; Stewart et al 1 984).

However, many treatment approaches are reporteci in the literature in case

series. Thus, the effect of treatment or specific treatment indications has not

been defined.

In addition to understanding how treatment affects outcorne, it is important

to understand how non-treatment factors affect outcorne. Fadors present at

baseline are of particular importance as they may allow dinicians to provide a

more accurate prognosis to their patients a initial consultation. Modifiable

factors such as adequacy of reduction may be useful to guide treatment

Furthemore, important prognostic variables should be wnsidered in the design

and analysis of clinical trials on management of DRF.

Prediction of outwme after distal radius fiadure has ben addressed in a

number of studies. Unfortunatety, the majority of the evidenœ in the literature

wmes from retrospedive studies Men using univariate analyses. Furthemore,

prediction is often addressed as a secondary or minor issue in a case series

analysis. FM Mis reason, followup times are often varÏable and recording of

potential predictors is mt systmaücC S o m of the prïmary advantages of

prospective cohort studies are that dinical measurements and follawup times

can be standardlzed and a priori selection of paentislr predidof variables is

possible- Muftivariate analyses are preferable in these type of studies because

the relative importance of different potential predidors car be ascertained.

A fumer problm in the definition af outcorne is that thete are no

standardized instruments for measurement of outcome. The most commonly

reported U~~ tcome saleu is the Gartiand and Werley (1 951 ) demerit system

which has the physician penalize the outcome score Men he fin& problems

with inovement, strength, radiographic changes, pain or defomity. While the

system is widely used and has helped physicians categorize outcomes with

some consistency, neither the reliability or validity of the system has been

demonstrated. Furthemore, it ha$ been stated that it is generous in assignhg

poor outcomes only to patients w*th severe problems related to their distal radius

fracture (Bradway et al 1989). Other outwmes that are sometimes reporteci

include gnp and range of motion attained or radiographic alignment on a fo l lw

up radiograph.

Despite the limitations in previous studies, it has been frequently reparted

that anatornic factors play a significant role in detemining outcome after DRF.

There remains controversy over wheaier the severity of anatornic interruption at

injury (Altissimi et al 1986; Bora et al 1989; Cooney & Berger 1993; Field et al

1992; McQueen 8 Caspers 1988; Roysarn 1993; Rubinovich 8 Rennie 1983;

Trumble et al 1994) or the residual anatornic defïcit Mer treatment ie quality of

reduction (Bayd 8 Home 1988; Jenkins 8 Mintowt Czyz 1988; McQueen IL

Caspers 1988; Solgaard 1988) is more aucial in detemining outcorne.

In a randomized trial of 150 patients with DRF. managed conservatively

using five different positions for cast immobilizatïon. no eff@ct of cast position

was reported. The authors reporteci that the quality of redmon was highly

dependent on the severity of the original displacement It w s also stated. that if

either dorsal angle or radial shift was correded, shortening tended to diminish.

Another prospective mal was designed to evaluate fundional bracing Mer

CollesJ m u r e s in 243 patients allocated to receive either a conventional

plaster cast, an above-elbow cast-brace or a belowelbow cast-braœ- Again, no

treatment effect vms obsenred. The fumtional result attained by a patient at

three months was related to the seventy of initial displacement (and to a lesser

extent, the anatomic result) (Stewrt et al 19û4). Fundional results wre rated

according to the scale desaibed by Sarmiento et al (1980) as a modification of

the Gartland and Weriey system (Gartland 8 Werley 1951 ) and anatomic results

wre based on the Lidstrom scale (1959). The consistency of treatment and

prospective f o l l m p in these randomized trials is a strength. The la& of

treatment effed makes these dinical trials similar to a ahoct study in that

standardized followup schedules and outcorne evaluations wre performed.

Other randomized trials Mich have found treatment Mects have also

addressed prognostic vanables as a secondary research question. ln a

randomized trial of 43 fractures which had displaced (Rom a total original group

of 1 01 Colles' Mures), patients were ranâomized to either remanipulation and

extemal fmtion or no intervention. M i l e extemal fixation resufted in better

anatomic position, fundion ~ w s no better than the control group. The authors

reported no airrelation between final anatomic and fwidional outcome (Roumen

et al 1 991 ). In contrast to this, another nndomized trial of extemal fixation

versus plaster casting for mrnrninuted Collesy fractures found that extemal

fixation produceci better anafomic and fundional results and that these were

correlated. (Howard et al 1 mg).

It is the perception af same authors h t anatomic factors are not that

important because patients with p r reducüon attain adequate outcornes or

because correlation betvuwn anatomic results and hnctional results are low

(Roumen et al 1991; Stein et al 1990). It has been reporteci that extemal faation

irnproves anatomic results in displaced fractures, but also results in a higher

complication rate, a f a o r HlhiCh may contribute fo the fadc of correiaion

be-n anatomic and findional results in cases treated with extemal fixation

(Schmalholz 1990). There is Hiide vanation in the complication rates reportecl

extemal fixation in case =-es, consequently the role of treatment

complications on outcome is not well unâerstaod. Some authors suggest that

extemal fwation impnwes anatomic results, but report that functional advantages

are minimal or late appearing. (Jenkins et al 1988; Roumen et al 1991 ). Other

authors report both superior anatomic and functional results with extemal

fixation as wmpared to consenrative management (Howard et al 1989). Overall,

most authors agree that extemal fixation improves anatornic alignrnent, but the

functional outcome Mer extemal fwation is more controvenial and may Vary

according to other factors,

Stewart et al (1 9û5) prospeaively studied 235 consenratïvely treated

displaced DRFs. Patient outcome was evaluated using Gartland and Werley's

"funciional outcome scale " at three and six months after fraaure. Accordhg to

this scale, three months after ftacture the functional outcorne was "faif and at

six months it was "goodn (69Oh excellent or good). Severe fractures were defined

as having a dorsal angle of r40°, radial shortening 2 1Omm or loss of radial

angle r 1 5 O . These patients were slower to recover. The anatomical result on a

radiographic outcome scale (Lidstrom 1959) was related to the quality of

redudion. Loss of reduction was a problem with 36 ftadures losing r 20" of

dorsal angulation duting splintage. Predicton of fuidional outcome w r e

evaluated by lwking for signifiant differenœs between subgroups. inaeased

age, being female, poor reduction (ie final anatomic reduction) and greater

fracture disp~acement resulted in poorer outcomes at three months, whereas

only age did so at six months. It was noted that older patients experienced a

greater loss of redudion wîth splintage which may suggest that it was the loss of

anatomic alignment not age that w s responsible for power outcornes in more

aged patients. It wms also stated that involvement of the ulnar styloid or intra-

articular cornminution did not predict outcome. Carpal tunnel syndrome

occumng as a complication of DRF was more Iikely to occur in aged patients

and in those with poor anatomic resuits-

Andher retrospective study observed that involvement of the radiwlnar

joint resulted in a graater incÏdence of pain, loss of supination range of motion

(ROM) and reduced grip strength. An ulnar styloid fracture was not related to

functional results (Roysam 1993). Porter and Stockley (1 987) shidied 1 1 5

patients w i i DRF treated conservatively. Cornminution, intra-artiwlar

involvement, malunion and seleaion for physiotherapy were associated with

poorer grip strength and ROM. A prospective study of 90 consecutive Colles'

fractures treated with closed redudion and plaster immobilizatîon evaluated the

prognostic value of radiographic parameters at baseline, immediataly aiter

reduction and again one week after later. The outcomes of interest were grip

strength and ROM. Radial shortening at baselîne. post-redudion and one week

after reduction wwe significantly related to movement and gnp strength (~0.21-

0.43). Dorsal tilt at one week was relatd to the amount of movement regained.

Othen agree that radial shortening is a predictor of outcome in dispfaced

DRF (Kaukonen et al 1988; Schmalholz 1989). Even minor axial shortening

wÏthout concomitant malalignrnent of the artiwlar surface has been shown to

carry an increased risk of permanent disability (Aro 8 Koivunen i99l).

Stemse logistic regression vms used to detemine predidors of

instability Mer DRF in a study of 267 patients treated with cast immobliization.

Axial shortening had the greatest progmstic pouiier in prediding instability.

Lidstrom's m u r e dassification scale. dorsal compression, cornminution and

age were also &und ta be signïficant predidors (Abbaszadegan et al 7989).

FrykmanJs classification system and radial compression were not related to

instabiiity. These authors calwlated that the pmbability of an acceptable

(radiographic) end result was less than 0.20 if radial shortening was Smm or

more. The impact of radial shortening has been substantiated by biomechanical

studies which have demonstratecl that radial shortening reduœs forearm rotation

capability (Bronstein et al 7997).

Some studies have evalwted long terni outcornes at 10 (Field et al 1992)

or even 30 years af€er fradure (Kopylov et al 1993). These studies support the

fact that articular incongniity and axial compression have a negative impact on

outcorne. Over time, articular incongniity can result in osteoarairitis.

Osteoarthritis ocaimng seven years Mer a DRF showed a significant

correlation to the initial displacement of the fracture and age of the patient

(Overgaard 8 Solgaard 1989).

Fracture classifications are widely used to describe fracture patterns. The

Older classification system has ben shomi to predict outcame 3.5 years after a

DRF for extra-artîcular fiadures treateâ with cast immobilization (Solgaard

1988). However, neither the Frykman nor A 0 fracture classifications were of

any use in predicting dinical outcome (detemiined by questionnaire) in 652

patients evalwted five years after a CollesJ fracture. In this study. haf the

patients had some residual impaiments and 8% had to give up leisure activities

or have work modifications (Flinkkila et al 1998). A study of 96 patients wïth

DRF suggested that six different classification systems (Older's, Castaing's,

Frykrnan's, Gartland's, Lidstrom's and Jenkinrs) were of no prognostic value in

predicting radiographic or functional outcornes or in distinguishing treatment

options (LenoMe et al 1996). m e r s agree that fiacture type does not predid

outcome Mer consetvative treatment (Altissimi et al f 986).

incongniity and radial shorteniog correlated outcorne, even after corr8ding

for the severity of the original injury. The most important factor was total

incongniity. Age did not correlate outcorne- Overall, these studies suggest

that restoration and maintename of anatmic alignmant of bony fiagmenfs is a

deteminant of outcome for fractures irrespective of fradure type or trealment

intervention-

Jupiter and Femandez (1 996) studied operative treatment of DRF and

found that age, tirne to operation, cornomitant injury of the upper extremity.

fracture of the ulnar styloid, radiwlnar index, ulnar angulation, m u r e

classification, cornminution and articular incongniity were not significantly

associated with outcome- The tM, factors found to be associateci with a poorer

outcome were osteoarthritis at followup (24-1 77 months) and loss of volar tilt ie

dorsal angulation-

As the majority of studies thaï investigated prognosis Mer DRF have

tended to fows on radiographic indicators of bony injury and the extent of

anatomic restoration with treatment, little is known about the role of patient

characteristics Trumble et al (1 994) point out that studies in which the

'outcorne measures" rely heavily on x-ray films add to the confusion because

adequate redudion can overshadow less promising funcüonal results. Field et

al (1 992) noted that grïp strength correlated poorly wïth functional score.

Rubinovicb et al (1 983) also found that radiographic results correlated with gnp

and pinch strength, but that regardless of radiologie appearance al1 patients had

similar levels of adivity, pain and motion. These studies suggest that functional

outcornes are not well d-bed by impairment ratings-

Despite the fact that impairment has been the focus of much of the

available literature addressing outcome after distal radius m u r e , the

measurement of irnpaiments Mer wist fiacture have not a h y s been

standardked, Furthemore, the extent to which impairments are related to

patient perwptions of pain and disability has received little attention. We

identifiecl issues in measurement of impairment where methodology

investigations were required. One problem w s that measurement of foream

rotation ROM was found to V a r y axwiderably betiriisen dinicians at different

centres. Documentation of a standardized me- and its associated reliability

was lacking. Therefore, w condudeci a study uutiich established high relia bility

for a defined measurement protocol (Armstrong et al q998). This study is

reported in Chapter 4 of this thesis. It has been suggested that grip endurance

wuld be an important fundional impaiment to measure. Therefore, we

investigated this as a potentially important outcome after distal radius fiadure-

Unfortunately, several methods and instruments were evaluated and as yet none produceci reliable results (MacDermid et al 1999). Therefore, this impaiment

was eliminated ftom hirther study- Another issue that required clarification was

the reliability of computerized hand evaluation systems. Computerized hand

evaluation systems have been introduced into clinical practiœs, including the

HULC, wïih little documentation of their reliability or validity. We established

that reliability of measures af grip strength and forearm rotation was at least as

high as that fw the standard goniorneter and Jamar deviœs (Armstrong et al

1998; MacDerrnid et al T999). These studies are reported in Chapters 4 and 5 of

this thesis. Other impairments in movement and dexterity evaluated in this study

were based on methods show by others to be reliable (Jebson et al 1969;

LaStayo & Wheeler 1994). These studies allowed us to measure mist related

impaiments with confidence in tfw reliability of Our measurement protocols.

The focus on radiographie predictors and impairment outcornes in the

Iiterature to date is a product of the evaluation philosophy of the time period

when these studies wwe conduded. More reœntly, the philosophy on

evaluation has moveâ to incorporate more patient focuseci outcornes, such as

patient-rated pain, function or quality of Ife. Reliable and valid outcome tools

which allow clinicians to obtain standardized patient ratings of mist related pain

and disability have recently been published (Macûemid et al 1998a). With an

instrument to measure wrist pain and disability it is now possible to detemine

factors that predict this outcorne.

The prirnary purpose of this paper was to detemine Mich baseline

patient charaderistics (age, sex, educaüon, secondary gain staw) anUor

radiographie variables (pre-redudion radial shartening , post-redudion radial

shortening, joint involvement ie AU fiadure type) are related to patient rated

pain and disability six months Mer DRF. A secondary purpose was to detemine

to what extent patient rated pain and disability are refated to Hnist impairnent

Methodi,

An OveMew of tfm Pmqmdve Study bsign

This study was a prospective observational study which entered an

inception cohort of patients with fractures of the distal radius at their initial visit

to orthopaedic clinic and evaluated the extent of pain and disability sixinonths

later using the Patient Rated Wrist Evaluation (PRWE). The PRWE evaluation

at six rnonths w s the pfimary outcorne measure and, therefore, the dependent

variable for the stemse multiple regression analysis- A pilot study was

performed on 50 patients to reduœ the number of potential patient

charaderistics. Smoking history, age, sex, energy of injury (three levels), and

dominance of injury were eliminated as potential predidors on the basis of this

pilot study (See Appendk 2). Secondary gain and educational level were

retained as potential predictors for this study because of statistical significanœ

in the pilot study, whereas age and sex w r e preserved due to their importance

as epidemiologic variables. The final list of independent variables analysed as

potential predidws. w r e age, sex, secondary gain, education, pre-reduction

radial shortening, post-redudim radial shortening and A 0 fracture type.

Multivariate stemse regression was used to detect which vafiables were

signifiant predidors Hihile controlling for the ef fes of other independent

variables. Impairment of M s t joint fundion was measured in the parameters of

range of motion (ROM), grip strength and dextenty. A global -st impa i rn t

score was calculateci by weighting three separate wrïst related physical

impairment tests (See Appendk 1). This global wrist impairrnent score was

correlated with the PRWE to detemine the strength of a-atim be-n

measured physical impairnient and patient ceporteci pain and disability.

Def inition of Independent Variables

Independent variabies that described smoking status (yedm and number

of pack years), a 3-point energy of injury scaie and wbther the injury was to a

dominant hand or not w r e ekninated fiom consideration due to their failure to

approach statistical signifcanœ in a pilot study of 50 patients. The study

protocol for the pilot study and the wrrent study were identical. See Appendix 2

for the specific results of this pilot study data analysis.

Based on decisions made fran pilot study and Iiterature review, seven

variables were evaluated in this cohort sludy. The variables undedined below

were entered into a forward-stepwise multiple regression analysis.

1 ) Patients who were involved in a legal case related to their DRF or had a daim

for compensation to the Workefs Safety lnsuranœ Board (VVSIS) were

classfied as having a positive secondarv gain motivation.

2) Acie was expressed in years.

3) Sex was classified as male or female,

4) intra-articular nature of fracture CAO WeJ HRS dassified as nonarticular,

partial artiwlar or fully artiaiiar (Muller et al 1990).

5) Radial shorteninq was measured as described by Warwick (1 993) on the pre-

reduction film to indicate the seveflty of the fracture bony displacement and p s t -

reduction to indicate the extent of anatomic restoration immediately Mer

treatment- The distance from the distal ulnar to distai radial surfaces was expressed in mm,

6) Education m s e x p r e d in a IO-level ordinal scale- It incfuded the follWng

levels: some grade schol, finished grade schwl, same high school, finished

high school, some college/technicaUdiploma program, finished collage,

technicalldiploma program, some university, finished university, some graduate

work at university. finished graduate wwk at university.

Sampie Sire Estimation

The primary question of this study was ans~liered using forniard stepwÏse

multiple regression analysis- Sample size estimation for regression analysis is

difficult due to a la& of accepted and readily available methods. This has been

attriboted to the diffiwlty in assesshg reasonable values for regression

coefficients and their standard emxs ( N o m 8 Streiner 1994). Thus,

researchen Men resort to theoretical arguments related to variablelsubject

ratios or to calwlation based on simple univariate statistics. As a result of the

ambiguity on the best way to estimate sample size for multiple regression

anal ysis, sample size estimation was perforrned using a variety of techniques.

The first approach to sampre size estimation was based on detennining

the sample size required to detect a difference in outcome between patients

and without secondary gain issues, based on a minimally clinically important

difFerenœ of 15 points (from the 100 point PRWE scale) and tha standard

deviations observeci in the pilot data ( ~ 2 0 ) . The calwlation of the sample size

requirement to detect a mean dimerence of 1 5 indicated that a total of 56

sobjects would be required (Robertsson et al 1990)-8

n= K2 a (k- ~ , l l x IV (an) + @-l ml o is the standard deviation of the underlying population. (v, p,) is the minimum

- .

important dinical difference, Q cumulative distribution fundion of the standardked normal deviate (1 -96 for a4.05 and 0.84 for p0.20). Thus the required , n (per group) is [(2 x W)/I 51 x (1 -96 +0.84)&28

The second appraadi ta sample size ca!culation was to estimate Vie

sample size required to detect a minimally important univafiate ~ I a t Ï o n (0.40

or 16% variation explained) (Lachin $981).e To detect a univariate correlation of

least 0.40, (a=0.05, fM.20) 47 subjects wwld be required (Glantz 1 992)-

The third approach to sample size estimation was to calculate the number

of subjeds required to detect a clinicaIly important differenœ in proportions ie

the proportion of patients Mth secondary gain issues who attained a good

outcorne versus the number of patients without secondary gain issues that

attained a good outcome. Differences be-n proportions tends to require

larger sample sizes than differenœs behAieen means. For this reason, a sample

size calailation based on proportions will provide consenrative estimates where

data are continwus. If a differenœ in proportion for patients with a "go&

outcome versus a "poor" outcorne is used as the mode1 and 0.80 venus 0.50 is

assumed to be a clinically significant differenœ in proportions, then 36 subjects

per group (total 72) would be required (Robertssan et al 1990).

The final approach to sample size estimation was to consider general

principles reported in literalure on multiple regression analysis. Age and sex

had been show in pilot analysis to be unrelated to outcorne but were ente&

into the analysis, because of their importance as epidemiologic variables-

Another five variables of cl inical interest (secondary compensation, education.

pre-reduction radial shortening, post-redudion radial shortening and joint

n= 3 + [(4(2, + 2, )2)/ (ln (1 +r)/(l-r))] n= sample size required; Z= score, rcorrelation coefficient. For a=0.05 Z,= 1-96 and for B=0.20 (power 80%) ZB-0.84; when r4.40, F 3 + [4(1.96 +0.84)2 1 (In 2.33)7= 47

l0 "= w1 -2, + @-' *)l r (PI x PJ+ (1 00-P3) 1 (P, 9J21 were p,= proportion of good outcornes in non-secondary gain group and h= proportion of gooâ outcornes in secondary gain group; a=0.05; p0.20; Q wmulative distribution fundion of the standardized normal deviate (1 -96 for a=0.05 and 0.84 for pr0.20) Thus the required , n (per group) is [(80 x 20 + 50 x 50)13d] x (1 -96 +O.-36

involvement) were aloo induded in the analyses. Thus, a total of seven

variables were entered into the stepwïse multiple regression analysà. Using 5

10 subjeds pet variable (entered into the analysis) as a comrnon rule of thumb

for rnultivarïate analyses, (Nman 8 Streiner 1994) a consmative estimate

of sample size requirements was 70 subjeds ( for seven variables).

As the consenrative estimates of sample size requkements suggested

that approximately 70 subjects m e requked for 80% power, this number was selected as the sample size for the study-

Testing Scheâub

The baselinel initial clinic visit occurs within the first week after a fracture

following primary care at an emergency department At this visit, the PRWE

was completed by the patient and dernographic data wwe obtained. Potential

predictors age, sex and educational level were obtained fram this baseline

information sheet (See appendix 5 for data collection forms). At six months

patients were re-evaluated, completing the PRWE, a follow-up questionnaire

on complications and Hlere tested for mist-related impairment-

Rec~itrnent

All patients with distal radius fractures attending the Hand and Upper

Limb Center for primary care ware identified by dinic Iists and attending

phpicians. All patients were enrolled in the outcame evaluation process, unless

they w r e unable to participate due to incampetence. Patients provided

infomed consent for the use of their results in this study. Testing was

incorporated into scheduled dinic visits and was a component of their r o m

care. A total of 70 patients wre tested.

Patients M o failedEurere unable to compry with th& scheduled medical

appointments were contacted by phone to detemine if they could reschedule

240

their appointment Ail patients in this study ampletecl their baseline and six-

month evaluations-

Impaimnt w n g

Patients were tested for physical impairment of the at their sixmonth

visit A global W s t impairment score was composeci of a weighted summation of

ROM, gr@ strength and dexterity scores. Cornparisons were made with the

unaffecteci exbemity or nonnative values-

ROM

ROM measures w r e perfomed on the N-K camputeflzed hand

evaluation systern. Wnst extensionMexion was measured WUMI a dorsal

placement protocol k n m to provide reliable results for dorsal placement of the

goniometer (LaStayo 8 WheeIer 1994). Radial and ulnar deviation was

perforrned accarding to an established protocol (NorkÎn 8 White 1995). One

axis was aligned with the third metacarpal and the other along the radius- The

axis of the goniometer was aligned over the center of the wrist joint in neutral-

Measures of ROM in pronationlsupination had been reporteci, by

clinicians at Our œnter, to be problematic in ternis of reliability- For this reason

a study was cornpleted to establish reliabfe rnethods for measuring foreann

rotation (Armstrong et al 1998)- From this study we w r e able to detemine that

reliable results iiiiere possible for pronationlsupination masures with both

inexperienced or experienced testers as long as a standard measurement

protocol was used. Pronation/ supination was performed by aligning the a m of

the goniornetet wïth the superior side of the am along specific anatomic

landmarks. This landmark w s the proximal wrist uease for supination and just

distal to the ulnar head in pronation. The other goniometer a m was aligned wÏth

the vertical plane. Ali measurements were recorded by f w t switch ~ Ï t h the N-K

and pfinted for data entry. A total sawe out of 30 points was allotted to active

ROM as describeci in Appendix ?.

Grip Strength

Grïp testing was perfomed with the NK Digitgrip device aaording to the

standard protocol desaibed by the Amer in Society for Hand Therapists (Fess

1992). Previws authors have desaibed reliable assesornent of grip strength

scores in normals and patients (MacDemid et al 1994) using this protocol-

Using this protocd and NK Digitgrip deviœ, we abtained high reliability

coefficients (MacDemid et al 1999). A grip strength score out of 40 points was

detemined as a ratio of the uninjured hand with the injured hand and adjusting

for dominanœ (adjustment factor =1.07) (See Appendix 5) (Mathiowetz et al

1985)

Dexterity

The checkers subtest of the Jebson's Hand Function Test was used to

evaluate radial hand dexterity (Jebson et al 1969). The dexterity component

was rated out of 15 points as a ratio of M8Cfed and unaff8Cfed sides and

adjusting for dominance (fa*- 1 -07).

Total Score

A total HBst impairment scare out of 85 points was computed as the total

of the three tests above. Forfurther background and disaission on the

development of an impainnent rating score see Appendix 1.

Descriptive statistics were calwlated fw the dependent variables. the

PRWE and the physical performance scores. The preliminary approach to

investigating the relationships behnieen baseline variables and outcome was to

examine univariate relationships. Pearson correlations between age, radial

shortenhg and education with the PRWE w r e calailated. In addition, analysis

of variance (NOVA) las used to chedc for signifiant differenœs in PRWE

scores between the sexes. between M u r e types and beheen patients

withiwithout sea~ndary gain (Table 2).

The definitive analysis of the relationship between baseline predicton

and s h o n t h PRWE scoces was fowud-$tepwbe mulopre Iinear regression

perfomed in BMDP sofhKare (2R). The F-to-enter was 4.0 and the F-to remove

was 3.9 ( a=0.05) (Table 3).

Because the goal of this study was prediaion of a dincial outcorne as

rated by patients, the focus was on the extent of variation explained by the

resultant model. Baseline factors were expaded to contnbute to outcome but not

fuliy explain it, as treatment interventions and complications are thought to be

important, but not controlled for in this study. Therefore, a meaningfui

contribution for baseline fadon was defined as combuting ~ 2 0 % to the

variation in outcome (ie six-month PRWE scores).

Data quality was verified by manual recheck of the data entry by two

persons. Observation of descriptive statistics for deviant values was also used

to veriry data entry. The stability of multiple regression results was verified by

ninning both forward and bachnrd procedures, by inspection of univariate

versus multivariate relationships and by repeating the analysis in another

statistical program.

Results

Ali subjects completed the PRWE at basdine and followwp. (See Table

1 ) A predominance of more "cornplex" artiwlar fractures was found in the sample

as a result of the fad that the centre is a tertiary care hand centre. (See Table

1). The results of univariate statistical analysis are in Tables 2. Baseline

variables, education and pre-redudion radial shortming were significantly

correlated 4th the skrnonth PRWE score. A slatistically significant difFerenœ

was noted in sixmonth PRWE scores (higher score means more pain and

disability) between patients *th secondary gain as a potential motivation (mean

= 37) versus those vvho were not on either Workers Compensation or involved in

legal action (mean=16).

Forward steWs8 multipb regressiun suggested tha€ secondary gain,

education and preredudion radial shortening were significant predidors of the

PRWE at six months after a DRF. The R-squared for the full mode1 was 0.25.

The total R-squared was comprised of secondary gain which contributed 0.16;

pre-reduction radial shortening which contributeci an additional 0.05 and

education which contnbuted an additional 0.04 to the total Rsquared.

Results from univariate association and multiple regtession analyses

agreed that neither age. sex, A 0 fracture type nor pst-redudion radial

shortening wwe predictors of outcorne. The correlation between sikmonth

patient rated wurisi pain and disability (sixinonth PRWE score) and Wt related

physical impairment at six=months was i--v-50. The individuai subtest

correlations were: grip ~ 0 . 4 6 , ROM r-0.41 and dexterity ~-0.31.

Discussion

Secondary gain and education level were found to be signifiant

predictors of six month pain and disability Mer a DRF in pilot study. These

effects were confimecl in the prïmary study suggesting that the is

consistent Radial shortening was found to be predidive of six month PRWE

scores in the main study, an efkct previowly reported M e n Me outcornes were

physical impairments such as grip or ROM (Eelma 8 McEMesh 1983; Lidstrom

1 959; Older et al 1965; van der Linden & Ericson 1 981 ; Villar et al 1987). The

variables of age, sex, smoking status and dominant injury were found not to be

significant predictors of six-month patient reported pain and disability in pilot

study. That sex and age did not influence ootcome was conflrmed in the main

study and agrws with reports by others flnimble et al 1994). The baseline

factors identifleci in this study all& one to explain 25% of the variation in

PRWE scores at six-months-

The outamie for this study was a newly developed sale for evaluation of

patient rating of wrïst pain and disability. The study population vms the spectnim

of DRF seen by a tertiary care speciaiked hand unit Thus, comparison studies

for which prediction ~ w s the study objedive are lacking. Few studies address

prediction and those that do tend to use retrospective followup wÏth variable

followup times and univariate methods. Tmmble et al (1994) reported an R-

squared of 0.40 for an injury score and 0.70 for total postoperative incongniity in

predicting a global outcome score comprised of ROM, gri@ strength and a pain

score. However, as ROM and strength are only moderately wrrelated to patient

reported pain and disability the study results do not directly apply. Furthemore,

their study addresseâ a select group of DRFs uvith cornminuteci, displaced intra-

articular inwngruity Hiho had failed closed manipulation and were al1 treated with

surgical stabilization and arüailar reconstruction at one centre. In this specific

subgroup of the DRF population, variabifity in a number of fadors including the

extent of soft tissue injury, treatrnent methods, fracture characteristics and

postoperative management were minimized. Thus the specific variables that

affect outcome in this specific subgroup may not be as related to outcome M e n

considering the full spectnim of DRF injury. This study suggests that different

factors may be more important in specifk subgroups of DRF and this is a

possibility wwthy af further investigation.

While 25% of outcome has been accounted for, 75% remains

unexplaineci, We must ask Hkiether this refiects the full extent to Mich baseline

factors influence six-month outcome or this study failed to detirte the tnie

importance of baselifle predictors. We expct that modifications of outcorne

capacity are adiieved by variables that occur after baseline. Treatment selection

and complications are expected to accwnt, in part, for outcomes- However,

there remains the possibility that the Vue importance of baseline fadors was

underestimated in Viis study. This may be pn'marily for two reasons: failure ta

select (or measure) important baseline variables. Som variables such as

dorsal angle w r e not induded in this analysis. The process for selecting

variables was detailed in Chapter 1 and *Il not be repeated here. H~~llliever, it is

entirely possible that the available Iiterature did not provide a sufficient basis to

select the best radiographie variables. Furthemore, some variables that are

known at baseline are not easily quantifieci and measurernent systems are not in

place. Therefore, they Hiare not selected or measured for pndical reasons. This

includes associated injuries th& ocairrd with a DRF fradure. While

conceptually, one might accept that amcornitant injury is likeiy to play some role

in outcome, our ability to detemiine this was limited by ouf inability to measure it-

Future studies wïll be required to address these issues and these are discussed

in the final chapter of this thesis.

A further potential problem in identifying the true role of baseline

variables is problems with measurement error or biases that could have Mected

our results. The use af a cohort study with standardized measurement

procedures minimizes the contribution of measurement error, Measutement

procedures were established through literature review and our own investigative

work on clinical measurement which is included in this thesis. The outcome

measures w r e known to be reliable and valid before they w r e included in the

study design.

Biases such as recall bias and the time dependency of recovery w r e

minimized by use of a prospective cohort study. Selection bias ~ w s rninimized by

inclusion of the study protocol into standard practice at the HULC. As al1 patients

consented for their dinical data to be used for this research study, the study

sample is representative of the patients oean at HULC.

Howver, the HULC is a highly specialized tertiary care unit Thus. these

results should only be generalked to other specialized hand units. The data

cannot be thought to be representative of the community level of care,

particularly in rural areas. The treatment approaches, type of patients and associated injury patterns aiuld Vary considerably betiiiiaen this study sample

and the wmmunity population.

While the majority of studies investigating factors which are prognostic of

outcorne after DRF have focuseci on anatomic considefaüon, this study

illustrated that the rnost infiuential factor a s seoondary gain. That secondary

gain influences outcome will make intuitive sense to many dinicians.

Furthemore, it supports the common practice of separating patients with

secondary gain issues from other patients Men presenting case series or

outcorne studies. This study not only documents the importance of secondary

gain after DRF but shows that it may exceed injury severity as an important

explanatory variable. Previous documentation of the eff-s of seaindary

(financial) gain on outwmes has been reported in fiterature on chronic pain

(Rohling et al 1 995). It might be assumed that chronic pain is a particularly rich

arena for secondary gain opportunists because a ladc of objective evidenœ of

disability is more cornmonplace in this field. However, this study illustrated that

secondary gain also had a strong impact on pain and disability reported Mer a

traumatic injury. Patients with issues of secondary gain reported more than

the pain and disabiiity as did the remainder of the study population. The

stepwïse multiple regression analysis indicated that the Mect of seaindary gain

remained strong even after wntrolling for injury severity.

One of the pnmary debates with respect to prognostication has been

whether anatomic positioning at injury or after reduction is most related to

outcorne. There are studies supprting both sides. Some studies find pre-

reduction radiographs contain influeMial predictors (Aitissimi et al 19û6; Bora et

al 1 989; Cooney 8 Berger 1993; Field et al 1 992; McQueen 8 Caspers 1 988;

Roysam 1993; Rubinovich & Rennie 1983; T m b l e et al 1994; Field et al 1992;

Field et al 1992) suggesting that the severity of the injury is important It should

be noted that injury severity is measured by bony displacement but

encompasses a variety of potmal unddying mechanisrns. It may be that the

associated soft tissue injuries such as nerve. tendon, circulatory and

iigamentous damage are the driving forces behind outmes. As there are no

systems in place for reporting these components of injury, their relative

importance has not been defined. If these soft tissue components of injury are

crucial to outcome, then bony displaœment rnay be a poor surrogate because

the force required to aeate a specific bony displacement Vary depending on

the bone quality and the direction of applied forces. Therafore, soft tissue

injuries will Vary between patients similar bony displaœment Determination

of the relative importance of soft tissue injuries requires methods of measuring

these components of a fracture. Our understanding of impairment and disability

after DRF wmuld be enhanced if this were possible.

It did appear from the present study that severity of injury was more

predictive than adequacy of reduction (represented through post-redudion radial

shortening). A spectnim of redudion may not have b e n observed in this study.

That is, poor reduction may affect outcome, but 'poor redudions" were not

observed in this study due to the use of %est knoliinr" treatment techniques in a

h ighly specialized upper extremity un+ Conversel y, the lack of correlation

between outcorne and post-redudion radial shortening may refled the fad that

even with adequate reduction, soft tissue injuries drive outcome and adequate

reduction of the fracture does not ensure adequate treatment for soft tissue

injury. Another potential explanation for the lack of significant relationship

between outwme and postiedudion radial shortening is that although better

reduction can be attanied by more invasive treatments, these treatment rnay be

associated with additional complications which affect outcorne. There has been

some suggestion that this does ocwr with extemal fmation ie redudion is better,

but there are a greater number of complications *th no evidence uf net

functional superïority (Roumen et al 1991)- However, randomized trials have

also indicated that extemal fixation is superïor to conservative management of

severely displaced fradures in young patients (Howd et al 1989). Future

studies Hihich detemine the effbct of different treatment options for different

fractures rnay help define the relative importance of treafment selection and final

anatomic position attained as determinants of outcorne.

One issue that must be considered h terms of defining predidors,

particularly radiographic predidon, is their reliability. It is known that la& of

reliability tends to reduœ the observed correlation betwen variables even when

a true relationship exists (Fleiss 9 986). Because radiographic classifications

have been $ h m to have relatively low reliability (Kreder et al 1996; MacDmid

et al 1998b). this may explain why this study and others (Altissimi et al 1986;

flinkkila et al 1998; Lenoble et al 1996) have failed to demonstrate a

relationship be-n M u r e type and outcorne. There wuld appear to be

room for improvement in the reliability of these classification systems Men one

considen the wide a m y of classification systems, the inconsistent usage during

cl in ical pradice and the minor emphasis on specific traininglevaluation for

residents on the dÏfFerent classification systems. We attempted to minimize the

effect of low inter-rater reliability in this study by using a single rater and Iïmiting

the classification level to the broader A 0 "type" which has been found to be

more reliable Vian further subdivisions of the A 0 system (Kreder et al 1996;

MacDermid et al 1998b). Future studies in this area should utilize a clearly

defined fracture classification and incorporate traininglevaluation of the

observers. Consideration may be given to moving away from fracture

classifications because of their inherent methodological flaws and towards

direct measurement of displaœment via radiographic measurements.

Measurement error may have contributeci to the non-significant correlations

observed between PRWE scores and pst-redudion radial shortening because

post-redudion radial shortening vms found to be less reliable than pre-reduction

radial shortening (Macûmid et al I998b).

Education was shown to have a positive M8d of six-month outcome after

a DRF in this study. Educaüon has been shown to affect health perception in a

variety of populations (Gudex et al 1996; Heistaro et al 1996). It has also been

shown to Med utility scores in patients rheumatoid arthritis even after

controlling for diseam severity ie variables for pain, disability, disease activity,

mood and side M8ds (Hurst et al 1997). The undeciying mechanisrn of how

education affects perceived pain and disability after a mist fiadure is not k n m .

The effect exists even after controlling for secondary gain and pre-redudion

radial shortening. It may be that education is sefving as a surrogate for

occupational demands, as those with lower levels of education rnay be employed

in jobs which require more manual labor and thus place higher stress on their

wrïsts. Higher education rnay result in lower physical demands during one's

occupation or more opportunity to modify wtxk demands. For example,

educated m e r s may be able to reassign elements of their job that are

physically stressing to others or rnay be able to relocate to new jobs more

easily. Education rnay afhct outcome through an effed on cornpliance with

rehabil itation; ie better education rnay be associated wÏth better compliance with

health instruction. Any, or all, of these fadors rnay contribute to the fact that

subjects with higher levels of education reported less pain and disability six- months after a DRF,

It is notewwVIy that the three baseline fadors found to be significant are

not modifiable fadors and cannot be used to guide treatment decisions at the

baseline clinic visit These three variables have value for assisting with

prognosis at a baseline clinical assessrnent They can also be controlled for in

future clinical trials to inaease the validity andlor efficiency of the clinical trial.

As only 25% of the outcome in DRF is explained by the three variables

detected in this study, further investigation is required to delineate additional

factors that contribute to patient outcornes. Radiographie ciassifications were

shown to have low reliability and predicüve ability. Houtever. it is possible that

other classification systems or refinement of wrrent systems wwld improve the

capabil ity of radiographic classification to predict outcorne- m e r pre-redudion

radiographic variables such as dorsal angle or radial may provide

additional bony Ïnjury information th* rnay be important to outcome. A soft-

tissue injury scale is needed to describe this aspect of DRF injury and to

determine the relative importance ofthis aspect of injury in detemhing outwme.

Furthemore, as the relative efFects of different treatrnents and the indications for

specific treatment hast not k e n defineci. much wwk needs to be done in this

area before the evolution of outwme d e r DRF can be fully appreüated. While

this study has helped define the relative importance of secondary gain, severity

of injury and patient education as predidorr of outcorne. it also illustrates that

many other prospective studies are required to build on this study and more fulfy

delineate the factors that predict outcorne Mer DRF.

Another factor that rnight wntribute to outcome is complications. It may

be that increased fracture severity is associateci with increased complication

rates. Complications occur after badine and thw were outside the scope of

the present study. Future studies should delineate the relative role of fracture

severity and complications. Howver, o n e again a method for defining fraaure

complications and their severity is required to investigate this. It is clear that a

number of variables not investigated in the present study must also contribute to

outcorne because the three factors found to be significant in the present study

explained only 25% of the variation in outcorne.

Many studies on prognosis have foaised on impairment outcornes,

primarily radiographic appearance. ROM and grip strength. The purpose of this

study was to detemine which factors were predidive of patient-reported pain

and disability after a DRF. Consequently, a secondary question that was

addressed was regarding the relationship betwaen obsewed physical

impairment and patient reported pain and disability. Measured physical

impaimnts wre not s û w g predidors of the amount of pain and disability

expected for patients with DRFs. The correlation of ~ 0 - 5 0 indiCates that the

varianœ in pain and disability explained by varianœ in physical impairment is

25%. Thus, measuring physical impairrnent tells the dinician no more about the

patient's perceived pain and disability than m l d estimation based on the

knowledge of the three basdine predidors found in this study (secondary gain

status, p r e r e d ~ ~ o n radial shortening and education levet). This suggests that

the extent to which physical impeu'mient ratiws can be translated ïnto

assumptions about disability is Iimited- This trend has also b e n observeû in

other orthopedic conditions (Bain et al 1995; MacDmid et al 1995). These

findings suggest it is necessary to masure both physical capacityfimpairrnent

and patient perceived pain and disability. Future research shoufd investigate

which of these outcornes determine other milestones such as 'retum to wrk-"

Concl usions

This study detennined that three baseline factors: secondary gain status,

education and pre-redudion radial shortenhg explain 25% of the varianœ in

patient reported pain and disability sixmonois after a DRF. Physical impairnient

at six-months only predided 25% of the varianœ in patÏent reported pain and

disability, suggesting that both should be rneasured to provide a more

comprehensive pictue of patients statu$ aRer a DRF. Further studies are

required to enhanœ our understanding of the relationships between patient

characteristics, injwy characteristics and treatment interventions in outcornes in

patients with DRF.

Abbaszadegan H, Jonsson U, von Sivers Kr Prediction of instability of Colles'

fractures. Acfa O~fhop Scand BO: 646-65û, 1989.

Altissimi M, Antenucci R, Fiacca Cl Mancini GB: Long terni results of

conservative treatmmt of fradures of the dista( radius. Clin Olaiop 206:

202-21 0, 1986-

Armstrong A, MacDennid JC, Chinchalkar S, Stevens RS, King G: Reliability of

range of motion measurments in the e l b a n d foream. J ShouM Ubow

Surg 7: 573580,1998.

Aro HT & Koivunen T: Minor axial shortening of the radius affects outcome of

Colles' M u r e treatment J Hand Surg Am 18: 392398, 1 991.

Bain GI, Roth JH, Pugh D, MacDemid JC: Matched hemiressction interposition

arthroplasty of the distal radioulnar joint J #and Surg Am =A: 944-950,

1995.

Bora RN, Culp RW, Osteman AL. Skiwen T: A flexible ~ Ï s t splint. J Hand Surg

14A: 574575,1989-

Boyd LG & Home JG: The outwme of fradures of the distal radius in young

adults. Injury 19: 97-1 00, 1988.

Bradway JK, Arnadio PC, Cooney WP: Open redudion and intemal fixation of

displaced, comminuted intra-articular fractures of the distal end of the

radius. J Bone JoM Sutg Am H : 839-847.19û9.

253

Bronstein Al, Trumble TE, Tenœr AF: The eff8ds of distal radius ftacture

malalignment on foream rotation: a cadaveric study. J Hand Suri Am

22: 258262,1997-

Carrozzella J & Stern PJ: Treatment of cornminuted dista1 radius fiadures with

pins and plaster. Hand Clin 4 399397,1988-

Catalano LW. Cole RJ, Gelberman RH: Displaceci intra-artiwlarfradures of €he

distal aspect of the radius. Long temi results in young adults after open

reductïon and intemal fixation- J &ne Joint Surg Am 79A: 1 290-1 302.

1997.

Cooney WP 8 Berger RA: Treatment of cornplex fractures of the distal radius-

Combined use of intemal and extemal fwation and arthrosccpic reduaion.

Hand Clin 9: 6û3-612, 1 993-

Dias JJ, Wray CC. Jones JM, Gregg PJ: The value of earfy mobilzation in the

treatment of Colles' M u r e s - J 8one Joint Surg Br 696: 463467, 19û7.

Eelma J 8 McEMesh EC: Colles fractures in young adults. Minn Med 487-490,

1983-

Fess EE: Gnp strength. In Casanova JS: Clinical Assessrnent

Rewrnmendations. Edlion 2nd. American Society of Hand Therapists,

Chicago. 1992.

Field J, Warwick D, Bannister GC, Gibsai AG: Long-term prognosis of displaced

Colles' fracture: a 1 O-year prospective review Injury 23: 529-532.1 992.

Fitoussi F, Ip WP, Chow SP: Treatment uf displaced intra-artiwlar farctures of

the distal end of the radius with plates. J &ne Jont Sutg Am 79A: 1303-

1312,1997.

Fleiss JL: Reliability of Measurement In Fleiss JL: The Design and Analysis of

Clinicai ErperimenfS- John Wiley and Son, Toronto. 3986.

Flinkkila T, Raatikainen 1, Hamalainen M: A 0 and Fry(mianls ciassifications of

Colles' fiadure- No prognostic value in 652 patients evaluated after 5

years. Acta Orfhop Scand 69: 77-81.1998.

Gartland J J 8 Werley CW: Evaluation of healed CollesB fractures. J Bone Joint

Surg Am 895-907.1951.

Glantz SA: Primer of Bio-statistics. Edition 3rd. McGrawHill inc, New York, NY.

1 992.

Gudex Cl Dolan P, Kind P. Williams A: Health state valuations from the general

public using the visual analogue scale. Qua1 l 3 5 Res 5: 521-531. 1996.

Heistaro S, Vartiainen E, Puska P: Trends in seff-rated health in Finfand 1 972-

1992. Prev Med 25: 625-632, 1996-

Home JG, Devane P, Purdie G: A prospective randomited trial of extemal

fixation and plaster cast immobilization in the treatment of distal radial

fractures. J Orthop Trauma 4: 3034.1990.

Howard PW, Stewart HD, Hind RE, Burke FD: Extemal fmtion or plaster for

severely displaced comminuted CollesB M u r e s ? A prospedive study of

anatomical and funaional results- J Bone Joint Surg Br74 : -73, 1989.

Hurst NP, Kind Pl Ruta 0, Hunter M, Stubbings A: Measwing health-related

qua1 ity of life in rheumatoid arthritis: validity, responsiveness and

rel iability of EuroQol (EQ4D)- BE J Rheumatol. 36: 551 -559, 1 997.

Jebson RH, Taylor N, TflBSCfimann RB, Trotter MJ. Howard LA: An objective and

standard- test of hand fiinction. Ardi Phys MW RehabiI 50: 31 1-319,

1 969.

Jenkins NH. Jones DG, Johnson SR, MintOWf-Czyz WJr Extemal fixation of

Colles' Mures. J Bone Jont Surg Br 698: 207-21 1.1987.

Jenkins NH, Jones DG, MintOWf Cryz WJ: Extemai fixation and recovery of

fundion following fiadures of the distal radius in ywng adults. InPIry 19:

235-238, 1 988-

Jenkins NH 8 Mintowt Cryz WJ: Mal-union and dysfuncüon in Colles' fiadure. J

Hand Sua Br 13: 291 -293.1 988-

Jupiter JB, Femandez DL, Toh CL, Fellman T. Ring D: Operative treatment of

volar intra-articular fractures of the distal end of the radius- J &ne Joht

Surg Am 78: 1 81 7-1 828.1 996.

ffiukonen JP, Portas M, Karaharju E: Anatomical results after distal foream

fractures. Ann Chir. GynaeooL TI: 21 -26, 7 988.

Knirk JL & Jupiter JB: Intra-artïcular fiactures of Vie distal end of the radius in

young adults. J Bone JoM Surg Am 68A: 647459.1986.

Kopylov P, Johnell O, Redlund-Johnell 1. Bengner U: Fractures of the distal end

of the radius in young adults: A 30-year f o l l m p . J Hand Surg Br 186:

4549,1993.

Kreder HJ, Hanel DP, McKee M. Jupiter JB, McGillivary G, Swiontkowski MF:

Consistency of A 0 fiadure classification for the distal radius. J &ne

Joint Surg Br 78B: 726-731.1996.

Lachin JM: Introduction to sample size detemination and pmer analysis for

clinical trials, Confrol Clin T a s 2: 93-1 3 3, 1981,

Larsen CF 8 Lauritsen J: Epidemiofogy of acute wîst trauma. Int J EpnlemW

22: 91 l-W6,1993-

LaStayo PC 8 Wheeler DL: Reliability of passive Hinst flexion and extension

goniornettic measurements: a mulücenter study. Phys Ther 74: 1 62-1 74,

1 994.

Lenoble E, Dumontier Cr Goutallier 0, Apoil A- [Fractures of the distal radius

with dorsal displacement: a comparative study of the predictive value of 6

classifications]. Rev Chir Omop Reparatn'cie Appar Mot 82: 3964û2,

1 996.

Lidstrom A: Fractures of the distal end of the radius: A clinical and statistical

study of end results. Acta Olfhop Scand Suppl4l: 1 959.

MacDemid JC, Aiyafi Tl Richards RS: Testietest reliability of static and

endurance gnp strength tests performed on the Jamar and NK

Unpublished wrk, 1999-

MacDemid JC, Bain GI, Rampersaud R, Roth JH: Fadors related to satisfaaion

in review of patients with hemirection arthroplasty or silastic thumb

arthroplasty. JHand T ' 8 : 50-50, 1995.

MacDermid JC, Ktamer JF, Woodbury MG, McFarlane RM, Roth JH: Interrater

reliability of pinch and grip strength measurements in patients with

cumulative trauma disorciers. JHand 7 7 1 ~ 7 : 10-14, 7994.

MacDemid JC, Marignani M, Hildebrandt K, Richards RS, Donner A, Bellamy N:

Reliability of radiographie measures and classification in patients with

distal radius M u r e s . Unpublished wok, 1998b.

257

MacDemid JC, Turgean Tl Beadle M. Richards RS. Roth JH: Patient rating of

wrïst pain and disability: A reliaMe and valid measurement tool. J Orthop

Trauma 12: 577-586, l998a-

Mathiowtt V, -man N, Volland G, Weber K, ûme M. Rogers S: Grip and

pinch strength: Normative data fw adults. Ard, Phys Med Rehabil66: 69-

74, 1985.

McQueen M 8 Caspers J: Colles frachire: does the anatomical result affea the

final fwidion? J Bone Joint Surg Br 7 0 649-651, lm.

Melone CP: Open treatment for displaced artiwlatftactures of the distal radius.

Clin Orthop 202: 1 03-1 1 1 , 1 986.

Muller ME, Nazarian S. Koch P. Schatzker J: The compreshensive clasification

of long bones. Springer-Verlag, New York 7990.

Norkin C & White DJ: Measurement of Joint Motion: A Guide to Goniometry. FA

Davis, Philadelphia. 1 995.

Noman GR 8 Streiner DL: Biostatistics. The Bare Essentials. Edition 1. Mosby,

St. Louis, MO, 1994,

Older TM, Stabler EV, Cassebaum WH: Colles fracture: evaluation and seledion

of therapy. J Trauma 5: 469-476, 1965.

Overgaard S & Solgaard S: Osteoarthritis M e r Colles' fradure. Oifhopedks 12:

413416,1989.

Porter M 8 Stockley 1: Fractures of the distal end of the radius. Intennediate and

end results in relation to radiologie parameten. Clin Oithop 220: 241 - 251,1987.

Robertsson GO, Jonsson GT, Sigurjonsson K Epidemiology of distal radius

fiactures in -land in 1985. Acfa Olthop Scand 6t: 457459, 1990-

Rodriquez-Merchan EC: Plaster cast versus percutamous pin fu<ation for

cornminuteci fractures of the distal radius in patients be-n 46 and 65

years of aga. J Oehop T i m a 11: 212-217,1997-

Rohling ML, Binder LM, LanghinrichsenRohling J: Money matters: A meta-

analytic review of the association between financial compensation and the

experieme of chronic pain. Heaffh Psych 14: 537547.1995.

Roumen RM, Hesp WL, Bruggink ED: Unstable Colles' hadures in efderiy

patients. A randomised trial of extemal fixation for redisplacement J

Bone Joint Surg Br 736: 307.31 1, l99l.

Roysam GS: The distal radioulnar joint in Colles' fractures. J Bone Joint Surg

Br 758: 5û-6û. 1993.

Rubinovich RM & Rennie WR: Colles' M u r e : end results in relation to

radiologie parameters. Can J Surg 26: 361-363, 1 9û3.

Ryu J, Cooney WP. Ashew W. An KN, Chao EYS: Functional ranges of motion

of the mist joint J Hand Surg Am 16A: 49-41 9,1991.

Sarmiento A, Zagorski JB, Sindair WF: Fundional bracing of Colles' fractures: A

prospective study of immoôilzaüon in supination vs pronation. Clin

Orthop 146: I75-183,198O.

Si=hrnalholz k- Closed reredudion af axial compression in Colles' fracture is

hardly possible. Acfa OHhop Scand 60: 57-59. 1989.

Schmalholz k- Extemal skeletal fmtion versus cernent fixation in the treatment

of redislocated Colles' fracture. Clin Orthop 236-24f, 1990.

259

Solgaard S: Fundion after distal radius fracture. Acta Ofthop Scand 59: 3942.

1988.

Steffen T, Eugster T, Jakob RP: Twalve years f o l l w p of fradures of the distal

radius treated with the A0 extemai fixatoc Injury25 Suppl4: SW,

1994,

Stein Hl Volpin G, Horesh 2, Hoerer D: Cast or extemal fmtion for M u r e of

the distal radius- A prospective study af 126 cases. Acta OrfC,op Scand

61: 453456,1990-

Stewart HD, lnnes AR, Burke FD: Functional cast-bracing for Colles' fractures: a

cornparison between cast-braüng and conventional plaster casts. J &ne

Joint Surg Br 666: 749753,1984.

Stewart HD, lnnes AR, Burke FD: Fadors affectin9 outcome of Colles fracture:

an anatornic and fucntional study- injury 16: 289-295, 1985.

Tnimble TE, Schmitt SR. Vedder NB: Factors Meeting fundional outcorne d

displaced intra- arh'cular distal radius fractures. J Hand Surg Am 19: 325

340,1994-

van der Linden W & Ericson R Colles fracturé: how should b displaœment be

measured and how should it be immobilized? J &ne Joint Surg Am 63A:

1 285-1 208,1981.

Villar RN, Marsh Dl Rushton hl, Greatorex RA: Three years mer Colles' fracture:

A prospective review. J Borie JoM Surg 696: 635638,1987.

WaMndc Dl Prothero O, Field J, Bannister G: Radiological measurement of

radial shortening in Colles' Fracture- J Hand Surg Br 188: 50-52. 1993.

Sex

Secondary Gain

Radial shortening mm - initial

Radial shortening mm post redudion

PRWE Score at badine, @=no wrist

pain or disability; 1 W=maximurn

score)

PRVVE Score six months after DRF

-

A (€xtra-artiwlar) 17%

B (Partial Artiwlar) 24%

C (Cornpiete Articuler) 59%

Finished grade saiool6%

Finished High school48%

Finished Col lege 28%

Finished University 40%

Finished graduate school4%

30% male; 70% female

14% Yes; 86% No

52 years (1 5 )

261

Table 2 Univan'ate Association 6etween P m d i c t ~ ~ and Outcorn (PRWE)

Variable

1 Age (years)

Pre-reduction radial shortening (mm)

Post-reduction radial shortening (mm) . t

Secondary Gain - 1 Sex

Pearson r (p values) I 0.03 (NS) 1

0.14 (NS)

DWeren- betwwn meam (p value) 1 - --

fernales 20, maies=19 (NS)

A= 19,8=16, C=21 (NS) 1

Table 3. Stepwïse Mulüpie Umar Regmssion Resub -- - - 1 p value STEP 1 F to enter Standard Coefficient

1 Education

1 Secondary Gain

Pre-reduction radial

shortening - - 1 Post-reduction radial

1 shortening

Variable Entered:

Secondary Gain

1 Education

Pre-reduction Radial

s hortening

Post-reduction radial

shortening

Variable Entered: Pre-

redudion radial

shortening

Sacondary gain: 0.41

(4-4)

Pre-reduction radial

shortening: 0.23 (-50)

R2=0.21

Education

Post-reduction radial

s hortening

Variable Entered:

Education

Secondary gain: 0.39

(4-4)

Pre-redudlon radial

shortening: 0.22 (0.49)

Education: -0.1 7 (0.80)

the independent variables secondary gain (p<0.0001) , prereduction radial

shortening (p<0.0095) and education level (p<O.û456). P values for age, se&

A 0 type, postieduction radial shortening wre > O 3 8 indicating they wre not

significant predictors.

CHAPTER 9: SUMMARY

The primary purpose of this study uias to cietennine which baseline

factors wre predidive of sixmonth patient rated pain and disability in patients

distai radius ftadllre. Because outcome had been largely framed in tetms of

irnpaiments in previous viiwk on distal radius ffadure, there was no outcorne

tool to measure wrïst-related pain and disability reporteci in the literature.

Furthemore, the relationship behrifeen mist impairment and pain and disability

had not been exploreci. Therefore, the preliminary requirement to conduding this

study was the develapment of an instrument to measure mis t .a ted pain and

disability (PRWE). This was accomplished (Macûennid 1996). The PRWE was

shown to be reliable in short and long term retest situations and on difFerent Wst

pathologies(MacDermid et al 1 99ûa). The PRWE ums s h m to be valid

because it w s able to detect clinical change and demonsbated appropriate

relationships to mist impaiments and to dher disabifity measures (even though

these wre not specific to the mist)(Macûerrnid et al 1998a). The PRWE was

shown ta be more responsive than other patient questionnaires in detecting

clinical recovery Mer DRF (MacDermid et al 1999b).

Measurement of wi~t impairment is a traditional focus of evaluation af€er

wrkt injury and was addressed as a secondary outcorne in this thesis.

Deficiencies in rnethodologic aspeds of measuring wrist impairnent were

evaluated as secondary questions in this thesis. Reliable methods of measuring

wrkt impairment in foream rotation (Armstrong et al 1998) and static gnp

strength (Macüemid et al 1999a) m e established. Gflp endurance was not

found to be a reliable measure of impairment (MacDermid et al 1999a). On the

basis of those results, grip endurance was eliminated as an indicator of ninst

impairment. Radiographic assessrnent of DRF is fundamental to dinical pradiœ

and yet the reliability of dassification of fractures was shown to be problematÏc

(MacDennid et a1 1998b). This thesis detmined th& radiographic measures

were more reliable tfmn classification systems- Issues in measurement may limit

the ability of radiographic variables to predict Mure outcornes.

Marty potentiat predidors wwe wnsidered Men designing this study.

Some were eliminated on the basis of lack of substantive support in the

literature. Smoking, age and sex were elimiriated on the basis &a 50 sub-

pilot study (Appendk 2). This pilot study also indicated that secondary gain and

education were significant predidors of sk-rnonths patient raed pain and

disability. A prospective stuây of 70 patients was conducfed using secandary

gain, preieduction radial shortening, post-reduction radial shortening, age, sex,

and fracture type (A0 dassification system) as potential predidors. Again, age

and sex were not significant predidors, whereas secondary gain and education

were. As well, preiedudion radial shortening was a signifiant predictor. The

total variance explained ~ i a s 25%. The most influential predictor las secondary

gain which explaineci 16% of the variance in outcorne- Patients secondary

gain issues reported more than the pain and disability than did those

without secmdary gain.

The correlation between measured physical impairnt (ROM. grip and

dexterity) and patient reported pain and disability so<-months after DRF was only

moderate (~0.50). Because impaiment measures such as grip and ROM were

shown to be highly reliable. the correlations observed suggest that impairment

and disability enwmpass d i f ie rd aspects of outcorne. KnOWIedge of baseline

secondary gain status. radial shortening and ducation provided as much

information on patient reportecl pain and disability six months later as did

rneasuring physical impairment at six montha As impairment and disability are

important, but separate, aimponents of patient status; both shouM be assessed

using reliable and valid scales where comprehensive assessrnent of recovery

after DRF is desired-

Specific Rscommedaüons Bîmd on Saidy Rndings

1. The PRWE is a 15-item questionnaires that a l l ~ patients to rate pain and

disability on 0-1 0 point scales It wes designed wiVi expert input to address

ci inical requirernents in assessrnent of patients with wrist injury. The PRWE is a

re!iabte, valid instrument aiat can be used to assess patient-rated wrist pain and

disabil ity.

2. The PRWE is more responsive than the DASH or SF-36 in detecting clinical

change after DRF. It should therefore, be inoorporated in dinical trials of wrM

interventions, in order to minimize sarnple size requirements. The selection of

an outcome tool for clinical practice depends on a vatiety of considerations, one

of which is responsiveness. The DASH had high responsiveness, can be used

wïth a variety of upper extremity pmblems and is a viable option for evaluating

DRF .

3- Standardization of methodology for assessing M s t impairment is essential.

Specific landmarking can al lw even inexperienced testers to achieve reliable

results for foream rotation ROM measures with a variety of instruments- The

least ermr is introduced if the same tester uses the same instrument on repeated

evaluations,

4. Measures of grip endurance have not been shown to be reliable and should

not be incorporated into clinical practiœ until reliable masurement protocols

have been identifieci. Protocois Hihere multiple repetitions are used to estimate

capacity may enhance reliability- Static grip can be measured reliably either

the traditional Jamar device or the computerized NK grip device.

5. Measures of w*st-related disability and physical impairment are moûerately

related, but separate mponents of outamie, and should both be measured in

evaluation of mist joint surgerylrehabilitation.

6. Radiographic dassificafions demonstrate poor to moderate inter-ater

reliability. The multitude of available scales, the iadc of systematic training or

evaluation of perfixmanœ and the lack of consistent utiiization in clinicai

practice may ccntn'bute to this problem. At present, they are not dependable

methods to desaibe fracture, guide treatment or predict outmmes.

7. Radiographic measures such as radial shortening and dorsal angle are not

routinely doaimentad in dinical pradice, but are reliable rneasures uf fiacture

severity and anatomic position after redudion. Routine usage of these

quantitative measures may enhanœ dinical practice.

8. Secondary gain, pre-reduction radial shortening and education status affect outcome and should be doaimenteci at badine and considered when reporting

outcorne results in cl in id studies,

Questions Raiseâ by the Present Stuûy

Each question anmered in this study raised new questions. The pnmary

purpose of Vie study was to find factors that predided six-month 'outcorne" ie

patient rated pain and disability. This outcome was seleded because of the

importance of patient-rated outcome. For this reason, a measure vms developed

to assess it Previous studies have tended to focus on impairments such as

radiographic appearanœ, strength or movement as outcornes. However, this

study demonstrated that impaiment and disability were only moderately related.

As both are important, future studies will be needed to investigate the predictors

of wrist impaiment Oîher outcomes are important and may have separate

predictors. For example, what determines ability to retum to vuwk? Haw is

retum to wwk related to measured physical impairments and patientieported

pain and disability? AnswerÏng tbese questions wïll help us understand the

effect of DRF from different perspedives and the relationships be-n

impairment, disability and handicap-

As only 25% of the variance in outcorne has been explained by factors

identifid in this study, fumer wwk needs to be done to identify other fadon

that are important Baseline fadors that wre not addressed in the present

study rnay be important. For example, saft tissue injury is suspeded to be

important Hmever, a &-tissue injury sale needs to be developed w that the

role of soft-tissue injury can be defined. m e r fradure severity measures such

as extent of cornminution or dorsal angle also rnay be important predidoro.

Adequacy of reduaion is thougM be a detenninant of outcorne (Eelma 8

McEIfresh 1983; Lidstrom 1959; Older et al 1 965; van der Linden & Er imn

1 98 1 ; Villar et al 1 987). This can be indicated by radiographic indicators

measured from films taken after fracture reduction- Postiedudion position rnay

be important where unacceptable reduction is attained. However, adequate

reduction rnay have been too unifonn in this study to demonstrate the

importance of in adequate reducüon. Furthemore. only one indicator of pst-

reduction position was evaluated in the present study. Other indicators of ps t -

reduction position such as s t e m rnay be important (Cooney et al 1979).

Problems the reliability of radiographic measures is an ongoing concem in

deterrnining the relative importance of post redudion anatomie position as a

determinant of outcame.

Other patient-specific charaderistics rnay be predictors of outcome.

Personality charaderistics and social situations rnay affect outcome, but are

not easily assessed in the clinic. îf one is ïnterested in building an explanatory

model to establish mechanisms that detemine outcorne devefopment, one rnight

wish to include these variables in future studies. Comorbidities rnay also be

important deteminants. Work to define important camarbidities and theïr

impact on outwme rnay provide further Ïnsight into how outcomes are achieved,

Once important baseline fadots are considered, then treatment benefits

and complications as predidors of outcome need to be defined- Furthemore, it

must be determinecl whether complications are predicted by injury severïty or

treatment interventions. Complications reiated to injury are not rnodifÏable,

whereas those related to trement may be. Given the vaiiety of treaùnent

options and the ladc of rïgid miteria fortteatmnt selectionl there exists an

opportunity to enhanœ outcome by detemining which treatrnents are able to

maintain redudion Mth least risk of clinically important complications.

The issue of treatment efficacy for speafic types of fractures needs to be

fumer delineated through randomized trials. While some randornized trials

have been published, there are still many gaps in our understanding of aiteria

for more aggressive interventions and the overall Mect of specific intenrentions.

A few basic surgical concepts on use of extemal fixation have been tested in

randomized trials (Howard et al f 989; Rodrïquez-Merchan 1997; Roumen et al

1 991 ; Schmalholz 1990; Sommerkamp et al 1994); as have a number of issues

regarding the timinglposition of immobilization (Dias et al 1 987; McAuliffe et al

1987). However, many wrrently used treatments have not been evaluated by

randomized trials such as arthrosmpic reduction and open reductionlinternal

fixation. Furthem~~e, the Iiterature on the role of physiotherapy on outcorne

afier DRF is extremely sparse. Very Iittle is k n m about the importance of

therapy after DRF, despite the fa& that a majority of patients are receiving it

The extent to Mich outcome can be reasonably explained Mer DRF is

not clear- The opportunity to enhanœ our understanding with respect to the

evolution of outcome after DRF is great because of the number of avenues

vvhich remain open to explorafion.

Sîxengtlw of th. Prawnt Study

1 Standardized dinical measurements wre formulated through literature

review and formal evaluations. This minimized the extent to which measurement

enor could contribute to failure to detect signifcéint relationships.

2. Followup mas complete. This negated any possible probiems with biases

attributed to loss-to fo l lmp.

3. H ighly skilled. specialized upper exûernity physicians (n=7) were utilized and

are thought to have pmvided %est )aioumu treatrnent Treatment variation

betwen clinicians is k n m to exist but is minimized to some extent in academic

group pradices by the cornmitment to teaching rounds, research etc.

4. A computerized evaluation system was used to measure impairment in gn'p

and ROM allowïng state of the art assessment of physical impairment and

routine calibration. This was thougM to contribute to reduced measurement

error.

5. The prospective design provided a standardized assessment and followup

protocol. Issues with recall bias and time dependency of outcornes were not a

consideration.

6. The incorporation of outcome evaluation into standard dinical practice

rn in imizes selection bias-

7. The development of a reliable, valid instrument for measuring patient rated

pain and disability a l l d this dimension of outcome to be evaluated.

8. A single rater w s used for radiographie evaluations to minimize the effects of

unreliability.

9. Patient characteristics that were found to be significant predictors in the pilot

study were confirmed in the main study-

Limitations of P m Study

1. Generalizability: The study population representativeness is unknown. It is

expeded that tertiary care centres have a preponderance of more difficult

fractures and that this study represents the profile of patients expeded at most

tertiary care centres. Treatment prÏnüples may vary in oaier areas or settings.

Therefore, the results of this study can only be applied to similar dinical .

situations. The distribution of patients might be quite diRerent in rural

community hospïtals. In partiwlar, it is thought that poor reductïon or loss of

reduction may be more prevalent in nonspeüalized centres and amirnunity

hospitals than occurs at the HULC. For this reason, p s t redudion variables or

variables Vlat predfd loss of reducüan may be mare important in these settings

than was found in the present study. HULC provides 'best knm* treatment

which is poorly defined for DRF. However, approaches in wmmunity hospitals

and university teaching centres are expeded to Vary.

2. Many potential predictors were not evaluated in aie present study. While an

R-squared of 25% is thought to indicate variables that are clinically important,

there remains a large unexplained variance. It is impossible to know whether

other baseline variables or modîfying variables such as treatment and

complications are responsible for significant unacaiunted for variations in

outcorne. Furthemore, the role of some aspects of DRF outcome prediction are

dficult to define due to measurement issues. Other aspects of injury may be

Meorized to be important but adequate systems to diagnose or quantify those

characteristics of injury are not in place. For examp te, it is ml y recent

increased utilization of arthroscopy in redudion of DRF, that the high incidence

of Iigamentous injury has been appreciated (Richards et al 1997). The clinicai

significance of these injuries, the natural history of recovery, and treatment

thresholds have not been defined- Systems to diagnose and quantify these

injuries are required before wa can fully defhe their importance in outcome after

DRF,

3. Some radiographic variables demonstrateci p r to moderate reliability.

Therefore, the tnre correlation behiiiean radiographic variables and outcome may

not have k e n establistied. Radiographs are diffiailt to standardize, not only in

ternis of their reading, but afso with respect to the way the radiograph is taken.

Issues such as mgnification and positioning can have appreciable effects on

the measurements achieved or the clarity of the view. This can be further

wmplicated in the post redudion stage Mer8 cas& or other immobilization

devises may obscure adequate viewing. For this reasan, measurement error is

a senous problem when depending on radiographs to represent anatomy. H i l e

anatomy may play a substantial role in outcome, the tnie extent of this

relationship is diwlt to decipher given the inherent timitaüons in radiographie

measurements,

4. Different raten were used for the physical impairnient evaluations. The

contribution of inter-tester enor was minimized through adherenœ to protocols

reported in the literature to be reliable or establishment of reliable protocols as

subquestions within the thesis. However, some measurement emor to due tester

variations was inevitable and rnay have contributeci to reduced correlation

between physical impainnents and patient ratings on the PRWE.

5. The study sarnple included a broad specüum of unselected patients treated at

the HULC. Treatment seledion was uncontrolled both at HULC and in the variety

of rehabilitation settings which patients attended during their recovery.

Prediction vvithin subgroups, where extraneous fadon wre reduced, may be

less problematic. Uncontrolled variables contribute to tandom error Mich can

impede the identification of tme predidor variables. However, an understanding

of what contributes to outcome in the spectnim of DRF is also important-

Directions for Future Research

This study provided some basic answen to questions on the

measurement and detemination of outcorne mer DRF. The process for routine

monitoring of potential predictors and outcames was set in place for this thesis

and continues as an ongoing outcome evaluation process for al1 patients at the

HULC. Many of the questions raised by this study m'Il be addressed in future

studies. More than 250 patients have entered this process. As the sample size

develops, additional questions wiil be addresseâ. Additional basdine predi-

that were not examined in the present study wîll be addressed in Mure studies-

The role of treabnent and complications wïll be investigated in the next cohort of

patients. The evaluation pmœss developed in this thesis also be used for

future randomized trlals on treatrnent

The work inioated as part of this thesis to evaluate clinical methodologies

also continues, A grobaf wist impairment measure needs to be developed mat

quant& important components of mist physical capability into a single

overall score. The wwk describecl in Appendk I lays groundwork, but further

development is required. Deteminhg uihich irnpainnents are most relaed to

fundion is essential, if impairment measuras are to Wed fundional outcornes.

Aitemate rnethods to assess gfip strength enduranœ must be developed and

tested. Then it must be determined if grip enduranœ is more related to fundion

than are the static measures presently used in the clinic. More responsive

methods for assessing dexterity rnusi be identified and incorporatecl into an

impaimient measure. Because of the importance of forearm rotation strength to

wrist hindion, a simple test to assess it should be developed for inclusion in a

M s t impaimient measure. This component of Mure research has been initiated

in that two PhD physical therapists with an interest in the wrist have agreed to

act as co-investigatars.

Ongoing wwk must be done to support the clinical use of the PRWE.

M i l e the PRWE has been accepted as part of Cedaron's outcorne software and

a number of centres are routinely using it, many dinicians do not know about it

or have access to it Answering requests for using it and providing instruction on

its use will be an ongoing proœss- Further research on its measurement

properties wïll be required to define its role in evaluation of odher wrist

pathologies.

The 'thesis that never ends" is an appropriate descriptor. Multiple studies

are being planned or in progrsss to address questions raised by this thesis.

Wnst fractures are cornmon. White they are not fatal, they can result in

prolonged or permanent pain and disability. For this reason, further research

which aids in ou? understanding of Mat predicts and limits impairment and

d isabil ity is warranteci-

References

Armstrong A, MacDerrnid JC, Chinchalkat S, Stevens RS. King G: Reliability of

range of motbn measurements in the elbow and foream. J Should Ubow

Surg 7: 93-580,1998.

Cooney WP, Linscheid RL, Dobyns JH: Memal pin fixation for unstable Colles'

fractures- J Bone Janf Surg Am 6iA: 84e845, t 979-

Dias JJ, Wray CC, Jones JM, Gregg PJ: The value of eady mobilzation in the

treafment of Colles' fiadures- J Bone Joint Sum Br696: 463467,1987.

Eelma J 8 McElfresh EC: Colles fractures in young adults. Minn Med 487490,

1983.

Howard PW. Stewart HD, Hind RE, Burke FD: Extemal fmtion or plaster for

severely displaced comminuted Colles' fradures? A prospective study of

anatomical and functional results. J &ne Joint Surg Br 71: 68-73, 1989.

Lidstrom A: Fractures of the distal end of the radius: A clinical and statisticai

study of end results. Acta Otthop Scand Suppl41: 1959.

MacDennid JC: Development of a scale for patient rathg of mist pain and

disability. J Hand Ther 9: 1 78-1 83, 1 996.

MacDerrnid JC, Alyafi T, Richards RS: Test-retest reliabil ity of static and

endurance grip strength tests performed on the Jamar and NK

Unpublished work, 1999a

MacDemid JC, Donner A, Richards RS, Bellamy N, Roth JH: Respansiveness of

the Sf-36, DASH, patient rated wrist evaluation and physical impairments

in evaluating recovery after a distal radius M u r e . In MacDemid JC:

276

Baseline Predictors of Pain and Disability Six-Months Follmhng Distal

Radius Fracture. University of Western Ontario, London, Ontario. 1999b.

MacDmid JC. Mangnani M, Hildebrandt K, Richards RS, Donner A, Bellamy N:

Reliability of radiographie measures and dassiflcation in patients with

distal radius fractures- Unpublished wrk, 1998b-

MacDennid JC, Turgeon T. Beadie Ml Richards RS, Roth JH: Patient rating of

Vlnist pain and disabiliw- A reliable and valid measurement tool. J Ofthop

Trauma 12: 577-586,199ûa-

McAulifbe TB, Hilliar KM, Coates CJ. Grange WJ: Eady mobifization of Colles'

fractures- A prospective trial. J Borie Joint Surg Br69B: 727-729.1987-

Older TM, Stabler EV, Cassebaum WH: Colles fracture: evaluation and selection

of therapy- J Trauma 5: 469-476, Tg65

Richards RS, Bennett JO, Roth JH, Milne KJ: Arthmscopic diagnosis of intra-

articular soft tissue injuries associateci wïth distal radial fradures. J Hand

Sutg Am 22: 772-776.1997-

Rodrïquez-Merchan EC: Plaster cast versus percuteneous pin f-on for

cornminuteci fractures of the distal radius in patients between 46 and 65

years of age- J Olthop Trauma 1 t : 21 2-21 7.1 997-

Rournen RM, Hesp WL, Bruggink ED: Unstable Colles' ftactures in elderly

patients. A randomised trial of extemal fixation for redisplacement J

Bone Joint Surg Br 738: 307-31 1, 1991.

Schmalholz A: Extemal skeletal fmtion versus cernent fo<ation in the treatrnent

of redislocated Colles' fracture- Clin Orthop 236-241 . 'lm.

Sommerkarnp TG. Seeman M, Silliman J, Jones A, Pattenon S. Walker J, et al :

Dynamic extemal fixation of unstable fradures d the distal part of the

radius. A prospedive, randomked camparison with static extemal fixation.

J &ne Joint Surg Am 76& 1149-1 161,1994.

van der Linden W 8 Ericson R: Colles fiachire: how should its displacement be

measwed and how shwld 1 be immobilized? J &ne Jdnf Surg Am 63A:

4 285-1 28%. W81-

Villar RN. Marsh Dl Rwhton N, Greatorex RA: Three years Mer Colles' fiacture:

A prospective review. J Bone J a n t Surg 698: 635438,1987.

278

APPENDK 1 : TWARDS A PERFORMANCE-6ASED SCALE FOR W S T

IMPAIRMENT

Introduction

Measurement of patient funaional ability after injury has traditionally

encompassed measures of movement and strength. A number af other

impairments such as in sensation, pnich strength and dexterity are less

frequently reporteci. Given that a number of components contribute to total

impairment of the wrïst joint, there is a need to have a simple scwing system

which would aliow clinicians to provide a global score for wïst impairment Such

a scale should provide a numerical rating of mist fundion that indudes

weighted components that cover a spectnim of physical impaiments relevant to

functional ability. A theoretical framewwk was created an impairment scale

using perfonnanœ-based testingbasedon a survey of expert dinicians

(MacDemid et al 1998) and a mode1 proposed by a subcommittee of the

lntemational Wnst Investigators. The suwey indicated that ROM, strength,

fundional tests and sensation were important concepts in the outcome M e r M s t

fracture. The International Wrïst Investigators (IW) had a subcommittee looking

ai development of a M s t M u r e outcome score. After three years of

consensus building, this cornmittee was able to achieve a draft outcome score

(Short 1994). They included ROM. grip strength, grip enduranœ and

complications as components of outcome.

Our approach to development of a wrist score was to evaluate data from

the IWI survey, a proposed outcorne scale Rom the IW and consensus

exercises ~ÏtMth the hand surgeons in o u area to choose potential components of

a physical impairnent sa le for the wrïst- Potentiai tests which could evaluate

these wmponents in a brief test that did mt require expensive equipment w r e

identifid so that the sale might be implemenfed in a variety of dinical

situations. Evidenœ fiorn the Iiterature or out am studies w r e used to make

final decisions on which tests should be induded. The tesk is not complete due

to d-rfficulties establishing brief, yet, reliabfe and vala measrirement protocols

for some aspects of impairnient Aich are thought to be important to Hirist

fundion. The progtess to date is sumrnarized in this paper.

Range of Motion

ROM is an important element of wrist function. From a theoreticai basis.

ROM is an indicator of joint fundion. Suwey data, (van der Linden 8 E~cson

1 981 ) previous mist scales (Gartiand 8 Werfey 1951 ) and the IWI

subcomrnittee (Short 1994) have agreed mat ROM is a primary cornponent of

wrïst status. Standard methods for measuflng wtist flexion/extension,

pronationlsupination and radial and ulnar deviations have been published

(Anonymous 1992; Norich 8 White 1995). Wrist extensiodfiexion is measured

wïth a dorsal or lateral placement protocof and is k n m to provide reliable

results (LaStayo 8 Wheeler 1994). Radial and ulnar deviation also has an

established protocol (Anonymous 1992). One axis is aligned with the third

metacarpal and the other along the radius. The amis of the goniorneter is aligned

over the œntre of the wrist joint in a neutral position. Our pilot studies indicated

that wist deviations wre measured reliabi1ity between clinicians whereas,

forearrn rotation was not Therefore we conduded a study Midi estabJished

that with a standardized protowl both experiencad and inexpriemeci dinicians

could measure foream rotation with standard and computerized goniometer.

This study determineci that reliabte measutes of foream rotation movement

could be attained wiai a standardized goniometer placement Pronationl

supination was perfomied by aligning the a m of the goniometer with the

superior side of the a m along specific anatomic landmarks. This landmark was

the proximai cxease for supination and just distal to the ulnar head in

pronation. The other goniorneter a m was aligried with the vertical plane.

Establishing reliable ROM protocoIs was fundamental to the proœss of

developing an impairnent scale, because of the importance of range of motion

measurement as an Micators of joint fundion-

A ROM grading scale w s baseci on a theoretical framevmrk developed

from articles Hihich studied range of motion used to perfom a variety of

functional adivities and consensus exercises with upper extremity surgeons on

draft scoring systems. The pwpose of the grading scale was to simplify mist

outcome evaluation and minimize the time required to perfom evaiuations by

requiring dinicians to masure only the affecteci side. Our subsequent studies

have suggested that while ROM t responsive in deteding dinical change, the

ordinal scale may be less responsive than reported by Amadio et al (1996)

suggesting that the use of raw data in Mure scales may be beneficial.

Grip Strength

Grip strength measures refiect another aspect of wrist fundion.

Theoretically, grip strength is related to mist fundion because the muscles that

perfon gripping cross the wrist joint and assist with stabilization of the wrïst

during grÏp adivities Grip strength is an indicator of muscle capability, but also

wilI be affeded by nerve dysfiindion or pain. Its importance is refleded in its

common use as an outcome masures in assessrnent af hand and w-st f'unction,

Furthemore, gr@ strength was wnsidered =very important" by the majority of the

surveyed members of IWI (MacOemid et al 1998). Grip strength is one of the

most frequently reported "outcornes" in studies reporting the resuits of DRF

treatment

Previow authors have describeci reliable assessrnent of grip strength

scores in nomals and patients using a protocol recommended by the Amencan

Society of Hand Therapists (MacDennid et al 1994). Using the instruments and

methods desaibed for this stuây, me obtainsd high reliability amfficients both in

patients (hilacDemiid et al 1994) and subjeds without hand pathology

(MacDennid et al 1996).

Grip strength scores based on amparisons with the uninvolved extremity

provide an estimate of the percentage loss in grip strength due to injury. We

incorporated a mectïon -or for the efkct of dominanœ on grip strength

ratios because the tendency towards higher strength on the dominant side would

tend to contribute to error in estimating impairment Men based on unadjusted

unaffected/Mected strength ratios (Mathi-tz et al 1985b; Petemn et al

1989). Grip strength n o m are avaiiabie and are usecl in place of the uninjured

side when there is bilateral pathology (Mathiowetr et al 1984).

Grip Endurance

Grip strength endurance was suggested by the IWI subcommittee (Short

1994) and the surgeons at wr centre to be an important indicator of patient

function. Ability to produœ repeated efforts may be more directly related to

ability to retum to wwk and daily adivlies than is isometnc gnp strength. While

at first glane. grip endurance is theoretical ly related ta mist fundion, there

were issues which cwld limit its ability to masure fundion in patients. Firstly, it

is possible that at low levels of strength patients would be able to produce this

low level of strength for a prolonged time and thus wwld have a nomal score on

endurance despite the fa& that their muscle capability mas far befw normal.

Secondly, as this measure has not been previously reported there is no data on

its relia bil ity or the range of nomal i t - Grip strength endurame impairment

would have to be based on a cornparison utAth an unafFected side. Normal

values would be requirad for cases Hihere pathology was bilateral. Wiou t

information on the reliability or normative values aie practical application of

endurance impairnent is premature.

Grip endurance impairment had a strong theoretical frame~l~rk but

measurement protocols were neither describecl or tested for reliability. For this

reason. we conduded a study whish evalueted several gcip endurance protacols

and compared the reliability of these tests with the more traditional static grip

strength measures (MacDermid et al 1996)- This stuây detemined that Mile

static grip strength tests had high reliability, the endurance tests did not

demonstrate acceptable relia bility. FurthemKlre, enduranœ capability was

highly variable between n o m l individuals, suggesting that a namm range af

normative values -Id be difflcult to establish, Normative data is essential to

outcorne scafes because in cases whem pathology is bilateral or a subsequent

injury ocairs to the other hand as normative values provide a comparator for

establishing impairment Our reliabil ity study suggested that grip endurance

was not a reliable rneasured component of impairment and thus should be

excluded from any global score until such time as reliability and validity can be

dernonstrateci-

Dextenty

Ability to manipulate objeds (dexterity) is a fundional ability dependent on

joint movement, muscle and nerve fundion. The three radial digits which are most

frequently used to manipulate srnall abjects are innervated by the median nerve

which is susceptible to compromise with distal radius fracture (Cwney et al 1 W0).

The IWI survey ranked functional tests as an extremely important component of

outwme rneasurernent (MacDemid et al 1998). A brief test vrhich mas sensitive to

median nerve fundion and ability to manipulate objects was required to provide

content validity to the impainnent $cale. Normative data w s required for

cornparison purposes when bilateral impaiments exist; consequently. a protocol

that met these requirements was identifid from the Iiterature. The "small objectsgU

subtest of the 3ebsonSs hand function test was select& to meet these criteria

(Jebson et al 1 969). After pilot reliability testing this item w s disqualifÏeâ due to

the problems wiai validity and reliability (MaCDennid 1996). The content validity of

the subtest as an indicator af rnedian nerve dexîefïty was compromiseci due the fa&

that presenœ or absence of fingemails had a merked Mec2 on test perfwmance.

The ability to pi& up fiat objects fiom a Rat su^- is faàlitated by having a

fingernail. As this trait does not refiect the content of the trait which is to be

measured and is not necessarily syrnmetïical, then its Mèct on a score was to

create variation unrelatecl to median m e funcüon-

Subsequently, two 0th- subtests of the Jebson's hand fundion test were

evaluated: the checkers and cards subtests (Jebson et al 1969). These tests ~iiiere

evaluated by observation of how the subtest was perfomied and by correlaüm of

subtest scores to wrist disability scores. Validïty of the cards subtest ~ w s

questioned in observing aie behaviow of the M s t M u r e patients who were able

to tum cards with very Iittle foream rotation. This obsenration was substantiated

by low correlations with patient reported disability. These findings suggested that

this subtest was not a good indication of foreann rotation during a functional task

and thus not an indicator of mist pathology impairment

The checkers subtest of the Jeb~on's hand hindion requires manoeuvrïng

four wooden checkers into a stack and thus loss of sensation to the radial digits,

loss of joint motion or muscle funcüon may result in impairecl ability to perform the

task It involves picking up wwden checkers using the fïngertips not the fingrnails

thus avoiding the problems expenenced with the small objects subtest. Acceptable

rel iabil ity and validity (relation to patient reported Wst disability) correlation

coefficients were observed for the checkers subtest It was concluded that the

Jebson's checker subtest was a bnef median-nerve-based functional test with

acoeptable face validity and measurement properties as a saeen for dextetity

problems after distal radius fracture. A functional test of foream rotation was also

thought of be relevant ta an overalf score ofvinist capability, but no appropriate test

was available in the cumnt literature.

Complications

A penalty for complication was proposed by the IW subcommittee as part

of their evaluation process. The view that uias agreed upon mer our consensus

exercises was that any signficant complications would Med impairment and thus

muld show up in d e r aspeds of performance- The designation of 'complication"

is otherwise quite subjective and the number of points of the penalty is difficult to

standardize given the multitude of complications and the range af severity Uiat

could occur within each type- A penalty based on subiedive rating of

complications would reduce the standardization of scorïng procedures and thus is

il! advised, from a meüwdologic standpoint,

St~ctural Criteria

The structural criteria for outcome assessrnt stated in the I WI survey was

that no more than 15 minutes could be dedicated to outcome assesment by most

clinicians (MacDemid et al 1998). That precluded the use of time consuming

clinical tests such as isokinetic testing or observer based functional or dexterïty

tests. It is quite possible that more reliable and responsive test elements could

have be seleded if time was not an important factor as a number of detailed

functional test have been developed for the upper extremity (Apfel 8 Carranza

1 992; Desosiers et al 1994; Ghiselli 1949; Jurgenson 1943; Mathimmtz et al 1 985a;

Stein 8 Yerxa 1 990; T ï n 8 Asher 1 948; Turgeon et al 1 998; van Lankveld et al

1996). The challenge in developing a physical impairment scale is to inciude the

relevant spectnim of physical capacities, a testing procedure that is reliable, valid

and responsive and can be completed in a tirneframe acceptable to clinical pradice

demands.

To develop a total performance-based outwme scale, content validity

requires that a reliable test of foreami rotation fundion be included in the test

battery. Further development will be required before a simple global score

indicating global Wst functional parforniance can be describeci. At present,

285

standardized methods have been established for mi& and fmarm ROM, grip

strength and dexterity screening- A sample form is in the appendices.

Summuy of D.vdopn#ntd Work

1 ) Grip strength is a reliable and valid component of evaluatïon ofmist impainnent

2) ROM meawres are reliable when perfomed by a standardized method and are

a valid component of evaluation of wrïst impairment

3) The Jebson Hand Fundion Test (checkers subtest) provides a simple, yet

reliable, indicatw of dextefity of the radial digits wuhich can be incorporated into a

global index of HRSt impairment,

4) A test which measures rotational capability of the foream must be developed to

provide a more comprehensive global score for overall uirist impairment.

5) Neither a complications penalty or a grip endurance impainnent score currently

has been descn'bed which demonstrates acceptable measurernent properties,

These measures should not be incaporateci into a global rating scale at this tirne.

Anonymous: Clinical Assessment Recommendations. (2) Amencan Society of

Hand Therapists, Chicago. 1992.

Amadio PC, Silverstein MD, l lstrup DM, Schleck CD, Jensen LM: Outcorne afbr

Colles M u r e : the relative responsiveness of three questionnaires and

physical examinaüon rneasures. J Hand Surg Br 21 : 781 -787. 1 996.

Apfel ER & Canaclfa J: Dexterity- In Casanova JS: Clinical Assessments

Recommendafions (2nd) American Society of Hand Therapists, chicago.

7 992.

Cooney WP. Dobyns JH. Linscheid RL: Complications of Colles' fiadures. J

Bone Joint Surg Am 62A: 613-61 9,1980.

Desosiers J, Bravo G, Hebert R, Dutil E, Mercier L: Validation of the box and

block test as a measufe of dextenty of elderiy people Reliability, validity

and noms studies. Ardr Phys AUéd Rehabi17S: 751 -755,1994

Gartland JJ 8 Wedey CW: Evaluation of healed Colles' ftactures. J &ne Joint

Surg A m 33A: 895907,1951-

Ghiselli EE: The Pennsylvannia Bimanual Work Sample. In Buros OK: fhird

Mental Measurements Yearbook Gryptran Press, Highland Park, N.J-

1949.

Jebson RH, Taylor N, Trieschmann RB, Trotter MJ, Howard LAr An objective and

standardized test of hand function. Atch Phys Med RehabiI50: 31 1-31 9,

1969.

J urgenson CE: Extension of the Minnesota Rate of Manipulation Test. J AppI

Psych 27: 164-1 69, 1 943.

LaStayo PC & Wheeler DL: Reliability of passive wist flexion and extension

goniometfic measurements: a multicenter study. Phys Ther 74: 1 62-1 74,

1 994.

MacDemid JC: Development of a scale for patient rating of wrïst pain and

disability. J Hand The? 9: 1 78-1 83, 1996.

MacDmid JC, Alyafi T, Richards RS: Reliability of static and endurance

measures of gr@. 1996.

MacDermid JC, Kramer JF, Woadbwy MG, Mdarlane RM, Roth JH: Interrater

reliability of pinch and grip Samgth measurements in patients with

cumulative trauma disorders- J Hand TherE 104 4,1994.

MacDermid JC, Turgeon T, Beadle M, Richards RS, Roth JH: Patient rating of

wrist pain and disability: A reliable and valid meawrment tool. J Orthop

Trauma 12: 577-586, 1998.

Mathiovuetz V, Kashman N. Volland Gr Weber K, Dow M, Rogers S: GrÏp and

pinch strength: Nomative data for adults. A E ~ Phys Med Rehabü 66: 69-

74,1985b.

Mathiowetz V, Volland G, Kashman N, Weber K: Adult noms for the box and

block test of manual dexterity. Am J Occup Ther 39: 386-391, 1989.

Mathiowetz V, Weber Y Volland G, Kashman N: Reliabilty and validity of grip

and pinch strength evaluations- J Hand Surg Am SA: 222426,1984.

Norkin C & White DJ: Measurernent of Joint Motion: A Guide to Goniometry. FA

Davis, Philadelphia. 1995.

Peterson P, Petrick M, Connor H, Conklin O: Grip Strength and Hand

Dominance: Challenging the 10% Rule. Am J Oocup Ther43: 444-447,

1989-

Short W I W Subcommittee progess on devlopment of a wrist score. 1994.

Stein C 8 Y e m EJ: A test of fine finger dexterity. Am J Occup W U : 499-

504,1990-

288

T i in J 8 Asher EJ: The Purdue Pegboard: Nomis and studies of reliability and

validity. J Appl Psjch 32: 234-247, 9 94û.

Turgeon T, MacDemid JC, Roth JH: Test-cetest reliability of the NK dexterity

test J Hand Ther 1998-

van der Linden W 8 Erïcson R: Colles ftacture: how should its displacement be

measured and h w should it be immobilized? J Borie Joint Sug Am 63A:

1285-1 288, t 981 -

van Lankveld W. van't Pad Bosch P, Bakker J, Terwïndt S. Franssen M, van Riel

P: Sequential Occupational Dexterity Assessrnent (SODA): a new test to

measure hand disability. J Hand Ther 9: 27-32.1 996.

APPENDUC 2: ResuM F m Pilot Study

This study was mnducted as describecf in chapter 8 using 50 patients with DRF.

Table 1. Subject Charadmistics

--

Sex

- - - - - -

Hand Dominance

injury Dominance

Smoking history

Education

WCB - - -

Litigation involved

Received physiotherapy

Average time to retum to work

Patient Reported Pain Oisability

Pain (û-50)

Total Pain and disability (0-1 00)

Distribution of Sample

14 males, 39 fernales

66% fail fram level(34% idsnaw) 25% fàll fhn motion or heigM 9% high energy impact injury

50 Right 3 Left

24 Dominant; 29 Non-dominant

55% never, 23% previous smokers; 23% current srnokers Average pack-years of smokers 18

2% some grade school 4% finished grade s&ml 15% some high school 32% completed high schwl 6% some college/technical 1 5% finished college/technical 9% some university 9% completed university 2% some graduate school 6% finished graduate school

6Or6

76%; Average 1 1 weeks (7) - - . -

1 O weeks (1 3)

Basel ine 1 six&mths

Table 2- Comtation Betnnren Variables And Outcorne

- - - - - - - -- ... - -

injury

Energy of injury

Smoking history

1 Education 4 revels 1 -0.02 1

4-08

4-05 1

4-04

Education (1 0 levels) 43-27'

1 Complication score* 1 0.34' 1 * statistically significant

Weeks of physiotherapy

" Complication score was a cumulative score for al1 complications found from a chart review once patients had cornpleted the study; scores wre assigned according to the follom'ng system: 1 ) Mild (1 point)- Complication noted on chart but no intenrentionldefinitive diagnosis required eg abnomal owelling, mild@ossible RSD (no blocks), CTS symptoms which resolved without testing; 2) Moderate (2 points)- Complication noted on chart, formai diagnostic procedures andlor treatrnent interventions performed eg possiblelmild RSD sent for 1-2 blocks. carpal tunnel symptoms *th nenm conduction studies and spl inting, pin tradc infection resolved with anübiotics, nerve injury with electrodiagnostic confirmation 3) Severe (3 points): Complication noted on chart. formal diagnosis and surgical treatment ie tendon rupture repaired surgically, carpal tunnel release. l o s of reduction with semndary reconstnidive procedure, pin i n f ' i o n Wth removal of hardware course of antibiotic; severe RSD with multiple Mocks and prolongeci ph ysiothera py

0-20

Table 3. ûiirences betwuen subgroup, (p values for ANOVA)

I Group 1

1 Males

I Nondominant injured

I Secondary gain

1 Srnoken

6-montti PRWE

6-monttr 1 PRWE

19-3

35-0

37

20-6

Dominant lnjured

Non-WCB

No secondary gain

Non-srnokers

20-9

18-7

0-73

0-1 7

17

19-3

0-01 *

0-85

Table 4. Results of Backward Stapwise Regr8ssion

Variables [P value

Sex - -- -

Age Sex Interaction

Dominant Ïnjured

Energy of injury

Smoking status

Pack years

Education (1 0 levels)

Secondary Compensation

Variable removed:

Dominant hand injured

Sex - - - - - -

Age Sex Interadion

Energy of injury

Smoking statw

Pack years --

Education (1 O levels)

Secondary Compensation

Variable removed:

energy of injury

0.57

Sex

Age Sex lnteradlan

Smoking statu~

Pack years

Education (1 O levels) - --

Secondary Compensation

Variable removeci:.

smoking

Sex

Age Sex Interaction -

Pack years

1 Education (1 O levels) 1 -3.4 1 -0.34 11-4 10.01

/ Secondary Compensation

rernoved: 1 1

-- - -

Age Sex Interaction -0.61 -1 -1 7

1 Education (1 O levels)

1 Secondary Compensation

Variable removed:

1 Education (1 0 levels)

1 Secondary Compensation

Variable removed:

-

sex interaction

1 Secondary Compensation

FINAL MODEL

Secondary Compensation

APPENDW 3: EfHICS APPROVALS

300

APPENWC I RELEASE TO USE ARTK=LES ACCEPTE0 FûR PUBUCATiON

FAX

APPENDIX 5: DATA COLLECTLON FORMS

is uwdto help us monitor )ioia recowy and f&ms f h t nrqy be rehed to your recovery- rfyou me unswe about hnv to answer a quesfrsfrion. or needhe@jWhgari tk f i ask us for A@- Yom Name Phone Number Date of Birth 43e Your HULC Physician Your S-e Femde O

Which hand do you nomrllp. wrïte witb? RightU Left L7 Which side arc you now mceiviag tnr#ment for? Right O Left

Did vou have an inimy? Yes Ci What is the injury?

How did you injure your arm? What was the date you h@ed it?

No Cl What is your problem? When did it start? (Date)

Which ofthe foiiowing treatments have y u had for vonr nruent ~robkm? Ci Spi& -, For how long'? Was t helpfbi? Yes O No Ci O Cast -, For how long? WasithelpW? YesCl No O O Surgery -+ m e ) Wasit helpfid? YesCi No O Physiotherapy -+ For how long? Was it helpfùi? Yes O No O O Medications -+ For how long? Was it heipfùt? Yes Ci No O Cl Injections + How -y‘? Was it helptùl? Yes Cl No O R ûther treatments?

Do you have othcr problems or injuries with either ofyour arms? Sameann Y e s U Descriibe No Ci Otherarm Yes U &scribe N o 0

Do you take pain medication for your present a m problem? Ci never a occasioaally

ddy O several times a day

Names of pain medications tned: Do you have rny medid problems? Heart problems U Arthritis ff Diabetes Ci Lung problems O Kihey problems 0 Qther q

What are your wrent medications and dosages?

Ifyou are fernale, have you gone through menopause? Yes U No Ci

Do you smoke? Yes O For how marry years'z How many packs pa day?

rquitu H ~ W long ago'I For how inany years did you smoke? How many packs per day did you smoke?

Do you consume akohot? CI never CI oCCaSiody L7 1-6 chinks pa week Ci 7-14 drinLs per week C i 15tdrlliks perweek

What is the highest level ofseho01 thit y00 eompleted? O Some grade school O Finished grade schaal

Some high school 0 Finished high school Cl Some college~tecbnicaVdipIoma program Cl FiniStied coUege/tecbniddiploma program ff Some University Ci Finished university C i Some graduate work at univetSity Cl Fiaished graduate work at &ersity

What occupation?

What is your m e n t employment status? [3 Full-the regularduties O Uaable to work because of injury Cl Part-time reguIar duties O Unable to work for other medical reasom Ci Full-tixne iight duties O Unemployed due to inab ' i to find a job i3 Part-the Light duties O Homemaker

cl Retired

Have you changed your job because of your present problan? Yes Cl No Cl 1s your a m problem a Worlcem Compemmtion Booed Case? Yes O No Ci PendMg O 1s your ann problan a case which mV01ves otigation O? Yes O No Cl Pending 0

307

These questions help us understand your problem and how it affects your daily life. Please answer every question @y circling the best number) based on your condition over the oast week If yoa are unsure about your answer, please give the best answer you caa.

Please rate your ability to per50rm the following activines. It doesn't matter which haad or ami you use, answer based on your ability, regardless of how you perform the task.

4. Prepare a meal 1 2 3 4 5 l

5. Push open a heavy 1 2 3 4 5 1 door

6. Place an object on a 1 2 3 4 5 shelf above your head

7. Do heavy household 1 2 3 4 5 chores (Wash floors or walls)

8. Garden or do yard 1 2 3 4 5 work

9- Make a bed- 1 2 3 4 5

1 0. Cany a shopping 1 2 3 4 5 bag or b r i e f i -

1 1 - Carry a heavy Mect 1 2 3 4 5 (over 10 lbs).

12. Changea l ighhb o v e r h d

-

17, ReaeatiOOal a-es which require W e art (cardpfaying, lmiahg etc.)

18, Recreat id actnrities r n w h i c h y ~ u w ~ - force or impact througb your am, s h d k or hand ( go& hammering, - a-)

22. During the papt week, to wbat extent bas you arm, shouider or hand problem intediireci with your normal d a 1 activities with f&y, niends, neighbours or groups? (Circle number)

Not at AR

1 h

Sfi ! !wY

2

Modtrrrtdy

3

Quite as Bit

L 4

Extrcmdy

5

309

23. hiring the past w e wae you iimïted in your work or other r&ar M y adVities as

PIease rate the seventy of the foiiowinp: symptoms ir

None Mild

a result of your a m , shoulda or hand problem? fCïrcle numbers)

I 24. Arm, shodder or hand pain

Not Limited at All

I

25- Ann, shoulder or hand pain when you performeù any specific activityactivity

26. Tingling (pins and needles) in your arm, shoulder or hand

s h m b Limited

2

I 27. Weakness is your arm, shoulder or hand-

28- StBùess in your am^,

shoulder or hand

Modemtely Limited

3

the Im week (cïrcle numl t

29. During the past week, how much difiiculty have you bad sleeping because ofthe pain in your arm, shoulder or hand? (Circle number)

V ~ V Limitui

4

Unabie

5

No DifIicul ty

30. I feel Iess capable, less confident or less useful because ofmy a m , shoulder or hand problem. (Cirele n u m k )

Mild DifficdQ

Smngb Disagree

1

Modemte Daffcpilty

k

Disagree

2

Neither rgte or Disagrec

3

Agrtt

4

Strongly A l P t

5

me questtons below will help us understand how much dz@cuityyou have had with y w n>rjst in the pasf week.. Y m will be descnbmgyour mer-e wnst symptoms mter the rn iurra on O scale of 0-20. PIease provide an answer for ALL questions. Ifyou did wt perform an activity, phse ESTIMATE the piin or difncuïty you wouid eIpee+

RATE YOUR P M None Worst

At rest 0 1 2 3 4 5 6 7 8 9 1 0

When Lifüng a heavy objet% I O 1 2 3 4 5 6 7 8 9 1 0

When it is a . its worst 1 0 1 2 3 4 5 6 7 8 9 1 0

How ofien do you have pain? 0 1 2 3 4 5 6 7 8 9 1 0 Never Aiw ay s 1

Rare the q f c l E ~ ~ y m experrexperrencedpifonni&g each of t k ifems Iised below, over the hepad week by Cn.chg t k m m k thrrt c d é ~ e x p e r r & ~ ) l o r a mcuïiy on O

scde of 0-10. A zem (0) meamym did not experrexperrience any d@fZc11ity d a îen 0 meam it was so @cuIfyou were unubIe go ab il ut 4n.

Tuni a door knob using my affected 0 1 2 3 4 5 6 7 8 9 1 0

Cut meat using a M e in my affeaed U band

1 Fasten buttons on my shïrt 1 0 1 2 3 4 5 6 7 8 9 1 0

Il Use bathroomtissue withmyafFé*ed 1 O 1 2 3 4 5 6 7 8 9 10

1 B. USUAL ACTIVITIES Rate the amaint of&flar&pu eqeriencedperfomingyuur nsnd acfnrl-ries in

euch of the meas Iisled belaw, over the pî week by crichg the mmtber thor best riemibesyaa drtcufty on O scak of 0-10. By "rcsiul activkes", we mean tk

Y 2. Household work (cleaning, &enance)

3. Work bour job or uraia everyday 1 ° 1 2 3 4 5 6 7 8 9 1 0

INSTRUCTIONS: This survey asks for yourviews about yoar health. This ùiformation d help keep track of how you féd and how well you are able to do your usual activitïes. Aaswer every question by markiag the appropriate box. Ifyou are unsure about how to answer a question, please give the best aoswer you can

Name: Date:

2. Cornoared to one vear -O, how would you rate your heehh in general now? 1) Much better now than one year ago 2) Somewhat better mw than one year ago 3) CI About the same as one year ago 4) a Somewhat worse now than one year ago

3. nie following items are about activities you might do dm@ a typical day. Does vour heaith now LUmt vou in these actïvities? Ifso, how much? (Merk 1 box on each he)

1, Ycs, 2, Yu, ACTlWHES Limited Limitai

A Lot A Little

a) Vigorous activities7 such as mdng, lifting

heavy objects, participahg in strenuous sports

b) Moderate actIvities, such as moving a tabfe,

pushing a vacuum cleaner7 bowling or playing

golf

c) L i g or carrying gr-es

d) CMing severai fJights of stairs

3, No, Not Limitcd Atm

I 3- Continueci -------- H O ~ much does your healîh now iimit ~ o u ?

- 1, Yu,

Limiteâ A Lot

-- - -

1 g) WaNcing more thin a mik I

2- Yes, Limitai A Little

3 - No, Not Limi#ed At AR

the foilowïng proMems with your work or 0th- rrgular daily actiVities as a r d ofvour ~hvsicd heaith? (JWuk 1 box on each he)

II a) Cut down on the amount of time you spent on work or other activities

b) Accompüsbcd k s s than you would Iike

. c) Were limited in the kiad of work or other actnnties

1 d) Had diffïcaity perfiotming the work or other activities

I 5. hiring the past 4 weeks, have you had any of the foiiowiag problems with your work or

other regular daS, actMties as a redt of any ernotiod oroblems (such as feehg depressed or amüous)? (Mark 1 box on each l k )

a) Cut d o m on the inoant of timt ~ O U spent on wo other actMties

6. During the ~ a s t 4 weeks, to wbat extent has your phys id health or emotiod problems mtdaed with your n o r d soàal &-n'es with fmdy, fiends, neighbours or groups? @&k one box)

1) NotataIl 4) 0 Quite abit 2) 0 Slightïy 5 ) a Extremefy

I 7. How nnich bodilv pain bave you had duniig the past 4 weeLs? @fa& one box) 1) None 4) a Moderate 2) Verymild 5) Severe

8. During the past 4 weeks, bw much did pain iaterfere with your no& work (ùicluding both work outside the home and housework)? @îkk one box)

1) 0 Not at all 4) aquiteabit 2) A littIe bit

9. These questions are about how you f d and how thiogs have been with you dUnaq the ~ a s t 4 weeks. For each question, plaise give the one m e r that cornes closest to the way you have been feelug- How much of the time during the past 4 weeks ....

a) Did you feel hii of pep?

b) Have you been a very newous person?

c) Have you felt so d o m in the dumps that nothhg could cheer you up?

d) Have you fek caim and peacem?

Il e) Did you have a lot of

I energy ?

A Litde None

of of the tht Time

Time

10. hiring the past 4 weeks, how much ofthe thne has yourphvsid health or emoîiod problems interfird with your social (iike visitiiig with fiiends, relatives, etc-)? (Mark one box)

1) 0 AIl of the time 4) A W e ofthe time 2) 0 Most of the time 5) 0 None of the the 3) 0 Some of the the

1 1- How TRUE or FALSE is each of the foilowing statements for you? II (Mark 1 box on each he)

a) 1 seem to get sick a little easier than other people

b) 1 am as heaithy as anyone 1 know

-- - - -

c) 1 expect my heaith to get worse

d) Mv health is excellent

WRIST IMPAIRMENT SCORE NAME DATE 1 M M - 3 0 Score according to scale beiow- (When ROM is not avdabie fiom neutd, subtract I/i the number of degrees away nom neutrai f?om avallable ROM, ie. Exteaston h m 20-40 = 20 degrees-10 degree penalty for distance away nom neutmi, AROM score @en for 10 degrees is 1 ph.)

supination

Score

O

1

2

3

4

5

6

Points

O

1

2

3

4

Fmge~ Fiexion Score (maximm 4 points)

More than one hger with extreme los of m o v e m e ie- flexion contracture>60 O ador extension lags 0 f X S O

One fïnger with extreme loss ofmovement , ie. flexion contrachire >60° and /or extension lags of>4S0

One or more fhgers with moderate Loss of movemeat, ie. flexion contracture >3O-6ûo andor extension Iags of304 O

One or more f igas with slight loss of movement, ie. up to 30" in flexion or exteasion

TOTAL AROM score = Sum of the above /30

II GRIP STRENGTB (mionum 40 poùits) - Jimar Position 2 - Average of 3 aials 1) Anécted Si& = Dominaatside score= A n é c t e d ~ x 4 0 =

unanécted grïp x 1-07

2) Anected Si& = Non- ~ S ü k S c o E e = Aaected~xQOx1.07 = u-gnp

1) Mecteci Si& = Domhant Side ScUre = Udected time x 15 = Affixied time x 1-07

AROM: GRIP STRENGTB: CHECKERDEXTERITY:

TOTAL=

cr/e-zimSbbz*