Post on 03-Jun-2020
transcript
1
The trajectory of recovery and the inter-relationships of symptoms, activity and
participation in the first year following total hip and knee replacement
AM Davis1, AV Perruccio2, S Ibrahim3, S Hogg-Johnson3,
R Wong4, DL Streiner5, DE Beaton6, P Côté7, MA Gignac7,
J Flannery8, E Schemitisch9, NN Mahomed10, EM Badley7
1. Division of Health Care and Outcomes Research, Toronto Western Research Institute;
Departments of Rehabilitation Science and Health Policy, Management and Evaluation,
University of Toronto, Toronto, Canada
2. Department of Orthopedic Surgery, Harvard Medical School and OrACORe, Brigham &
Women's Hospital, Boston, MA
3. Institute for Work and Health, Toronto, Canada
4. Division of Health Care and Outcomes Research and Arthritis Community Research and
Evaluation Unit, Toronto Western Research Institute, Toronto, Canada;
5. Departments of Psychiatry, University of Toronto, Toronto, Canada and McMaster University,
Hamilton, Canada
6. Keenan Research Institute, St. Michael’s and Departments of Occupational Therapy and
Rehabilitation Science, University of Toronto, Toronto, Canada
7. Division of Health Care and Outcomes Research and Dalla Lana School of Public Health,
University of Toronto, Toronto, Canada
*Manuscript
Post-Print (final draft post-refereeing)
2
8. Department of Medicine, Division of Physiatry, University of Toronto and Toronto
Rehabilitation Institute, Toronto, Canada
9. Division of Orthopaedic Surgery, University of Toronto and St. Michael’s, Toronto, Canada
10. Division of Orthopaedic Surgery, University of Toronto and The Arthritis Program, Toronto
Western Hospital, Toronto, Canada
Corresponding author: Aileen M. Davis, PhD
MP11-322, 399 Bathurst Street
Toronto, ON, Canada M5T 2S8
Tel: 416 603-5543
Fax: 416 603-6288
Email: adavis@uhnresearch.ca
Running title: Recovery following joint replacement
3
Abstract (249 words) Objective: Primary hip (THR) and knee (TKR) replacement outcomes typically include pain and
function with a single time of follow-up post-surgery. This research evaluated the trajectory of
recovery and inter-relationships within and across time of physical impairments (PI) (e.g.
symptoms), activity limitations (AL), and social participation restrictions (PR) in the year
following THR and TKR for osteoarthritis.
Design: Participants (hip: n=437; knee: 494) completed measures pre-surgery and at 2 weeks, 1,
3, 6 and 12 months post-surgery. These included PI (HOOS/KOOS symptoms and Chronic Pain
Grade); AL (HOOS/KOOS activities of daily living and sports/leisure activities); and, PR (Late
Life Disability and the Calderdale community mobility). RANOVA was used to evaluate the
trajectory of recovery of outcomes and the inter-relationships of PI, AL and PR were evaluated
using path analysis. All analyses were adjusted for age, sex, obesity, THR/TKR, low back pain
and mood.
Results: THR: age 31-86 years with 55% female; TKR: age 35-88 years with 65% female.
Significant improvements in outcomes were observed over time. However, improvements were
lagged over time with earlier improvements in PI and AL and later improvements in PR. Within
and across time, PI was associated with AL and AL was associated with PR. The magnitude of
these inter-relationships varied over time.
Conclusion: Given the lagged inter-relationship of PI, AL and PR, the provision and timing of
interventions targeting all constructs are critical to maximizing outcome. Current care pathways
focusing on short-term follow-up with limited attention to social and community participation
should be re-evaluated.
Key words: knee replacement, hip replacement, recovery, outcomes, path analysis
4
Introduction
Total hip (THR) and knee replacement (TKR) are well-established, effective
interventions for people with moderate to severe osteoarthritis (OA) resulting in large and
significant improvements in pain and functional limitations 1-6. Pain and function outcomes
usually are reported with separate subscales of a measure or measures representing individual
constructs and they are most often reported pre-surgery and at 6 or 12 months post-surgery 1, 6.
The few studies that have looked at recovery over multiple time points have only evaluated a
limited range of outcomes or have used a measure that combines constructs such as physical
function and social and leisure activities 7-9. Additionally, these outcomes have been evaluated
separately and studies have not considered concurrently how, for example, current pain affects
current physical function, future pain and or future physical function as people recover. Among
people with arthritis and other chronic conditions, cross-sectional studies have suggested that
current pain affects current function 10, 11 and, in longitudinal studies, prior pain has been shown
to impact future function 11. Hence, we would anticipate that outcomes such as pain and function
would be inter-related over the course of recovery following THR and TKR. However, as data
regarding the trajectory of recovery are limited, we do not understand the inter-relationships of
the pain and function trajectories. Understanding how and when in the trajectory of recovery
various outcomes affect each other is critical to identifying targets for treatment and maximizing
outcome for people with THR and TKR.
In addition, while the benefits from joint replacement (TJR) typically have been
quantified using standardized patient-reported outcomes that evaluate symptoms (mainly pain)
5
and function in day-to-day activities 12, recent studies indicate that people with arthritis also are
concerned about their return to higher demand activities and participation in social roles, leisure
pursuits and their community interactions 13, 14 (collectively referred to as participation). There is
a paucity of literature evaluating participation in people who have undergone TJR, particularly
using standardized patient-reported outcome measures. To our knowledge, no research to date
has evaluated the trajectory of recovery of participation following THR or TKR or the inter-
relationships of the various outcomes over time.
The purpose of this study was to describe, for the first time, the trajectory of recovery of
symptoms, daily activities and participation individually and to evaluate how these various
outcomes, relevant to people with THR and TKR, influence each other within and over time
during the first year following surgery.
Methods
Study design and setting
This prospective longitudinal study recruited participants between 2005 and 2008 who were
between the ages of 18 and 85 years from four tertiary care centers in Toronto, Canada who were
undergoing primary THR or TKR surgery for OA and subsequent rehabilitation based on a
standardized care pathway (Bone and Joint Health Network at
http://www.boneandjointhealthnetwork.ca/?sec_id=243&msid=3). The over arching model for
care once the decision for surgery is made is based on same day admission for surgery, four day
acute hospital stay with discharge home for 8 visits of home-based therapy over 6 weeks or a
6
three day acute hospital stay with discharge to inpatient rehabilitation for 7 days. Criteria for
inpatient rehabilitation include meeting two of the following: inability to walk one city block or
15 minutes pre-surgery (gait aids are allowed); unstable cardiac disease; or, no social supports.
The rehabilitation care maps are based on daily (acute care) or weekly plans that include
components of assessment, interventions, outcome targets and ‘red flags or warnings that would
require notification of concerns to the surgeon. The care maps are not prescriptive but rather
guidelines within which the rehabilitation staff have the flexibility to meet individual patient
needs.
Those undergoing revision arthroplasty or hemi arthroplasty were excluded. Exclusion
criteria also included joint replacement for trauma or malignancy. Participants required sufficient
fluency in English to complete the self-report questionnaires. Informed consent was obtained in
accordance with the ethics review boards that approved the study at the participating institutions.
Consenting participants completed pre-surgery questionnaires within two weeks prior to
surgery at their pre-admission clinic visit and then at 2 weeks, 1, 3, 6 and 12 months post-surgery
by mail. The proportion completing questionnaires for THR and TKR respectively relative to
these times was: within three weeks pre-surgery: 87.7, 87.4; post surgery 10 to 21 days: 88.0,
84.8; 4-6 weeks: 88.7, 86.0; 12 to 16 weeks: 91.7, 90.2; 24 to 28 weeks: 81.7, 75.4; and, 50 to 54
weeks: 79.1, 75.2. In addition to the standardized patient reported outcome measures described
below, age, sex, obesity based on body mass index (BMI) >30, education, work status (full-time,
part-time or not working), THR or TKR, comorbidity (based on a no/yes response to the listing
on the American Academy of Orthopedic Surgeons questionnaire) 15, 16 and presence or absence
of low back pain were recorded pre-surgery on the self-report questionnaire.
7
Measures
Our choice of outcome measures was guided by the World Health Organization
framework of International Classification of Functioning, Disability and Health (ICF) 17. The
ICF framework is a biopsychosocial model describing human functioning through the capture of
body structure and function, activity and participation in the context of a person’s social and
physical environment. These constructs of the ICF are defined as follows 17. Impairment of body
structure or function is a loss or abnormality in body structure or physiological function
(including mental functions). Activity limitations are difficulties an individual may have in
executing activities. An activity limitation may range from slight to severe deviation in terms of
quality or quantity in executing the activity in a manner or to the extent that is expected of people
without the health condition. Participation restrictions are problems an individual may
experience in involvement in life situations and roles. Personal or environmental contextual
factors may facilitate or hinder performance across ICF constructs.
The Physical impairments construct included the Hip Disability (HOOS) 18 and Knee
Injury (KOOS) 19 and Osteoarthritis Outcome Score pain subscales and the Chronic Pain Grade
20, 21. The HOOS- and KOOS-pain subscales assess the extent of pain during activities such as
‘walking on a flat surface’ and ‘going up and down stairs’. The HOOS-pain and KOOS-pain are
10- and 9-item scales respectively with response options ranging from 0-‘none’ to 4-‘extreme’
and scores are summed. The Chronic Pain Grade measures pain intensity based on the responses
8
to three questions with scores ranging from 1-‘no pain’ to 10-‘pain as bad as it could be’. The
scores are summed.
Impairment of mood included fatigue measured by the Profile of Mood States (POMS)
subscale 22 and the anxiety and depression subscales of the Hospital Anxiety and Depression
Scale (HADS) 23. The POMS is a frequently used measure of mainly mental (as opposed to
physical) fatigue and has been used in studies across a range of chronic conditions 22. The HADS
has been widely used in community-based and outpatient populations 24. Fatigue was evaluated
through 5 items scored 0-‘not at all’ to 4-‘extremely’; scores are summed. The anxiety and
depression subscales both consist of 7 items. There are four response options scored 0-3, some of
which are reverse scored, ranging from none to maximum experience of the item. The total score
is the sum of items scores and ranges from 0-21 where higher scores indicate more anxiety and
depression.
The Activity Limitation construct was captured by measures commonly used in TJR
samples; the HOOS/KOOS Function in daily living subscale (which includes the same items as
The Western Ontario and McMaster University Osteoarthritis Index (WOMAC) 2 function
subscale) and the HOOS/KOOS Function in sport and recreation subscale 18, 19. The
HOOS/KOOS Function in daily living subscale evaluates an individual’s basic mobility and
activities of daily living (e.g., walking on flat ground, rising from sitting, climbing stairs, etc.) in
17 questions with response options ranging from 0-‘not at all difficult’ to 4-‘extreme difficulty’.
The HOOS/KOOS Function in sport and recreation subscale evaluates more demanding
activities (e.g. twisting on a loaded leg, squatting, etc.) in 4 and 5 items respectively with the
9
same response options. For both subscales, scores are summed and converted to a 0-100 score.
Participation Restrictions as a construct were captured by the Late Life Disability
Instrument (LLDI) 25 and a measure evaluating community mobility. For the LLDI, respondents
rate: a) the frequency; and, b) the extent to which they feel limited in their ability to personally
perform 16 socially defined life tasks expected of an individual (e.g. social, leisure, exercise and
household and personal management roles) within a typical social, cultural and physical
environment on a 1-‘completely’ to 5-‘not at all’ scale. Additionally, respondents completed a
community mobility measure, adapted from the Calderdale Rheumatic Disablement Survey 26,
that assessed the extent to which a respondent’s chronic condition limited their mobility or
ability to travel within their community with 4 items scored 1-‘none’ to 5-‘can no longer do’. For
both the LLDI and mobility measures item scores are summed.
In addition to reporting the individual measures, we also created summary measures for
each of impairments (symptoms), activity limitations, and participation restriction constructs. For
ease of comparison, all impairment, activity limitation and participation measures were
transformed to a 0-10 scale with higher scores indicating worse health/more difficulty. Summary
variables were constructed for each of physical impairments (score range 0-20), impairment of
mood (score range 0-30), activity limitations (score range 0-20) and participation restrictions
(score range 0-30) by summing the individual measure scores within each ICF construct.
Analysis
10
Descriptive analyses were conducted for all variables using the Statistical Analysis
System (SAS) v. 9.2 software. Data were checked to identify outliers and examine the
distributional properties using the skewness and kurtosis values. The distribution of all variables
approximated normality. Nevertheless, the final model in our path analysis, evaluating the inter-
relationships of impairments, activity limitations and participation restrictions within and over
time that is described later, was estimated using Maximum Likelihood with a mean adjusted chi-
square test statistic to take account of any small deviations from normality 27.
In addition to reporting the individual outcome measure and construct scores at each time
point descriptively, the data for impairments, activity limitations and participation restrictions
constructs were graphed over time. In this case, the summary scores were converted to a 0-10
scales and graphed relative to zero-centered pre-surgery construct scores. Multivariate repeated
measures analysis of variance was conducted to confirm statistically significant improvements
over time for each construct after testing for model assumptions 28. The model was adjusted for
age, sex, THR/TKR, obesity and low back pain and the pre-surgery status for the given
construct.
Finally, we used path analysis to evaluate how impairments, activity limitations and
participation restrictions influenced each other within and across time, adjusting for age, sex,
THR/TKR, obesity and low back pain. A diagram that summarizes the hypothesized
relationships among the constructs is shown in Figure 1. Based on clinical knowledge we
anticipated that the time of recovery would also vary by construct. We expected earlier
11
improvements in impairments, followed by activity limitations and participation restrictions
improving later post-surgery. If present, this time effect would be reflected in the magnitude of
the path coefficients.
It should be noted that we initially conceptualized impairments to include both physical
impairments and mood. However, initial descriptive and correlational analyses (correlational
data not shown) indicated that physical impairments and mood were likely separate constructs.
As mood demonstrated little change over time and based on prior literature that found mood to
be related to outcomes such as symptoms and function 29, pre-surgery mood was included as a
covariate in our analyses.
Model analyses were conducted using Mplus 5.21 30. As suggested in the literature31,
several indices were examined to determine overall model fit, including Root Mean Square Error
of Approximation (RMSEA), Comparative Fit Index (CFI), Tucker-Lewis Index (TLI) and
Standardized Root Mean square Residual (SRMR). Good fit was supported by: RMSEA ≤0.05
with a 90% upper confidence limit <0.08 and non-significant p-value; SRMR≤0.08; and, CFI and
TLI ≥0.95 32-34. Once we established model fit, this model was the baseline model against which
the testing of the cross-sectional and longitudinal stability of the relationships between the
constructs of physical impairments, activity limitations and participation restrictions was
initiated. Accordingly, as described in the supplementary data, we compared nested models using
the chi-square difference test.
12
The direct maximum likelihood (DML) estimation method 35 was used to handle missing
data, assuming data were ‘missing at random’. Parameter estimates generated using DML are
consistent and efficient 36. Overall, we had 85% complete data for the outcome measures.
Results
The mean age of the sample (n=931) was 64 years (range 31-88) and 60% were female.
Forty-six percent of the sample had more than high school education and 34% were working
either part-time or full-time prior to surgery. Fifty-three percent of participants were obese. The
most frequent comorbidities were hypertension (44%), low back pain (22%), diabetes (12%) and
respiratory disease (9%). Table 1 provides descriptive statistics for the entire sample and for
those with THR (n=437) and TKR (n=494) separately.
As anticipated and is typical of people having THR or TKR, the sample had physical
impairments, activity limitations and participation restrictions as well as impairment of mood
prior to surgery (Table 2).
Multivariate repeated measures analysis of variance demonstrated that there was a
statistically significant time effect (p<0.0001) for all outcome constructs. The largest
improvements occurred through 3 months post-surgery for all ICF constructs although there were
some incremental improvements through twelve months post-surgery as shown in Table 3. The
trajectories of recovery for THR and TKR based on standardized scores are shown in Figures 2
and 3. People with THR had rapid improvement in physical impairment in the first 2 weeks post-
13
surgery; minimal change in mood; more gradual improvement in activity; and, early worsening
of participation with subsequent rapid improvement through 3 months post-surgery. In contrast,
people with TKR experienced gradual improvement in physical impairments over 3 months post-
surgery with ongoing small improvements; little change in mood through the first post-surgical
month and then a very small improvement; little change in activity limitations over the first
month after which there is more rapid improvement; and, worsening of participation restrictions
through the first month with rapid improvement through 3 months and subsequent smaller
incremental gains.
The proportion of total change in physical impairments, activity limitations and
participation restrictions from 3 to 12 months post-surgery was 6%, 20% and 18% respectively
for THR patients. For TKR patients, the proportions were 18%, 28% and 27% respectively for
physical impairments, activity limitations and participation restrictions over the same period.
Mood demonstrated little change and stabilized quickly post-surgery for THR patients while
11% of the total change occurred between 3 and 12 months post-surgery for those with TKR. Of
note, participations restrictions increased through the first month following surgery for both THR
and TKR participants and then began to improve. Additionally, those who were younger, male,
had a THR, were not obese and who did not have low back pain had statistically significantly
better outcomes over time (p<0.02 for all covariates). The exception was for participations
restrictions where age and joint replaced were not statistically significant (details not shown).
We next tested the inter-relationships of physical impairments, activity limitations and
participation restrictions. The hypothesized model did not display overall good fit. However,
after adjusting for covariates (age, sex, THR/TKR, obesity, low back pain and mood) and
14
subsequently, including longer-term effects of activity limitations at 3 months to 12 months post-
surgery and from pre-surgery participation restrictions to each post-operative time, the model
displayed good fit (see supplementary material including table S1).
Testing the time-dependency of the inter-relationships of physical impairments, activity
limitations and participation restrictions, the analyses showed that the relationships between the
constructs persisted over time, although their magnitudes varied. Adjusted for covariates and
with equality constraints from previous activity limitations to current participation restrictions
and from previous physical impairments to current activity limitations, Figure 4 displays the final
model results with standardized coefficients. To summarize, the final model depicted in Figure 4
shows the following patterns:
1) physical impairments improve rapidly and then stabilize over time;
2) activity limitations improve rapidly but more slowly than physical impairments;
3) participation restrictions continue to improve over one year post-surgery although less
so in the later months;
4) participation restrictions improve more slowly than physical impairments and activity
limitations;
5) in the longer term, prior activity limitations influence future activity limitations;
6) pre-surgery participation restrictions influence future participations restrictions; and,
7) change in a construct influences future status of another construct. The negative
coefficients on the diagonals in Figure 4 represent the effect of the change that occurs over time
in one construct on the status of another construct such that the larger the improvement in, for
example, physical impairments between times 1 and 2, the less activity limitations at time 2, etc.
15
Therefore, the ICF constructs were inter-related within time and there also were
simultaneous direct and indirect effects demonstrated among (and between) different constructs
over time. Also, the magnitude of the coefficients demonstrated a lagged time effect for the
constructs as anticipated.
Discussion
To our knowledge, this is the first work that has evaluated participation restrictions as an
outcome, in addition to physical impairment or symptoms and activity limitations, in people with
THR and TKR in the first year following surgery. Additionally, the work for the first time
simultaneously evaluated the inter-relationship of these constructs within and between post-
operative time periods. As we hypothesized, the inter-relationships of the constructs, specifically
physical impairment, activity limitations and participation restrictions, are not stable over time.
The implications of these findings for outcome measurement are significant when examining
change in outcome. That is, the time of measurement needs to be considered, as does the status
and change of other relevant outcomes, when interpreting a given construct. Importantly, our
work also suggests that the type and timing of rehabilitation interventions that address all
relevant constructs are critical for optimizing outcome in people recovering from THR and THR.
Wait time pressures over the past number of years have resulted in many institutions (including
the recruitment sites in this study) adopting standardized care pathways through the continuum
of care (acute care through rehabilitation) to facilitate efficiencies that allow management of
increased surgical volumes. These care pathways tend to focus on the short-term, maximizing
symptom relief, range of motion, strengthening and basic mobility with the majority of people
16
discharged from all rehabilitation services between 6 to 12 weeks post joint replacement 34 37.
Given the later improvements observed in participation restrictions and the incremental ongoing
improvements in physical impairments and activity limitations beyond 3 months post-surgery in
this work, periodic guidance and or changes to home rehabilitation programs beyond these short-
term care pathways should be evaluated in future work to determine if they hasten and further
enhance outcome.
The trajectory of recovery we observed is consistent with what is observed clinically and
confirmed our a priori hypothesis in that physical impairments improve sooner than activity
limitations, although both improve early in the post operative period, whereas participation
restrictions increase immediately post-surgery only starting to show improvement at 3 months
post-surgery. In this sample, we also noted that mood was associated with very early post-
surgery outcomes. We suspect that this reflects the anxiety related to having surgery and this is
supported by others who have found that people undergoing general surgery and knee
replacement experience anxiety related to their surgery 38, 39.
While others have used different patient-reported outcomes and or times of follow-up in the
first year following surgery compared to our study, the recovery patterns we observed for
physical impairments and activity limitations are similar to those reported in the literature.
Bachmeier et al. found that change in WOMAC pain, stiffness and function subscales scores was
largest at 3 months post-surgery and that THR patients had greater pain relief overall than TKR
patients 7. Their results, while reported at three-month intervals in the first year post-surgery,
must be interpreted with caution as the sample attrition over the year of reporting was
17
approximately 50%. Zimmerman reported outcomes at 2, 6 and 12 months in people having
cemented vs. uncemented THR and found little improvement in pain and function after 2 months
post-surgery 8. Kennedy similarly found that most improvement in function occurred between 3
and 4 months post-surgery in people with TKR 9. Their work included the Lower Extremity
Functional Scale, a self-report measure that combines activity limitation and participation
restriction items into a single score, as well as the 6-minute walk test.
All of our analyses included adjustments for age, sex, TKR/THR, obesity, low back pain,
and mood. Consistent with the literature, we found that outcomes were better for people with
THR 7, and that females 29, 40 and those who were older 41, had more comorbidities including
obesity 40, 42 and low back pain 43 generally had poorer outcomes.
In choosing our outcomes, while we used measures that are commonly reported for THR
and TKR patients 12, we deliberately chose measures that represented activity limitations and
participation restrictions as separate entities. The ICF framework itself does not separate activity
and participation 17. However, a number of authors have argued that activity and participation are
distinct and should not be combined 44-46 and still others have demonstrated that activity and
participation are two distinct constructs 47, 48. Although not the intent of this work, our results
also confirm this distinction between activity and participation based on the differing patterns of
recovery and their inter-relationships.
We recognize that there continues to be much debate about the definition of participation
and how the construct should be measured 49. As such, some may criticize the LLDI as a
18
measure of participation restrictions and suggest that the personal and social roles as
operationalized by the measure more closely represent activity limitations, albeit of higher
demand than those of the HOOS and KOOS. However, when this study began, the LLDI was
deemed the best available measure. Additional measures of participation restrictions have since
been developed 14 although they are yet to be used in people with joint replacement. Given the
identified impact of arthritis on participation 14 and the importance of it as an outcome for joint
replacement 13, we would recommend that participation be included as a separate outcome in
future studies of patients undergoing total joint replacement.
The main limitation of this study was that recruitment included patients who had their
surgery in academic, tertiary care centers. This may limit the generalizability of the results for
those who have their joint replacement in community hospitals. However, comparison of pre-
surgery and outcome scores on the WOMAC between patients treated in academic (two of which
were recruitment sites for this current work) and community-based hospitals have demonstrated
no difference 50.
In conclusion, this work reported on the trajectory of recovery in the first year following
THR and TKR and showed that although the greatest improvement in physical impairments,
activity limitations and participation restrictions occurs by 3 months post-surgery, up to 28% of
the total improvement occurs between 3 and 12 months post-surgery depending on the outcome
construct. Additionally, while physical impairments, activity limitations and participation
restrictions are inter-related within and across time, the inter-relationships among constructs are
not stable over time. As such, recovery time and the impact of one outcome on another outcome,
19
need to be considered in interpreting outcome. The results have implications for rehabilitation
following hip and knee replacement. Specifically, provision and appropriate timing of
rehabilitation interventions that target all of outcomes are critical to maximizing outcomes.
Acknowledgements
The authors would like to acknowledge the following orthopaedic surgeons who permitted
access to their patients for study recruitment:
Dr. Allan Gross, Mount Sinai Hospital, Toronto, Canada
Dr. David Backstein, Mount Sinai Hospital, Toronto, Canada
Dr. James P. Waddell, St. Michael’s Hospital, Toronto, Canada
Dr. Jeffery Gollish, Sunnybrook Health Sciences Centre, Toronto, Canada
Dr. Hans Kreder, Sunnybrook Health Sciences Centre, Toronto, Canada
Dr. Roderick Davey, Toronto Western Hospital, University Health Network, Toronto, Canada
This work was written solely by the authors.
Contributions
Contributions of the authors are as follows:
Conception and design: Davis, Badley, Streiner, Gignac, Ibrahim, Hogg-Johnson, Perruccio, Beaton, Flannery, Schemitisch, Mahomed Analysis and interpretation of the data: Ibrahim, Hogg-Johnson, Perruccio, Davis, Badley, Streiner, Gignac, Beaton, Côté, Flannery, Schemitisch, Mahomed Drafting of the article: Davis, Ibrahim, Perruccio, Hogg-Johnson, Wong Critical revision of the article for important intellectual content: Davis, Badley, Ibrahim, Hogg-Johnson, Perruccio, Streiner, Beaton, Gignac, Flannery, Schemitisch, Mahomed
20
Final approval of the article: Davis, Badley, Streiner, Gignac, Ibrahim, Hogg-Johnson, Perruccio, Streiner, Wong, Beaton, Gignac, Flannery, Schemitisch, Mahomed Provision of study materials or patients: Schemitsch, Mahomed Statistical expertise: Hogg-Johnson, Ibrahim, Perruccio Obtaining of funding: Davis Administrative, technical, or logistic support: Wong, Davis Collection and assembly of data: Wong, Ibrahim
Dr. Aileen Davis, adavis@uhnresearch.ca, assumes responsibility for the integrity of the work as
a whole, from inception to finished article.
Role of the funding source
This work was supported by an operating grant (number 77518) from the Canadian Institutes of
Health Research. The funding sponsors had no role in the conduct, interpretation or
dissemination of this work.
Competing interests
None of the authors have any competing interests in relation to this work.
21
References
1. Fortin PR, Clarke AE, Joseph L, Liang MH, Tanzer M, Ferland D, et al. Outcomes of
total hip and knee replacement: preoperative functional status predicts outcomes at six
months after surgery. Arthritis Rheum 1999;42:1722-8.
2. Bellamy N, Buchanan P, Goldsmith CH, Campbell J, Stitt LW. Validation study of
WOMAC: a health status instrument for measuring clinically important patient relevant
outcomes following total hip or knee arthroplasty in osteoarthritis. J Ortho Rheumatol
1988;1:95-108.
3. Chang RW, Pellisier JM, Hazen GB. A cost-effectiveness analysis of total hip
arthroplasty for osteoarthritis of the hip. Jama 1996;275:858-65.
4. Harris WH, Sledge CB. Total hip and total knee replacement (2). N Engl J Med
1990;323:801-7.
5. Laupacis A, Bourne R, Rorabeck C, Feeny D, Wong C, Tugwell P, et al. The effect of
elective total hip replacement on health-related quality of life. J Bone Joint Surg Am
1993;75:1619-26.
6. Judge A, Cooper C, Williams S, Dreinhoefer K, Dieppe P. Patient-reported outcomes one
year after primary hip replacement in a European Collaborative Cohort. Arthritis Care
Res (Hoboken) 2010;62:480-8.
7. Bachmeier CJ, March LM, Cross MJ, Lapsley HM, Tribe KL, Courtenay BG, et al. A
comparison of outcomes in osteoarthritis patients undergoing total hip and knee
replacement surgery. Osteoarthritis Cartilage 2001;9:137-46.
8. Zimmerman S, Hawkes WG, Hudson JI, Magaziner J, Hebel R, Towheed T, et al.
Outcomes of surgical management of total HIP replacement in patients aged 65 years and
22
older: cemented versus cementless femoral components and lateral or anterolateral versus
posterior anatomical approach. Journal of Orthopaedic Research 2002;20:182-91.
9. Kennedy DM, Stratford PW, Riddle DL, Hanna SE, Gollish JD. Assessing recovery and
establishing prognosis following total knee arthroplasty. Phys Ther 2008;88:22-32.
10. Sale JE, Gignac M, Hawker G. The relationship between disease symptoms, life events,
coping and treatment, and depression among older adults with osteoarthritis. J Rheumatol
2008;35:335-42.
11. Hawker GA, Gignac MA, Badley E, Davis AM, French MR, Li Y, et al. A longitudinal
study to explain the pain-depression link in older adults with osteoarthritis. Arthritis Care
Res (Hoboken) 2010.
12. Riddle DL, Stratford PW, Bowman DH. Findings of extensive variation in the types of
outcome measures used in hip and knee replacement clinical trials: a systematic review.
Arthritis Rheum 2008;59:876-83.
13. Rastogi R, Davis AM, Chesworth BM. A cross-sectional look at patient concerns in the
first six weeks following primary total knee arthroplasty. Health Qual Life Outcomes
2007;5:48.
14. Gignac MA, Backman CL, Davis AM, Lacaille D, Mattison CA, Montie P, et al.
Understanding social role participation: what matters to people with arthritis? J
Rheumatol 2008;35:1655-36.
15. Sangha O, Stucki G, Liang MH, Fossel AH, Katz JN. The Self-Administered
Comorbidity Questionnaire: a new method to assess comorbidity for clinical and health
services research. Arthritis Rheum 2003;49:156-63.
23
16. Hunsaker FG, Cioffi DA, Amadio PC, Wright JG, Caughlin B. The American academy
of orthopaedic surgeons outcomes instruments: normative values from the general
population. J Bone Joint Surg Am 2002;84-A:208-15.
17. WHO, Organization. WH. International Classification of Functioning, Disability and
Health. Geneva., World Health Organization. 2001.
18. Nilsdotter AK, Lohmander LS, Klassbo M, Roos EM. Hip disability and osteoarthritis
outcome score (HOOS)--validity and responsiveness in total hip replacement. BMC
Musculoskelet Disord 2003;4:10.
19. Roos EM, Roos HP, Lohmander LS, Ekdahl C, Beynnon BD. Knee Injury and
Osteoarthritis Outcome Score (KOOS)--development of a self-administered outcome
measure. J Orthop Sports Phys Ther 1998;28:88-96.
20. Von Korff M, Ormel J, Keefe F, Dworkin S. Grading the severity of chronic pain. Pain
1992;50:133-49.
21. Von Korff M, Jensen M, Karoly P. Assessing global pain severity by self-report in
clinical and health services research. Spine 2000;25:3140-51.
22. O'Connor PJ. Evaluation of four highly cited energy and fatigue mood measures. J
Psychosom Res 2004;57:435-41.
23. Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta Psychiatr Scand
1983;67:361-70.
24. Herrmann C. International experiences with the Hospital Anxiety and Depression Scale--
a review of validation data and clinical results. J Psychosom Res 1997;42:17-41.
24
25. Jette AM, Haley SM, Coster WJ, Kooyoomjian JT, Levenson S, Heeren T, et al. Late life
function and disability instrument: I. Development and evaluation of the disability
component. J Gerontol A Biol Sci Med Sci 2002;57:M209-16.
26. Badley EM, Tennant A. Impact of disablement due to rheumatic disorders in a British
population: estimates of severity and prevalence from the Calderdale Rheumatic
Disablement Survey. Ann Rheum Dis 1993;52:6-13.
27. Finney SJ, DiStefano C. Non-normal and categorical data in structural equation
modeling. Greenwich, CT, Information Age Publishing 2006.
28. Misangyi VF, LePine JA, Algina J, Goeddeke F, Jr. The Adequacy of Repeated-
Measures Regression for Multilevel Research: Comparisons With Repeated-Measures
ANOVA, Multivariate Repeated-Measures ANOVA, and Multilevel Modeling Across
Various Multilevel Research Designs. Organizational Research Methods 2006;9:5-28.
29. Jones CA, Beaupre LA, Johnston DW, Suarez-Almazor ME. Total joint arthroplasties:
current concepts of patient outcomes after surgery. Rheum Dis Clin North Am
2007;33:71-86.
30. Muthen LK, Muthen BO. Mplus: Statistical Analysis with Latent Variables User's Guide.
5th Edition. Los Angeles, CA, Muthen & Muthen 1998-2009.
31. Bollen KA, Long JS. Introduction. In K. A. Bollen & J.S. Long. Testing structural
equation models. Newsbury Park, CA, Sage Publications 1993.
32. Hu LT, Bentler PM. Fit indices in covariance structure modeling: sensitivity to
underparameterized model misspecification. Psychological Methods 1998;3:424-53.
33. Hu LT, Bentler PM. Cutoff criteria for fit indexes in covariance structure analysis:
conventional criteria versus new alternatives. Structural Equation Modeling 1999;6:1-55.
25
34. Browne MW, Cuceck R. Alternative ways of assessing model fit. In K. A. Bollen & J.S.
Long. Testing structural equation models. Newsbury Park, CA, Sage Publications 1993.
35. Enders CK. Applied missing data analysis. New York, The Guilford Press 2010.
36. Allison PD. Missing data techniques for structural equation models. Journal of Abnormal
Psychology 2003;112:545-57.
37. Bone and Joint Canada.,. http://www.boneandjointcanada.com/?sec_id=336&msid=3; .
38. Moerman N, van Dam FS, Muller MJ, Oosting H. The Amsterdam Preoperative Anxiety
and Information Scale (APAIS). Anesth Analg 1996;82:445-51.
39. Lingard EA, Riddle DL. Impact of psychological distress on pain and function following
knee arthroplasty. J Bone Joint Surg Am 2007;89:1161-9.
40. Bourne R, Mukhi S, Zhu N, Keresteci M, Marin M. Role of obesity on the risk for total
hip or knee arthroplasty. Clin Orthop Relat Res 2007;465:185-8.
41. Jones CA, Voaklander DC, Johnston DW, Suarez-Almazor ME. Health related quality of
life outcomes after total hip and knee arthroplasties in a community based population. J
Rheumatol 2000;27:1745-52.
42. Krushell RJ, Fingeroth RJ. Primary Total Knee Arthroplasty in Morbidly Obese Patients:
a 5- to 14-year follow-up study. J Arthroplasty 2007;22:77-80.
43. Naylor JM, Harmer AR, Heard RC. Severe other joint disease and obesity independently
influence recovery after joint replacement surgery: an observational study. Aust J
Physiother 2008;54:57-64.
44. Nordenfelt L. Action theory, disability and ICF. Disabil Rehabil 2003;25:1075-79.
45. Jette AM, Haley SM, Kooyoomjian JT. Are the ICF Activity and Participation
dimensions distinct? J Rehabil Med 2003;35:145-9.
26
46. Badley EM. Enhancing the Conceptual Clarity of the Activity and Participation
Components of the International Classification of Functioning, Disability, and Health.
Social Science & Medicine 2008;66:2335-45.
47. Pollard B, Dixon D, Dieppe P, Johnston M. Measuring the ICF components of
impairment, activity limitation and participation restriction: an item analysis using
classical test theory and item response theory. http://www.hqlo.comt/content/7/1/41.
48. Jette AM, Tao W, Haley SM. Blending activity and participation sub-domains of the ICF.
Disabil Rehabil 2007;29:1742-50.
49. Dijkers MP. Issues in the conceptualization and measurement of participation: an
overview. Arch Phys Med Rehabil 2010;91:S5-16.
50. Gandhi R, Tso P, Davis A, Mahomed NN. Outcomes of total joint arthroplasty in
academic versus community hospitals. Can J Surg 2009;52:413-6.
27
Figure Captions Figure 1. Hypothesized model of the inter-relationships within and among physical
impairments, activity limitations and participation restrictions within and across time in
people with primary hip and knee replacement.
PI=Physical impairments; AL=Activity limitations; and, PR=Participation restrictions
Figure 2. Trajectory of recovery in the year following total hip replacement (n=437) for
physical impairments, mood, activity limitations and participation restrictions.
Constructs have been graphed as change relative to a zero-centred pre-surgery score. All
constructs are standardized 0-10 scores where 0 represents less of the construct such that
negative change represents improvement. Time 0=pre-surgery; times 1, 2, 3, 4, 5 represent 2
weeks and 1, 3, 6 and 12 months post-surgery respectively.
PI=Physical impairments; MI=Mood impairments; AL=Activity limitations; and,
PR=Participation restrictions
Figure 3. Trajectory of recovery in the year following total knee replacement (n=494) for
physical impairments, mood, activity limitations and participation restrictions.
Constructs have been graphed as change relative to a zero-centred pre-surgery score. All
constructs are standardized 0-10 scores where 0 represents less of the construct such that
negative change represents improvement.
Time 0=pre-surgery; times 1, 2, 3, 4, 5 represent 2 weeks and 1, 3, 6 and 12 months post-surgery
respectively.
28
PI=Physical impairments; MI=Mood impairments; AL=Activity limitations; and,
PR=Participation restrictions
Figure 4. Final model of the inter-relationships of physical impairments, activity limitations
and participation restrictions within and across time.
The model is adjusted for covariates of age, sex, hip versus knee replacement, obesity, low back
pain and mood. Values on pathways are completely standardized coefficients.
Time 0=pre-surgery; times 1, 2, 3, 4, 5 represent 2 weeks and 1, 3, 6 and 12 months post-surgery
respectively.
PI=Physical impairments; AL=Activity limitations; and, PR=Participation restrictions
29
Table 1: Description of Participants Pre-surgery with Total Hip (THR) and Total Knee
(TKR) Replacement
Hip
(n=437) Knee (n=494)
Total sample (n=931)
Mean age, sd (range) 63, 12 (31-86) 65, 10 (35-88) 64, 11 (31-88)
Sex: female 240 (55%) 321 (65%) 561 (60%)
BMI: 25-29.9 >30 Missing
148 (34%) 153 (35%) 136 (31%)
188 (38%) 222 (45%) 84 (17%)
336 (47%) 375 (53%) 220 (24%)
Education: some university 228 (53%) 199 (40%) 427 (46%)
Lives alone 91 (20%) 121 (25%) 212 (23%)
Paid work (FT/PT)* 161 (37%) 155 (31%) 316 (34%)
Comorbidity: Cardiac Hypertension Lung disease** Diabetes Low back pain
26 (6%) 240 (55%) 61 (14%) 74 (17%) 114 (26%)
35 (7%) 173 (35%) 25 (5%) 35 (7%) 89 (18%)
61 (6%) 413 (44%) 86 (9%) 109 (12%) 203 (22%)
*Paid work (FT/PT) includes those working full-time or part-time ** includes chronic obstructive lung disease and asthma
30
T
able
2: T
otal
hip
and
kne
e re
plac
emen
t out
com
e m
easu
re d
escr
iptiv
es o
ver
time
Pr
e-su
rger
y Po
st-s
urge
ry
Mea
sure
mea
n (s
d)
2 w
eeks
m
ean
(sd)
1
mon
th
mea
n (s
d)
3 m
onth
s m
ean
(sd)
6
mon
ths
mea
n (s
d)
12 m
onth
s m
ean
(sd)
T
otal
gro
up (n
=931
)
Ph
ysic
al Im
pair
men
ts (P
I)
(0-2
0)
11.3
(3.8
) 7.
6 (4
.1)
5.9
(3.8
) 3.
9 (3
.4)
3.5
(3.5
) 2.
8 (3
.2)
Impa
irm
ent o
f Moo
d (M
I)
(0-3
0)
9.6
(5.3
) 9.
6 (5
.4)
7.8
(5.2
) 6.
0 (4
.9)
6.0
(5.1
) 5.
8 (4
.9)
Act
ivity
Lim
itatio
ns (A
L)
(0-2
0)
14.2
(2.9
) 13
.5 (2
.6)
11.8
(3.0
) 9.
1 (3
.8)
8.2
(4.2
) 7.
6 (4
.2)
Part
icip
atio
n R
estr
ictio
ns
(PR
) (0-
30)
12.9
(5.0
) 17
.6 (5
.5)
13.7
(5.6
) 7.
8 (5
.1)
6.8
(4.9
) 6.
4 (4
.8)
TH
R (n
=437
)
Ph
ysic
al Im
pair
men
ts (P
I)
HO
OS
pain
(0-1
0)
5.1
(1.8
) 2.
7 (1
.7)
2.0
(1.4
) 1.
3 (1
.3)
1.2
(1.4
) 1.
1 (1
.3)
Chr
onic
Pai
n G
rade
(0-1
0)
6.1
(2.2
) 3.
0 (2
.1)
1.9
(1.7
) 1.
3 (1
.6)
1.2
(1.6
) 1.
0 (1
.5)
Tot
al P
I (0-
20)
11.2
(3.6
) 5.
8 (3
.6)
3.9
(2.8
) 2.
7 (2
.7)
2.4
(2.8
) 2.
1 (2
.7)
Impa
irm
ent o
f Moo
d (M
I)
POM
S fa
tigue
(0-1
0)
4.1
(2.6
) 3.
8 (2
.3)
2.68
(2.1
) 2.
2 (2
.1)
2.3
(2.2
) 2.
1 (2
.0)
HA
DS
anxi
ety
(0-1
0)
3.0
(1.9
) 2.
2 (1
.7)
1.9
(1.6
) 1.
6 (1
.7)
1.7
(1.7
) 1.
8 (1
.7)
HA
DS
depr
essi
on (0
-10)
2.
6 (1
.7)
2.6
(1.8
) 1.
9 (1
.5)
1.5
(1.4
) 1.
3 (1
.4)
1.3
(1.4
) T
otal
MI (
0-30
) 9.
7 (5
.2)
8.5
(5.0
) 6.
5 (4
.6)
5.2
(4.4
) 5.
2 (4
.6)
5.2
(4.5
) A
ctiv
ity L
imita
tions
(AL
) W
OM
AC
phy
sica
l (0-
10)
5.3
(1.8
) 4.
1 (1
.8)
2.9
(1.5
) 2.
0 (1
.4)
1.6
(1.5
) 1.
5 (1
.5)
HO
OS
recr
eatio
n an
d le
isur
e (0
-10)
8.
6 (1
.6)
8.6
(1.3
) 7.
6 (1
.8)
5.5
(2.5
) 4.
7 (2
.7)
4.3
(2.7
)
Tot
al A
L (0
-20)
13
.9 (3
.1)
12.7
(2.7
) 10
.5 (2
.9)
7.5
(3.7
) 6.
4 (3
.9)
5.8
(3.9
) Pa
rtic
ipat
ion
Res
tric
tions
(PR
) LL
DI f
requ
ency
(0-1
0)
4.0
(1.5
) 5.
6 (2
.0)
4.9
(1.7
) 3.
4 (1
.6)
3.1
(1.6
) 3.
0 (1
.5)
LLD
I lim
itatio
n (0
-10)
4.
0 (2
.1)
5.6
(2.3
) 4.
2 (2
.3)
2.01
(2.1
) 1.
6 (1
.9)
1.6
(1.9
) C
alde
rdal
e co
mm
unity
m
obili
ty (0
-10)
5.
2 (2
.4)
6.3
(2.7
) 4.
2 (2
.7)
2.0
(2.1
) 1.
6 (1
.9)
1.5
(1.8
)
Tot
al P
R (0
-30)
13
.1 (5
.1)
17.6
(5.3
) 13
.1 (5
.8)
7.2
(5.1
) 6.
2 (4
.8)
5.9
(4.5
)
31
TK
R (n
=494
)
Phys
ical
Impa
irm
ents
(PI)
K
OO
S pa
in (0
-10)
5.
2 (1
.2)
4.2
(1.8
) 3.
6 (1
.7)
2.4
(1.7
) 2.
1 (1
.80)
1.
7 (1
.6)
Chr
onic
Pai
n G
rade
(0-1
0)
6.1
(2.4
) 5.
1 (2
.2)
4.1
(2.2
) 2.
5 (2
.1)
2.3
(2.2
) 1.
8 (2
.0)
Tot
al P
I (0-
20)
11.3
(3.9
) 9.
4 (3
.7)
7.8
(3.6
) 5.
0 (3
.5)
4.4
(3.7
) 3.
5 (3
.4)
Impa
irm
ent o
f Moo
d (M
I)
POM
S fa
tigue
(0-1
0)
3.9
(2.8
) 4.
7 (2
.6)
3.8
(2.5
) 2.
77 (2
.3)
2.8
(2.5
) 2.
7 (2
.4)
HA
DS
anxi
ety
(0-1
0)
3.1
(1.9
) 2.
8 (2
.1)
2.6
(1.9
) 2.
1 (1
.8)
2.1
(2.0
) 2.
1 (1
.9)
HA
DS
depr
essi
on (0
-10)
2.
5 (1
.7)
3.1
(2.0
) 2.
7 (1
.9)
1.9
(1.7
) 1.
8 (1
.7)
1.6
(1.6
) T
otal
MI (
0-30
) 9.
5 (5
.4)
10.5
(5.6
) 9.
1 (5
.4)
6.7
(5.2
) 6.
6 (5
.5)
6.4
(5.2
) A
ctiv
ity L
imita
tions
(AL
) K
OO
S ph
ysic
al (0
-10)
4.
9 (1
.8)
4.6
(1.9
) 3.
6 (1
.7)
2.5
(1.6
) 2.
3 (1
.7)
2.1
(1.6
) K
OO
S re
crea
tion
and
leis
ure
(0-1
0)
9.3
(1.2
) 9.
5 (0
.9)
9.1
(1.4
) 7.
9 (2
.1)
7.5
(2.4
) 7.
0 (2
.6)
Tot
al A
L (0
-20)
14
.2 (2
.7)
14.1
(2.4
) 12
.8 (2
.6)
10.5
(3.3
) 9.
9 (3
.7)
9.1
(3.8
) Pa
rtic
ipat
ion
Res
tric
tions
(PR
) LL
DI f
requ
ency
(0-1
0)
3.9
(1.5
) 5.
3 (2
.1)
4.8
(1.8
) 3.
5 (1
.6)
3.2
(1.6
) 3.
1 (1
.6)
LLD
I lim
itatio
n (0
-10)
3.
8 (2
.1)
5.3
(2.5
) 4.
3 (2
.3)
2.4
(2.2
) 2.
0 (2
.0)
1.8
(1.9
) C
alde
rdal
e co
mm
unity
m
obili
ty (0
-10)
5.
0 (2
.3)
6.8
(2.4
) 5.
1 (2
.6)
2.7
(2.2
) 2.
3 (2
.2)
2.0
(2.2
)
Tot
al P
R (0
-30)
12
.7 (4
.9)
17.6
(5.6
) 14
.2 (5
.4)
8.4
(5.1
) 7.
4 (4
.9)
6.8
(4.9
)
POM
S=Pr
ofile
of M
ood
Stat
es
HA
DS=
Hos
pita
l Anx
iety
and
Dep
ress
ion
Scal
e H
OO
S=H
ip d
isab
ility
and
Ost
eoar
thrit
is O
utco
me
Scal
e K
OO
S=K
nee
inju
ry a
nd O
steo
arth
ritis
Out
com
e Sc
ore
LLD
I=La
te L
ife D
isab
ility
Inst
rum
ent
32
Tab
le 3
: Cha
nge
of IC
F C
onst
ruct
s ove
r T
ime
Pre-
surg
ery
to 1
2 m
onth
s po
st-s
urge
ry
Pre
-sur
gery
to 3
m
onth
s pos
t-su
rger
y
3 m
onth
s to
12 m
onth
s po
st-s
urge
ry
ICF
Con
stru
ct
Mea
n ch
ange
M
ean
chan
ge
Perc
enta
ge
of to
tal
chan
ge
Mea
n ch
ange
Pe
rcen
tage
of
tota
l ch
ange
TH
R (n
=437
)
Ph
ysic
al Im
pair
men
ts
(0-2
0)
9.1
8.5
93
0.6
6
Impa
irm
ent o
f Moo
d (0
-30)
4.
5 4.
5 10
0 0.
01
0
Act
ivity
Lim
itatio
ns
(0-2
0)
8.0
6.4
80
1.6
20
Pa
rtic
ipat
ion
Res
tric
tions
(0
-30)
7.
2 5.
9 82
1.
3 18
TK
R (n
=494
)
Phys
ical
Impa
irm
ents
(0
-20)
7.
8 6.
3 81
1.
4 18
Impa
irm
ent o
f Moo
d (0
-30)
3.
1 2.
8 90
0.
3 11
Act
ivity
Lim
itatio
ns
(0-2
0)
5.1
3.7
72
1.4
28
Pa
rtic
ipat
ion
Res
tric
tions
(0
-30)
5.
9 4.
4 74
1.
6 27
33
Supplementary data: Approach to Path Analysis to Obtain the Final Model
Table S1 shows the model fit for our hypothesized model, the model including covariate
adjustment and our final baseline model as described in the text of the manuscript.
Nested models were tested against the baseline model. Five sets of constraints (i.e. equality
constraints over time) were sequentially introduced into the model and tested against the baseline
model (M3) using chi-square difference testing. Where no statistically significant difference
between one model and the next was found, this was interpreted as stability in the respective
relationships over time. In these cases, the equality constraints were retained in the model. On
the other hand, where a significant difference was detected, this was interpreted as a non-stable,
or time-dependent relationship over time. Equality constraints were not retained in this case.
The set of five equality constraints tested were as follows:
Mc1: a model where paths from previous physical impairment to current activity limitation
were constrained to be equal across time;
Mc2: a model where paths from previous activity limitation to current participation
restriction were constrained to be equal across time;
Mc3: a model where paths from current physical impairment to current activity limitation
were constrained to be equal across time.
Mc4: a model where paths from current activity limitation to current participation restrictions
were constrained to be equal across time;
Mc5: a model where paths from current physical impairment to current participation
restriction were constrained to be equal across time; and,
34
As shown in table S1, the model comparisons indicated that there was no degradation in
model fit with the introduction of across time equality constraints (Mc1 versus baseline: Δχ2
=3.6, df =4, p-value =.4628; and, Mc2 versus baseline: Δχ2 = 5.0, df =4, p-value =.2873).
However, the within time equality constraints were not equal i.e. the inter-relationships of the
constructs were not stable over time as shown in table S1 (i.e. Mc3, Mc4 and Mc5 each versus
baseline).
With the final model established, the model was re-assessed using Maximum Likelihood
with a mean adjusted chi-square test statistic to ensure robustness against any non-normality 30.
35
Tab
le S
1. G
oodn
ess-
of fi
t for
diff
eren
t mod
els b
ased
on
path
ana
lysi
s
χ2
(df)
C
FI
TL
I R
MSE
A
SRM
R
Mod
el
com
pari
son
∆χ2
(df)
P-
valu
e
M1:
Hyp
othe
size
d m
odel
14
57
(200
) .8
95
.863
.0
86
.129
--
--
M2:
Initi
al m
odifi
ed m
odel
11
60
(196
) .9
20
.893
.0
76
.102
M
1 vs
. M2
297.
0 (4
) <.
0001
M3:
Initi
al fi
nal m
odel
(B
asel
ine
mod
el)
781.
3 (1
86)
.950
.9
30
.061
.0
84
M2
vs. M
3 37
8.7
(10)
<.
0001
Add
ing
cons
trai
nts t
o Ba
selin
e M
odel
M
c1: P
ath
from
pre
viou
s ph
ysic
al im
pairm
ent t
o cu
rren
t act
ivity
lim
itatio
n co
nstra
ined
784.
9 (1
90)
.950
.9
32
.060
.0
84
Mc2
vs.
M3
3.6
(4
) .4
628
Mc2
: Pat
h fr
om p
revi
ous
activ
ity li
mita
tion
to c
urre
nt
parti
cipa
tion
rest
rictio
n co
nstra
ined
786.
3 (1
90)
.950
.9
32
.061
.0
84
Mc1
vs.
M3
5.0
(4
) .2
873
Mc3
: Pat
h fr
om c
urre
nt
phys
ical
impa
irmen
t to
curr
ent a
ctiv
ity li
mita
tion
cons
train
ed
796.
8 (1
90)
.949
.9
31
.061
.0
84
Mc5
vs.
M3
15.6
(5
) .0
081
Mc4
: Pat
h fr
om c
urre
nt
activ
ity li
mita
tion
to c
urre
nt
parti
cipa
tion
rest
rictio
n co
nstra
ined
904.
6 (1
91)
.941
.9
19
.066
.0
89
Mc3
vs.
M3
123.
3 (5
) <.
0001
Mc5
: Pat
h fr
om c
urre
nt
phys
ical
impa
irmen
t to
curr
ent p
artic
ipat
ion
rest
rictio
n co
nstra
ined
806.
1 (1
91)
.949
.9
30
.061
.0
84
Mc4
vs.
M3
24.8
(5
) .0
002
36
Fina
l Mod
el
789.
9 (1
94)
.950
.9
33
.060
.0
84
Fina
l vs.
M3
8.6
(8)
.377
2
Fina
l (M
LR)
731.
7 (1
94)
.949
.9
32
.057
.0
84
- -
-
χ2 =
chi-s
quar
e, d
f = d
egre
es o
f fre
edom
; CFI
= C
ompa
rativ
e Fi
t Ind
ex; T
LI=
Tuck
er-L
ewis
Inde
x; R
MSE
A =
Roo
t Mea
n Sq
uare
Err
or o
f A
ppro
xim
atio
n; S
RM
R=
Stan
dard
ized
Roo
t Mea
n Sq
uare
Res
idua
l.
Not
e:
All
mod
els w
ere
adju
sted
for a
ge, s
ex, k
nee/
hip
repl
acem
ent,
obes
ity, l
ow b
ack
pain
and
moo
d
Figu
re 1
Impa
irmen
t
Impa
irmen
t
Impa
irmen
t
Impa
irmen
t
Impa
irmen
t
Impa
irmen
t Ti
me
0
Tim
e 1
Ti
me
2
Tim
e 3
Ti
me
4
Tim
e 5
Act
ivity
Act
ivity
Act
ivity
Act
ivity
Act
ivity
Act
ivity
Li
mita
tions
Lim
itatio
ns
Li
mita
tions
Lim
itatio
ns
Li
mita
tions
Lim
itatio
ns
Tim
e 0
Ti
me
1
Tim
e 2
Ti
me
3
Tim
e 4
Ti
me
5
Parti
cipa
tion
Pa
rtici
patio
n
Parti
cipa
tion
Pa
rtici
patio
n
Parti
cipa
tion
Pa
rtici
patio
n R
estri
ctio
ns
R
estri
ctio
ns
R
estri
ctio
ns
R
estri
ctio
ns
R
estri
ctio
ns
R
estri
ctio
ns
Tim
e 0
Ti
me
1
Tim
e 2
Ti
me
3
Tim
e 4
Ti
me
5
Figu
res
Figu
re 2 -1
0-8-6-4-20246810
01
23
45
Tim
e
PI MI
AL PR
Figu
re 3 -1
0-8-6-4-20246810
01
23
45
Tim
e
PI MI
AL
PR
Figu
re 4
PI
.6
06
-.2
97
.
661
-.
287
.5
07
-.2
14
.51
2
-.181
.424
-.
183
.3
37
.
.2
10
.
.106
.
198
.
200
.1
06
.458
-.15
3
.
657
-.
145
.52
0
-.1
82
.454
-
.245
.3
99
-.
268
.335
.1
43
.299
.120
.131
.403
.728
.711
.699
.730
.258
.74
1 .6
09.3
35 .4
61
.242
.587
.7
12
.532
.6
81
Phys
ical
Impa
irmen
t
Act
ivity
Li
mita
tion
Parti
cipa
tion
Res
trict
ion
Pre-
Surg
ery
2-W
eeks
1-
Mon
th3-
Mon
ths
6-M
onth
s12
-Mon
ths