The clinimetric properties of the world health organization disability assessment schedule II in early inflammatory arthritis

Authors


Abstract

Objective

To assess the clinimetric properties of a new health-related quality of life (HRQOL) instrument, the World Health Organization Disability Assessment Schedule II (WHODAS II), in patients with early inflammatory arthritis.

Methods

Internal consistency as well as criterion, construct, and discriminative validity of the WHODAS II were assessed in 172 patients with early inflammatory arthritis who completed the WHODAS II, the Medical Outcomes Study Short Form 36 (SF-36), and other measures of disease severity, functioning, pain, depression, and resource use. Test–retest reliability of the WHODAS II was assessed by having a subset of 20 patients complete the WHODAS II a second time, 1 week after the first assessment.

Results

The WHODAS II had high internal consistency (Cronbach's α = 0.96 for patients working or in school and 0.93 for patients not working or in school). Test–retest intraclass correlation coefficients of the WHODAS II total score and subscales ranged from 0.82–0.96. The WHODAS II total score was strongly correlated with the SF-36 physical component score (Kendall's tau-b 0.51, P < 0.001) and moderately correlated with the SF-36 mental component score (tau-b 0.43, P < 0.001). WHODAS II correlations with disease outcomes ranged from Kendall's tau-b 0.15–0.55. The WHODAS II significantly differentiated between every aspect of disease severity assessed with the exception of measures of health resource use.

Conclusion

The WHODAS II is a valid and reliable measure of HRQOL in cross-sectional studies of patients with early inflammatory arthritis. Research is still required to investigate potential item redundancy and determine its usefulness in longitudinal studies.

INTRODUCTION

Rheumatoid arthritis (RA) is a disease of unknown etiology, characterized by joint inflammation and progressive deformity affecting 1% of the population (1, 2). It is increasingly clear that there is a window of opportunity wherein early aggressive therapy may prevent damage (3–7). Early inflammatory arthritis clinics have now been established to understand more about the disease in its early stages to promote early referral and to prompt treatment (8–14). Measuring health-related quality of life (HRQOL) has become standard in many outcome studies of various kinds of arthritis. Many studies in patients with RA, and now in those with early inflammatory arthritis, use the Medical Outcomes Study Short Form 36 (SF-36) to assess HRQOL (15–25). The SF-36 is a self-report measure with 8 domains assessing physical and emotional impairment and activity limitation. However, the SF-36 does not measure certain core aspects of HRQOL outlined by the World Health Organization (WHO) such as cognitive performance, sexuality, and participation in society (26, 27).

The International Classification of Functioning, Disability and Health (ICF) is the classification of functioning and disability used by the WHO. It systematically organizes the consequences of disease into 3 dimensions: body functions and structure (symptoms and impairment), activities, and participation (28–35). Several recent studies support the ICF as a useful framework for understanding the complex relationships among the diverse consequences of chronic conditions (36–38). The World Health Organization Disability Assessment Schedule II (WHODAS II) is an HRQOL instrument developed to operationalize the core dimensions of the ICF (39). It is a multidimensional instrument that has 6 domains: understanding and communicating, getting around, self care, getting along with others, household and work activities, and participation in society. The WHODAS II is distinguished from other HRQOL instruments because it is based on an international classification system, is applicable across cultures, and treats all disorders at parity when determining the level of functioning.

The WHODAS II has now been evaluated and found to be valid in some rheumatic conditions, in particular ankylosing spondylitis (40), back pain (41), and musculoskeletal conditions (defined as low back pain, RA, and osteoarthritis) (42); however, it has not yet been evaluated in patients with early inflammatory arthritis.

We believe that the WHODAS II may be a useful tool for evaluating HRQOL in patients with early inflammatory arthritis. The aim of this study was to assess the clinimetric properties of the WHODAS II in early inflammatory arthritis. For criterion validity, we hypothesized that the WHODAS II should correlate well with another measure of HRQOL; in this case, the SF-36. For construct validity, we hypothesized that the WHODAS II should correlate well with other measures of disease status in early inflammatory arthritis, specifically disease activity, pain, functioning, depression, and health care utilization (convergent validity), and should discriminate between early inflammatory arthritis patients with better and worse disease (discriminative validity).

PATIENTS AND METHODS

A total of 172 patients with early inflammatory arthritis were recruited from the McGill Early Arthritis Registry, which enrolls patients with 1 or more joints with synovitis for ≥6 weeks and <1 year with a diagnosis of either RA or undifferentiated inflammatory arthritis, with no other cause for their synovitis and no structural evidence of a previous episode of inflammatory arthritis. Patients are referred to the McGill Early Arthritis Registry by 21 rheumatologists in the greater Montreal area, are age ≥18 years, speak English or French, and agree to periodic physical and laboratory examinations and to complete a number of self-report questionnaires assessing disability, pain, psychosocial factors, and health services use. The participating rheumatologists, each of whom works in a private office or outpatient hospital clinic, were asked to recruit every new early inflammatory arthritis patient who fit the inclusion criteria consecutively. Those who agreed to participate were then referred to the registry. All patients who fit the criteria at the time of the registry baseline assessment and who consented to be a part of the registry were enrolled. Enrollment began in January 2004 and for this study, ended in September 2006.

Test–retest reliability.

All patients completed the WHODAS II for the first time at their baseline registry visit. Patients were approached consecutively to assess test–retest reliability. A subset of 20 patients who agreed were also given a second copy of the WHODAS II in a self-addressed stamped envelope with instructions indicating that they were to complete it again 1 week later and return the form by mail thereafter. This time period was chosen to reduce the possibility that rapid changes in health status could influence reporting.

Validity.

One way of conceptualizing the assessment of construct validity is to consider it as the testing of hypotheses of what a valid instrument would and would not correlate with (43). We hypothesized that a valid HRQOL questionnaire in early inflammatory arthritis would 1) correlate well with an established HRQOL instrument already used in this population such as the SF-36 (criterion validity), 2) correlate well with indices of disease activity such as joint inflammation, disability, pain, patient and physician global assessment of disease severity, depression, and recent resource use (construct validity), and 3) discriminate between early inflammatory arthritis patients with a better or worse disease process (discriminative validity).

Study instrument.

The short-form, self-administered, 36-item English version of the WHODAS II questionnaire was used. This instrument consists of 36 Likert-formatted questions divided into 6 domains: understanding and communicating, getting around, self care, getting along with people, life activities, and participation in society. The WHODAS II has 2 variations of a total score that can be combined according to the WHODAS II syntax provided by the WHO. For people working or in school, the WHODAS II total score is the weighted sum of all 36 items, whereas for people not working or in school the WHODAS II total score is the weighted sum of 32 items. The final score ranges from 0 (best) to 100 (worst). The WHODAS II was translated into Canadian French based on methodology that has been previously published (44, 45), and the clinimetric properties of the Canadian French version will be reported elsewhere.

Criterion measure.

The SF-36 is a widely used measure of HRQOL (26, 27) consisting of 8 domains: physical functioning, social functioning, role limitations related to physical problems, role limitations related to emotional problems, mental health, vitality, bodily pain, and general health perceptions. Each domain can be scored separately or grouped and summarized into a physical component score and a mental component score. Each subscale and summary score is weighted and standardized and ranges from 0 (worst) to 100 (best).

Construct measures.

Disease activity was assessed using a joint count (the number of tender and swollen joints as defined by the American College of Rheumatology [ACR] joint count of 66 swollen joints and 68 tender joints) (46, 47), acute-phase reactants (C-reactive protein [CRP] level), and the Disease Activity Score in 28 joints (DAS28) using the CRP (48).

Physical functioning was measured with the Stanford Health Assessment Questionnaire Disability Index (HAQ DI) (49, 50), which is a measure developed specifically for use in arthritis and has excellent psychometric properties. The HAQ DI consists of 20 items assessing 8 domains, including dressing, standing, eating, walking, toileting, reach, grip, and instrumental activities. The HAQ DI provides a mean total score ranging from 0 (without any difficulty) to 3 (unable to do).

Pain was assessed using the short-form McGill Pain Questionnaire (MPQ) (51, 52). The MPQ provides a summary score for sensory pain items, a summary score for affective pain items, a total pain score, a visual analog scale score, and a present pain–intensity index. For the purposes of the present study, only the total pain score was used. The MPQ total score is the sum of sensory and affective items, ranges from 0 (no pain) to 45 (severe pain), and has been demonstrated to have excellent reliability.

Patient global assessments of activity, fatigue, and pain were assessed using an 11-point numeric rating scale ranging from 0 (best) to 10 (worst) with reference to the past week. Physicians were asked to use the same type of 11-point numeric rating scale to rate the patients' overall disease activity.

Depression was assessed using the Center for Epidemiologic Studies Depression Scale (CES-D) (53, 54). The CES-D is a 20-item self-report scale that asks the patient to answer how frequently from 0 (rarely or none of the time) to 3 (most or all of the time) they have experienced each item in the past week. The CES-D cutoff score for depression is ≥16; however, a higher cutoff score of ≥19 is often used in studies of RA to control for symptom overlap (55).

Health services use was estimated by the number of patient self-report doctor visits, tests taken, and hours spent visiting physicians in the past 4 months (56, 57).

Statistical analysis.

Descriptive statistics were used to summarize the baseline characteristics of early inflammatory arthritis patients. Internal consistency was assessed using Cronbach's alpha, where values <0.70 indicate that individual items are providing an inadequate contribution to the overall scale and values >0.90 suggest that there may be a high level of item redundancy (58). Confidence intervals (CIs) for Cronbach's alpha were obtained using a nonparametric bootstrap as suggested by Efron and Tibshirani (59).

Test–retest reliability was evaluated using a 2-way random effects model, a single-measure intraclass correlation coefficient (ICC) to assess the degree of agreement between repeat WHODAS II subscales, and total scores.

Criterion and construct validity were evaluated using nonparametric Kendall's tau-b correlation coefficients because the WHODAS II score distributions were not normal. In general, correlation coefficients <0.3, 0.3–0.49, and >0.5 represent small, medium, and large correlations, respectively (60).

Discriminative validity was assessed by constructing dichotomous subsets of various early inflammatory arthritis outcome variables to identify patients with better and worse disease. A DAS28 CRP threshold of 5.1 was used to distinguish between mild and moderate-severe disease activity (61), a HAQ DI threshold of 1 was used to distinguish between mild and moderate-severe levels of disability, and a CES-D cutoff of 19 was used to distinguish depressed from nondepressed patients (55). Because no validated thresholds were available for patient and physician global assessments of disease severity, SF-36 summary scores, pain, or resource use, median values were used as cutoffs between better and worse disease. Nonparametric comparisons were made using the Mann-Whitney U test for each subset.

All statistical analyses were performed with SPSS, version 13 (SPSS, Chicago, IL) and the R statistical package (Foundation for Statistical Computing, Vienna, Austria).

Sample size justifications.

The minimum sample size to measure test–retest reliability was determined by assuming an ICC of 0.85. Therefore, a sample of 20 patients would be required to obtain a 90% CI with a total width of 0.25 (0.125 on either side) (62, 63). Patients were approached consecutively for the test–retest reliability and those who agreed to complete the second questionnaire were included in the sample.

Ethical considerations.

All patients in the McGill Early Arthritis Registry signed an informed consent form and the Early Arthritis Registry was approved by the Institutional Review Boards of McGill University and the Sir Mortimer B. Davis Jewish General Hospital.

RESULTS

Patient characteristics.

Of the 172 patients included in this study, 68.6% were women (Table 1). The mean ± SD age of the patients was 56.2 ± 15.0 years and 54.1% were working or in school. The mean ± SD symptom duration was 7.0 ± 4.0 months and the mean ± SD DAS28 CRP score was 4.7 ± 1.6. Approximately 19% of patients met the ACR criteria for RA (64) and the median WHODAS II score was 19.8 (interquartile range 25.7) with a skewed distribution (Figure 1). The greatest limitations appeared in life activities and mobility (Table 1).

Table 1. Baseline characteristics of patients with early inflammatory arthritis (n = 172)*
 Value
  • *

    Values are the mean ± SD unless otherwise indicated. WHODAS II = World Health Organization Disability Assessment Schedule II; IQR = interquartile range; CRP = C-reactive protein; DAS28 = Disease Activity Score in 28 joints; HAQ DI = Health Assessment Questionnaire Disability Index; MPQ = McGill Pain Questionnaire; CES-D = Center for Epidemiologic Studies Depression Scale.

  • Paid, nonpaid, self employed, or full-time student.

Age, years56.2 ± 15.0
Women, frequency (%)118 (68.6)
Disease duration, months7.0 ± 4.0
Postsecondary education, frequency (%)106 (61.6)
Working or in school, frequency (%)93 (54.1)
WHODAS II total score if working or in school (n = 93), median (IQR)17.0 (26.0)
WHODAS II total score if not working or in school (n = 79), median (IQR)25.0 (25.0)
WHODAS II composite total score (n = 172), median (IQR)19.8 (25.7)
 WHODAS II subscale scores, median (IQR) 
  Understanding and communicating5.0 (25.0)
  Getting around31.3 (50.0)
  Self care10.0 (20.0)
  Getting along with people8.3 (25.0)
  Life activities, household35.0 (40.0)
  Life activities, working or in school (n = 93)17.9 (37.5)
  Participation in society25.0 (25.0)
CRP level, mg/liter26.8 ± 50.3
DAS28 score (n = 152)4.7 ± 1.6
Functioning, HAQ DI score0.87 ± 0.78
Pain, MPQ score7.6 ± 7.9
Patient global assessment of disease severity 
 Activity4.7 ± 2.8
 Fatigue5.0 ± 2.7
 Pain4.7 ± 2.8
Physician global assessment of disease severity3.8 ± 2.5
Depression 
 CES-D score14.1 ± 10.2
 Patients with CES-D score >1947 ± 27.3
Resource use in the last 4 months (n = 137) 
 Physician visits6.6 ± 6.4
 Tests taken5.3 ± 3.2
 Hours spent with physicians15.7 ± 15.2
Figure 1.

Baseline distribution of the World Health Organization Disability Assessment Schedule II (WHODAS-II) total score in 172 patients with early inflammatory arthritis. Median (interquartile range) is 19.8 (25.7).

Internal consistency and test–retest reliability.

Internal consistency was assessed in the entire sample of 172 patients. For people working or in school, the internal consistency of the 36-item WHODAS II total score was very high (Cronbach's α = 0.96, 95% CI 0.95–0.97). For people not working or in school, internal consistency of the 32-item WHODAS II total score was also very high (Cronbach's α = 0.93, 95% CI 0.91–0.95). The alpha values for the WHODAS II subscales ranged from 0.71 for getting along with people to 0.94 for life activities at work or in school (Table 2). Test–retest reliability was assessed in 20 subjects. The characteristics listed in Table 1 were compared between this subset and the rest of the group, although statistical analyses were not performed because of the relatively small sample size. The groups appeared similar except that patients in the test–retest group were slightly younger (49.2 years versus 56.7 years), were mostly women (90% versus 67.1%), were more likely to be working or in school (80% versus 53%), and were more likely to score above a cutoff of 19 on the CES-D questionnaire for depression (45% versus 27%). The WHODAS II scores were in general somewhat better in the test–retest group. The single-measure ICC for the WHODAS II total score between time 1 and time 2 was 0.94 (95% CI 0.86–0.98). The ICCs for the WHODAS II subscales ranged from 0.82 for participation in society to 0.96 for getting around (Table 3).

Table 2. Reliability. Internal consistency of the World Health Organization Disability Assessment Schedule II (WHODAS II) subscales and total scores (n = 172)
SubscaleCronbach's α
Understanding and communicating0.88
Getting around0.89
Self care0.82
Getting along with people0.71
Life activities 
 Household0.90
 Working or in school0.94
Participation in society0.85
WHODAS II 36-item total score, working or in school (n = 93)0.96
WHODAS II 32-item total score, not working or in school (n = 79)0.93
Table 3. Test–retest reliability. ICCs of the WHODAS II subscales and total scores (n = 20)*
 Single-measure ICC95% CI
  • *

    ICC = intraclass correlation coefficient; WHODAS II = World Health Organization Disability Assessment Schedule II; 95% CI = 95% confidence interval.

Understanding and communicating0.850.67–0.94
Getting around0.960.91–0.99
Self care0.820.59–0.92
Getting along with people0.910.79–0.96
Life activities  
 Household0.910.79–0.96
 Working or in school0.960.88–0.98
Participation in society0.830.62–0.93
WHODAS II 36-item total score, working or in school (n = 16)0.960.90–0.99
WHODAS II 32-item total score, not working or in school (n = 4)0.900.12–0.99
WHODAS II total score, working/in school and not working/in school (n = 20)0.950.89–0.98

Criterion validity.

The WHODAS II total score was strongly correlated with the SF-36 physical component score (τ = −0.51, P < 0.001) and moderately correlated with the SF-36 mental component score (τ = −0.43, P < 0.001). Similarly, the WHODAS II subscales were all moderately to strongly correlated with the subscale domains and summary scores of the SF-36 (data not shown).

Construct validity.

As predicted, the WHODAS II correlated well with multiple other measures of disease status in early inflammatory arthritis (Table 4). The WHODAS II was moderately correlated with tender (τ = 0.34, P < 0.001) and swollen joint counts (τ = 0.17, P = 0.002), CRP (τ = 0.20, P = 0.001), and the DAS28 CRP (τ = 0.39, P < 0.001). The WHODAS II was moderately to strongly correlated with self reports of pain (τ = 0.41, P < 0.001), disability (τ = 0.55, P < 0.001), patient and physician global assessments of disease severity (tau-b range 0.34–0.47, P < 0.001), and depression (τ = 0.52, P < 0.001). Correlations between the WHODAS II and recent resource use were weak but significant, ranging from τ = 0.15 (P = 0.014) for the number of physician visits to τ = 0.20 (P = 0.002) for the number of hours spent seeing physicians in the last 4 months.

Table 4. Construct validity. Nonparametric correlation coefficients of the WHODAS II with measures of disease status in early inflammatory arthritis*
 Kendall's τ-bP
  • *

    WHODAS II = World Health Organization Disability Assessment Schedule II; NS = not significant; CRP = C-reactive protein; DAS28 = Disease Activity Score in 28 joints; HAQ DI = Health Assessment Questionnaire Disability Index; MPQ = McGill Pain Questionnaire; CES-D = Center for Epidemiologic Studies Depression Scale.

Age, years0.01NS
Disease duration, months−0.03NS
Tender joint count0.34< 0.001
Swollen joint count0.170.002
CRP level, mg/liter (n = 152)0.200.001
DAS28 score (n = 152)0.39< 0.001
Functioning, HAQ DI score0.55< 0.001
Pain, MPQ score0.41< 0.001
Patient global assessment of disease severity  
 Activity0.39< 0.001
 Fatigue0.39< 0.001
 Pain0.47< 0.001
Physician global assessment of disease severity0.34< 0.001
Depression, CES-D score0.52< 0.001
Resource use in the last 4 months, number (n = 137)  
 Physician visits0.150.014
 Tests taken0.160.009
 Hours spent with physicians0.200.002

Discriminative validity.

Dichotomous disease groups were constructed for every disease outcome variable to test whether the WHODAS II could differentiate between early inflammatory arthritis patients with better and worse disease (Table 5). The WHODAS II significantly differentiated between every aspect of disease severity assessed, with the exception of resource use in the past 4 months.

Table 5. Discriminative validity. Comparison of WHODAS II total score with subsets of indices of disease severity in early inflammatory arthritis*
Indices of disease severityNo.WHODAS II score, median (IQR)P
  • *

    WHODAS II = World Health Organization Disability Assessment Schedule II; IQR = interquartile range; CRP = C-reactive protein; DAS28 = Disease Activity Score in 28 joints; HAQ DI = Health Assessment Questionnaire Disability Index; MPQ = McGill Pain Questionnaire; CES-D = Center for Epidemiologic Studies Depression Scale; NS = not significant.

  • Subset cutoffs calculated using median scores.

Tender joint count  < 0.001
 ≤108912.3 (16.6) 
 >107630.3 (23.4) 
Swollen joint count  0.013
 ≤48415.2 (25.8) 
 >48124.5 (28.8) 
CRP level, mg/liter  0.007
 ≤8.87414.1 (26.5) 
 >8.89123.9 (26.7) 
DAS28 CRP score  < 0.001
 <5.18511.3 (16.0) 
 ≥5.18030.3 (23.6) 
Functioning, HAQ DI score  < 0.001
 <110011.1 (15.8) 
 ≥16533.7 (20.1) 
Pain, MPQ score  < 0.001
 ≤58810.6 (14.7) 
 >57731.5 (22.8) 
Patient global assessment of disease severity   
 Activity  < 0.001
  ≤510413.7 (23.4) 
  >56129.3 (24.7) 
 Fatigue  < 0.001
  ≤58612.7 (19.3) 
  >57926.4 (25.6) 
 Pain  < 0.001
  ≤59613.0 (15.8) 
  >56933.0 (23.9) 
Physician global assessment of disease severity  0.001
 ≤39013.1 (18.5) 
 >37530.2 (24.6) 
Depression, CES-D score  < 0.001
 <1912115.2 (19.5) 
 ≥194439.4 (23.5) 
Resource use in the last 4 months, number   
 Physician visits (n = 137)  NS
  <57717.9 (27.2) 
  ≥56020.8 (24.9) 
 Tests taken (n = 137)  NS
  <58217.2 (23.9) 
  ≥55521.7 (27.0) 
 Hours spent seeing physicians (n = 137)  NS
  <105916.0 (25.5) 
  ≥107822.2 (25.4) 

DISCUSSION

The WHODAS II demonstrated good reliability and validity in patients with early inflammatory arthritis. Internal consistency was excellent for both the WHODAS II 36-item score used for people working or in school and the WHODAS II 32-item score used for people not working or in school (Cronbach's α = 0.96 and 0.93, respectively). However, alpha values >0.90 may represent item redundancy (43). Further investigation of the internal consistency of the WHODAS II subscales and total scores after removing highly correlated items is warranted.

Test–retest reliability of the WHODAS II, measured with the ICC, was excellent for both those working or in school and those not working or in school, as well as for all subscale scores, indicating that the WHODAS II is a highly reliable measure across time.

The WHODAS II correlated moderately to strongly with the criterion measure, the SF-36. It exhibited good construct validity because it also correlated well with all of the early inflammatory arthritis disease status measures, including tender and swollen joint counts, CRP, DAS28 CRP, disability, pain, patient and physician global assessments of disease severity, depression, and resource use. Moreover, the WHODAS II discriminated between every disease outcome subset assessed except for health resource use in the past 4 months, which supports its discriminative validity.

Validity of an instrument is population specific (65, 66). Our cohort of patients with early inflammatory arthritis includes both patients with RA according to ACR criteria (64), and those who do not meet these criteria but who have developed a form of polyarthritis, labeled in this study as having undifferentiated inflammatory arthritis. It is important to have a valid HRQOL measure for this population.

A few studies have already used the WHODAS II, mostly in the field of psychiatry (67, 68). A few studies have demonstrated that it is also valid in arthritis-related conditions, namely ankylosing spondylitis and back pain. Our findings are consistent with those studies and are unique in that to our knowledge, this is the first study investigating the reliability and validity of the WHODAS II in early inflammatory arthritis.

This study does have several limitations. The WHODAS II was administered to patients with early inflammatory arthritis and therefore the validity of the WHODAS II cannot be generalized to other patient populations. In particular, our early inflammatory arthritis cohort differs from patients with established RA because many of our patients may not develop persistent erosive disease (69–71). Therefore, other studies will be required before generalizing these results to the RA population. Also, it is possible that using a subsample to perform part of the analysis, in this case the test–retest group, may affect generalizability of the results in that certain subgroups of patients may not be representative of the entire sample. We compared the characteristics of the 20 patients who were included in the test–retest analysis with those from the rest of the sample and did find some differences, although the small size of the test–retest group precluded performing statistical analyses. We also have not yet assessed the sensitivity to change of the WHODAS II. Nevertheless, others have found that it is responsive to change in patients with certain rheumatic conditions (40). In addition, the ability of an instrument to detect clinically meaningful differences can be assessed cross-sectionally by comparing groups that are different in terms of some disease-related variable (72). Our discriminative analysis, which clearly showed the ability of the WHODAS II to distinguish between subsets of patients with less and more severe disease, is such an example. Therefore, we believe that the WHODAS II will be found to be responsive over time. Ongoing longitudinal studies in our patients will allow us to confirm this.

In conclusion, this study has demonstrated that the WHODAS II is a valid and reliable measure of HRQOL in cross-sectional studies of patients with early inflammatory arthritis. Research is still required to investigate potential item redundancy and determine the usefulness of the WHODAS II in longitudinal studies.

AUTHOR CONTRIBUTIONS

Dr. Baron had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study design. Baron, Hudson, Steele.

Acquisition of data. Baron, Schieir, Hudson, Berkson, Couture, Fitzcharles, Gagné, Garfield, Gutkowski, Kang, Kapusta, Ligier, Mathieu, Ménard, Starr, Stein, Zummer.

Analysis and interpretation of data. Baron, Schieir, Hudson, Steele, Kolahi.

Manuscript preparation. Baron, Schieir, Steele.

Statistical analysis. Baron, Schieir, Steele.

ROLE OF THE STUDY SPONSORS

The funding sources had no role in the design of the study, analysis of the data, preparation of the manuscript, and decision to submit for publication.

Ancillary