Measures of Work Disability and ProductivityRheumatoid Arthritis Specific Work Productivity Survey (WPS-RA), Workplace ActivityLimitations Scale (WALS), Work Instability Scale for Rheumatoid Arthritis (RA-WIS),Work Limitations Questionnaire (WLQ), and Work Productivity and Activity ImpairmentQuestionnaire (WPAI)

KENNETH TANG,1 DORCAS E. BEATON,1 ANNELIES BOONEN,2 MONIQUE A. M. GIGNAC,3 ANDCLAIRE BOMBARDIER4

INTRODUCTION

The impact of arthritis on work is an area of increasingresearch interest and a growing number of outcome mea-sures to quantify such impact have become available inrecent years. Recent reviews from an Outcome Measures inRheumatology initiative have broadly identified 24 instru-ments in this area (1,2), though only 11 had been used inarthritis to date. Previous studies have shown that thesemeasures only moderately correlate with each other (3–10), therefore it is important to recognize that availableinstruments offer distinct perspectives on health-relatedwork impacts. For example, some measures are aimed atexamining degree of difficulties with specific workplaceactivities, while others are designed to quantify the extentof absenteeism (e.g., number of days off work) and/or

presenteeism (e.g., being at work but working at reducedproductivity, also referred to as “at-work productivityloss” or “at-work disability”). Also available are instru-ments focused on assessing related concepts such as work“performance,” “efficiency,” “instability,” or degree of “in-terference” at work. For the purpose of this review, wehave adopted a broader approach and consider a diverserange of available measures that offer varying perspectivesand approaches to quantifying the impact of health prob-lems on work (Table 1).

Beyond their diverse conceptual foci, existing instru-ments also differ in terms of their scope of measurement(e.g., impact on employment work versus nonpaid workand/or leisure activities), disease attribution (e.g., diseasespecific versus generic), length (e.g., number of sectionsand items), and recall period. Some are designed as “mod-ular” instruments that assess work impacts using a seriesof global rating scales and discrete items (often organizedinto multiple sections) that are generally not intended tobe summative. Others are classic “psychometric” mea-sures consisting of summative items that inform diseaseimpact on different specific aspects of work (contributes toan overall construct). Specific impacts examined at theitem level may include problems meeting the physicaldemands of work, challenges associated with time man-agement, difficulties maintaining interpersonal relation-ships at work, cognitive concerns (e.g., worries about con-tinued employability), and/or issues related to symptomcontrol/exacerbation and fatigue. The diversity of avail-able measures is a strength in this growing field and hasaccommodated the growing interest to apply these toolsfor a broad range of purposes in arthritis. For example,these measures have been used to examine the epidemiol-ogy (e.g., population trends, determinants) of work disabil-ity/participation in the arthritis population, to evaluate theeffectiveness of clinical or workplace interventions (e.g.,clinical trials), and also, to estimate the economic costs ofhealth-related work productivity loss at the societal level.

Five specific measures were selected for a detailed re-view in this article. These were chosen on the basis of 2

Mr. Tang is recipient of a Canadian Institutes of HealthResearch PhD Fellowship, Canadian Arthritis NetworkGraduate Award, and Syme Fellowship from the Institutefor Work & Health. Dr. Beaton is recipient of a CanadianInstitutes of Health Research New Investigators award. Dr.Bombardier is recipient of a Canada Research Chair inKnowledge Transfer for Musculoskeletal Care.

1Kenneth Tang, MSc(PT), MSc, Dorcas E. Beaton, PhD:St. Michael’s Hospital, Institute for Work & Health, andUniversity of Toronto, Toronto, Ontario, Canada; 2AnneliesBoonen, MD, PhD: Maastricht University Medical Centerand Caphri Research Institute, Maastricht, The Nether-lands; 3Monique A. M. Gignac, PhD: Toronto Western Re-search Institute, University Health Network, and Universityof Toronto, Toronto, Ontario, Canada; 4Claire Bombardier,MD: University Health Network, Institute for Work &Health, and University of Toronto, Toronto, Ontario, Can-ada.

Dr. Boonen has received consultancy fees, speaking fees,and/or honoraria (less than $10,000 each) from Abbott andPfizer.

Address correspondence to Kenneth Tang, MSc(PT), MSc,Mobility Program Clinical Research Unit, Li Ka Shing Insti-tute of St. Michael’s Hospital, 30 Bond Street, Toronto, On-tario, Canada. E-mail: ken.tang@utoronto.ca.

Submitted for publication January 31, 2011; accepted inrevised form May 10, 2011.

Arthritis Care & ResearchVol. 63, No. S11, November 2011, pp S337–S349DOI 10.1002/acr.20633© 2011, American College of Rheumatology

MEASURES OF FUNCTION

S337

main criteria: the availability of measurement evidencespecific to the arthritis population (primarily osteoarthritisor inflammatory arthritis), and evidence of previous re-search application in arthritis populations beyond psycho-metric testing. By these selection criteria, both modularand psychometric measures were represented. Measureschosen included the Rheumatoid Arthritis Specific WorkProductivity Survey (11), Workplace Activity LimitationsScale (12), Work Instability Scale for Rheumatoid Arthritis(13), Work Limitations Questionnaire (14), and Work Pro-ductivity and Activity Impairment Questionnaire (15). Itshould also be recognized that disease-specific variants of2 of the 5 selected measures are available for ankylosingspondylitis (i.e., the Work Productivity and Activity Im-pairment Questionnaire for Ankylosing Spondylitis [16],and the Ankylosing Spondylitis Work Instability Scale[17]); however, these measures will not be presented inthis review. Also, we have emphasized psychometric evi-

dence primarily in arthritis populations, notwithstandingthat evidence in nonarthritis soft tissue musculoskeletalconditions (e.g., low back pain, upper-extremity disorders)and other nonmusculoskeletal disorders are also availablefor a number of these measures.

RHEUMATOID ARTHRITIS SPECIFIC WORKPRODUCTIVITY SURVEY (WPS-RA)

Description

Purpose. Measures the impact of rheumatoid arthritis(RA) on the productivity of employment work, householdwork, and daily activities. The original version of WPS-RAwas published in 2009, primarily intended for use in clin-ical trials (11).

Content. Assesses the number of days of work absence(absenteeism), number of days with reduced work produc-

Table 1. Summary of measures of role functioning and productivity at work, sorted by year of original publication

Author, year (ref.) MeasureMeasure

type*No. items/sections† Scope‡

Osterhaus et al, 1992 (18) Osterhaus Technique (OST) M 4 A, PReilly et al, 1993 (15) Work Productivity and Activity Impairment-

General Health (WPAI)§M 6 A, P

Van Roijen et al, 1996 (19) Health and Labor Questionnaire (HLQ) M 14 A, PEndicott et al, 1997 (20) Endicott Work Productivity Scale (EWPS) Ps 25 PKopec et al, 1998 (21) Occupational Role Questionnaire (ORQ) Ps 8 PBrouwer et al, 1999 (9) Quantity and Quality Method (QQ) from the

Productivity and Disease QuestionnaireM 2 P

Amick et al, 2000 (22) Work Role Functioning-26 (WRF) Ps 26 PLerner et al, 2001 (14) Work Limitations Questionnaire-25 (WLQ-25)§ Ps 25 PAltshuler et al, 2002 (23) Life Functioning Questionnaire (LFQ) M 16 A, PKoopman et al, 2002 (24) Stanford Presenteeism Scale-6 (SPS-6) Ps 6 PKumar et al, 2003 (25) Health-Related Productivity Questionnaire Diary

(HRPQ-D)M 9 A, P

Kessler et al, 2003 (26) World Health Organization Health and WorkPerformance Questionnaire (HPQ)§

M 13 A, P

Gilworth et al, 2003 (13) Work Instability Scale for Rheumatoid Arthritis(RA-WIS)§

Ps 23 P

Goetzel et al, 2003 (27) Work Productivity Short Inventory (WellnessInventory by Pfizer) (WPSI)

M 4 A, P

Shikiar et al, 2004 (28) Health and Work Questionnaire (HWQ) M 5 PTurpin et al, 2004 (6) Stanford Presenteeism Scale-13 (SPS-13) Ps 13 A, PGignac et al, 2004 (12) Work Activity Limitations Scale (WALS)§ Ps 12 PStewart et al, 2004 (29) Work and Health Interview-The American

Productivity Audit (WHI)Ps 7 A, P

Burton et al, 2004 (30,31) Work Limitations Questionnaire-8 (WLQ-8) Ps 8 PBeaton et al, 2005 (32) Work Limitations Questionnaire-16 (WLQ-16) Ps 16 PMunir et al, 2005 (33) Work Limitations Questionnaire, 6 items (WLQm) Ps 6 PFeuerstein et al, 2005 (34) Workstyle Scale-Long Version (WSL) Ps 91 PFeuerstein et al, 2006 (35) Workstyle Scale-Short Version (WSS) Ps 32 PVan Roijen, 2007 (36) Short Form-Health Labor Questionnaire (SF-HLQ) M 11 A, POsterhaus et al, 2009 (11) Rheumatoid Arthritis Specific Work Productivity

Survey (WPS-RA)§M 9 A, P

* Modular measure (M) consists of series of global rating scales (e.g., visual analog scales or numeric rating scales) and discrete items generally notintended to be summative; contents may be organized into multiple sections within a questionnaire. Psychometric measure (Ps) consists of summativeitems that contributes to overall concept (can be multidimensional, i.e., consisting of subscales).† For modular measures organized by sections, the number of sections (not items) are indicated; count excludes sociodemographic items/sectionsunless directly pertaining to employment work (e.g., current work status, employment income).‡ Absenteeism (A) examines work status and/or extent of time/frequency of being off work (e.g., sick leave). Presenteeism (P) examines on-the-jobimpact (e.g., productivity loss associated with reduced work efficiency, or degree of workplace activity limitations).§ Identified as a candidate Outcome Measures in Rheumatology work productivity outcome measure (2).

S338 Tang et al

tivity, and degree of interference on work productivity dueto RA (presenteeism).

Number of items. 9 questions (Q), organized into 3 sec-tions. Section I: employment status, type of work (Q1);section II: employment work (days of work missed [Q2],days with productivity reduced by at least half [Q3], de-gree of interference on work productivity [Q4]); and sec-tion III: household work and activities (days of nonpaidwork missed [Q5], days with productivity reduced by atleast half [Q6], days missed family, social, or leisure activ-ities [Q7], days with outside help [Q8], degree of interfer-ence on [nonpaid] work productivity [Q9]).

Response options/scale. Unique for each individualquestion. Q1: employed outside of home (yes/no), type ofwork (8 options); Q2, Q3, and Q5–Q8: number of days(count data); Q4 and Q9: global rating scale, 0–10 (0 � nointerference, 10 � complete interference).

Recall period for items. 1 month.Endorsements. None. (The WPS-RA is 1 of 6 measures

identified as a candidate Outcome Measures in Rheuma-tology work productivity outcome.)

Examples of use. The WPS-RA has been applied as asecondary study end point to evaluate effects of certoli-zumab pegol in RA (37,11), and also as an outcome toevaluate the effects of certolizumab pegol with methotrex-ate in RA (38).

Practical Application

How to obtain. Request for permission to use theWPS-RA should be made to Global Health Outcomes Re-search, UCB Pharma. A copy of the WPS-RA questionnairecan be obtained at URL: http://www.biomedcentral.com/content/supplementary/ar2702-S1.doc.

Method of administration. Currently intended for inter-viewer administration, but there are future plans to de-velop a self-administered version.

Scoring. Each of 9 questions is scored individually (i.e.,8 separate outcome scores [Q1 not considered an out-come]), item scores are not intended to be combined.

Score interpretation. Values for Q2, Q3, and Q5–Q8reflect number of days impacted; for Q4 and Q9, higherscores indicate greater interference of RA on work. Cutpoints and normative values are not yet established.

Respondent burden. Low.Administrative burden. Low, assuming interviewer ad-

ministration.Translations/adaptations. English.

Psychometric Information

Method of development. The WPS-RA was developedthrough a literature review on work productivity associ-ated with RA or other chronic health conditions (e.g.,migraine, depression) (11). Patients/workers were not di-rectly involved in scale development, but item selectionemphasized patient-centered considerations (e.g., rele-vance, burden).

Acceptability. In a randomized trial of 220 patientswith RA, Osterhaus et al (11) reported relatively low miss-ing frequencies of 0.5% (Q6), 1.4% (Q8), 1.8% (Q9), and

0% for all other scale items. Q2, Q3, and Q5–Q8 providecount data and may have propensity for floor effect. Read-ability of WPS-RA items is high.

Reliability. Test–retest or interrater evidence not yetavailable, internal consistency testing is not applicable.

Validity. Modest evidence to date. Relationship of theWPS-RA with other work-specific measures has yet to beassessed, but evidence of known group differences againstgeneral health indicators is available. In Osterhaus et al(11), WPS-RA item scores for Q2–Q4 (employment work)were shown to differ (P � 0.05) between known groups (�third quartile versus � first quartile) based on scores fromthe Health Assessment Questionnaire disability index(HAQ DI), Short Form-36 physical component summaryscore (SF-36 PCS), and the SF-36 mental component sum-mary score (MCS). Item scores for Q5–Q9 (household workand daily activities) were also shown to differ (mostly P �0.05) between known groups (� third quartile versus �first quartile) based on scores from the HAQ DI, SF-36 PCS,and SF-36 MCS.

Ability to detect change. There are no known reports todate on the sensitivity of the WPS-RA to known changes inwork disability or productivity, but effect sizes have beenassessed in persons with RA in a 24-week certolizumabpegol trial (11). For clinical responders based on the Amer-ican College of Rheumatology (ACR20) criteria for 20%improvement (i.e., individuals showing a large change),standardized response mean (SRM) �0.8 (large effect size)was observed for Q4, Q6, and Q9, SRM 0.5–0.8 (moderateeffect size) was shown for Q3 and Q5, and SRM �0.5(small effect size) was shown for Q2, Q7, and Q8. ForACR20 nonresponders (i.e., individuals showing smallchange, no change, or deterioration), SRM �0.5 was evi-dent for all WPS-RA items (Q2–Q9). In the same trial,clinical responders based on improvements in the HAQ DI(minimum clinically important difference [MCID] 0.22,i.e., a large change in disability) had an SRM �0.8 for Q4and Q6, SRM �0.5 for Q3 and Q5, and SRM �0.5 for Q2,Q7, and Q8. Among HAQ DI nonresponders (i.e., smallchange, no change, or deterioration), SRM �0.5 was ob-served for all items (Q2–Q9). Some care is needed wheninterpreting SRMs reported for nonresponders as these arederived from assessing a pool of individuals who haveundergone varying degrees of change in the trial.

Critical Appraisal of Overall Value to theRheumatology Community

Strengths. Provides broad coverage of impacts on sev-eral work domains; questionnaire considers work impactsfrom the perspectives of both absenteeism and presentee-ism (at-work productivity) as well as impacts on bothemployment work and nonpaid activities (e.g., householdwork, social and leisure activities).

Caveats and cautions. Patients/workers were not di-rectly involved in conceptualization or scale development.Count data (number of days) are gathered for 6 of the 9questionnaire items (proper statistical treatment required).Evidence to support application beyond RA is not yetavailable. There is a large number of “outcomes” (i.e., 8separate scores derived, not intended to be summated/

Work Disability S339

combined). Q2–Q4 pertains specifically to employmentwork, therefore will not be relevant for persons who aretemporarily unemployed at the time of data collection (i.e.,lead to missing values). Number of “days with productiv-ity reduced by at least half” (Q3 and Q6) could be chal-lenging to appraise and may overlook days where there isalso considerable impact but productivity is reduced byless than half. Overall, there is limited psychometric test-ing against work-specific indicators (to support validityand responsiveness) to date, but it should be consideredthat the WPS-RA is a relatively new measure with muchpromise.

Clinical usability. Good potential, although more stud-ies to establish clinical parameters (e.g., MCID, patientacceptable symptomatic state) are needed. The WPS-RAcould be potentially useful for providing a more completeview of disease impact (i.e., on both paid and unpaidwork) to facilitate decision making on clinical manage-ment and issues around job modifications and work/lifebalance. It is also a highly feasible measure given lowrespondent and administrative burden.

Research usability. Research usability is promising.There is emerging evidence of its application and psycho-metric performance from clinical trials in RA. Low admin-istrative and respondent burden.

WORKPLACE ACTIVITY LIMITATIONSSCALE (WALS)

Description

Purpose. Measures limitations experienced while per-forming workplace activities, including difficulties associ-ated with upper-extremity functioning, lower-extremityfunctioning, concentration at work, and the pace andscheduling of work. The WALS is intended for arthritispopulations. Original publication in 2004 (12).

Content. Assesses difficulties with mobility, prolongedsitting and standing, lifting, working with hands, crouch-ing, bending or kneeling, reaching, scheduling, workhours, pace of work, concentration, and meeting currentjob demands. Respondents are asked to respond to ques-tions assuming that assistance from others or gadgets andequipment are not available.

Number of items. 11-item and 12-item versions of theWALS are available. An item asking about “difficultiesconcentrating on work” is not included in the 11-itemversion. This item was added to the 12-item version basedon patient feedback.

Response options/scale. Four-point Likert scaling: fromno difficulty (score � 0) to not able to do (score � 3). Notapplicable to my job and difficulty unrelated to arthritisresponse options are also available (both scored 0).

Recall period for items. “In general” or “typically.”Endorsements. None. (The WALS is 1 of 6 measures

identified as a candidate Outcome Measures in Rheuma-tology work productivity outcome.)

Examples of use. Has been used in samples with inflam-matory arthritis (IA; e.g., rheumatoid arthritis [RA], psori-atic arthritis), osteoarthritis (OA), and lupus. Examined as

a factor associated with arthritis-related work changes,work transitions and job accommodations (12,39–41), be-havioral coping efforts (42), arthritis–work spillover (43),chronic job stress and strain (44), and disclosure of chronicdisease in the workplace (45).

Practical Application

How to obtain. Information on the WALS can be ob-tained free of charge, from Monique Gignac, PhD, To-ronto Western Research Institute at the University HealthNetwork, Toronto, Ontario, Canada. E-mail: gignac@uhnres.utoronto.ca.

Method of administration. Self- or interviewer admin-istration.

Scoring. Can be expressed as mean of all scale items(range 0–3) or as a summed total score (summed scorerange 0–33 for 11-item version or 0–36 for 12-item ver-sion); mean values can be imputed for up to 2 missingitems.

Score interpretation. Higher scores indicate greaterworkplace activity limitations. Preliminary data on cutoffvalues have been examined (41).

Respondent burden. Low.Administrative burden. Minimal.Translations/adaptations. English.

Psychometric Information

Method of development. Items were based on a reviewof the literature and were modeled after the Health Assess-ment Questionnaire (HAQ) and modified to be specific toworkplace tasks and activities. Patients/workers were notdirectly involved in the development of the scale.

Acceptability. In a sample of 250 workers with eitherRA or OA, Beaton et al (3) reported WALS summed scoresto be available in 234 patients (6% completed �10 of the11 scale items) with 0% at floor score (WALS � 0) and3.4% at ceiling score (WALS � 33). Dhanhani (40) re-ported �1% missing values for employed and not em-ployed groups with lupus. Readability of WALS items ishigh.

Reliability. For the 11-item WALS, Gignac et al (12,43)and Gignac (42) reported a Cronbach’s alpha range from0.78–0.81 (n � 349–491 with OA or IA) over 4 timepoints, each 18 months apart. In a sample of 250 workerswith either RA or OA, Beaton et al (3) reported a Cron-bach’s alpha of 0.87. For the 12-item WALS, a Cronbach’salpha of 0.81 has been reported in a sample of 292 patientswith either OA or IA (44). Dhanhani (40) reported a Cron-bach’s alpha of 0.86 for an employed sample with lupusand 0.80 for those not working.

Validity. In a sample of 250 workers with either RA orOA, Beaton et al (3) reported a correlation range of r �0.43–0.66 against a series of work-oriented constructs(self-reported global items); the WALS also showed mod-erate-to-high correlations against other work-specific mea-sures, including the 6-item Stanford Presenteeism Scale(r � 0.66), Endicott Work Productivity Scale (r � 0.55),Work Instability Scale for Rheumatoid Arthritis (r � 0.77),and Work Limitations Questionnaire Index (r � 0.61).

S340 Tang et al

Moreover, the WALS also showed moderate correlationsagainst the HAQ disability index (r � 0.66), self-ratedarthritis severity (r � 0.62), and self-rated pain intensity(r � 0.67), which might be considered somewhat higherthan expected, given these were not work-specific indica-tors. None of the comparators tested in this study might beconsidered a gold standard reference indicator, but over-all, there is moderate support for its construct validity.

Ability to detect change. In Beaton et al’s sample of 250workers with either RA or OA (3), moderate responsive-ness of the 11-item WALS to 1-year improvements (stan-dardized response mean [SRM] �0.79) and 1-year deteri-orations in work ability (SRM 0.50) were found (ranked1st out of 5 at-work measures compared in the study),but smaller effect sizes for 1-year improvements (SRM�0.37) and 1-year deterioration in work productivity(SRM 0.18) were observed (ranked 2nd out of 5 at-workmeasures compared). In this analysis, single item globalindices of change were used to provide “reference” indi-cators of change, and individuals showing varying magni-tudes of change were pooled in the analysis (i.e., no strat-ification of individuals based on magnitude of change).Additional studies to examine the responsiveness of theWALS to more defined changes (i.e., “smaller” versus“larger”) in workplace activity limitations could be infor-mative.

Critical Appraisal of Overall Value to theRheumatology Community

Strengths. This is a feasible measure that captures manyof the key perspectives of workplace activity limitations,and has shown very good psychometric performance over-all. Included in the scale is an assessment of potentialdifficulties related to mobility (around workplace, to andfrom work), which is often omitted in other work out-comes. It also offers a difficulty unrelated to arthritis re-sponse option (scored as 0), which is also quite unique. Todate, available psychometric evidence has been gatheredfrom samples with OA, IA, and lupus, suggesting potentialutility across different arthritis conditions.

Caveats and cautions. The WALS has yet to be exam-ined to date as a true study end point or applied in clinicaltrials. There is also limited current information on inter-pretation of scores or cut points. Users should be aware ofthe 2 different versions of the measure, which have showncomparable levels of internal consistency (comparabilityof other psychometric properties not yet known). Recalltime frame is not a specified time, but is “in general” or“typically.”

Clinical usability. The WALS has good potential. It is afeasible tool that can inform impact of disease at the “ac-tivity” level (specific to the workplace). When used inconcert with other health indicators, it has the potential tohelp guide clinical decisions related to management strat-egies (e.g., the need for additional therapeutic or work-place interventions) and vocational recommendationsand/or decisions (e.g., sick leave); establishing clinicalparameters (e.g., minimum clinically important difference,patient acceptable symptomatic state) in future studieswill be useful to this end.

Research usability. Research usability is good, with lowrespondent burden and minimal administrative burden.

WORK INSTABILITY SCALE FORRHEUMATOID ARTHRITIS (RA-WIS)

Description

Purpose. Measures the extent of work instability (WI),which is defined as “a state in which the consequences ofa mismatch between an individual’s functional abilitiesand the demands of his or her job can threaten continuingemployment if not resolved” (13). Originally developedspecifically for rheumatoid arthritis (RA).

Content. This is a psychometric (summative) measurewith items covering a broad range of specific work-relatedissues (e.g., symptom control, time management, task dif-ficulties at work, cognitive distresses due to concernsabout future employability) that may signify a functionalability/job demands mismatch.

Number of items. 23.Response options/scale. Dichotomous response options:

yes/no.Recall period for items. “At the moment.”Endorsements. None. (The RA-WIS is 1 of 6 measures

identified as a candidate Outcome Measures in Rheuma-tology work productivity outcome.)

Examples of use. RA-WIS was assessed as a secondaryoutcome in a clinical trial of anti–tumor necrosis factoradalimumab on RA (46), and as an outcome for comparingoccupational therapy versus usual care in RA (47). Macedoet al (48) examined the relationships between the DiseaseActivity Score using 28-joint counts, Health AssessmentQuestionnaire (HAQ) scores, and RA-WIS (as study out-come). The RA-WIS also showed ability to predict arthri-tis-related work transitions (e.g., disability leave of ab-sence, reducing work hours, or job changes) amongworkers with RA or osteoarthritis (OA) within 1 year (49).

Practical Application

How to obtain. The RA-WIS is copyrighted to the Psy-chometric Laboratory for Health Sciences, University ofLeeds; further information available at URL: www.leeds.ac.uk/medicine/rehabmed/psychometric/Scales3.htm.

Method of administration. Patient administration (self-report).

Scoring. The RA-WIS is scored by summing responsesfrom all 23 scale items (scale range 0–23), no instructionsfor missing value available; conversion into interval-levelscaling (derived from Rasch analysis) is available for OA(50).

Score interpretation. Cut points have been establishedto differentiate levels of WI: low �10, moderate 10–17,and high �17 (13). This 3-level categorization has alsodemonstrated predictive validity for arthritis-related worktransitions within 1 year (49).

Respondent burden. Low. Easy to read.Administrative burden. Minimal.Translations/adaptations. Available in 18 languages.

Adaptation of the RA-WIS requires explicit permission

Work Disability S341

from Galen Research in Manchester. Cross-cultural valid-ity of the RA-WIS has been shown for English, Dutch,German, and French (51).

Psychometric Information

Method of development. Details of original develop-ment of the RA-WIS are reported in Gilworth et al (13).Qualitative interviews were conducted in individuals withRA to identify key themes (job flexibility, good workingrelationships, and symptom control) from which itemswere generated; 76 initial statements were identified aspotential items, which were reduced to 36 items based onability to discriminate against 5 levels of work instabilityas assessed by vocational experts, and then finally reducedto 23 items based on assessment of fit to the Rasch model(1-parameter item response theory approach).

Acceptability. In a sample of 250 workers with eitherRA or OA, Beaton et al (3) reported that RA-WIS scoreswere available in 223 patients (scores were not calculatedif �10% of items were missed), of whom 9.4% had thefloor score (RA-WIS � 0), while 0.4% were at the ceilingscore (RA-WIS � 23). Overall, readability of RA-WIS itemsis high.

Reliability. Gilworth et al (13) reported a test–retest cor-relation of r � 0.89 (n � 51); in a sample of 250 workerswith RA or OA (3), Kuder-Richardson Formula-20 (KR-20)for the RA-WIS was 0.91 and item-total correlation rangedfrom 0.34–0.71. KR-20 was 0.93 for 130 patients with OAwithin this sample (50).

Validity. In a sample of 250 workers with RA or OA (3),the level of correlation (Spearman’s rank correlation coef-ficient) against work-oriented constructs (self-reportedglobal items) was r � 0.54–0.74, which was consideredbest among 5 at-work measures compared in this study.The RA-WIS also showed moderate-to-high correlationsagainst other work-specific measures, including the Work-place Activity Limitations Scale (r � 0.77), 6-item StanfordPresenteeism Scale (r � 0.69), Endicott Work ProductivityScale (r � 0.64), and Work Limitations Questionnaire In-dex (r � 0.61), and correlated moderately with the HAQ(r � 0.66), self-rated arthritis severity (r � 0.62), and painintensity (r � 0.67). Among workers with OA (n � 130)(50), the range of correlation (Spearman’s rank correlationcoefficient) against work-oriented constructs (self-reportedglobal items) was r � 0.55–0.77, and moderate-to-highcorrelations with the HAQ (r � 0.70), arthritis severity (r �0.75), and pain intensity (r � 0.79) were also found. Ofnote is that the level of correlations between the RA-WISand health (non–work specific) measures from these stud-ies appeared to be somewhat higher than might be ex-pected (i.e., comparable to the level of correlation againstwork-specific measures). Tang et al (50) demonstrated thatthe scoring structure of the RA-WIS shows adequate fit tothe expectations of the Rasch model with only minor mod-ifications, and therefore its summed score may be consid-ered compatible for transformation into interval-level scal-ing. Proper fit to the Rasch model requires the pattern ofitem response to satisfy a number of criteria, including: 1)approximation to the Guttman structure, 2) demonstratingthe lack of differential item functioning, as well as 3)

providing evidence of unidimensionality and local inde-pendence of items. These criteria were met when RA-WISwas tested in OA.

Ability to detect change. In a sample of 250 workerswith either RA or OA (3), high responsiveness of the RA-WIS to 1-year improvements (standardized response mean[SRM] �0.64) and 1-year deteriorations in work ability(SRM 0.88) were found (ranked 2nd out of 5 at-work mea-sures compared in the study), but only small-to-negligibleeffect sizes for 1-year improvements (SRM �0.29) anddeteriorations in work productivity (SRM 0.00) werefound (ranked 5th out of 5 at-work measures compared inthe study). In a sample of 130 patients with OA, large effectsizes were observed in the RA-WIS for 1-year deteriora-tions in intrusiveness of arthritis on work (SRM 1.05), andalso for 1-year improvements in intrusiveness of arthritison work (SRM �0.78) (50). In both of these studies, singleitem global ratings of change were used to provide “refer-ence” indicators, and it should be considered that individ-uals demonstrating various magnitudes of change werepooled into the same analysis. Further studies to examinewhether the RA-WIS is similarly responsive to more de-fined changes (“smaller” versus “larger”) in work instabil-ity could be informative.

Critical Appraisal of Overall Value to theRheumatology Community

Strengths. “Work instability” (functional ability/job de-mands mismatch) is a unique concept among availablework outcomes; evidence of ability to predict for futurework transition outcomes in RA and OA suggests that itcould have a potential role in risk prognostication, inaddition to potential applications as a study end point.Patient preference for the RA-WIS over 3 other at-workdisability measures was shown in a study that sampledinjured workers with upper-extremity musculoskeletaldisorders (4), likely due to the relative simplicity of itemstatements and/or the dichotomous response options.Overall psychometric evidence is very good in both RAand OA. Alternative versions of the scale have been devel-oped for other rheumatic conditions like ankylosing spon-dylitis (Work Instability Scale for Ankylosing Spondylitis)(17).

Caveats and cautions. Some concerns about the limitedresponse options (i.e., lack of a middle-ground option be-tween “yes” and “no”) have been suggested by workerswith RA or OA (Beaton et al: unpublished observations).Some items appeared to exhibit some redundancy (50).

Clinical usability. The RA-WIS is feasible and is a mea-sure that has been well received by patients with arthritisor other musculoskeletal conditions. It has potential forclinical use for risk prognostication of adverse future workoutcomes. Support for the predictive validity of the pro-posed RA-WIS cut points has been shown, which could beapplied for risk stratification (low versus moderate versushigh WI). Further psychometric evidence and establishingclinical parameters (e.g., minimum clinically importantdifference, patient acceptable symptomatic state) will behelpful for clinical interpretation of scores.

S342 Tang et al

Research usability. Research usability is excellent. It isa versatile tool (intended concept could be of interest aseither a prognostic factor or study end point). There isminimal administrative and respondent burden, unlessscore calibrations are used (i.e., conversion of summedscores to interval-level scores).

WORK LIMITATIONS QUESTIONNAIRE (WLQ)

Description

Purpose. Measures the on-the-job impact of chronichealth conditions and treatment with a focus on assessinglimitations while performing specific job demands. Origi-nal 25-item version (WLQ-25) was published in 2001 (14),and several shortened versions have been tested or appliedin studies in workers with various musculoskeletal disor-ders: WLQ-16 (32), WLQ-8 (30,31), and a 6-item version(33). Another variant is the Work Role Functioning scale(WRF-26) (22), which has similar purpose and content, butis reversed in conceptual orientation (i.e., assesses level ofrole functioning, not limitations). The current review willfocus mainly on the WLQ-25.

Content. The WLQ-25 can be organized into 4 domains:time management (TM), addresses difficulty with handlinga job’s time and scheduling demands; physical demands(PD), examines ability to perform job tasks that involvebodily strength, movement, endurance, coordination, andflexibility; mental-interpersonal demands (MI), addressescognitively demanding tasks and on-the-job social interac-tions; and output demands (OD), concerns reduced workproductivity (14).

Number of items. 25 total scale items, divided into 4subscales: TM (5 items), PD (6 items), MI (9 items), and OD(5 items).

Response options/scale. Three of the 4 subscales (TM,MI, OD) examine proportion of time with difficulty: “noneof the time (0%),” score � 0; “a slight bit of the time,”score � 1; “some of the time (50%),” score � 2; “most ofthe time,” score � 3; “all of the time (100%),” score � 4;plus a “does not apply to my job” option (treated as miss-ing, no score). The PD subscale has reverse instructionsand examines proportion of time without difficulty (sameresponse options provided).

Recall period for items. 2 weeks for the WLQ-25. Thisvaries with other versions.

Endorsements. None. (The WLQ-25 is 1 of 6 measuresidentified as a candidate Outcome Measures in Rheuma-tology work productivity outcome.)

Examples of use. Rohekar and Pope (52) applied theWLQ-25 as an outcome to assess work disability in sero-negative spondylarthritis. Allaire et al (53) applied theWLQ-25 to examine the impact of rheumatoid arthritis(RA) on work disability among 5,419 older workers (agerange 55–64 years) with RA. Lerner et al (54) recentlydeveloped a method to impute work productivity impactusing other health variables, following an examination ofthe relationship between the WLQ-25 and an array of pain,functioning, and general health measures. Tang et al (4)provided a head-to-head comparison of the psychometricperformance of 4 at-work disability measures (including

WLQ-16) among injured workers with upper-extremitydisorders. Zhang et al (55) examined the comparability ofmethods to estimate productivity loss based on 4 differentinstruments (includes conversions based on the WLQ In-dex). Associations between medical conditions (31), ar-thritis (56), and health risks (57) with work limitations hadbeen examined using the WLQ-8.

Practical Application

How to obtain. The WLQ is copyrighted: Work Limita-tions Questionnaire 1998, by The Health Institute, LernerD, Amick B 3rd, GlaxoWellcome. The WLQ is providedfree of charge for noncommercial applications.

Method of administration. Patient administration (self-report).

Scoring. Multiple approaches have been recommendedto score the WLQ-25. Subscales are scored by multiplyingthe mean of subscale items by 25 (range 0–100, 100 � mostlimitations, response orientation needs to be reversed forthe PD subscale), missing response for up to 50% of sub-scale items is allowable for subscale scoring. Scales arescored by multiplying the mean of all scale items by 25(range 0–100, 100 � most limitations, response orientationneeds to be reversed for the PD subscale), missing responsefor up to 50% of subscale items is also allowable for scalescoring (Amick BC: unpublished observations). Weightedindex scoring (WLQ Index) is calculated from subscalescores using a weighted formula based on an analysis ofthe relationship between WLQ scores and actual employeeproductivity loss relative to healthy employees (58).Scores from all 4 subscales are needed to calculate theWLQ Index, computer scoring is necessary. Formulas forcalculating the WLQ Index and conversion to productivityloss estimates versus healthy controls (range 0–25%) arepublished in a technical report available from the devel-opers (58).

Score interpretation. Normative WLQ subscale means(SE) have been reported in a small sample of healthyworkers recruited from Massachusetts (37): PD � 4.5 (1.4);TM � 7.2 (3.1); MI � 10.6 (2.7); OD � 7.2 (2.8). A 10%increase in WLQ subscale score has been proposed toequate a productivity decline of 4–5% (59).

Respondent burden. Moderate. Lerner et al (60) re-ported an administration time of approximately 30 min-utes for workers with osteoarthritis (OA), and approxi-mately 15 minutes for healthy controls. Some workershave expressed concern over the flipping of instructionsfor different sections of the questionnaire (i.e., the TM, MI,and OD subscales ask about amount of time with difficulty,while the PD subscale asks about amount of time withoutdifficulty).

Administrative burden. Moderate. Computer scoring isrequired for handling/imputation of missing values, andfor calculating the WLQ Index.

Translations/adaptations. Over 30 official languagetranslations.

Psychometric Information

Method of development. Content and format of theWLQ-25 were developed from focus groups (workers with

Work Disability S343

chronic conditions), cognitive interviews, and an alternateform comparison (stem/response) (14); 70 job demand-level limitation items and 7 dimensions were originallygenerated, which were reduced to 25 items through cog-nitive interviews (14).

Acceptability. Lerner et al (60) reported �1% missingdata in OA, floor effect � 20.4–25.8% (subscales), andceiling effect � 0.9–2.2% (subscales). In workers with RA,Walker et al (61) reported a missing proportion of 3.1–21.8% for individual scale items and 3.1–5.6% for sub-scale scores, and as a result, the WLQ Index score wasunavailable in 10.1% of the sample due to missing data. Ina study that recruited 250 workers with RA or OA, Beatonet al (3) reported 5.6% (WLQ Index) and 19.9–35.5% (sub-scales) of the sample had the floor score, while 0% (WLQIndex) and 1.2–3.3% (subscales) of the sample had theceiling score.

Reliability. Among workers with OA, Lerner et al (60)reported the following item-to-total correlation coeffi-cients and Cronbach’s alpha for the 4 WLQ-25 subscales:PD: 0.72–0.82, � � 0.93; TM: 0.79–0.92, � � 0.95; MI:0.81–0.92, � � 0.97; OD: 0.82–0.89, � � 0.96. In RA,Walker et al (61) reported a Cronbach’s alpha range of0.83–0.88. In a pooled sample of workers with RA or OA,Beaton et al (3) reported the following item-total correla-tions and Cronbach’s alpha for the 4 WLQ-25 subscales:PD: 0.38–0.63, � � 0.77; TM: 0.60–0.76, � � 0.86; MI:0.50–0.90, � � 0.94; OD: 0.61–0.81, � � 0.88.

Validity. In OA (60), each of the 4 WLQ-25 subscalescores showed a logical and statistically significant asso-ciation (P � 0.001 in analysis of variance [ANOVA] F test)with self-reported arthritis severity (4 levels: poor to verygood); specific WLQ-25 subscales also showed linear as-sociations (subscales demonstrating P � 0.05 in ANOVA Ftest in parentheses below) with level of arthritis pain (PD,TM, MI, OD), joint stiffness (PD, TM, MI, OD), functionallimitations due to arthritis (PD, TM), Short Form 12 (SF-12) physical component score (PCS; PD), self-reportedwork productivity (OD), and work absences (OD). In RA(61), the WLQ Index showed low-to-moderate correlationswith the SF-36 mental component score (r � �0.60), SF-36PCS (r � �0.49), fatigue visual analog scale (VAS; r �0.50), pain VAS (r � 0.46); Health Assessment Question-naire (HAQ; r � 0.56), HAQ-II (r � 0.54), depression score(r � 0.46) and anxiety score (r � 0.41) from the ArthritisImpact Measurement Scale, days with limited activities(r � 0.38), and days unable to work (r � 0.29). Wolfe et al(62) reported similar levels of correlation between theWLQ Index and HAQ (r � 0.57), HAQ-II (r � 0.55), mod-ified HAQ (r � 0.55), SF-36 PCS (r � 0.50), and VAS pain(r � 0.47) when tested in a population of workers with RA.In a sample of 250 workers with either RA or OA (3), theWLQ Index showed moderate correlations (r � 0.49–0.67)against a series of work-oriented constructs (self-reportedglobal items), and also against the HAQ (r � 0.49), arthritisseverity (r � 0.42), and self-rated pain intensity (r � 0.48).In this study, the WLQ Index also showed moderate-to-high correlations against other work-specific measures, in-cluding the Workplace Activity Limitations Scale (r �0.61), 6-item Stanford Presenteeism Scale (r � 0.63), En-dicott Work Productivity Scale (r � 0.61), and Work Insta-

bility Scale for Rheumatoid Arthritis (r � 0.67). Overall,relationships between the WLQ-25 and various health-and work-related indicators were by and large in line withexpectations, providing strong support for its constructvalidity.

Ability to detect change. In a sample of 250 workerswith either RA or OA (3), levels of responsiveness of theWLQ Index to 1-year improvements (standardized re-sponse mean [SRM] �0.28) and 1-year deteriorations inwork ability (SRM 0.20) were modest (ranked 5th out of 5at-work measures compared in the study). Varying levelsof responsiveness to 1-year improvements (SRM �0.64)compared to 1-year deteriorations in work productivity(SRM 0.08) were also reported (tied for 3rd out of 5 at-workmeasures compared). In this study, single item global rat-ings of change were applied as “reference” indicators, andindividuals experiencing varying levels of changes werepooled into the same analysis (i.e., no stratification ofindividuals based on magnitude of change). Further eval-uations are needed to examine level of responsiveness ofthe WLQ-25 against more defined magnitudes of change(“smaller” versus “larger”) in work limitations.

Critical Appraisal of Overall Value to theRheumatology Community

Strengths. The WLQ has a solid foundation of develop-ment, which has led to well-defined domains that havebeen consistently applied in various studies to assess theimpact of health conditions on specific aspects of work.The breadth of potential work limitations examined ishigh, and should have strong relevance for many job typesand health conditions. To date, this is one of the mostwidely used measures in the field, with studies in RA, OA,and several other musculoskeletal conditions; cumulativeevidence of its construct validity is excellent. The WLQIndex can be converted to provide an estimate percentageof productivity loss, a unique property among availablemeasures in the field, which is potentially useful to bridgemeasurement needs for dual clinical/economic costingpurposes.

Caveats and cautions. Several variations of the measureexist in the literature (WLQ-25, WLQ-16, WLQ-8, WRF-26)and approaches to score are unique among tools and acrossexisting studies. The different versions also vary in termsof recall period, specific wording of items, and whetherspecific sections (i.e., PD subscale) have reverse orienta-tion within the full questionnaire. Care must be takenwhen making comparisons of results across versions. Forthe WLQ-25, reverse instructions for the PD domain mayconfuse respondents and could be a source of error. Thereis also a generous allowance of missing data (up to 50% ofmissing items may be imputed).

Clinical usability. Moderate administrative and respon-dent burden should be considered. More studies toestablish clinically meaningful cut points (e.g., minimumclinically important difference, patient acceptable symp-tomatic state) could help improve clinical usability.

Research usability. Excellent. Very good-to-excellentpsychometric evidence in arthritis as well as in otherclinical populations. High potential for use for economic

S344 Tang et al

costing purposes given the time orientation of responseoptions, and purported relationship between WLQ indexand level of work productivity reported by Lerner et al(59).

WORK PRODUCTIVITY AND ACTIVITYIMPAIRMENT QUESTIONNAIRE (WPAI)

Description

Purpose. Measures the effect of health and symptomseverity on work productivity and nonwork activities.Two versions are available: general health (WPAI:GH) orspecific health problem (WPAI:SHP), the latter is designedsuch that it can be modified for any health problem byspecifying the disease/condition of interest in the ques-tions (i.e., to derive disease-specific versions of the scale).The original WPAI:GH was published in 1993 (15); ap-proach to scoring the questionnaire has changed since theoriginal publication.

Content. Examines the extent of absenteeism, presen-teeism, and impairment in daily activities attributable togeneral health (WPAI:GH) or a specific health problem(WPAI:SHP).

Number of items. 6 questions (Q), each with uniqueresponse options: current employment status (Q1), num-ber of hours missed due to health problem (Q2), number ofhours missed due to other reasons (Q3), hours actuallyworked (Q4), degree to which health affected productivitywhile working (Q5), degree to which health affected regu-lar (nonwork) activities (Q6). Items are not intended to besummative.

Response options/scale. Q1: yes/no, Q2–Q4: number ofhours (count data), Q5: global rating scale, 0–10 (0 �health problems had no effect on my work, 10 � healthproblems completely prevented me from working), Q6:global rating scale, 0–10 (0 � health problems had noeffect on my daily activities, 10 � health problems com-pletely prevented me from doing my daily activities).

Recall period for items. 7 days.Endorsements. None. (The WPAI:GH is 1 of 6 measures

identified as a candidate Outcome Measures in Rheuma-tology work productivity outcome.)

Examples of use. Stockl et al (63) applied the WPAI:GHas a secondary outcome to evaluate the effect of arheumatoid arthritis (RA) disease treatment managementprogram. In a study by Zhang et al (55), productivity costsassociated with arthritis estimated with the WPAI:GHwas compared against estimates made from 3 othermeasures (Health and Labor Questionnaire [HLQ], TheHealth and Work Performance Questionnaire, and WorkLimitations Questionnaire-25). Haibel et al (64) appliedthe WPAI for ankylosing spondylitis (AS; WPAI:SpA) asan outcome measure to examine the efficacy of infliximabtherapy in nonsteroidal antiinflammatory drug–refractorypatients with AS. Associations among demographic,health-related, and treatment factors with the level of workproductivity (WPAI:SpA) have been recently examined inpatients with AS (65).

Practical Application

How to obtain. The WPAI is available at URL: www.reillyassociates.net/Index.html. Permission and fees arenot required to use the WPAI.

Method of administration. Self-administered or inter-viewer administered.

Scoring. Detailed information is provided in the aboveweb site; 4 outcome (OC) scores can be derived from theWPAI: OC1, percent work time missed due to health, Q2/(Q2 � Q4) (percentage of absenteeism); OC2, percent im-pairment while working due to health, Q5/10 (percentageof presenteeism); OC3, percent overall work impairmentdue to health, Q2/(Q2 � Q4) � [(1 � Q2/(Q2�Q4)) �(Q5/10)]; OC4, percent activity impairment due to health,Q6/10.

Score interpretation. For all 4 outcomes, greater scores(range 0–100%) indicate greater impact of health, clini-cally important cut points not yet established.

Respondent burden. Minimal.Administrative burden. Low, only basic calculations

needed.Translations/adaptations. Translated in more than 80

languages (see URL: www.reillyassociates.net/WPAI_Translations.html). The WPAI:SHP can theoretically beadapted to any specific disease or health problem. Psycho-metric evidence is available for a wide range of diseases.Recently, a version of WPAI for AS has been developed(WPAI:SpA) (16).

Psychometric Information

Method of development. WPAI items were generatedfrom 3 sources (15): 1) review of work productivity litera-ture, 2) comments from patients with allergic rhinitis onan interviewer-administered version of the WPAI itemsfrom a series of clinical studies, and 3) cognitive debriefingof subjects following interviewer administration and self-administration of WPAI items to determine final wording.

Acceptability. In the original study that included a sam-ple of workers with RA or other musculoskeletal disorders(15), up to 21% of the sample had missing data on WPAIquestions when the measure was self-administered, butminimal missing data when the questions were inter-viewer administered. The WPAI:SpA had �10% missingwhen self-administered (16).

Reliability. In the original study by Reilly et al (15),only a modest range of Pearson’s correlation coefficient(0.71–0.75) was found in a test–retest comparison ofWPAI:GH fielded in workers with nonspecific health prob-lems within the same day (at least 4 hours later). Level ofagreement (e.g., intraclass correlation coefficients) be-tween test–retest scores was not reported in this study.

Validity. Some support for the construct validity of theWPAI:GH in RA was provided in Zhang et al (66), althoughthe scale appeared to correlate more strongly to healthstatus indictors than work-specific comparators where theopposite might be expected. OC1 showed a correlation ofr � 0.56 with the number of absent workdays in the past 3months (question adapted from the Productivity and Dis-ease Questionnaire), OC2 showed a correlation of r � 0.39

Work Disability S345

with the number of hours lost due to presenteeism (itemfrom the HLQ), and OC4 showed a correlation of r � 0.39with the number of hours getting help on unpaid workactivities (item from the HLQ). In this study, 3 of 4WPAI:GH outcomes (OC2, OC3, and OC4) also showedmoderate-to-high correlations (r � 0.67–0.77) against aseries of health status outcomes (function, pain, patientglobal estimate on health impact, fatigue, and patientglobal assessment of disease activity). In an adalimumabversus placebo clinical trial in AS (16), WPAI:SpA out-comes were shown to be able to discriminate between“higher” and “lower” scores (split on the basis of medianscore in the sample) in the Bath Ankylosing SpondylitisDisability Activity Index (BASDAI), the Ankylosing Spon-dylitis Quality of Life (ASQOL), Short Form 36 (SF-36)physical component summary, SF-36 mental componentsummary, and Health Utility Index-3 at P � 0.05. Evidenceof known group differences is consistent with a prioriexpectations in the trial, although it should be recognizedthat the comparators applied were not work-specific indi-cators; additional studies to specifically examine the re-sponsiveness of the WPAI to “true” changes in work pro-ductivity would be informative.

Ability to detect change. There are no known reports onthe responsiveness of the WPAI against comparable indi-cators of work productivity to date, but effect sizes forpersons with AS in a 24-week adalimumab versus placebotrial have been reported by Reilly et al (16). In this study,among clinical responders based on improvements in theBASDAI (�1.96 decrease in score, i.e., a large change indisease activity), standardized response means (SRMs) forWPAI outcomes were OC1: �0.25; OC2: �0.86; OC3:�0.89; OC4: �1.29. Among nonresponders (i.e., smallchange, no changes, or deterioration), SRMs were OC1:�0.14; OC2: �0.52; OC3: �0.54; OC4: �0.39. Among clin-ical responders based on changes in the ASQOL (�1.8decrease in score, i.e., a large change in quality of life),SRMs were OC1: �0.31; OC2: �0.89; OC3: �0.94; OC4:�1.18. Among ASQOL clinical nonresponders (i.e., smallchange, no change, or deterioration), SRMs were OC1:�0.11; OC2: �0.46; OC3: �0.38; OC4: �0.40. It is impor-tant to exercise care when interpreting SRMs reported for“nonresponders” as this is derived from a pool of individ-uals who have undergone varying degrees of change overthe course of the trial.

Critical Appraisal of Overall Value to theRheumatology Community

Strengths. The WPAI is designed to have generalizabil-ity to a broad range of occupations/diseases, and evidenceof reliability and validity is available in many musculosk-eletal and nonmusculoskeletal conditions. Item content ishighly consistent across different versions of the measure(WPAI:GH versus WPAI:SHP), which should facilitatecomparison of outcome scores in different studies (onlydisease attribution of items differ between versions). It hasan intuitive method to score, is compatible with economiccosting (orientation of response is based on amount of timeaffected), and has low respondent burden.

Caveats and cautions. Four separate outcome scores arederived from the questionnaire (not intended to be sum-mated/combined). There have been important changes inthe approach to score the WPAI questionnaire since itsoriginal development.

Clinical usability. The WPAI has good potential. Estab-lishing clinical parameters (e.g., minimum clinically im-portant difference, patient acceptable symptomatic state)will be helpful for clinical interpretability. Administrativeand respondent burdens are low.

Research usability. Research usability is good. Appli-cability and good psychometric performance of the WPAIhas been shown in several clinical trials with patients withAS.

DISCUSSION

The current review has revealed a moderate level of evi-dence to date to support the psychometric properties of 5selected measures of work disability and productivity inarthritis/musculoskeletal populations. It is important torecognize that the Work Productivity and Activity Impair-ment Questionnaire and Work Limitations Questionnairewere developed as generic (not disease specific) instru-ments, and there exists additional evidence in the litera-ture to support their psychometric performance in otherpopulations that have not been reviewed in detail in thisarticle. On the other hand, evidence of psychometric per-formance for the Workplace Activity Limitations Scale,Work Instability Scale for Rheumatoid Arthritis, and inparticular the Rheumatoid Arthritis Specific Work Produc-tivity Survey has only emerged in the past few years, sincethese are relatively new measures.

We believe there is room for continued growth in thisarea of research. Specific measurement attributes requir-ing further research include examinations of test–retestreliability, validation against work-specific constructsand/or productivity data (where relevant), and assess-ments of responsiveness to more defined magnitudesof change (e.g., sensitivity to “smaller” versus “larger”known changes). Also, data on clinically relevant param-eters (e.g., minimum clinically important difference) arescarce to date and will be important to establish to helpevaluate treatment efficacy in research trials and longitu-dinal observational studies.

Issues on the interpretability of scores derived frommeasures of work disability and productivity are also ofemerging interest. An important concept to recognize isthat the extent of disease impact on work is ultimately afunction of both the person and his or her work context(i.e., environmental factors) and the manner in which theyinteract (2,67). At the individual level, a change in scorecould reflect a change in the person’s capacity to work,and/or a change in the demands of the job, for example, inthe case where a work transition (e.g., job modifications,reduced work hours) has taken place to allow a personwith arthritis to function better at work. To provide a morecomplete understanding of the bases of change over time,users may consider fielding additional instruments thatcan offer insights into the work context (e.g., job type,

S346 Tang et al

work status, contractual hours, availability of workplacesupport) to supplement outcome measures designed toquantify the level of work disability and productivity.

Overall, the current diversity of available measures inthis field is impressive. While the availability of a widerange of instruments can provide users with many options,some care is important when selecting an outcome to meetthe needs of a particular research study or clinical pur-pose. The specific work-related concept or measurementperspective being sought, the availability of supportingpsychometric evidence, and pragmatic considerations(e.g., applicability, feasibility) should be concurrently con-sidered. In addition to the summary of evidence providedin the current article, users may also consider additionalfindings and insights from a number of recent studies (3,4)that have examined the head-to-head psychometric per-formance of multiple work measures in arthritis/musculo-skeletal populations to help inform the selection of out-comes in future research or clinical applications.

ACKNOWLEDGMENTSThe authors would like to thank members of the OutcomeMeasures in Rheumatology (OMERACT) Work Productiv-ity Special Interest Group, Canadian Arthritis NetworkWork Productivity Group, and I-CAN-Work Alliance forcontributions to this work. We also acknowledge QuenbyMahood and Denise Linton for providing assistance withliterature search and manuscript preparation.

AUTHOR CONTRIBUTIONS

All authors were involved in drafting the article or revising itcritically for important intellectual content, and all authors ap-proved the final version to be published.

REFERENCES

1. Escorpizo R, Bombardier C, Boonen A, Hazes JM, Lacaille D, Strand V,et al. Worker productivity outcome measures in arthritis. J Rheumatol2007;34:1372–80.

2. Beaton D, Bombardier C, Escorpizo R, Zhang W, Lacaille D, Boonen A,et al. Measuring worker productivity: frameworks and measures.J Rheumatol 2009;36:2100–9.

3. Beaton DE, Tang K, Gignac MA, Lacaille D, Badley EM, Anis AH, et al.Reliability, validity, and responsiveness of five at-work productivitymeasures in patients with rheumatoid arthritis or osteoarthritis. Arthri-tis Care Res (Hoboken) 2010;62:28–37.

4. Tang K, Pitts S, Solway S, Beaton D. Comparison of the psychometricproperties of four at-work disability measures in workers with shoulderor elbow disorders. J Occup Rehabil 2009;19:142–54.

5. Lavigne JE, Phelps CE, Mushlin A, Lednar WM. Reductions in individ-ual work productivity associated with type 2 diabetes mellitus. Phar-macoeconomics 2003;21:1123–34.

6. Turpin RS, Ozminkowski RJ, Sharda CE, Collins JJ, Berger ML, BillottiGM, et al. Reliability and validity of the Stanford Presenteeism Scale.J Occup Environ Med 2004;46:1123–33.

7. Ozminkowski RJ, Goetzel RZ, Chang S, Long S. The application of twohealth and productivity instruments at a large employer. J Occup En-viron Med 2004;46:635–48.

8. Meerding WJ, IJzelenberg W, Koopmanschap MA, Severens JL, BurdorfA. Health problems lead to considerable productivity loss at workamong workers with high physical load jobs. J Clin Epidemiol 2005;58:517–23.

9. Brouwer WB, Koopmanschap MA, Rutten FF. Productivity losses with-out absence: measurement validation and empirical evidence. HealthPolicy 1999;48:13–27.

10. Sanderson K, Tilse E, Nicholson J, Oldenburg B, Graves N. Whichpresenteeism measures are more sensitive to depression and anxiety?J Affect Disord 2007;101:65–74.

11. Osterhaus JT, Purcaru O, Richard L. Discriminant validity, responsive-ness and reliability of the rheumatoid arthritis-specific Work Produc-tivity Survey (WPS-RA). Arthritis Res Ther 2009;11:R73.

12. Gignac MA, Badley EM, Lacaille D, Cott CC, Adam P, Anis AH. Man-aging arthritis and employment: making arthritis-related work changesas a means of adaptation. Arthritis Rheum 2004;51:909–16.

13. Gilworth G, Chamberlain MA, Harvey A, Woodhouse A, Smith J,Smyth MG, et al. Development of a work instability scale for rheuma-toid arthritis. Arthritis Rheum 2003;49:349–54.

14. Lerner D, Amick BC 3rd, Rogers WH, Malspeis S, Bungay K, Cynn D.The Work Limitations Questionnaire. Med Care 2001;39:72–85.

15. Reilly MC, Zbrozek AS, Dukes EM. The validity and reproducibility ofa work productivity and activity impairment instrument. Pharmaco-economics 1993;4:353–65.

16. Reilly MC, Gooch KL, Wong RL, Kupper H, van der Heijde D. Validity,reliability and responsiveness of the Work Productivity and ActivityImpairment Questionnaire in ankylosing spondylitis. Rheumatology(Oxford) 2010;49:812–9.

17. Gilworth G, Emery P, Barkham N, Smyth MG, Helliwell P, Tennant A.Reducing work disability in ankylosing spondylitis: development of awork instability scale for AS. BMC Musculoskelet Disord 2009;10:68.

18. Osterhaus JT, Gutterman DL, Plachetka JR. Healthcare resource and lostlabour costs of migraine headache in the US. Pharmacoeconomics1992;2:67–76.

19. Van Roijen L, Essink-Bot ML, Koopmanschap MA, Bonsel G, Rutten FF.Labor and health status in economic evaluation of health care: theHealth and Labor Questionnaire. Int J Technol Assess Health Care1996;12:405–15.

20. Endicott J, Nee J. Endicott Work Productivity Scale (EWPS): a newmeasure to assess treatment effects. Psychopharmacol Bull 1997;33:13–6.

21. Kopec JA, Esdaile JM. Occupational role performance in persons withback pain. Disabil Rehabil 1998;20:373–9.

22. Amick BC 3rd, Lerner D, Rogers WH, Rooney T, Katz JN. A review ofhealth-related work outcome measures and their uses, and recom-mended measures. Spine (Phila Pa 1976) 2000;25:3152–60.

23. Altshuler L, Mintz J, Leight K. The Life Functioning Questionnaire(LFQ): a brief, gender-neutral scale assessing functional outcome. Psy-chiatry Res 2002;112:161–82.

24. Koopman C, Pelletier KR, Murray JF, Sharda CE, Berger ML, Turpin RS,et al. Stanford presenteeism scale: health status and employee produc-tivity. J Occup Environ Med 2002;44:14–20.

25. Kumar RN, Hass SL, Li JZ, Nickens DJ, Daenzer CL, Wathen LK.Validation of the Health-Related Productivity Questionnaire Diary(HRPQ-D) on a sample of patients with infectious mononucleosis:results from a phase 1 multicenter clinical trial. J Occup Environ Med2003;45:899–907.

26. Kessler RC, Barber C, Beck A, Berglund P, Cleary PD, McKenas D, et al.The World Health Organization Health and Work Performance Ques-tionnaire (HPQ). J Occup Environ Med 2003;45:156–74.

27. Goetzel RZ, Ozminkowski RJ, Long SR. Development and reliabilityanalysis of the Work Productivity Short Inventory (WPSI) instrumentmeasuring employee health and productivity. J Occup Environ Med2003;45:743–62.

28. Shikiar R, Halpern MT, Rentz AM, Khan ZM. Development of theHealth and Work Questionnaire (HWQ): an instrument for assessingworkplace productivity in relation to worker health. Work 2004;22:219–29.

29. Stewart WF, Ricci JA, Leotta C, Chee E. Validation of the work andhealth interview. Pharmacoeconomics 2004;22:1127–40.

30. Burton WN, Chen CY, Conti DJ, Pransky G, Edington DW. Caregivingfor ill dependents and its association with employee health risks andproductivity. J Occup Environ Med 2004;46:1048–56.

31. Burton WN, Pransky G, Conti DJ, Chen CY, Edington DW. The associ-ation of medical conditions and presenteeism. J Occup Environ Med2004;46 Suppl:S38–45.

32. Beaton DE, Kennedy CA. Beyond return to work: testing a measure ofat-work disability in workers with musculoskeletal pain. Qual Life Res2005;14:1869–79.

33. Munir F, Jones D, Leka S, Griffiths A. Work limitations and employeradjustments for employees with chronic illness. Int J Rehabil Res2005;28:111–7.

34. Feuerstein M, Nicholas RA, Huang GD, Haufler AJ, Pransky G, Robert-son M. Workstyle: development of a measure of response to work inthose with upper extremity pain. J Occup Rehabil 2005;15:87–104.

35. Feuerstein M, Nicholas RA. Development of a short form of the Work-style measure. Occup Med (Lond) 2006;56:94–9.

36. Van Roijen L. Short Form - Health and Labour Questionnaire. Institutefor Medical Technology Assessment, Erasmus MC, University Med-ical Centre Rotterdam; 2007.

37. Hazes JM, Taylor P, Strand V, Purcaru O, Coteur G, Mease P. Physicalfunction improvements and relief from fatigue and pain are associatedwith increased productivity at work and at home in rheumatoid arthri-

Work Disability S347

tis patients treated with certolizumab pegol. Rheumatology (Oxford)2010;49:1900–10.

38. Kavanaugh A, Smolen JS, Emery P, Purcaru O, Keystone E, Richard L,et al. Effect of certolizumab pegol with methotrexate on home and workplace productivity and social activities in patients with active rheuma-toid arthritis. Arthritis Rheum 2009;61:1592–600.

39. Gignac MA, Cao X, Lacaille D, Anis AH, Badley EM. Arthritis-relatedwork transitions: a prospective analysis of reported productivity losses,work changes, and leaving the labor force. Arthritis Rheum 2008;59:1805–13.

40. Dhanhani A. The workplace challenges of patients with lupus. Gradu-ate Department of the Institute of Medical Science, University ofToronto; 2010. URL: http://hdl.handle.net/1807/25407.

41. Gignac MA, Cao X, Tang K, Beaton DE. An examination of arthritis-related work place activity limitations and intermittent disability overfour-and-a-half years and its relationship to job modifications andoutcomes. Arthritis Care Res (Hoboken) 2011;63:953–62.

42. Gignac MA. Arthritis and employment: an examination of behavioralcoping efforts to manage workplace activity limitations. ArthritisRheum 2005;53:328–36.

43. Gignac MA, Sutton D, Badley EM. Reexamining the arthritis-employ-ment interface: perceptions of arthritis-work spillover among em-ployed adults. Arthritis Rheum 2006;55:233–40.

44. Gignac MA, Sutton D, Badley EM. Arthritis symptoms, the work envi-ronment, and the future: measuring perceived job strain among em-ployed persons with arthritis. Arthritis Rheum 2007;57:738–47.

45. Gignac MA, Cao X. “Should I tell my employer and coworkers I havearthritis?” A longitudinal examination of self-disclosure in the workplace. Arthritis Rheum 2009;61:1753–61.

46. Bejarano V, Quinn M, Conaghan PG, Reece R, Keenan AM, Walker D,et al. Effect of the early use of the anti–tumor necrosis factor adali-mumab on the prevention of job loss in patients with early rheumatoidarthritis. Arthritis Rheum 2008;59:1467–74.

47. Macedo AM, Oakley SP, Panayi GS, Kirkham BW. Functional and workoutcomes improve in patients with rheumatoid arthritis who receivetargeted, comprehensive occupational therapy. Arthritis Rheum 2009;61:1522–30.

48. Macedo A, Oakley S, Gullick N, Kirkham B. An examination of workinstability, functional impairment, and disease activity in employedpatients with rheumatoid arthritis. J Rheumatol 2009;36:225–30.

49. Tang K, Beaton DE, Gignac MA, Lacaille D, Zhang W, Bombardier C,and the Canadian Arthritis Network Work Productivity Group. TheWork Instability Scale for rheumatoid arthritis predicts arthritis-relatedwork transitions within 12 months. Arthritis Care Res (Hoboken) 2010;62:1578–87.

50. Tang K, Beaton DE, Lacaille D, Gignac MA, Zhang W, Anis AH, et al.The Work Instability Scale for Rheumatoid Arthritis (RA-WIS): does itwork in osteoarthritis? Qual Life Res 2010;19:1057–68.

51. Gilworth G, Emery P, Gossec L, Vliet Vlieland TP, Breedveld FC,Hueber AJ, et al. Adaptation and cross-cultural validation of the rheu-matoid arthritis work instability scale (RA-WIS). Ann Rheum Dis 2009;68:1686–90.

52. Rohekar S, Pope J. Assessment of work disability in seronegative spon-dyloarthritis. Clin Exp Rheumatol 2010;28:35–40.

53. Allaire S, Wolfe F, Niu J, Lavalley M, Michaud K. Work disability andits economic effect on 55–64-year-old adults with rheumatoid arthritis.Arthritis Rheum 2005;53:603–8.

54. Lerner D, Chang H, Rogers WH, Benson C, Schein J, Allaire S. A methodfor imputing the impact of health problems on at-work performanceand productivity from available health data. J Occup Environ Med2009;51:515–24.

55. Zhang W, Gignac MA, Beaton D, Tang K, Anis AH. Productivity lossdue to presenteeism among patients with arthritis: estimates from 4instruments. J Rheumatol 2010;37:1805–14.

56. Burton WN, Chen CY, Schultz AB, Conti DJ, Pransky G, Edington DW.Worker productivity loss associated with arthritis. Dis Manag 2006;9:131–43.

57. Burton WN, Chen CY, Conti DJ, Schultz AB, Pransky G, Edington DW.The association of health risks with on-the-job productivity. J OccupEnviron Med 2005;47:769–77.

58. Lerner D, Rogers WH, Chang H. Scoring the work limitations question-naire (WLQ) and the WLQ index for estimating work productivity loss.Technical Report. Boston: The Health Institute, Tufts-New EnglandMedical Center; 2003.

59. Lerner D, Amick BC 3rd, Lee JC, Rooney T, Rogers WH, Chang H, et al.Relationship of employee-reported work limitations to work produc-tivity. Med Care 2003;41:649–59.

60. Lerner D, Reed JI, Massarotti E, Wester LM, Burke TA. The WorkLimitations Questionnaire’s validity and reliability among patientswith osteoarthritis. J Clin Epidemiol 2002;55:197–208.

61. Walker N, Michaud K, Wolfe F. Work limitations among workingpersons with rheumatoid arthritis: results, reliability, and validity ofthe work limitations questionnaire in 836 patients. J Rheumatol 2005;32:1006–12.

62. Wolfe F, Michaud K, Choi HK, Williams R. Household income andearnings losses among 6,396 persons with rheumatoid arthritis. J Rheu-matol 2005;32:1875–83.

63. Stockl KM, Shin JS, Lew HC, Zakharyan A, Harada AS, Solow BK, et al.Outcomes of a rheumatoid arthritis disease therapy management pro-gram focusing on medication adherence. J Manag Care Pharm 2010;16:593–604.

64. Haibel H, Song IH, Rudwaleit M, Listing J, Hildemann S, Sieper J.Multicenter open-label study with infliximab in active ankylosingspondylitis over 28 weeks in daily practice. Clin Exp Rheumatol 2008;26:247–52.

65. Maksymowych WP, Gooch KL, Wong RL, Kupper H, van der Heijde D.Impact of age, sex, physical function, health-related quality of life, andtreatment with adalimumab on work status and work productivity ofpatients with ankylosing spondylitis. J Rheumatol 2010;37:385–92.

66. Zhang W, Bansback N, Boonen A, Young A, Singh A, Anis AH. Validityof the work productivity and activity impairment questionnaire: gen-eral health version in patients with rheumatoid arthritis. Arthritis ResTher 2010;12:R177.

67. Sandqvist JL, Henriksson CM. Work functioning: a conceptual frame-work. Work 2004;23:147–57.

S348 Tang et al

Su

mm

ary

Tab

lefo

rW

ork

Dis

abil

ity

and

Pro

du

ctiv

ity

Mea

sure

s*

Mea

sure

Purp

ose/

cont

ent

Met

hod

ofad

min

istr

atio

nR

espo

nden

tbu

rden

Adm

inis

trat

ion

burd

enSc

ore

inte

rpre

tati

onR

elia

bili

tyev

iden

ce†

Val

idit

yev

iden

ce†

Abi

lity

tode

tect

chan

ge†

Stre

ngth

sC

auti

ons

Rh

eum

atoi

dA

rth

riti

sS

pec

ific

Wor

kP

rod

uct

ivit

yS

urv

ey(W

PS

-R

A)

Imp

act

ofR

Aon

emp

loym

ent

wor

k,h

ouse

hol

dw

ork,

dai

lyac

tivi

ties

Inte

rvie

wer

adm

inis

tere

dL

owL

owQ

1:n

/a;Q

2,Q

3,Q

5–Q

8:n

um

ber

ofd

ays

affe

cted

;Q4,

Q9

(0�

no

inte

rfer

ence

,10

�co

mp

lete

inte

rfer

ence

)

TR

T:n

otye

tes

tabl

ish

ed;

IC:n

/a

Mod

est

(RA

)M

odes

t(R

A)

Con

sid

ers

both

abse

nte

eism

and

pre

sen

teei

sm;

con

sid

ers

imp

acts

onbo

them

plo

ymen

tan

dh

ouse

hol

dw

ork

Eig

ht

sep

arat

eou

tcom

esc

ores

are

der

ived

(not

inte

nd

edto

besu

mm

ativ

e)

Wor

kpla

ceA

ctiv

ity

Lim

itat

ion

sS

cale

(WA

LS

)

Ph

ysic

alfu

nct

ion

ing,

acti

vity

lim

itat

ion

sat

wor

k

Sel

f-or

inte

rvie

wer

adm

inis

tere

d

Low

Min

imal

Ran

ge0–

33(1

1it

emW

AL

S)

or0–

36(1

2it

emW

AL

S),

hig

her

scor

em

ean

sm

ore

wor

kpla

ceac

tivi

tyli

mit

atio

ns

TR

T:n

otye

tes

tabl

ish

ed;

IC:g

ood

(OA

,IA

)

Ver

ygo

od(O

A,R

A)

Ver

ygo

od(O

A,R

A)

Bri

efto

olth

atex

amin

eske

yas

pec

tsof

wor

kpla

ceac

tivi

tyli

mit

atio

ns;

com

pre

hen

sive

resp

onse

key

wh

ich

incl

ud

es“n

otap

pli

cabl

e”an

d“d

iffi

cult

ies

un

rela

ted

toar

thri

tis”

opti

ons

Rec

all

tim

efr

ame

isn

otsp

ecifi

ed,b

ut

is“i

nge

ner

al”

or“t

ypic

ally

”

Wor

kIn

stab

ilit

yS

cale

for

Rh

eum

atoi

dA

rth

riti

s(R

A-

WIS

)

WI:

deg

ree

ofm

ism

atch

betw

een

fun

ctio

nal

abil

itie

san

djo

bd

eman

ds

Sel

f- adm

inis

tere

dL

owM

inim

alto

low

(in

terv

al-

leve

lsc

ore

tran

sfor

mat

ion

avai

labl

e)

Ran

ge0–

23,2

3�

hig

hes

tW

I;p

rop

osed

cut

poi

nts

:�10

�lo

wW

I,10

–17

�m

oder

ate

WI,

�17

�h

igh

WI

TR

T:

exce

llen

t(R

A);

IC:

exce

llen

t(R

A,O

A)

Ver

ygo

od(R

A,O

A)

Ver

ygo

od(R

A,O

A)

Isa

un

iqu

eco

nce

pt

amon

gav

aila

ble

wor

kou

tcom

em

easu

res

inth

efi

eld

;p

atie

nt

pre

fere

nce

for

this

scal

esh

own

inar

thri

tis/

mu

scu

losk

elet

alp

opu

lati

ons

Dic

hot

omou

sre

spon

seop

tion

sm

aybe

too

coar

se;s

ome

item

sap

pea

red

red

un

dan

t

Wor

kL

imit

atio

ns

Qu

esti

onn

aire

(WL

Q)

Mea

sure

sp

rop

orti

onof

tim

ew

ith

dif

ficu

lty

per

form

ing

vari

ous

asp

ects

ofw

ork

Sel

f- adm

inis

tere

dM

oder

ate

Mod

erat

eR

ange

0–10

0(4

subs

cale

s),1

00�

mos

tli

mit

atio

ns;

10%

incr

ease

insu

bsca

lesc

ore

equ

ival

ent

to4–

5%lo

ssin

wor

kp

rod

uct

ivit

y

IC:v

ery

good

(RA

,OA

)E

xcel

len

t(R

A,O

A)

Goo

d(R

A,O

A)

Wel

l-es

tabl

ish

edd

omai

nst

ruct

ure

;val

idat

edac

ross

dif

fere

nt

mu

scu

losk

elet

ald

isea

sean

dch

ron

ich

ealt

hp

opu

lati

ons;

WL

QIn

dex

can

beco

nve

rted

top

rovi

de

anes

tim

ate

of%

pro

du

ctiv

ity

loss

(aga

inst

hea

lth

yco

ntr

ols)

Mu

ltip

leva

rian

tsof

WL

Qan

dsc

orin

gap

pro

ach

esex

ist,

cau

tion

isre

quir

edw

hen

com

par

ing

fin

din

gs;

reve

rsed

inst

ruct

ion

sfo

rth

ep

hys

ical

dem

and

ssu

bsca

levs

oth

ersu

bsca

les

Wor

kP

rod

uct

ivit

yan

dA

ctiv

ity

Imp

airm

ent

Qu

esti

onn

aire

(WP

AI)

Ext

ent

ofab

sen

teei

sm,

pre

sen

teei

sm,

and

imp

airm

ents

ond

aily

acti

viti

es

Sel

f-or

inte

rvie

wer

adm

inis

tere

d

Min

imal

Low

4d

iffe

ren

tou

tcom

esc

ores

(eac

hw

ith

ran

ge0–

100,

ind

icat

ing

%im

pai

rmen

t)

TR

T:m

odes

t(R

A);

IC:

n/a

Mod

est

(RA

,AS

);ad

dit

ion

alev

iden

ceav

aila

ble

inot

her

pop

ula

tion

s(n

otap

pra

ised

her

e)

Goo

d(R

A,A

S)

Evi

den

ceof

reli

abil

ity

and

vali

dit

yh

asbe

ensh

own

acro

ssm

any

dis

ease

pop

ula

tion

s;lo

wre

spon

den

tbu

rden

;hig

hp

oten

tial

for

econ

omic

cost

ing

app

lica

tion

s

Fou

rou

tcom

esc

ores

are

der

ived

;exa

min

esgl

obal

imp

act

(dif

ficu

ltie

sw

ith

spec

ific

asp

ects

ofw

ork

not

asse

ssed

)

*R

A�

rheu

mat

oid

arth

riti

s;Q

�qu

esti

on;n

/a�

not

app

lica

ble;

TR

T�

test

–ret

est;

IC�

inte

rnal

con

sist

ency

;OA

�os

teoa

rth

riti

s;IA

�in

flam

mat

ory

arth

riti

s;W

I�

wor

kIn

stab

ilit

y;A

S�

anky

losi

ng

spon

dyl

itis

.†

Sp

ecifi

car

thri

tis/

rheu

mat

icp

opu

lati

onin

dic

ated

inp

aren

thes

esw

her

eev

iden

ceis

avai

labl

e.

Work Disability S349