ORIGINAL ARTICLE TARA SYMONDS, 1 BERNADETTE HUGHES, 1 SHANMEI LIAO, 2 QIUQING ANG, 2 AND NICHOLAS BELLAMY 3 INTRODUCTION

Transcription

1 Arthritis Care & Research Vol. 67, No. 11, November 2015, pp DOI /acr VC 2015, American College of Rheumatology ORIGINAL ARTICLE Validation of the Chinese Western Ontario and McMaster Universities Osteoarthritis Index in Patients From Mainland China With Osteoarthritis of the Knee TARA SYMONDS, 1 BERNADETTE HUGHES, 1 SHANMEI LIAO, 2 QIUQING ANG, 2 AND NICHOLAS BELLAMY 3 Objective. To establish the reliability, validity, and sensitivity to change of the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) among Chinese subjects with osteoarthritis (OA) of the knee, living in mainland China. Methods. A multicenter, randomized, double-blind, placebo-controlled, parallel-group study was conducted for validation of the electronic personal digital assistant version of the WOMAC Numerical Rating Scale (NRS) 3.1 Index in China. A total of 287 subjects with OA of the knee were randomized to receive either meloxicam (15 mg) or placebo. Psychometric properties of the WOMAC were evaluated by estimating the reliability, validity, and sensitivity to change. Equivalence of the electronic version was also compared with the paper version. Results. Intraclass correlation coefficients for the WOMAC pain, stiffness, and physical function subscales were 0.81, 0.76, and 0.85, respectively, indicating good test retest reliability. Similarly, internal consistency was strong (Cronbach s alpha for the 3 WOMAC subscales was 0.84, 0.86, and 0.96, respectively). Pearson s correlation coefficients for WOMAC pain and Short Form 36 health survey (SF-36) bodily pain, as well as WOMAC physical function and SF-36 physical functioning domains were >0.4, indicating convergent validity, whereas the coefficients for all 3 WOMAC domains with SF-36 mental health and mental health component scores were <0.4, indicating divergent validity. There was strong discriminant validity between healthy volunteers and OA patients. The effect sizes of change from baseline to week 12 in WOMAC subscale scores were large, demonstrating sensitivity to change. Equivalence between paper and electronic versions was very high. Conclusion. The culturally and linguistically validated Chinese version of the WOMAC NRS 3.1 for mainland China is psychometrically robust in its validity, reliability, and sensitivity to change for patients with OA of the knee. INTRODUCTION Osteoarthritis (OA) was designated as one of the key conditions for special attention during the World Health Organization s Bone and Joint Decade ( ) (1 3). It is the most common joint disease for individuals middle-aged and older and is a major cause of pain and disability in many countries worldwide (4). OA is characterized by joint pain, stiffness, an increased potential for muscular atrophy, and bone deformity. The prevalence of OA increases with age, affecting nearly 75% of people ages.65 years. The incidence of OA also intensifies with age, specifically after age 40 years (5,6), and OA affects women more frequently than men (7). Most of the OA disability burden is attributable to the hips and knees. In fact, OA is the precipitating diagnosis for more than 90% of the increasing number of total hip or knee joint replacement operations being undertaken worldwide. Some of the Supported by Pfizer. 1 Tara Symonds, PhD (current address: Clinical Outcomes Solutions, Folkestone, UK), Bernadette Hughes, PhD: Pfizer, Walton Oaks, UK; 2 Shanmei Liao, PhD, QiuQing Ang, MD, PhD: Pfizer Research and Development, Shanghai, China; 3 Nicholas Bellamy, MD, DSc, FRACP: School of Medicine, University of Queensland, Herston, Queensland, Australia. Dr. Symonds owns stock or stock options from Pfizer. Dr. Hughes owns stock or stock options from Pfizer. Dr. Shanmei owns stock or stock options from Pfizer. Dr. Ang owns stock or stock options from Pfizer. Dr. Bellamy is the registered copyright holder and registered trademark holder for the WOMAC Index and receives licensing fees for the WOMAC Index, AUSCAN Index, and associated user guides. Address correspondence to Tara Symonds, PhD, Strategic Lead, Clinical Outcomes Solutions Ltd, Folkestone, Kent, CT194RH, UK. tara.symonds@clinoutsolutions.com. Submitted for publication July 8, 2014; accepted in revised form May 19,

2 1554 Symonds et al Significance & Innovations To date there are no psychometric data published in the peer-reviewed literature on the electronic Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) Numerical Rating Scale (NRS) 3.1 translation for the Peoples Republic of China. This study establishes the psychometric properties of the WOMAC Index among Chinese subjects with osteoarthritis (OA) of the knee. The WOMAC NRS 3.1 Simplified Chinese for mainland China Index demonstrated strong validity and reliability in patients with OA of the knee. Further, the measure was sensitive to change on an electronic personal digital assistant medium. other factors that are associated with the presence of OA include elevated body mass index (BMI), low socioeconomic status, ethnicity, lower levels of education, and negative behavioral influences, each associated with a higher degree of prevalence and impairment (5). Although OA is highly prevalent in the US and Europe, there has been an increase in OA in Asian countries over the past years (7). In China, around 50 million elderly people ages.60 years are estimated to have OA (8). The prevalence of OA in Asia is expected to rise further as the population ages, and researchers believe the percentage of people ages $65 years will more than double in the next 3 decades, from 6.8% in 2008 to 16.2% in 2040, thereby resulting in increased overall community health care burden, including medical and indirect costs (lost wages and productivity) (7). OA and other rheumatic conditions seldom cause death, but they have a substantial impact on health. For this reason, health-related quality of life (HRQOL) measures such as the Short Form 36 health survey (SF-36) (9) and the Arthritis Impact Measurement Scale (10) are more informative about their impact than related mortality rates (11). Limitations imposed by functional disability of OA can impact every aspect of HRQOL, including an individual s ability to be independent, along with social life, relationships, and emotional wellbeing (12). Thus, quantification of functional disability or HRQOL more broadly in OA patients plays a key role in describing, predicting, and measuring the severity of OA. Several instruments are currently available for measuring different aspects of HRQOL in OA patients, including function and the experience of pain. The Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC), a 24-item self-report questionnaire that addresses joint pain, stiffness, and physical function related to OA of the knee and hip, is one of the most widely used condition-specific instruments across various countries in the world (13 15). To enable comparisons between patient-reported outcome assessments made in different countries, the questionnaire needs to be validated and adapted for use in different cultures. The WOMAC index has been widely used in clinical trials and is extensively validated in and translated into over 100 languages (12). The WOMAC scale has demonstrated good psychometric properties of reliability (intraclass correlation coefficient [ICC] $0.7, Cronbach s a $0.8) and convergent validity when correlated with a variety of similar measures (14,16). With regards to the Asian population, the WOMAC index has been translated into language forms for Bangladesh (Bangla), Hong Kong (Cantonese), India (Bengali, English, Gujarati, Hindi, Kannada, Malayalam, Marathi, Punjabi, Tamil, Telugu, and Urdu), Japan (Japanese), Malaysia (Cantonese, English, Malay, Tamil), Philippines (Cebuano, English, Tagalog), Singapore (English, Mandarin), South Korea (Korean), Taiwan (Mandarin), and Thailand (Thai) (17). To date there are no psychometric data published in the peer-reviewed literature on the electronic WOMAC Numerical Rating Scale (NRS) 3.1 translation for the Peoples Republic of China. The primary objective of this study was to establish the reliability, validity, and sensitivity to change of the WOMAC NRS 3.1 Index among Chinese subjects with knee OA living in mainland China. PATIENTS AND METHODS Study design. A multicenter, randomized, double-blind, placebo-controlled, parallel-group methodology study was conducted for validation of the WOMAC NRS 3.1 Index in China with meloxicam as the active treatment arm. The WOMAC index has been used to assess subjects with OA of the hip or knee and consists of 24 items in subscales measuring 3 concepts: pain, stiffness, and physical function. The WOMAC pain subscale comprises 5 questions, the WOMAC physical function subscale has 17 questions, and the WOMAC stiffness subscale includes 2 questions. These questions measure the amount of pain, degree of difficulty in completing various day-to-day activities, and stiffness experienced in the hip or knee, respectively, in the previous 48 hours. Usually subject self-administered, the index is amenable to electronic data capture formats (18 20). A total of 287 mainland Chinese subjects with OA of the knee were randomized to 2 treatment groups in a 1:1 ratio of meloxicam (15 mg) or placebo. A total of 52 healthy volunteers were recruited to evaluate the discriminant validity of the WOMAC NRS 3.1 Simplified Chinese for mainland China version by comparing the values of all WOMAC subscales and total score between the OA subjects and healthy volunteers at screening (visit 1). All subjects were recruited from urban centers in China. The study had a screening period (beginning up to 21 days prior to randomization), which included an entry criteria evaluation (visit 1), test retest assessment (visit 2), and a 7-day initial pain assessment period (visit 3), followed by a 12-week treatment period. Subjects ages years, diagnosed with OA of the index knee, based on American College of Rheumatology criteria (21) with radiographic confirmation (Kellgren/ Lawrence grade $2) were included in the study. Furthermore, subjects were included only if they were willing

3 Chinese WOMAC Validation Study 1555 Table 1. Summary of patient-reported outcomes administered at different time points* Questionnaires Screening visit 1 Screening visit 2 Screening visit 3 Baseline Week 2 Week 4 Week 8 Week 12 Daily Average Pain x x x x x x x x WOMAC x x x x x x x x PGIC x PGAO x x x x x x x SF-36 x x x EQ-5D x x x * WOMAC 5 Western Ontario and McMaster Universities Osteoarthritis Index; PGIC 5 Patient Global Impression of Change; PGAO 5 Patient Global Assessment of Osteoarthritis; SF-36 5 Short-Form 36 health survey; EQ-5D 5 EuroQol 5-domain questionnaire. and able to discontinue all current analgesic therapy for OA relief and met each of the following criteria: 1) WOMAC pain subscale $4 in the index knee at screening with or without regularly taking analgesic medication; 2) for treatment-naive subjects not regularly taking pain medication during the month prior to screening (defined as,4 days/week), having an NRS pain score of $4 at screening and at baseline, based on at least 4 daily diary entries over a 7-day period; 3) for subjects currently taking pain medication, an increase of $1 point in the daily NRS pain score in the index knee following washout for subjects regularly taking pain medications during the month prior to screening; 4) WOMAC physical function subscale score $4 in the index knee at baseline; and 5) a Patient Global Assessment of OA (PGAO) response of fair, poor, or very poor at baseline (PGAO is a 1-item global measure that asks considering all the ways your osteoarthritis in your knee affects you, how are you doing today? ). To be enrolled as a healthy volunteer, subjects had to have no history of any chronic pain condition in the last 6 months and no history of other diseases that may involve the index knee, including inflammatory joint disease, crystalline disease (gout or pseudogout), endocrinopathies, metabolic joint diseases, lupus erythematosus, rheumatoid arthritis, joint infections, neuropathic disorders, avascular necrosis, Paget s disease, or tumors. The healthy volunteers were only required to have a single visit where they had to complete the WOMAC and PGAO using an electronic personal digital assistant (PDA); the screen size was similar to a standard smartphone. All participating subjects meeting the screening criteria were required to complete an e-diary (the same PDA device as the healthy volunteers), which captured the following data: 1) the daily 11-point pain NRS to assess average knee pain; 2) assessments of the WOMAC NRS 3.1 from screening through week 12 or early termination (subscale scores were calculated according to the WOMAC manual (22), with higher scores indicating more pain, more stiffness, or worse function for each of the subscales); and 3) the PGAO score, from a single question with 5 grades, where 1 5 very good and 5 5 very poor, collected at all clinic visits except visit 3. Other scales were also administered to patients on paper. The SF-36 acute version (SF-36v2) (9) is a self-administered questionnaire that measures each of the following 8 health concepts: physical functioning, role limitations due to physical problems, social functioning, bodily pain, mental health, role limitations due to emotional problems, vitality, and general health perception. The survey has been validated for use in China (23). Two component scores are also derived from the 8 subscale scores, physical component summary (PCS) and mental component summary (MCS) scores. Since response options differ and raw scores also differ across the domains, the 8 health concepts are calibrated on a scale of 0 100, with higher scores reflecting better subject status. Standard scoring was used for PCS and MCS. In addition to the SF-36v2, paper format questionnaires included the Patient s Global Impression of Change (PGIC). The PGIC questionnaire was only administered during visit 2 of screening and measured change in a subject s overall status on a scale where 1 5 very much improved and 7 5 very much worse. This questionnaire was used to determine subjects who stated limited change (no change or minimal improvement/worsening) for test retest reliability assessment. Table 1 outlines the time points at which the patient-reported outcome measures were administered in the study. Since the WOMAC validation was to be completed electronically, an assessment of equivalence with the paper version was also of interest. To allow this comparison, patients completed a paper copy of the WOMAC after they had completed all the other measures at visit 1 (i.e., electronic WOMAC, pain NRS, PGAO, SF-36, and EuroQol 5- domain questionnaire). Cross-cultural adaptation. Three Chinese linguists, bilingual in both Chinese and English, were tasked to create the WOMAC NRS 3.1 Simplified Chinese for mainland China Index translation from the English version (one of these translators resided in China). Tandem forward translations, harmonization, and independent back-translation resulted in a harmonized translation in Chinese, which was then reviewed by a survey research expert. Only a few issues were raised, in relation to physical function items. For item 16, the phrase panty hose and stockings was not relevant and was translated as problems with putting on socks that are either ankle-, calf- or knee-high. Item 20 concerns getting in and out of the bathtub, but use of bathtubs is not very common in China, so this item was amended to getting in and out of the bathtub/shower.

4 1556 Symonds et al Table 2. Demographic characteristics* FAS Characteristics Placebo (n 5 143) Meloxicam (n 5 144) VS (n 5 356) HS (n 5 52) Age, years (21.2) (0.7) 1 (0.7) 3 (0.8) 14 (26.9) (2.8) 12 (8.3) 20 (5.6) 7 (13.5) (30.1) 44 (30.6) 104 (29.2) 12 (23.1) (42.7) 60 (41.7) 150 (42.1) 8 (15.4) (23.8) 27 (18.8) 79 (22.2) 0 Mean 6 SD Range Sex Men 37 (25.9) 22 (15.3) 77 (21.6) 15 (28.8) Women 106 (74.1) 122 (84.7) 279 (78.4) 37 (71.2) Weight, kg Mean 6 SD Range BMI (kg/m 2 ) Mean 6 SD Range Height, cm Mean 6 SD Range * Values are number (percentage) unless indicated otherwise. FAS 5 full analysis set; VS 5 validation analysis set; HS 5 healthy subjects; BMI 5 body mass index. Defined as weight/(height ) 2 Cognitive debriefing was performed by bilingual interviewers in China, with 10 native speakers who were diverse in age, sex, and education level. Generally the subjects found the translation easy to understand and did not recommend any changes. The study was approved by the ethics committee of each study site, and all research was carried out in compliance with the Helsinki Declaration and related Chinese regulations. Statistical analysis. Three analysis sets were used in the validation analyses. 1) The validation analysis set (VS) comprised all subjects who were non-screen failure at visit 1 and returned for visit 2. This analysis set was used for all validation analyses except for the sensitivity-to-change analysis. 2) The full analysis set (FAS) comprised all randomized subjects who received at least 1 dose of study medication. This analysis set was used for the sensitivityto-change analysis. 3) The healthy subjects set contained all healthy subjects. This analysis set was used for the discriminant validity analysis. Data were analyzed using SAS statistical software, version 9.2 for UNIX. Descriptive statistics (mean 6 SD, median, minimum, maximum, percentage missing, and percentage at ceiling and floor) were calculated for all WOMAC subscales and total scores. Psychometric properties of the WOMAC were evaluated by estimating the reliability (test retest reliability and internal consistency), validity (convergent/divergent and discriminant), and sensitivity to change. Test retest reliability of the WOMAC index was determined using ICCs, carried out by a mixedeffects model with no fixed effect and with subject as the only random term for a subset of subjects from the VS population who answered no change, minimally improved, or minimally worse to the PGIC question. The timeframe between completions was 7 12 days. A score of $0.6 was considered an indication of good test retest reliability (24). Internal consistency reliability was measured by Cronbach s alpha in the VS population. Cronbach s alpha was calculated for each subscale score of the WOMAC at visit 1 and a score of $0.7 was regarded as acceptable internal reliability. Correlations between WOMAC domains and the SF-36 domains at visit 1 were calculated using the VS population. As a priori hypotheses, convergent validity of the WOMAC, i.e., the strength of association between items purported to measure similar concepts, was defined as r $ 0.4 (25) in comparisons between the WOMAC pain subscale and the SF-36 bodily pain subscale, and the WOMAC physical function subscale with the SF-36 physical functioning subscale and SF-36 PCS score. The PCS is a composite score of all 8 domains, with greater weight given to the physical functioning, role-physical, bodily pain, and general health domains. In contrast, divergent validity, i.e., weaker relative association between items purported to measure different concepts, was defined as r, 0.4 between all 3 WOMAC domains and SF-36 mental health and MCS scores. The MCS is a composite score of all 8 domains, with greater weight given to the vitality, social functioning, role-emotional, and mental health domains. Pearson s correlation coefficients were estimated

5 Chinese WOMAC Validation Study 1557 Table 3. Baseline characteristics of the FAS population (n 5 287)* Characteristics Placebo (n 5 143) Meloxicam (n 5 144) Body site assessed Right knee 67 (46.9) 73 (50.7) Left knee 76 (53.1) 71 (49.3) K/L grade (70.6) 103 (71.5) 3 35 (24.5) 32 (22.2) 4 7 (4.9) 9 (6.3) Baseline WOMAC pain No. subjects Mean 6 SD Median (range) 5.8 ( ) 5.8 ( ) Baseline WOMAC stiffness No. subjects Mean 6 SD Median (range) 5.5 ( ) 5.5 ( ) Baseline WOMAC physical function No. subjects Mean 6 SD Median (range) 5.8 ( ) 6.1 ( ) Average pain in the index knee at baseline No. subjects Mean 6 SD Median (range) 6.0 ( ) 6.0 ( ) PGAO at baseline Fair 77 (53.8) 82 (56.9) Poor 59 (41.3) 51 (35.4) Very poor 4 (2.8) 6 (4.2) Missing 3 (2.1) 5 (3.5) * Values are the number (percentage) unless indicated otherwise. FAS 5 full analysis set; K/L 5 Kellgren/ Lawrence; WOMAC 5 Western Ontario and McMaster Universities Index; PGAO 5 Patient Global Assessment of Osteoarthritis. Collected by personal digital assistant during initial pain assessment period. in all analyses. Discriminant validity was assessed by comparing subscale scores and total WOMAC score between OA subjects in the VS group and healthy volunteers at screening by 2-sample t-test. Sensitivity to change was assessed by comparing the difference between all WOMAC subscale scores at week 12 from baseline for the meloxicam arm, using mixed-effects models for repeated measures in the FAS population. Given the known efficacy of meloxicam, if the effect size of change from baseline in WOMAC subscales in the meloxicam arm was at least 0.2 or more (based on Cohen s definition of a small effect size) (26), the size of change would indicate that the translated WOMAC was responsive to active treatment. To explore the relationship between the electronic WOMAC and the paper WOMAC, Pearson s correlation coefficient was calculated along with ICCs for the subscales and total score from all patients in the VS population. No prespecified criterion was defined, since this was an exploratory exercise. RESULTS A total of 473 subjects were screened across 24 centers in China. Of those, 356 subjects met the screening criteria and were included in the VS. A total of 287 subjects received study treatment (placebo, n 5 143; meloxicam, n 5 144). Only 26 subjects (17 from the placebo arm and 9 from the meloxicam treatment arm) discontinued the study. In addition, 52 healthy subjects were recruited to evaluate the discriminant validity of the WOMAC scale. All subjects in the FAS population (n 5 287) were diagnosed with OA with a median duration since diagnosis of 1.9 and 1.1 years, for the placebo and meloxicam arms, respectively. Most patients (95.8%, n 5 275) used the electronic PDA for completing the WOMAC index at all study time points. The rest completed a paper version for at least 1 time point when the PDA was not available. Demographic characteristics of the FAS and the VS populations are summarized in Table 2. More than 70% of subjects recruited into the study were women. Age, weight, BMI, height, and sex distribution were all comparable among the FAS placebo group, FAS meloxicam group, and VS group. Baseline characteristics, including body site (right or left knee), WOMAC pain, stiffness, and physical function subscale scores at baseline, and PGAO at baseline for FAS, were also comparable between placebo and meloxicam groups and are summarized in Table 3. Assessment of psychometric properties. Reliability. ICCs were tested on 348 subjects who stated no changes or min-

6 1558 Symonds et al Table 4. Pearson s correlation coefficient between WOMAC (n 5 356) and SF-36 domains at visit 1 (validation analysis set)* SF-36 Pain WOMAC Physical function Stiffness General health perception Role-emotional Role-physical Social functioning Vitality Bodily pain Mental health Physical functioning Mental component summary Physical component summary * WOMAC 5 Western Ontario and McMaster Universities Osteoarthritis Index; SF-36 5 Short-Form 36 health survey. Scored in the opposite direction to the WOMAC. imal changes on the PGAO. All WOMAC subscale scores were larger than the prespecified cutoff point 0.6 (0.81, 0.76, and 0.85 for the WOMAC pain, stiffness, and physical function subscales, respectively), indicating good test retest reliability. Internal consistency was acceptable, with Cronbach s alpha for the 3 WOMAC subscales at 0.84, 0.86, and 0.96, respectively, which exceeded the prespecified criterion of.0.7. Validity. Absolute values of Pearson s correlation coefficients for both WOMAC pain and SF-36 bodily pain domains, as well as WOMAC physical function and SF-36 physical function domains, were.0.4, which indicated a moderate to strong relationship and thus demonstrated convergent validity. The absolute value of the correlation coefficient of the WOMAC physical function and SF-36 physical component was slightly lower than 0.4 at Absolute values of Pearson s correlation coefficients of all 3 WOMAC domains with SF-36 mental health and SF-36 mental health component scores were less than 0.4, which indicates divergent validity, as predicted. All correlations between WOMAC and SF-36 are shown in Table 4. In addition, the Pearson s correlation coefficients between WOMAC pain and physical function, pain and stiffness, and physical function and stiffness were 0.83, 0.68, and 0.75, respectively. The scores for WOMAC pain, stiffness, and physical function subscales, and total score at visit 1 for subjects in the VS population and for healthy volunteers were statistically significantly different (i.e., mean 6 SDs for WOMAC pain, physical function, and stiffness were , , and , respectively; P, for all), thereby indicating good discriminant validity of the WOMAC. Sensitivity and correlation of WOMAC scores. The effect size of change from baseline to week 12 in WOMAC subscale scores was larger than 0.2 (1.7, 1.4, and 1.7 for pain, stiffness, and physical function, respectively), which indicated that the WOMAC index was sensitive to change. Pearson s correlation coefficients and ICCs of WOMAC subscale scores and total scores were all above 0.9, which indicated that e-diary and paper scores were highly associated (Table 5). DISCUSSION The WOMAC NRS 3.1 Simplified Chinese for mainland China Index demonstrated strong psychometric properties in patients with OA of the knee. Similar to other culturally validated measures (27), test retest reliability assessed by ICC was good for the 3 subscales of WOMAC (all around 0.8), indicating a stable response in Chinese patients with OA of the knee. Given the 7 12 day interval between test and retest, and the dynamic nature of symptom severity in knee OA, which can fluctuate even within a 24-hour period (28), the ICC values are very good. Indeed, in some other validation studies, test retest reliability has been somewhat lower (28). Internal consistency coefficients (Cronbach s alpha) were also good for the 3 subscales (.0.80), indicating an acceptable level of reliability for group comparisons (29). These findings are similar to those found for the Thai (30) and Singapore (31) versions of the WOMAC. Similarly, convergent validity was generally demonstrated as predicted, except for the slightly lower correlation (0.39) between WOMAC physical function and SF-36 PCS score. The assumption was that since the PCS mainly captures physical function, it would be correlated with the WOMAC domain at the 0.4 level or above. However, given the fact that the general health and role-functioning domains are incorporated in the PCS, together with the physical function domain, this incorporation may have compromised that association. This contention is further supported by the observed correlation of 0.51 between the SF-36 physical function domain and the WOMAC physical function domain. Likewise, the lower-level correlations between WOMAC subscale scores and the MCS were predicted, because the WOMAC does not measure the main focus of the MCS (i.e., mental health). The general pattern of correlation observed between WOMAC and SF- 36 pain and physical function subscales was as expected. The strongest correlation of WOMAC pain was, as expected, with SF-36 bodily pain (20.47), and the strongest correlation of WOMAC function was, as expected, with SF-36 physical function (20.51). The 5 WOMAC Table 5. Relationship between electronic WOMAC (n 5 356) and paper version at visit 1 (validation analysis set)* WOMAC subscales Pearson correlation coefficient ICC Pain Stiffness Physical function Total * WOMAC 5 Western Ontario and McMaster Universities Osteoarthritis Index; ICC 5 intraclass correlation coefficient.

7 Chinese WOMAC Validation Study 1559 situation-specific pain questions and 17 WOMAC physical function questions are knee-specific and conditionspecific, respectively, and are considered over the previous 48 hours. By contrast, the single SF-36 global pain question (recall period the previous week) and 10 SF-36 physical function questions (recall period the previous week) do not refer to any specific joint, or in the case of the pain question, to any particular anatomic area, and in no instance refer to any specific condition such as OA. In this study, similar albeit slightly lower levels of correlation were observed between WOMAC pain and SF-36 function (20.45), and between WOMAC function and SF- 36 bodily pain (20.48). Since the 2 questionnaires are very different in concept, content, composition, scaling, and recall period, the apparent strength of association noted may be related to these differences in this particular group of patients. In particular, concomitant involvement by OA in other lower extremity joints, or nonarticular comorbidity (either painful or painless) might be expected to influence SF-36 scores, and consequently their strength of association (either greater or lesser), with WOMAC scores. Differences in the strength of association observed between pain and physical function (WOMAC 0.83 versus SF and 20.48, respectively) is compatible with this explanation. These issues notwithstanding, the important contribution of pain severity to the level of disability is acknowledged as 1 of several modulating factors. Other factors may include, but are not limited to, psychological, coping, helplessness, determination/motivation, personality, and vitality factors. Overall, the pattern and level of correlations observed support the convergent and divergent validity of the WOMAC NRS 3.1 Simplified Chinese for mainland China Index. Furthermore, the measure was able to discriminate between healthy individuals and those with OA. Finally, a reduction in pain severity was observed in the meloxicam group at week 12 from baseline, indicating that the translated WOMAC Index was sensitive to change and responsive to meloxicam treatment in OA patients. Similar findings were reported for the Spanish version of the WOMAC Index (effect size range 1 1.8) in a study looking at patients with hip or knee OA, specifically patients who had to undergo hip or knee replacement (32). Equivalence of the PDA electronic version to the paper version was demonstrated. A similar result has been shown using a randomized order of presentation design in a comparison of a mobile phone based version of the WOMAC NRS 3.1 Index and a paper version (33). However, a limitation of the current study was that completion on paper or electronic media first or second was not randomized. Due to the complexity of the study, we could not arrange that patients all completed the electronic version prior to completing the paper version, which occurred after completion of all the other patient-reported outcomes measures assessed in the study. In conclusion, the culturally and linguistically validated WOMAC NRS 3.1 Simplified Chinese for mainland China is deemed to be psychometrically robust in its validity, reliability, and sensitivity to change on an electronic PDA medium. ACKNOWLEDGMENT Editorial support was provided by Shalaka Marfatia of PharmEDGE and was funded by Pfizer. AUTHOR CONTRIBUTIONS All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be submitted for publication. Dr. Symonds 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 conception and design. Symonds, Hughes, Liao, Ang. Acquisition of data. Symonds, Hughes, Liao, Ang. Analysis and interpretation of data. Symonds, Hughes, Liao, Ang, Bellamy. ROLE OF THE STUDY SPONSOR Drs. Symonds, Hughes, Liao, and Ang were employees of Pfizer at the time of the study and were involved in the study design, the data analysis and interpretation, and the writing and editing of the manuscript. The content of this article was not contingent on approval by Pfizer. REFERENCES 1. World Health Organization. World Health Report 2002: reducing risks, promoting healthy life. Geneva: WHO; URL: 2. Brooks PM, Hart JA. The bone and joint decade: [letter]. Med J Aust 2000;172: Chen FP, Chang CM, Hwang SJ, Chen YC, Chen FJ. Chinese herbal prescriptions for osteoarthritis in Taiwan: analysis of National Health Insurance dataset. BMC Complement Altern Med 2014;14: Cook C, Pietrobon R, Hegedus E. Osteoarthritis and the impact on quality of life health indicators. Rheumatol Int 2007;27: Arthritis Foundation. Osteoarthritis fact sheet URL: 6. Lawrence RC, Felson DT, Helmick CG, Arnold LM, Choi H, Deyo RA, et al. Estimates of the prevalence of arthritis and other rheumatic conditions in the United States: part II. Arthritis Rheum 2008;58: Fransen M, Bridgett L, March L, Hoy D, Penserga E, Brooks P. The epidemiology of osteoarthritis in Asia. Int J Rheum Dis 2011;14: Li NH. Epidemiologic features of osteoarthritis in middleaged and elderly population. Chinese J Clin Rehabil 2005;9: In Chinese. 9. Ware JE, Kosinski M, Dewey JE. How to score version 2 of the SF-36 (r) health survey. Lincoln (RI): Quality Metric Incorporated; Meenan RF, Gertman PM, Mason JH. Measuring health status in arthritis: the Arthritis Impact Measurement Scales. Arthritis Rheum 1980;23: Doward LC, McKenna SP. Defining patient-reported outcomes. Value Health 2004;7:s Moskowitz RW. The burden of osteoarthritis: clinical and quality-of-life issues. Am J Manag Care 2009;15:S Bellamy N, Buchanan WW. A preliminary evaluation of the dimensionality and clinical importance of pain and disability in osteoarthritis of the hip and knee. Clin Rheumatol 1986;5: Bellamy N, Kean WF, Buchanan WW, Gerecz-Simon E, Campbell J. Double-blind randomized controlled trial of sodium meclofenate (Meclomen) and diclofenac sodium

8 1560 Symonds et al (Voltaren): post validation reapplication of the WOMAC Osteoarthritis Index. J Rheumatol 1992;19: Bellamy N, Buchanan WW, Goldsmith CH, Campbell J, Stitt LW. Validation study of WOMAC: a health status instrument for measuring clinically important patient relevant outcomes to anti-rheumatic drug therapy in patients with osteoarthritis of the hip or knee. J Rheumatol 1988;15: McConnell S, Kolopack P, Davis AM. The Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC): a review of its utility and measurement properties. Arthritis Rheum 2001;45: American College of Rheumatology. Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC). URL: rheumatology.org/practice/clinical/clinicianresearchers/outcomes_ Instrumentation/Western_Ontario_and_McMaster_Universities_Ost eoarthritis_index_%28womac%29/. 18. Theiler R, Spielberger J, Bischoff HA, Bellamy N, Huber J, Kroesen S. Clinical evaluation of the WOMAC 3.0 OA Index in numeric rating scale format using a computerized touch screen version. Osteoarthritis Cartilage 2002;10: Bischoff-Ferrari HA, Vondechend M, Bellamy N, Theiler R. Validation and patient acceptance of a computer touch screen version of the WOMAC 3.1 osteoarthritis index. Ann Rheum Dis 2005;64: Bellamy N, Wilson C, Hendrikz J, Patel B, Dennison S. Electronic Data Capture (EDC) using cellular technology: implications for clinical trials and practice, and preliminary experience with the m-womac Index in hip and knee OA patients. Inflammopharmacology 2009;17: Altman R, Asch E, Bloch D, Bole G, Borenstein D, Brandt K, et al. Development of criteria for the classification and reporting of osteoarthritis: classification of osteoarthritis of the knee. Arthritis Rheum 1986;29: Bellamy N. WOMAC 3.1 user guide IX. Brisbane (Australia); Li L, Wang HM, Shen Y. Chinese SF-36 health survey: translation, cultural adaptation, validation and normalization. J Epidemiol Community Health 2003;57: Fayers PM, Machin D. Quality of life: the assessment, analysis and interpretation of patient-reported outcomes. 2nd ed. Hoboken (NJ): Wiley; Juniper EF, Gordon HG, Roman J. How to develop and validate a new health-related quality of life instrument. In: Spilker B, editor. Quality of life and pharmacoeconomics in clinical trials. Philadelphia: Lippincott-Raven; pp Cohen J. Statistical power analysis for the behavioral sciences. 2nd ed. Hillsdale (NJ): Erlbaum; Konstantinidis GA, Aletras VH, Kanakari KA, Natsis K, Bellamy N, Niakas D. Comparative validation of the WOMAC osteoarthritis and Lequesne algofunctional indices in Greek patients with hip or knee osteoarthritis. Qual Life Res 2014; 23: Bellamy N, Sothern RB, Campbell J. Rhythmic variations in pain perception in osteoarthritis of the knee. J Rheumatol 1990; 17: Tuzun EH, Eker L, Aytar A, Daskapan A, Bayramoglu M. Acceptability, reliability, validity and responsiveness of the Turkish version of WOMAC osteoarthritis index. Osteoarthritis Cartilage 2005;13: Kuptniratsaikul V, Rattanachaiyanont M. Validation of a modified Thai version of the Western Ontario and McMaster (WOMAC) osteoarthritis index for knee osteoarthritis. Clin Rheumatol 2007;26: Xie F, Li SC, Goeree R, Tarride JE, O Reilly D, Lo NN, et al. Validation of Chinese Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) in patients scheduled for total knee replacement. Qual Life Res 2008;17: Escobar A, Quintana JM, Bilbao A, Azkarate J, Guenaga JI. Validation of the Spanish version of the WOMAC questionnaire for patients with hip or knee osteoarthritis: Western Ontario and McMaster Universities Osteoarthritis Index. Clin Rheumatol 2002;21: Bellamy N, Wilson C, Hendrikz J, Whitehouse SL, Patel B, Dennison S, et al. Osteoarthritis index delivered by mobile phone (m-womac) is valid, reliable, and responsive. J Clin Epidemiol 2011;64: