Validity of the Childhood Health Assessment Questionnaire Is Independent of Age in Juvenile Idiopathic Arthritis

Transcription

1 Arthritis & Rheumatism (Arthritis Care & Research) Vol. 51, No. 4, August 15, 2004, pp DOI /art , American College of Rheumatology ORIGINAL ARTICLE Validity of the Childhood Health Assessment Questionnaire Is Independent of Age in Juvenile Idiopathic Arthritis JACQUES POUCHOT, 1 EMMANUEL ECOSSE, 2 JOËL COSTE, 2 AND FRANCIS GUILLEMIN, 3 FOR THE FRENCH QUALITY OF LIFE STUDY GROUP AND THE PAEDIATRIC RHEUMATOLOGY INTERNATIONAL TRIALS ORGANISATION Objective. To determine whether the Childhood Health Assessment Questionnaire (CHAQ) is valid for the comparison of different age subgroups and for longitudinal studies in juvenile idiopathic arthritis (JIA). Methods. A CHAQ was administered to 306 children with JIA. Rasch analyses were used to compare the difficulty of each of the 30 items of the questionnaire for children of 2 age groups (>10 years old and <10 years old). Results. Independent of the physical disability level assessed by the Rasch model, 8 of the 30 items (27%) of the CHAQ were rated significantly different in the 2 age groups. Despite this age-related variation in item difficulty, the impact on the CHAQ disability index using its original scoring system remained low (about 0.25 points on a scale of 0 3). Conclusion. The difficulty of 8 of 30 items of the CHAQ depends on the respondent s age. Nevertheless, the design of the CHAQ and its scoring system remove most of the expected physical development bias. KEY WORDS. Childhood Health Assessment Questionnaire; Validity; Psychometry; Rasch analysis; Juvenile idiopathic arthritis. Supported in part by a European grant for Juvenile Idiopathic Arthritis Research: BMH4-CT Jacques Pouchot, MD: Hôpital Louis Mourier, Colombes, France; 2 Emmanuel Ecosse, PhD, Joël Coste, MD, PhD: Hôpital Cochin, Paris Cedex, France; 3 Francis Guillemin, MD, PhD: Ecole de Santé Publique, Vandoeuvre Cedex, France. Address correspondence to Jacques Pouchot, MD, Service de Médecine Interne, 178 rue des Renouillers, Colombes, France. jacques.pouchot@lmr.ap-hop-paris.fr. Submitted for publication April 28, 2003; accepted in revised form October 19, INTRODUCTION Juvenile idiopathic arthritis (JIA) is a heterogeneous group of inflammatory rheumatic disorders of unknown etiology. Significant progress has come from the recent development of internationally agreed classification criteria (1). The various subtypes of JIA differ in their clinical manifestations and prognoses, but many patients have a chronic disease course (2 4). Although disease outcomes have considerably improved in recent years compared with earlier reports (5,6), patients with JIA may still develop joint destructions and deformities, chronic uveitis and impaired vision, and delayed growth (3,4). The need to assess the global outcome of chronic diseases in a comprehensive manner has gained wide acceptance over the last 2 decades. The assessment of functional ability was recently included in the core set of outcome variables recommended for use in JIA clinical trials (7). The Childhood Health Assessment Questionnaire (CHAQ) (8), which was adapted from the Stanford Health Assessment Questionnaire (9), is one of the most widely used questionnaires to assess physical function in JIA patients. The CHAQ and its culturally adapted versions have been shown to be valid and reliable in JIA (8,10 18). The CHAQ is usually completed by the parents, but a child version is also available for children 8 years old. The questionnaire comprises 30 items and measures 8 functional areas: dressing and grooming, arising, eating, walking, hygiene, reach, grip, and activities. The CHAQ takes 10 minutes to complete, and scoring is easily obtained by hand calculation in 2 minutes (8). In chronic pediatric diseases, a specific problem arises from the potential for physical development to act as a confounding factor that may render hazardous the comparison of groups of children of different ages. Indeed, in children, age is strongly related to physical development stage and should be strongly considered when looking at the ability to perform a given functional activity. Nevertheless, it has been claimed that the CHAQ overcomes this drawback due to its design and method of scoring, meaning that it can be used in longitudinal clinical studies or to 519

2 520 Pouchot et al compare the scores of different age groups (8). Indeed, when adapting the HAQ for use in children with JIA, Singh et al (8) first added new questions so that each of the 8 functional areas contained at least 1 question that is relevant to children of all ages (between 1 and 19 years old). Second, a new response modality not applicable was added for the activities that a child is unable to perform because of developmental immaturity. Finally, during the scoring process, within each functional area, the item with the highest disability score determines the score of that area. With this scoring system, each child s score is based on their performance for age-appropriate activities, and therefore, older children are scored on a more complex activity than their younger counterparts and their scores can be considered comparable. The objective of this study was to verify that the psychometric properties of the CHAQ overcome the physical development bias, thus allowing the use of this questionnaire for longitudinal clinical and epidemiologic studies, and for comparison of groups of children of different ages with JIA. PATIENTS AND METHODS Patients and settings. All consecutive children from January 1999 to May 2000 with systemic, polyarticular (5 or more joints affected), extended, or persistent JIA satisfying the Durban criteria (1) were included in the study. The onset type was determined 6 months after the onset of symptoms. All the children were recruited from the outpatient pediatric clinics of 16 pediatric referral centers by the French members of the Paediatric Rheumatology International Trials Organisation, and included both newly diagnosed and established patients. During this cross-sectional survey, a validated French version of the CHAQ (18) was completed by 1 parent for each included patient. Written informed consent was obtained from at least 1 parent of all participating children. Measures. The response modalities of the CHAQ are 4-point ordinal scales (0 no difficulty, 1 some difficulty, 2 much difficulty, 3 unable to do). An additional response modality (not applicable for age) can be used to identify the activities that a child is unable to do because he/she is too young. In addition to its 30 items, the CHAQ comprises few additional questions concerning the aids or devices that would be needed to perform the evaluated activities and that are used in the scoring system. Within each of the 8 areas of activity, the item obtaining the highest (most severe disability) value gives the score for the area of activity. When a device or the aid of a person is required to perform a given activity, the score is at least 2 for the related area of activity. For a given functional area, the response in at least 1 item is sufficient to obtain a score for this area. Indeed, some questions are not adapted for the youngest children (i.e., cut fingernails, cut his/her own meat, run errands and shop). The global disability index is obtained by calculating the mean of the 8 functional areas. The score can range from 0 (no disability) to 3 (maximum disability). The score distribution is not continuous and comprises 25 equivalent levels of points. We also collected data concerning age, sex, disease duration, joint scores (swollen joints, painful joints, joints with limited range of motion, and active joints), erythrocyte sedimentation rate (ESR), presence of uveitis, antinuclear antibodies, and rheumatoid factor. Overall disease activity was rated by the physician on a 10-cm visual analog scale. Statistical analysis. Because the scores of the CHAQ (8 functional areas and the disability index) were skewed, nonparametric analyses were performed. Quantitative data were compared using the Kruskall-Wallis test, and Spearman correlation coefficients were computed. All the data were analyzed with the use of SAS statistical software (SAS Institute, Cary, NC). A P value 0.05 was considered statistically significant. All tests were 2 sided. Rasch analyses for ordinal scales were performed. The Rasch analysis is based on item response theory (19) and has recently been used to study several functional assessment questionnaires in rheumatology (20 24). The Rasch analysis is used to verify the unidimensionality of the scale (i.e., all the items of the CHAQ measure a latent trait, namely the physical function of children with JIA). This is assessed by calculating a statistical measure of the fit of individual items with the Rasch model (chi-square test with a limit value for significance set at 0.01 as recommended because of multiple testing). Furthermore, the Rasch analysis provides an estimate of individual item difficulty. Indeed, items of a questionnaire usually differ from each other in the level of difficulty to perform (for example it is more difficult to cut meat or to open jars than to lift a cup or a glass to mouth). The difficulty of an individual item, also called item location or calibration in the Rasch analysis, is expressed on a linear scale of physical disability (the latent trait measured by the CHAQ) and defines its position on a continuum from less difficult to more difficult items to perform. The measure of difficulty is expressed on a logit scale (natural log of an odds ratio). In the current analyses, items with negative values are more difficult to perform than those with positive values. It is also possible to display the difficulty of an item by its item characteristic curve. This is the expected item value (in the original scoring system of the CHAQ graded from 0 [no physical disability] to 3 [maximum physical disability]) as a function of the location on the continuum of functional disability assessed by the Rasch model. According to the claims of the CHAQ (8), items should ideally function similarly across age groups. We investigated this by looking for a systematic bias due to the physical development stage of growing children. On the basis of an expected difference in developmental motor skills and to obtain roughly equal samples, the children were divided in 2 groups: 10 years or older (151 children) and 9 years or younger (156 children). First, the rate of not applicable answers was calculated for each group. The between-group difference in difficulty of each item of the scale was then calculated, and the effect of age was

3 Validity of the CHAQ 521 Table 1. Demographic and clinical characteristics of the 306 children with juvenile idiopathic arthritis* Juvenile idiopathic arthritis subtypes Systemic n 70 Polyarticular n 66 Extended n 77 Persistent n 93 Age, years Girls, no. (%) 37 (53) 54 (82) 68 (88) 79 (85) Disease duration, years ESR Physician VAS, Number of swollen joints Number of painful joints Number of joints with limited range of motion Number of active joints Persistent systemic features, no. (%) 34 (51) Antinuclear antibodies, no. (%) 2 (3) 25 (38) 61 (79) 72 (77) Rheumatoid factor, no. (%) 0 10 (15) 4 (5) 1 (1) Uveitis, no. (%) 0 4 (6) 23 (30) 24 (26) * Data presented as mean SD unless otherwise noted. ESR erythrocyte sedimentation rate; VAS visual analog scale. Global evaluation of disease activity on a 10-cm visual analog scale. tested by analysis of variance (significance level was set at 0.01 because of multiple testing). A graphic analysis was used to compare the item characteristic curves for the 2 age groups. For a given item, if the 2 curves were not reasonably close, this meant that the item was not interpreted in the same way in the 2 groups. Rasch analyses were performed using the RUMM statistical program (RUMM laboratory, Murdoch University, Perth, Western Australia). RESULTS The demographic and clinical characteristics of the children included in the study are presented in Table 1. Three hundred six patients with JIA were enrolled, including 70 (23%) with systemic onset, 66 (22%) with polyarticular onset, 77 (25%) with an extended subtype, and 93 (30%) with a persistent subtype. In 85% of cases, the CHAQ was completed by the mother (mean SD age years) and in the remaining 15%, by the father (mean age years). The results obtained for the 8 functional areas of the CHAQ and the disability index are presented in Table 2 for all JIA children and for the various JIA subtypes. Because many of the JIA children were very young, especially in the extended group, the not applicable answer was commonly selected, particularly in 2 functional areas (dressing and grooming [16%], and activities [14%]). Nevertheless, because the functional area score can be calculated when at least 1 item is available, missing scores were very low (0 3% for the 8 functional areas). The response patterns were skewed toward normal functional ability. Table 2. The 8 functional areas and the disability index of the Childhood Health Assessment Questionnaire (CHAQ) using original scoring Juvenile idiopathic arthritis (JIA) subtypes Systemic n 70 Polyarticular n 66 Extended n 77 Persistent n 93 P* All JIA children n 306 CHAQ functional areas Dressing and grooming Arising Eating Walking Hygiene Reach Grip Activities Disability index * Kruskall-Wallis test. Scores can range from 0 to 3, lower scores indicate better functional ability; values are expressed as means standard deviations; scores were obtained using the original scoring system. Disability index is the mean score of the 8 CHAQ functional areas.

4 522 Pouchot et al Table 3. Results of the Rasch analysis of the Childhood Health Assessment Questionnaire Functional areas and items of the CHAQ Location (difficulty)* 2 Fit of item P Difference difficulty Age effect P Number of not applicable 1 9 years > 10 years Dressing and grooming Dress, including tying shoelaces and doing buttons Shampoo his/her hair Remove socks Cut fingernails Arising Stand up from a low chair or floor Get in and out of bed or stand up in crib Eating Cut his/her own meat Lift a cup or glass to mouth Open a new cereal box Walking Walk outdoors on flat ground Climb up five steps Hygiene Wash and dry entire body Take a tub bath (get in and out of tub) Get on and off the toilet or potty chair Brush teeth Comb/brush hair Reach Reach and get down a heavy object Bend down to pick up clothing or a piece of paper from the floor Pull on a sweater over his/her head Turn neck to look back over shoulder Grip Write or scribble with pen or pencil Open car doors Open jars which have been previously opened Turn faucets on and off Push open a door when he/she has to turn a door knob Activities Run errands and shop Get in and out of car or toy car or school bus Ride bike or tricycle Do household chores (e.g., wash dishes, take out trash, vacuuming, yardwork, make bed, clean room) Run and play * Location of the items on the physical disability continuum estimated by the Rasch model (items with negative values are more difficult to perform than those with positive values). Differences in difficulty of items between the 2 groups of children ( 10 years minus 10 years). Items that do not fit adequately the item response theory model. Items that significantly varied in difficulty with age. Unidimensionality of the CHAQ. The unidimensionality of the questionnaire was assessed by measuring the fit of each item with the Rasch model (Table 3). Four items (13%) did not fit the model of item response theory: cut fingernails and ride bike or tricycle had the poorest fit, with When the with much difficulty and unable to do response modalities were combined, the fit of the items improved (not shown). Item difficulty and variation of item difficulty with age. The locations (difficulty) of individual CHAQ items on the continuum of physical disability estimated by the Rasch

5 Validity of the CHAQ 523 younger children. For the remaining 5 items, the rates of unable to do answers were high for children of 9 years or less, and the younger children always had higher items scores than the older children. Figure 1. Role of the age grouping on the difficulty of items in the Rasch analysis of the 3 items of the Childhood Health Assessment Questionnaire (CHAQ) with the highest effects. In each panel, the thick curve depicts the theoretical characteristic curve of the selected item for the entire group of children with juvenile idiopathic arthritis: the greater the disability (estimated by the Rasch model), the higher the expected score (in the original scoring system) for the considered item. The 2 thin curves in each panel were obtained for the 2 groups of children ( 10 years of age [ ] and 10 years of age [Œ]). For a similar location on the physical disability continuum estimated by the Rasch model, the older children [ ] are expected to obtain higher scores than the younger children, which could possibly translate into a higher disability index of the CHAQ using its original scoring system (see text). model are presented in Table 3. This table also shows the effect of age on the interpretation of individual items of the CHAQ. For 8 items, there was a statistically significant variation of difficulty between the 2 age groups. This difference was particularly high for 3 items (Figure 1). These 3 items were each in a different functional area ( stand up from a low chair or floor in the arising area, run and play in the activities area, and bend down to pick up clothing or a piece of paper from floor in the reach area). There were almost no unable to do answers for these 3 items and the average expected scores were higher (more severe disability) for the older children than for the Consequence on the disability index score of the CHAQ. We graphically evaluated the effect of the variation of item difficulty across age on the disability index of the CHAQ for the 3 items having the greatest effects (Figure 1). The figure shows that the older children always (all along the continuum of the latent trait, namely physical disability) had a higher score than the younger children, but that this difference never exceeded 1 point, even for the item with the greatest age effect ( stand up from a low chair or floor that belongs to the arising functional area). This item is located at 0.59 on the disability continuum, whereas the only other item in this functional area ( get in and out of bed or stand up in crib ) is much easier to perform ( 0.18; Table 3). Therefore, because the functional area score is determined by the most difficult item (i.e., stand up from a low chair or floor for this functional area), the maximum effect of this variation of item difficulty with age would be to increase the arising score by 1 point, which would in turn increase the disability index of the CHAQ by points (1/8). The same reasoning can be applied to the other items for which difficulty varied with age. The item run and play, which has a location at 1.09 and a very low rate of not applicable answers, may also bias the score. Again, with a graphically determined conservative estimate, the score for the activities area would be biased by a maximum of 1 point, which would also result in a point bias in the disability index of the CHAQ in favor of the older children. The third item bend down to pick up clothing or a piece of paper from floor should not influence the score of its functional area because it is located at 0.15 and the most difficult item in the reach functional area is located at Finally, the statistically significant variation in difficulty with age for the remaining 5 items should minimally bias the disability index in the opposite direction (younger children had higher expected scores than the older children for these 5 items). Therefore, a conservative estimate shows that the maximum bias induced by the variability of item difficulty with age is about 0.25 points over a range of 0 to 3 for the disability index of the CHAQ. Sensitivity of Rasch analysis to disability level in JIA. Using the Rasch model, the disability index and the scores of each of the 8 functional areas of the CHAQ were computed for the 4 JIA subgroups. As it was observed with the original scoring (Table 1), the scores obtained using the Rasch model clearly discriminated between the 4 groups (Table 4). Construct validity of the CHAQ in the 2 selected age groups. The construct validity of the CHAQ was examined in further detail by analyzing in the 2 age groups the correlations between the disability index of the CHAQ and the various joint counts, the overall disease activity rated by the physician on a 10-cm visual analog scale, and the

6 524 Pouchot et al Table 4. The 8 functional areas and the disability index of the Childhood Health Assessment Questionnaire (CHAQ) using the Rasch model scoring Juvenile idiopathic arthritis (JIA) subtypes Systemic n 70 Polyarticular n 66 Extended n 77 Persistent n 93 P* All JIA children n 306 CHAQ functional areas Dressing and grooming Arising Eating Walking Hygiene Reach Grip Activities Disability index * Kruskall-Wallis test. The lower the score the better the functional ability; values are expressed as means standard deviations; scores were obtained using the Rasch model. Disability index is the mean score of the 8 CHAQ functional areas. ESR (Table 5). As expected, all the correlation coefficients were significantly associated with the scores of the CHAQ. More importantly, the correlation coefficients were of a similar magnitude in the 2 age groups. DISCUSSION The results of this study show that the CHAQ is suitable for long-term longitudinal clinical and epidemiologic studies, and for comparing children of different ages with JIA. This supports the claims of Singh et al (8) that the CHAQ seems to be free of physical developmental bias. These results may be particularly important because age at onset differs according to the subtype of JIA, JIA frequently being diagnosed at a younger age than polyarticular JIA (4,25). Table 5. Correlations between the disability index of the Childhood Health Assessment Questionnaire (CHAQ) and various criteria of disease activity according to the age group of children with juvenile idiopathic arthritis Disability index of the CHAQ Children 1 9 years* (n 156) Children >10 years* (n 151) Number of swollen joints Number of painful joints Number of joints with limited range of motion Number of active joints Physician global assessment Erythrocyte sedimentation rate * Spearman correlation coefficients; P for all. Physician global evaluation of disease activity on a 10-cm visual analog scale. The Rasch analysis revealed that the difficulty of 8 of 30 items of the CHAQ differed with age, and that this variation was large for 3 items from 3 different functional areas (arising, reach, and activities). The score for a given item can be viewed as a function of normal ability, taking into account the physical development stage of the child, and the physical disability due to JIA. It could be assumed that the normal functional ability for an activity would be higher in older children. However, for the 3 items that showed the highest level of variation of difficulty with age, the older children (10 years or more) had higher scores than the younger children on average, suggesting that these items were more difficult to perform for older children (i.e., at the same physical disability level, they obtained more severe scores than their younger counterparts). This would increase the CHAQ score of older children in the related functional area using the original scoring system despite a similar disability level. This paradox may possibly be explained by carefully examining the wording of these 3 items, each of which seems to explore more than a single concept. For example, for the run and play item, it is possible that parents of older children may consider the run activity, whereas parents of younger children may concentrate on the play activity. If this is indeed correct, then one can argue that running may be a more difficult activity than playing. A similar explanation is tenable for the apparent paradoxical behavior of the 2 other items. A modification of the wording of these 3 items could help to limit this bias. The bias due to the variation of item difficulty across age seems to be very low and should not exceed 0.25 points, which leads to an overestimation of the disability index of the CHAQ (using the original scoring system) for children aged 10 years. In our study, the observed difference in the disability index score of the CHAQ between the subtype (younger children) and the other subtypes could be partly explained by this bias; however, the mean difference of 0.4 points (Table 2) exceeded the

7 Validity of the CHAQ 525 difference that could be due to physical development bias. In longitudinal long-term studies, this bias is conservative as it could theoretically mask a minimal improvement in disability. The responsiveness of the CHAQ has only been studied in a small number of studies, most of which were short term (13,17,26). In a small group of children with juvenile dermatomyositis, the disability index score of the CHAQ decreased (showing improvement) from 0.97 to 0.34 after 3 months of corticosteroid treatment (26). Two studies found that the effect size, a calibrated index of sensitivity to change, was small (0.28) or moderate (0.51) in children who improved, and moderate (0.54 and 0.60) in those who deteriorate (13,17). In an open study, JIA patients improved by 0.4 points after 6 months of treatment with methotrexate (0.8 to 0.4 with an effect size of 0.3) (27). The correlations between the CHAQ and several variables (various joint counts, global assessment by physician, and ESR) were similar in the 2 studied age groups (Table 5), which provide further evidence for the assertion that the CHAQ validity is independent of age in JIA. Some limitations of our study should be underlined. Notably, it is possible that the relatively low number of children studied did not allow perfect fitting of all items to the Rasch model. However, another study using the Rasch model to examine the unidimensionality of the CHAQ showed a misfit rate of 13% of the functional areas (28). The floor effect, which is important for most of the items of this questionnaire, also makes modeling difficult. The age cutoff of 10 years was chosen on the basis of an expected difference of physical development stage between the 2 groups, but was also based on the need to have similar group sizes. Therefore, the age effect should be studied more thoroughly to confirm the results of our study. The variation of item difficulty with age should also be studied with the child self-administered version of the CHAQ. It should be stressed that the problem of variation of item difficulty with age is probably not unique to CHAQ, but may be a common limitation of health questionnaires designed to assess disability in chronic rheumatic diseases in children (29 32). Obviously, such physical development bias is not encountered in the adult population. The CHAQ is available in many languages, which renders it very interesting for international studies (10 18). This questionnaire is easy to administer and very well accepted by the patients and their parents. The CHAQ is reliable and valid, and the results of our study add further psychometric evidence supporting the opinion that this instrument is valid for long-term clinical and epidemiologic studies, and for comparing groups of children of different ages. The CHAQ explores the same concept and has the same structure as the HAQ, so that these 2 instruments may be used successively over time and their data pooled, as was done in a recent study (33). ACKNOWLEDGMENT We are indebted to Mrs. A. Verdier for her help with the development of the French version of the CHAQ. REFERENCES 1. Petty RE, Southwood TR, Baum J, Bhettay E, Glass DN, Manners P, et al. Revision of the proposed classification criteria for juvenile idiopathic arthritis: Durban, J Rheumatol 1998; 25: Andersson Gäre B, Fasth A. The natural history of juvenile chronic arthritis: a population based cohort study. I. Onset and disease process. J Rheumatol 1995;22: Andersson Gäre B, Fasth A. The natural history of juvenile chronic arthritis: a population based cohort study. II. Outcome. J Rheumatol 1995;22: Guillaume S, Prieur AM, Coste J, Job-Deslandre C. Long-term outcome and prognosis in -onset juvenile idiopathic arthritis. Arthritis Rheum 2000;43: Ansell BM, Wood PHN. Prognosis in juvenile chronic polyarthritis. Clin Rheum Dis 1976;2: Dequeker J, Mardjuadi A. Prognostic factors in juvenile chronic arthritis. J Rheumatol 1982;9: Giannini EH, Ruperto N, Ravelli A, Lovell DJ, Felson DT, Martini A. Preliminary definition of improvement in juvenile arthritis. Arthritis Rheum 1997;40: Singh G, Athreya BH, Fries JF, Goldsmith DP. Measurement of health status in children with juvenile rheumatoid arthritis. Arthritis Rheum 1994;37: Fries JF, Spitz PW, Young DY. The dimensions of health outcomes: the Health Assessment Questionnaire, disability and pain scales. J Rheumatol 1982;9: Fantini F, Corvaglia G, Bergomi P, Gattinara M, Gerloni V, Lomater C, et al. Validation of the Italian version of the Stanford Childhood Health Assessment Questionnaire for measuring functional status in children with chronic arthritis. Clin Exp Rheumatol 1995;13: Arguedas O, Andersson-Gäre B, Fasth A, Porras O. Development of a Costa Rican version of the Childhood Health Assessment Questionnaire. J Rheumatol 1997;24: Goycochea-Robles MV, Garduno-Espinosa J, Vilchis-Guizar E, Ortiz-Alvarez O, Burgos-Vargas R. Validation of a Spanish version of the Childhood Health Assessment Questionnaire. J Rheumatol 1997;24: Garcia-Garcia J-J, Gonzalez-Pascual E, Pou-Fernandez J, Singh G, Jimenez R. Development of a Spanish (Castillian) version of the Chilhood Health Assessment Questionnaire: measurement of health status in children with juvenile chronic arthritis. Clin Exp Rheumatol 2000;18: Moroldo MB, de Cunto C, Hübscher O, Liberatore D, Palermo R, Russo R, et al. Cross-cultural adaptation and validation of an Argentine Spanish version of the Stanford Childhood Health Assessment Questionnaire. Arthritis Care Res 1998;11: van der Net J, Prakken ABJ, Helders PJM, Ten Berge M, van Herwaarden M, Sinnema G, et al. Correlates of disablement in polyarticular juvenile chronic arthritis: a cross-sectional study. Br J Rheumatol 1996;35: Len C, Goldenberg J, Bosi Ferraz M, Hilario MOE, Oliveira LM, Sacchetti S. Crosscultural reliability of the Childhood Health Assessment Questionnaire. J Rheumatol 1994;21: Flato B, Sorskaar D, Vinje O, Lien G, Aasland A, Moum T, et al. Measuring disability in early juvenile rheumatoid arthritis: evaluation of a Norwegian version of the Childhood Health Assessment Questionnaire. J Rheumatol 1998;25: Pouchot J, Larbre JP, Lemelle I, Sommelet D, Grouteau E, David L, et al. Validation of the French version of the Childhood Health Assessment Questionnaire (CHAQ) in juvenile idiopathic arthritis. Joint Bone Spine 2002;69: Nunnaly JC, Bernstein IH. Psychometric theory. 3rd ed. New York: McGraw-Hill; Wolfe F, Kong SX. Rasch analysis of the Western Ontario MacMaster questionnaire (WOMAC) in 2205 patients with osteoarthritis, rheumatoid arthritis, and fibromyalgia. Ann Rheum Dis 1999;58: Ryser L, Wright BD, Aeschlimann A, Mariacher-Gehler S. A new look at the Western Ontario and McMaster Universities

8 526 Pouchot et al Osteoarthritis Index using Rasch analysis. Arthritis Care Res 1999;12: Wolfe F, Hawley DJ, Goldenberg DL, Russell IJ, Buskila D, Neumann L. The assessment of functional impairment in fibromyalgia (FM): Rasch analysis of 5 functional scales and the development of the FM Health Assessment Questionnaire. J Rheumatol 2000;27: Wolfe F. Which HAQ is best: a comparison of the HAQ, MHAQ and RA-HAQ, a difficult 8 item HAQ (DHAQ), and a rescored 20 item HAQ (HAQ20): analyses in 2491 rheumatoid arthritis patients following leflunomide initiation. J Rheumatol 2001;28: White LJ, Velozo CA. The use of Rasch measurement to improve the Oswestry classification scheme. Arch Phys Med Rehabil 2002;83: Woo P, Wedderbum LR. Juvenile chronic arthritis. Lancet 1998;351: Feldman BM, Ayling-Campos A, Luy L, Stevens D, Silverman ED, Laxer RM. Measuring disability in juvenile dermatomyositis: validity of the Childhood Health Assessment Questionnaire. J Rheumatol 1995;22: Ruperto N, Ravelli A, Falcini F, Lepore L, Buoncompagni A, Gerloni V, et al. Responsiveness of outcome measures in juvenile chronic arthritis. Rheumatology (Oxford) 1999;38: Tennant A, Kearns S, Turner F, Wyatt S, Haigh R, Chamberlain MA. Measuring the function of children with juvenile arthritis. Rheumatology (Oxford) 2001;40: Lovell DJ, Howe S, Shear E, Hartner S, McGirr G, Schulte M, et al. Development of a disability measurement tool for juvenile rheumatoid arthritis: The Juvenile Arthritis Functional Assessment Scale. Arthritis Rheum 1989;32: Howe S, Levinson J, Shear E, Hartner S, McGirr G, Schulte M, et al. Development of a disability measurement tool for juvenile rheumatoid arthritis: The Juvenile Arthritis Functional Assessment Report for children and their parents. Arthritis Rheum 1991;34: Wright FV, Law M, Crombie V, Goldsmith CH, Dent P. Development of a self-report functional status index for juvenile rheumatoid arthritis. J Rheumatol 1994;21: Wright FV, Kimber JL, Law M, Goldsmith CH, Crombie V, Dent P. The Juvenile Arthritis Functional Status Index (JASI): a validation study. J Rheumatol 1996;23: Ruperto N, Levinson JE, Ravelli A, Shear ES, Link Tague B, Murray K, et al. Longterm health outcomes and quality of life in American and Italian inception cohorts of patients with juvenile rheumatoid arthritis. I. Outcome status. J Rheumatol 1997;24: APPENDIX A: STUDY GROUP MEMBERS Members of the French Quality of Life Group C. Brégeon, J. Coste, D. Dumont, L. Engerran, J. Goguel, F. Guillemin, J.P. Larbre, J.S. Le Quintrec, F. Mas, M. Marty, S. Perrot, J. Pouchot, J. Sany. French members of the Paediatric Rheumatology International Trials Organisation (PRINTO) C. Barbier, M. Bost, L. David, A. Duquesne, C. Job Deslandre, M. Fischbach, S. Gandon Laloum, L. Goumy, E. Grouteau, C. Guyot, MH Guyot, S. Jean, I. Kone Paut, E. Le Gall, I. Lemelle, L. de Lumley, F. Mazingue, P. Pillet, E. Plouvier, AM. Prieur, P. Quartier, D. Sommelet.