CORRELATES OF FUNCTIONAL DISABILITY IN EARLY RHEUMATOID ARTHRITIS: A CROSS-SECTIONAL STUDY OF 706 PATIENTS IN FOUR EUROPEAN COUNTRIES

Transcription

1 British Journal of Rheumatology 1996;35: CORRELATES OF FUNCTIONAL DISABILITY IN EARLY RHEUMATOID ARTHRITIS: A CROSS-SECTIONAL STUDY OF 706 PATIENTS IN FOUR EUROPEAN COUNTRIES L. M. SMEDSTAD,*t T. MOUM,* F. GUILLEMIN,}: T. K. KVTEN.t M. B. FINCH, T. P. B. M. SUURMEIJERU and W. J. A. VAN DEN HEUVEL^J *Department of Behavioural Sciences in Medicine, University of Oslo, ^Oslo City Department of Rheumatology, Diakonhjemmet Hospital, Oslo, Norway, \School of Public Health, University of Nancy, Nancy, France, ^Rheumatology Unit, Green Park Health Care Trust, Musgrave Park Hospital, Belfast and ^Northern Centre for Health Care Research, University of Groningen, The Netherlands SUMMARY In this cross-sectional study of 706 European patients with rheumatoid arthritis (RA) of < 4 yr duration, we examined possible correlates of functional disability assessed by the Health Assessment Questionnaire. First, we examined a subsample of 237 Norwegian patients. The Ritchie index, sex, age, erythrocyte sedimentation rate (ESR) and disease duration correlated significantly with disability, whereas serum rheumatoid factor, hand X-ray changes and educational level did not. Subsequently, we cross-validated these findings in a similar sample of 469 French, Dutch and Northern Irish patients. The results supported the Ritchie index, sex, ESR and disease duration as significant correlates of disability, whereas rheumatoid factor, age and education were not significantly correlated with disability. The correlation between X-ray changes and disability could not be cross-validated. The main findings of this study are that female sex correlates significantly with disability even early in the course of RA, whereas the rheumatoid factor does not. KEY WORDS: Rheumatoid arthritis, Disability, Prognosis, Age, Sex, Rheumatoid factor. PROGNOSTICATION of future outcome has proved a difficult concept in the rheumatological literature [1]. The early identification of factors associated with progressive disease in rheumatoid arthritis (RA) remains a major challenge to clinicians and researchers. RA is a potentially disabling disease and runs an unpredictable course. There is evidence that disability develops very early in the course of RA, and that the disease may be more amenable to treatment at an early stage [2]. Strategies for the treatment of RA in the 1990s advocate earlier recognition of progressive disease, accompanied by earlier treatment interventions [3]. Accordingly, the identification of factors indicative of a poor functional outcome early in the course of RA is crucial for tailoring treatment interventions and preventing irreversible joint damage and subsequent functional disability. In most prognostic studies on RA, two factors are consistently identified as predictors of a poor prognosis, i.e. seropositivity [4-11] and female sex [10, 12-14]. Less consistent associations with a poor outcome are found for the presence of erosions [9, 12, 15, 16], older age [12, 13], persistently elevated erythrocyte sedimentation rate (ESR) [6, 10], high joint activity scores [8, 9], long duration of the disease [13] and low levels of formal education [17]. The aim of this cross-sectional study was to identify variables associated with functional disability early in the course of RA in a cross-cultural European setting. Identical inclusion criteria were applied in four Submitted 20 June 1995; revised version accepted 4 March Correspondence to. L. M. Smedstad, Department of Behavioural Sciences in Medicine, PO Box 1111 Blindern, N-0317 Oslo, Norway. countries, i.e. France, The Netherlands, Northern Ireland and Norway, giving a total sample of 706 patients with RA of 0-4 yr duration (mean 2.0 yr). By examining one of the four study samples, we generated working hypotheses as to which of the study variables could explain a poor functional status. Subsequently, we cross-validated these results in the rest of the sample. The study aims were: (1) to generate hypotheses as to which variables significantly explain disability as opposed to variables that have no significant correlation to disability in a sample of 237 Norwegian patients with early RA; (2) to cross-validate these findings in a similar sample of 469 patients from France, The Netherlands and Northern Ireland. PATIENTS AND METHODS Patients Patients were sampled from well-defined geographical regions in four European countries. The inclusion criteria were set by the EURIDISS project (European Research on Incapacitating Diseases and Social Support) [18], of which this study is a part. The sampling procedure, as well as the samples, have been described in detail elsewhere [19-21]. The inclusion criteria were as follows: residence in the study area, age yr, diagnosis of RA according to the 1987 ARA criteria [22] and disease duration 0-4 yr. Exclusion criteria were the presence of other incapacitating diseases, stage IV according to Steinbrocker's functional class [23] or probable loss to follow-up (e.g. moving foreseen). The non-response rate varied from 12% in The Netherlands to 30% in France British Society for Rheumatology

2 SMEDSTAD ET AL:. CORRELATES OF DISABILITY IN RA 747 Measures Data on disease duration and current medication were collected from the medical files of the patients, whenever available, and/or from medical history taking. Joint tenderness, assessed by the Ritchie articular index [24], number of s.c. nodules and Steinbrocker's functional class [23] were recorded by a rheumatologist or a rheumatology research nurse. Blood samples were drawn for the analysis of ESR and rheumatoid factor (RF). The analyses of IgM-RF were performed at one centre for all samples, applying an ELISA technique and WHO international standard reference RF preparations [25]. In the Norwegian sample, radiographic changes were assessed by one experienced radiologist on anterioposterior hand radiographs according to the Larsen score [26]. The dependent variable, i.e. self-reported functional disability, was assessed by the Health Assessment Questionnaire (HAQ) [27], which has been validated in Dutch [28], French [29] and Swedish [30], which is very similar to the Norwegian language. This disability questionnaire examines eight dimensions of activities of daily living. The item with the highest score within each dimension is selected. A sum score representing the average value of the eight dimensions is computed. This sum score may have any value from zero (able to do without any difficulty) to three (unable to do). The Cronbach's a of the HAQ sum score, reflecting its internal consistency, was 0.89 in the pooled sample. The level of formal education was recorded and coded according to the International Standard Classification of Education [31], and subsequently recoded into levels of education ranging from 1 (< 7 yr of education) to 7 (> 16 yr of education). Data analyses National differences with respect to categorical and continuous variables were examined by x 1 tests and one-way analyses of variance (ANOVAs), respectively. To counteract the possibility of a type I error, the latter were corrected for the multiple comparisons performed by means of the modified least significant difference tests; a procedure available in the software program SPSS. To explore the multivariate effects of possible correlates of disability represented by the HAQ score, linear multiple regression analyses were applied. A forward stepwise selection of independent variables was chosen to identify variables that contributed significantly to the explained variance in HAQ score. This procedure wasfirstconducted in the Norwegian sample, and subsequently repeated in the rest of the sample in order to examine the cross-cultural validity of the variables selected in the first regression procedure. Possible interaction effects between pairs of independent variables were explored by entering multiplicative terms in the regression equation. For all analyses, two-sided /-tests and 5% levels of significance were chosen. RESULTS Patient characteristics of the four samples are presented in Table I. There were no significant differences between the four countries with respect to sex ratio or age at inclusion. Disease duration was significantly longer in Norway than in The Netherlands (2.3 and 1.8 yr, respectively). The level of education was significantly lower in France compared to the three other countries, and significantly higher in Norway than in The Netherlands. Table II presents clinical variables in the four national samples. The percentage of seropositive patients ranged from 62% in France to 81% in The Netherlands. However, the differences in seropositivity between the countries were not significant at the 5% level. The Ritchie index and ESR did not differ significantly between the four samples. The percentage of patients currently using diseasemodifying anti-rheumatic drugs (DMARDs) differed significantly between Norway (52% on DMARDs) and the other three countries (68% in France, 69% in The Netherlands and 78% in Northern Ireland, respectively). Hand X-ray changes were present in 55% of the Norwegian sample. The mean HAQ score ranged from 0.93 in Norway and France to 1.20 in Northern Ireland. These differences in HAQ scores were not, however, statistically significant when correcting for the multiple comparisons performed. Table III presents the results from the forward stepwise regression analyses in the Norwegian sample (/j = 237) with the HAQ score as the dependent variable. The Ritchie index, sex, ESR, age at inclusion and disease duration were all selected as significant % Females Age(yr) Disease duration (yr) Level of education 1= <7yr 7 = > 16 yr TABLE I Patient characteristics by country,* mean values and (S.D.) Norway n = (13.6) 2.3 (1.2) 3.6 (1.5) France n= (11.7) 1.9 (1.1) 2.1 (1.4) The Netherlands n = (11.8) 1.8(1.1) 3.0 (1.0) Northern Ireland n (11.8) 1.9(1.1) 3.0 (1.6) P, non-significant difference at the 5% level. Differences with respect to continuous variables were examined with correction for multiple comparisons with modified least significant difference tests.

3 748 BRITISH JOURNAL OF RHEUMATOLOGY VOL. 35 NO. 8 % Seropositive patientsf % of patients with nodules Ritchie index ESR % on DMARDs Steinbrocker functional capacity % I/II/III HAQ disability index (0-3) TABLE II Clinical variables by country,* mean values and (S.D.) Norway n = (6.0) 26 (20) 52 9/72/ (0.64) France n= (7.3) 29 (26) 68 16/70/ (0.68) The Netherlands n = (10.3) 28 (24) 69 9/74/ (0.80) Northern Ireland (10.2) 28(27) 78 22/73/ (0.78) 0.002, non-significant difference at the 5% level; ESR, erythrocyte sedimentation rate; DMARDs, disease-modifying anti-rheumatic drugs; HAQ disability index. Health Assessment Questionnaire disability index. Differences with respect to continuous variables were examined with correction for multiple comparisons with modified least significant difference tests. trheumatoid IgM factor measured by the WHO standard method and in the same laboratory for all four samples. correlates of disability, explaining a total of 44% of the adjusted variance in HAQ score. Sex was scored as 1 for males and 2 for females, implying that female sex was associated with higher disability. With a level of significance at P < 0.05, the remaining eligible independent variables were not entered into the equation, indicating that they would not reach this level if entered. These variables, i.e. level of education, X-ray changes and RF, and the corresponding P values at the last step of the forward procedure, are also listed in Table III. We repeated the above analyses omitting the X-ray changes from the model, as this variable was not available for the rest of the European sample. This resulted in the same variables being selected as significant correlates of disability, i.e. the Ritchie score, sex, ESR, age at inclusion and disease duration, whereas education and RF again failed to contribute significantly to the explained variance in disability. To cross-validate the above findings, forward stepwise regression analyses were performed in one pooled sample (n = 469) containing the French (n = 135), Dutch (n = 283) and Northern Irish (n = 51) patients (Table IV). Disability measured by the HAQ score was still the dependent variable, and the set of eligible independent variables was the same as in the analyses of the Norwegian sample with the exception of X-ray changes, which were not available. Repeatedly, the Ritchie index, ESR, disease duration and sex were selected as significant correlates of disability, explaining totally 40% of the variance in HAQ score, whereas age at inclusion was not entered into the model. Consistent with the results in the Norwegian sample, neither level of education nor RF was significantly correlated with disability. In both of the above regression models, the RF was coded as a dichotomous variable in terms of presence or absence. We then repeated the analyses including the RF as a continuous variable in the set of eligible independent variables. The results were similar to those presented in Tables III and IV, i.e. the RF was still not entered into any of the two models, whereas the same variables as above were selected as significant correlates of disability. Since the RF was consistently not selected as a significant correlate of disability in any of the two models of regression, we wanted to compare seropositive and seronegative patients in the total sample (n = 706). Disability was slightly, but not significantly, higher in seropositive than seronegative patients (HAQ score 1.03 and 0.94 respectively, P = 0.177). Sixty-five per cent of seropositive patients were currently using DMARDs compared to 54% of seronegative patients, whereas the corresponding numbers for oral steroid medication were 23 and 22%, respectively. TABLE III Forward stepwise regression model, Norwegian sample (n = 237). Dependent variable: HAQ disability index Variables entered into the equation Ritchie index Sex 1 males, 2 = females ESR Age at inclusion Disease duration Variables not entered into the equation Education X-ray changes Rheumatoid factor (dichotomous) Partial r* P HAQ disability index. Health Assessment Questionnaire disability index; ESR, erythrocyte sedimentation rate. 'Partial correlation coefficient. TABLE IV Forward stepwise regression model, French, Dutch and Northern Irish samples (n «469). Dependent variable: HAQ disability index Variables entered into Ritchie index ESR Disease duration Sex 1 - males, 2 = Variables not entered Education Age at inclusion Rheumatoid factor the equation females into the equation (dichotomous) Partial r* , P < HAQ disability index, Health Assessment Questionnaire disability index; ESR, erythrocyte sedimentation rate. *Partial correlation coefficient

4 SMEDSTAD ET AL.: CORRELATES OF DISABILITY IN RA 749 Using multiplicative terms in a regression equation including all variables selected in our two stepwise procedures, we explored, in the pooled sample of 706 patients, the possibility that the Ritchie index shows differential effects on the HAQ depending on the sex and age of the patient. No significant second-order interactions were identified by the multiplicative terms involving the Ritchie index and sex, and the Ritchie index and age, respectively. Although ANOVAs revealed no significant differences in HAQ scores between the four countries when correcting for the multiple comparisons performed (Table II), we also wanted to explore possible national differences in disability with a multivariate approach. Forward linear regression analyses with the HAQ score as the dependent variable were performed in the pooled sample of 706 patients. Independent variables were country (entered as a set of dummies) as well as the correlates of disability identified above. Controlling for the Ritchie index, ESR, sex, disease duration and age, the dummy variables indicating country of residence contributed as a set significantly to the variance in HAQ score (F= 4.85, P = 0.002). Using Norway as the reference sample, patients in The Netherlands and Northern Ireland had significantly higher mean HAQ scores, whereas the mean HAQ score in France was slightly, but not significantly, lower than in the Norwegian reference sample. DISCUSSION We consider the four national samples to be sufficiently similar to permit comparisons between one sample and the other three. Furthermore, we regard our samples as representative of early RA with respect to demographic and clinical variables (Tables I and II). The percentage of seropositive individuals is characteristic of samples identified in hospital settings [32]. The most striking finding of this study is the consistent lack of a correlation between the RF and functional disability. This goes contrary to a substantial body of follow-up studies of RA. In a prospective study by Sherrer et al. [12] with an average follow-up time of 11.9 yr, the presence of RF was one of the best predictors of disability. Woolf and colleagues [5] found that a positive RF at presentation of RA indicated a poor outcome in terms of disability 5 yr later. In a cohort study of 128 patients with RA, Suarez-Almazor and co-workers [7] found that the presence of RF was related to more severe outcomes in terms of joint counts and radiological scores assessed after 6-7 yr disease duration, van Zeben and colleagues [8] demonstrated a positive RF as one of the best predictors of outcome after 6 yr follow-up. In a prospective study of early RA over 15 yr, Corbett and co-workers [9] confirmed seropositivity as one of the early indicators of a poor outcome. Finally, in a review of the literature by van der Heijde et al. [10] in 1988, they concluded that most reports state the presence of RF as a strong predictor of an unfavourable course. In contrast to many other studies on RF in which comparisons are undermined by discrepancies in methodology, the measurement of the IgM-RF in the present study was performed at one centre and with WHO standard reference RF preparations, thus ensuring reliable RF measures. In this study of 706 patients with early RA, neither the presence nor the titre of RF was able to explain current disability. The probable explanation for this lack of correlation between disability and the RF is the short disease duration of the present sample (range 0-4 yr, mean 2.0 yr). This implies that the well-documented link between the RF and an accelerated deterioration in physical function may be impossible to identify early in the course of the disease. Another possible contribution to the lack of correlation between the RF and disability is that seropositive patients in the study were treated more aggressively with DMARDs, thus theoretically reducing their level of disability. The second variable which failed to explain disability in both regression models was the level of formal education. This is opposite to the findings of Callahan and Pincus [17], who demonstrated that education serves as a significant marker of clinical status in RA. Their results are in keeping with a study by Leigh et al. [13], who found that education was negatively associated with rapid deterioration in physical function. However, Corbett et al. [9] found no difference in functional deterioration between patients of different socioeconomic classes; a measure which overlaps with educational level to a considerable extent. Age as a risk factor of increased disability is a controversial issue in the rheumatological literature [10]. This is reflected in the present study where age at inclusion was significantly related to disability in the Norwegian sample, whereas it failed to be identified as a significant correlate of disability in the second regression model with French, Dutch and Northern Irish patients. In a study by Sherrer et al. [12], age was the most powerful single predictor of disability, whereas Suarez-Almazor and colleagues [7] did not find any relationship between age at onset and prognosis of RA. Some authors have argued that disability in older individuals probably includes the effects of co-morbid conditions and senescence itself [13], as opposed to a more severe disease in the elderly. Based on the first regression model with Norwegian patients, we hypothesized that hand X-ray changes would not correlate with disability. As this variable was not available for the rest of the sample, we were not, however, able to test this hypothesis. The lack of a relationship between erosive changes and disability is contrary to several studies of long-lasting disease [9, 12]. Although the mean disease duration of our sample may be too short to identify a relationship between erosions and disability, X-ray changes were actually present in 55% of the 237 patients. Furthermore, there is evidence that erosive changes occur within the first 2yr of the disease [15, 16], thus supporting the relevance of examining the correlation between X-ray changes and disability even early in the course of RA.

5 750 BRITISH JOURNAL OF RHEUMATOLOGY VOL. 35 NO. 8 The Ritchie score and the ESR were significantly correlated with disability in both of the regression models. Since disease activity and impairment drive the HAQ [33], these findings are not surprising. There is substantial evidence in the literature that measures reflecting the disease process are related to disease outcome in RA [6, 8-10]. Furthermore, disease duration was consistently correlated with functional disability even in this sample of patients with a maximum disease duration of 4 yr. Our finding is in agreement with a study on disability and disease duration by Leigh and collegues [13], in which a significant association between disease duration and HAQ score was found. In the present study, females consistently scored higher than males on functional disability. Theoretically, the current measure of disability, i.e. the HAQ, may be biased against women, thus augmenting the impression of severity of RA in women. However, most studies relate female sex to a poorer disease prognosis, even when outcome measures other than the HAQ are used [10], supporting the view that the sex discrepancy in disability observed in our study reflects a true difference. Of the two factors commonly accepted as predictors of a poor prognosis, i.e. seropositivity and female sex, only the latter was significantly correlated with disability in this cross-sectional study of early RA. These results indicate that while the disease duration of the present sample was too short to demonstrate increased disability among seropositive individuals, it was sufficiently long to demonstrate elevated levels of disability among females. Indirectly, this gives evidence for the relative strength of these two variables as early correlates of disability, indicating that the effect of sex on disability is more pronounced than the effect of the RF at an early stage of the disease. CONCLUSION Early identification of risk factors of disability in RA is a challenge to clinicians and researchers. In this cross-sectional study of early RA, the RF, traditionally considered as a predictor of a bad prognosis, was not significantly correlated with disability, whereas female sex, disease activity measures and duration of disease were. Early identification of high- versus low-risk patients with RA implies that aggressive and potentially toxic treatment can be reserved for those individuals who are most likely to fare badly, whereas low-risk patients may be candidates for less, rather than more, aggressive therapy. ACKNOWLEDGEMENTS The authors would like to thank Y. Jansson (Norway), S. Briancon, A. Gaucher and J. Pourel (France), A. L. Bell and P. Kane (Northern Ireland), and M. A. van Leeuwen, M. van Rijswijk, S. van der Burg, H. Lim and F. Speerstra (The Netherlands). The EURIDISS study is supported internationally by the COMAC-Health Services Research; in Norway by the Research Council of Norway, the Norwegian Rheumatism Association, the Legacy of Marie and Else Mustad, the Legacy of Grete Harbitz, Anders Jahre's Research Foundation and Gythfeldt's Research Foundation; in France by a contract IERM CRE and the French Society of Rheumatology; and in The Netherlands by 'Het Nationaal Reumafonds' and the Ministry of Welfare, Health and Cultural Affairs. REFERENCES 1. Kirwan JR. A theoretical framework for process, outcome and prognosis in rheumatoid arthritis. J Rheumatol 1992;19: Wolfe F, Hawley DJ, Cathey MA. Clinical and health status measures over time: prognosis and outcome in rheumatoid arthritis. / Rheumatol 1991; 18: Pincus T. The paradox of effective therapies but poor long-term outcomes in rheumatoid arthritis. Semin Arthritis Rheum 1992;21(suppl. 3): Spector TD, Hart DJ, Powell RJ. Prevalence of rheumatoid arthritis and rheumatoid factor in women: evidence for a secular decline. Ann Rheum Dis 1993; 52: Woolf AD, Hall ND, Goulding NJ et al. Predictors of the long-term outcome of early synovitis: a 5-year follow-up study. Br J Rheumatol 199I;30t Rasker JJ, Cosh JA. Course and prognosis of early rheumatoid arthritis. Scand J Rheumatol 1989;79(suppl.): Suarez-Almazor ME, Soskolne CL, Saunders D, Russell AS. Outcome in rheumatoid arthritis. A 1985 inception cohort study. / Rheumatol 1994;21: van Zeben D, Hazes JMW, Zwinderman AH, Vandenbroucke JP, Breedveld FC. Factors predicting outcome of rheumatoid arthritis: results of a followup study. / Rheumatol 1993;29: Corbett M, Dalton S, Young A, Silman A, Shipley M. Factors predicting death, survival and functional outcome in a prospective study of early rheumatoid arthritis over fifteen years. Br J Rheumatol 1993;32: van der Heijde DMFM, van Riel PLCM, van Rijswijk MH, van de Putte LBA. Influence of prognostic features on the final outcome in rheumatoid arthritis: a review of the literature. Semin Arthritis Rheum 1988; 17: Weisman MH. Prognosis in rheumatoid arthritis. Curr Opin Rheumatol 1990;2: Sherrer YS, Bloch DA, Mitchell DM, Young DY, Fries JF. The development of disability in rheumatoid arthritis. Arthritis Rheum 1986;29: Leigh JP, Fries JF, Parikh N. Severity of disability and duration of disease in rheumatoid arthritis. J Rheumatol 1992;19: Thompson PW, Pegley FS. A comparison of disability measured by the Stanford Health Assessment Questionnaire Disability Scales (HAQ) in male and female rheumatoid outpatients. Br J Rheumatol 1991;30: Fuchs HA, Kaye JJ, Callahan LF, Nance PE, Pincus T. Evidence of significant radiographic damage in rheumatoid arthritis within the first 2 years of disease. J Rheumatol 1989;16: Eberhardt KB, Rydgren LC, Pettersson H, Wollheim FA. Early rheumatoid arthritis onset, course, and outcome over 2 years. Rheumatol Int 1990; 10:

6 SMEDSTAD ET AL.: CORRELATES OF DISABILITY IN RA Callahan LF, Pincus T. Formal education level as a significant marker of clinical status in rheumatoid arthritis. Arthritis Rheum 1988;13: EURIDISS Group: European Research on Incapacitating Diseases and Social Support. Int J Health Sci 1990;l: Smedstad LM, Vaglum P, Kvien TK, Mourn T. The relationship between self-reported pain and sociodemographic variables, anxiety, and depressive symptoms in rheumatoid arthritis. J Rheumatol 1995;22: Guillemin F, Suurmeijer T, Krol B et al. Functional disability in early rheumatoid arthritis: description and risk factors. / Rheumatol 1994;21: Suurmeijer TPBM, Doeglas DM, Briancon S et al. The measurement of social support in the 'European Research on Incapacitating Diseases and Social Support': the development of the Social Support Questionnaire for Transactions (SSQT). Soc Sci Med 1995; 40: Arnett FC, Edworthy SM, Bloch DA et al. The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum 1988;31: Steinbrocker O, Traeger CH, Battcrman RC. Therapeutic criteria in rheumatoid arthritis. J Am Med Assoc 1949;14(k Ritchie DM, Boyle JA, Mclnnes JM et al. Clinical studies with an articular index for the assessment of joint tenderness in patients with rheumatoid arthritis. Q J Med 1968;37: van Leeuwen MA, Westra J, van Riel PLCM, Limburg PC, van Rijswijk MH. IgM, IgA and IgG rheumatoid factors in early rheumatoid arthritis. Scand J Rheumatol 1995;24: Larsen A, Dale K, Eek M. Radiographic evaluation of rheumatoid arthritis and related conditions by standard reference films. Ada Radiol Diagn 1977; 18: Fries JF, Spitz P, Kraines RG, Holman HR. Measurement of patient outcome in arthritis. Arthritis Rheum 198O;23: van der Heijde DMFM, van Riel PLCM, van de Putte LBA. Sensitivity of a Dutch Health Assessment Questionnaire in a trial comparing hydroxychloroquine vs sulphasalazopyrine. Scand J Rheumatol 1990;19: Guillemin F, Briancon S, Pourel J. Validity and discriminant ability of the HAQ Functional Index in early rheumatoid arthritis. Disabil Rehabil 1992;14: Ekdahl C, Eberhardt K, Andersson SI, Svensson B. Assessing disability in patients with rheumatoid arthritis. Scand J Rheumatol 1988;17: International standard classification of education (ISCED). Paris: Unesco, Division of Statistics on Education, Pincus T, Callahan LF. Reassessment of twelve traditional paradigms concerning the diagnosis, prevalence, morbidity and mortality of rheumatoid arthritis. Scand J Rheumatol 1989;79(suppl.): Wolfe F, Kleinheksel SM, Cathey MA, Hawley DJ, Spitz P, Fries JF. The clinical value of the Stanford Health Assessment Questionnaire Functional Disability Index in patients with rheumatoid arthritis. J Rheumatol 1988;15: