Outcome in adults with juvenile idiopathic arthritis: A quality of life study




To evaluate quality of life (QOL) in adults with juvenile idiopathic arthritis (JIA), using validated measures of functional disability and generic health status, and to quantify their educational attainment and employment status.


The adult rheumatology departmental database was used to identify patients. Functional disability and generic health status/QOL were assessed by the Health Assessment Questionnaire (HAQ) and the Short Form 36-item health profile (SF-36), respectively. Educational achievement and employment status were assessed by questionnaire.


Complete data were available for 82 of the 101 patients identified. The median age of patients was 30 years, and the median disease duration was 21 years. No deaths were recorded. All subtypes of JIA were represented. Thirty-nine percent of patients had active disease (based on the physician global assessment scale score). The median HAQ score was 1.125 (range 0–3). SF-36 scores for bodily pain, general health, physical functioning, vitality, emotion, and social isolation were significantly worse in patients compared with controls, and this trend increased with increasing age of the patients and disease duration. The SF-36 mental summation scores of patients were low compared with those of controls, for all subtypes of JIA, and this finding was independent of the degree of functional disability (by HAQ and SF-36 physical summation scores). The educational attainment of patients was comparable to that of local controls, but unemployment rates for patients were 3-fold higher than those for controls.


This is the largest study in which the SF-36 was used to assess generic health status and QOL in adults with JIA. Many patients had active disease in adulthood, and although the physical outcome of adults with JIA is relatively good, a profound effect on generic health status and QOL was demonstrated for all types of JIA. Furthermore, despite excellent educational attainment, there was a high rate of unemployment among patients.

Juvenile arthritis is a heterogeneous group of diseases of unknown etiology, many of which are clinically and genetically distinct from chronic arthritis in adults. Juvenile idiopathic arthritis (JIA) is the current term that has been proposed to unify the previous classifications of juvenile chronic arthritis (JCA) and juvenile rheumatoid arthritis (JRA) (1).

The outcome of juvenile arthritis is generally regarded as “good,” with many children experiencing a spontaneous remission. However, clinical experience suggests that JIA is not a benign disease, and a review of the literature indicates that the outcome in adults is variable. At least one-third of children will have ongoing active disease into their adult years (2–23), and many will have some limitation in their activities of daily living (15, 16, 19, 20). However, studies of the outcome of juvenile arthritis, as summarized in Table 1, often are not comparable because of the different definitions of disease (i.e., JRA or JCA) and markedly different treatment regimens (reflecting changes in the approach to management over several decades). In addition, such studies often include patients who were recruited from tertiary centers and therefore are biased toward severe disease. Furthermore, different measures of outcome have been used, and the duration of followup is variable. Nonetheless, these outcome studies do show that approximately one-third of patients have marked functional disability according to the Steinbrocker criteria (24), and many patients (10–45%) have ongoing active disease at 10 years of followup.

Table 1. Outcome studies in juvenile arthritis
Author, year (ref)No. of patientsFollowup, yearsSteinbrocker class III/IV, % of patientsNo. of deaths (%)
Lindbjerg, 1964 (3)7510.5235 (7)
Laaksonen, 1966 (4)505>164824 (5)
Jeremy, 1968 (5)4618 (range 5–20)24No data given
Schaller, 1972 (6)1247.512No data given
Goel, 1974 (7)1001–2049 (9)
Hill, 1976 (8)5814.5333 (5)
Ansell, 1976 (9)59>152217 (29)
Hanson, 1977 (10)1235–2528No data given
Stoeber, 1981 (11)43315 (range 10–22)4129 (7)
Levinson, 1991 (14)101>1517No data given
Gare, 1995 (15)143750 (0)
Ruperto, 1997 (19)77151 (1.3)
Peterson, 1997 (20)44>20No data given
Flato, 1998 (16)54108No data given
Zak, 2000 (23)6512.4 (mean)112 (3.1)
Oen, 2002 (22)3926.7 (range 0.2–20.8)2.5No data given

It is important to note that although many patients do not have clinical evidence of active joint disease during their adult years, there may well be other sequelae of previous inflammation or treatment that can have a detrimental effect on outcome (e.g., joint failure requiring joint replacement surgery, visual loss resulting from uveitis, growth retardation, osteoporosis) (25), and the psychological impact of juvenile arthritis can be marked (18). Recent outcome studies in chronic disease have focused on the global outcome from the patient's perspective. Few published studies to date have addressed the impact of JIA on quality of life (QOL), and these are summarized in Table 2. Although such studies demonstrate that the QOL for patients with JIA often is “good,” they are difficult to judge objectively, because the number of patients frequently is small, and they differ in terms of methodology (with differences in the measures of QOL, definitions of juvenile arthritis, source of patients [reflecting referral bias], duration of followup, and source of controls [e.g., siblings or the general population]).

Table 2. QOL studies in adults with juvenile arthritis*
Author, year (ref)No. of patientsFollowup, yearsQOL measureSummary
  • *

    QOL = quality of life; AIMS = Arthritis Impact Measurement Scale; GHQ = General Health Questionnaire; BDI = Beck Depression Inventory; MAA = Mental Adjustment to Arthritis score; QOLS = QOL scales; GAS = Global Assessment Scale; HAQ = Health Assessment Questionnaire; SF-36 = Short Form 36.

Hill, 1976 (8)5815Interview“Good” quality of life
Doherty, 1988 (17)2513AIMSWorse compared with controls
David, 1994 (18)4320GHQ, BDI, MAAMore anxiety and depression compared with controls
Ruperto, 1997 (19)11815QOLS77% “delighted or pleased” with outcome
Flato, 1998 (16)7210GASNo difference in QOL in those with active disease or those in remission
Peterson, 1997 (20)4325HAQ, SF-36SF-36 scores worse compared with controls

This study of long-term outcome in adults with JIA is the largest to date in which validated tools were used to assess functional disability and generic health status and QOL. We also include educational attainment and employment status as measures of outcome.


There has been a regional pediatric rheumatology service in Newcastle since 1972 (26), with the children's and adolescents' clinics at the Royal Victoria Infirmary. Since 1996, continuity of care into the adult years has been provided through the adult JIA clinic at Freeman Hospital, within the adult rheumatology department. In this study, the adult rheumatology department database was used to identify patients, by means of the umbrella term “juvenile arthritis.” Details that were collected included demographic information, disease subtype at onset (reclassified using JIA criteria) (1), disease duration, interventions (including joint surgery), and current medications.

As part of their routine clinical care, patients were assessed by one of us (HEF), using a 10-cm visual analog scale (VAS), the physician global assessment scale of disease activity (PGAS; in which 0 represents no disease activity and 10 represents severe disease activity), and joint counts (i.e., the number of active and/or restricted joints). Patients completed a self-rated 10-cm VAS for pain, based on their pain experience in the week before the clinical assessment. The Health Assessment Questionnaire (HAQ) (27) was used as an arthritis-specific measure of functional disability, and the Medical Outcomes Study 36-item Short Form (SF-36) (28–31) was used as a measure of generic health status and QOL; both of these tools are validated and are used extensively in rheumatology.

Additional questionnaires regarding educational experience and employment status were designed with the help of the hospital audit department and were completed by patients at the time of their clinical assessment or via a postal survey (with one reminder sent, when appropriate). The questionnaires included items regarding current employment status and highest level of education attained.

Data were collated on a Microsoft Access database, and results of the SF-36 were analyzed using commercially available scores for age- and sex-matched controls (29). The SF-36 physical summation score (PSS) and mental summation score (MSS) were developed as a way to meaningfully reduce the number of comparisons in clinical studies from 8 scales to 2 summary measures (30, 31). Data were analyzed using the statistical software package Arcus BioStat, version 1.1, in conjunction with Microsoft Excel. To compare the median scores between the study groups, we used the Mann-Whitney U test for nonparametric data; for comparisons between different disease subtypes, we used Kruskal-Wallis one-way analysis of variance.



From a total of 101 adult patients identified, complete data were available for 82 (68 of whom were female) (Table 3). The median age was 30 years (range 17–68) years, with a median disease duration of 21 years (range 3–61 years). No deaths were recorded in this group. Patients' disease subtypes were reclassified according to JIA criteria, using available data that was obtained at the time of presentation. JIA subtypes were represented as follows: oligoarticular onset (21 patients, 13 with persistent disease and 8 with extended disease), polyarticular onset (32 patients, 12 of whom were rheumatoid factor [RF] positive), systemic onset (12 patients), psoriatic (7 patients), and enthesitis-related (10 patients). Uveitis resulting in visual impairment occurred in 3 patients (an extended oligoarticular pattern in 2 patients and a polyarticular pattern in 1 patient, who was RF negative), none of whom are registered as being blind or partially sighted. All patients were able to complete the questionnaires independently. Twenty-six patients had undergone joint replacement surgery, which included hips (n = 41), knees (n = 26), elbows (n = 3), ankles (n = 2), shoulder (n = 1), and multiple finger joints (n = 1). The 19 patients who were identified from the database but for whom complete data were not available were similar to the study patients in terms of sex distribution and age (median age 30 years, range 19–46 years), and there was no significant difference in the distribution of disease subtypes.

Table 3. Patient demographics*
  • *

    HAQ = Health Assessment Questionnaire; VAS = visual analog scale.

  • Active joint disease is defined as a physician global assessment scale score >0.

Response rate, no.82
Age, median (range) years30 (17–68)
Sex, no. women/men68/14
Age at diagnosis, median ± SD years8.72 ± 4.47
Disease duration, median (range) years21 (3–61)
No. (%) patients with active joint disease32 (39)
HAQ score, median (range)1.125 (0–3)
Self-reported pain score on VAS, mean (range)1.54 (0–7)

Outcome measures.

Disease activity.

The PGAS showed that 32 (39%) of 82 patients had active joint disease (defined as a PGAS score >0). The median PGAS score was 0.85 (range 0–6.9). The median active joint count was 13 (range 1–64), and the median restricted joint count was 10 (range 0–68). There was a correlation between the PGAS and the number of active joints (parametric Pearson's correlation coefficient 0.4). Of the 32 patients with polyarticular-onset disease, 23 had active disease (according to the PGAS score), and 10 were RF positive. Only 2 patients with RF-positive polyarticular disease had inactive disease according to the PGAS score or the active joint count.

Disease-modifying antirheumatic drugs (DMARDs) were currently being used by 23 of the 82 patients (methotrexate [n = 12], sulfasalazine [n = 8], hydroxychloroquine in combination with methotrexate [n = 1], oral gold [n = 1], cyclosporine [n = 1], and oral corticosteroids were currently being used by 7 patients). Among the patients using DMARDs, 17 still had some evidence of active disease according to the PGAS score (median score 1.5; range 1–7). Nonsteroidal antiinflammatory drugs were being used on a regular basis by 49 patients, and analgesics were being used by 29 patients (4 of whom were taking analgesics only). A total of 26 patients were taking no regular medication. Many patients reported no pain, and the median self-reported pain score on a 10-cm VAS was 1.54 (range 0–7).

Joints that were involved (whether active or restricted) at the time of the study included cervical spine (68% of patients), knees (64%), hips (58%), wrists (62%), small joints of the hands (metacarpophalangeal 55%, proximal interphalangeal 55%, distal interphalangeal 43%), elbows (51%), temporomandibular joints (TMJs) (48%), ankles (28%), and feet (22%). Symmetric small and large joint involvement was evident in 26 of 32 patients with polyarticular-onset JIA. A pattern of asymmetric large and small joint involvement was evident in 12 patients, who had the following disease patterns at presentation: psoriatic (n = 3), enthesitis-related (n = 4), extended oligoarticular onset (n = 3), and polyarticular onset (n = 2). An additional 3 patients (1 male) had developed classic ankylosing spondylitis; all 3 of these patients had oligoarticular-onset JIA and were HLA–B27 positive.

Functional disability and QOL.

The median HAQ score for the study group was 1.125 (range 0–3; the higher the score, the greater the degree of functional disability). There was a significant difference in the HAQ scores according to JIA subtype (P = 0.004 by Kruskal-Wallis one-way analysis of variance). The lowest scores were observed in patients with oligoarticular-onset disease (median score 0.125, range 0–1.375), and the higher scores were observed in patients with systemic-onset (median score 1.8, range 0.25–3) and polyarticular-onset disease (median score 1.5, range 0–3) subtypes. For the 32 patients with active disease according to the PGAS score, the median HAQ score (1.5, range 0–3) was greater than that for patients with inactive disease (0.75, range 0–3) (P = 0.03 by Mann-Whitney U test).

The SF-36 data of patients compared with those of healthy age- and sex-matched controls are shown in Table 4. The data show a major impact of JIA on generic health status, with patients having statistically significantly lower (i.e., worse) scores for all physical domains (physical functioning, vitality, bodily pain, general health) and for social functioning and emotional role in the mental domains, compared with age- and sex-matched controls. Fifteen patients had an SF-36 score of 0 (i.e., worst possible functioning) in any domain; however, no patient scored 0 in all domains, and only 1 patient scored 0 in more than 2 domains. Table 5 shows the PSS, MSS, and HAQ scores of the patients, stratified by age.

Table 4. QOL as measured by the SF-36 in adults with juvenile arthritis and age- and sex-matched controls*
  • *

    Values are the mean score (range) for each domain. The higher the score, the better the quality of life for that particular domain. P values were determined by the Mann-Whitney U test. QOL = quality of life; SF-36 = Short Form 36; NS = not significant.

 Physical functioning52.81 (0–100)92.37 (74–95)<0.0001
 Vitality49.94 (0–95)64.15 (58–68)<0.0001
 Bodily pain50.86 (12–100)83.72 (73–87)<0.0001
 General health50.19 (5–97)75.08 (65–77)<0.0001
 Physical role61.56 (0–100)88.7 (72–92)NS
 Social functioning76.08 (0–100)89.31 (84–91)0.047
 Emotional role76.67 (0–100)83.37 (80–84)<0.0001
 Mental health71.19 (0–96)73.05 (70–74)NS
Table 5. Functional disability (HAQ) and summated quality of life (SF-36) scores in adults with juvenile arthritis
 Age <30 years (n = 50)Age 31–45 years (n = 24)Age >45 years (n = 8)
  1. Values are the mean ± SD. Statistical analysis by Kruskal-Wallis one-way analysis of variance showed no statistical significance when comparing PSS, MSS, or HAQ scores between age groups. HAQ = Health Assessment Questionnaire; SF-36 = Short Form 36; PSS = physical summation score; MSS = mental summation score.

PSS40 ± 234 ± 233 ± 7
MSS52 ± 348 ± 451 ± 6
HAQ score0.9 ± 0.41.3 ± 0.32.2 ± 0.8

There was a trend, albeit not significant, that generic health status, as reflected by the SF-36 scores, was lower with increasing age of the patients; this trend was largely attributable to lower PSS scores rather than lower MSS scores. A similar trend, again not significant, showed that the HAQ scores were higher (i.e., worse functional disability) in older patients. Table 6 shows the HAQ, PSS, and MSS scores, stratified according to JIA subtype. There were significant differences in physical disability outcome between the subtypes of JIA, as assessed by the PSS (P < 0.0001, by Kruskal-Wallis), and the HAQ (P = 0.0004, by Kruskal-Wallis). The most physical disability (i.e., the lowest PSS scores and the highest HAQ scores) was observed in patients with polyarticular-onset (RF positive or RF negative), systemic-onset, or psoriatic disease. The least physical disability was observed in patients with oligoarticular-onset JIA. The PSS scores, but not the MSS scores, correlated with the HAQ scores. For the 37 patients with HAQ scores <1, the mean (± SD) PSS score was 45 ± 2, which was significantly different from the mean PSS score for the 35 patients with HAQ scores >1 (29 ± 2; P < 0.0001 by Mann-Whitney U test). The MSS scores of patients were low compared with those of controls for all subtypes of JIA and irrespective of the HAQ score, which suggests that the psychosocial impact of JIA is marked in patients with all subtypes of disease, regardless of the degree of physical disability.

Table 6. HAQ scores, PSS, and MSS according to JIA subtype at disease onset*
  • *

    Values are the mean ± SD. P values were determined by Kruskal-Wallis one-way analysis of variance. HAQ = Health Assessment Questionnaire score; PSS = physical summation score; MSS = mental summation score; JIA = juvenile rheumatoid arthritis; RF = rheumatoid factor. Variation between JIA subtypes was significant for the HAQ score and the PSS, but not for the MSS.

  • P = 0.0004.

  • P < 0.0001.

  • §

    P not significant.

Enthesitis/spondylarthropathy0.98 ± 0.6740 ± 448 ± 8
Extended oligoarticular1.1 ± 0.9840 ± 1752 ± 48
Oligoarticular0.25 ± 0.2245 ± 552 ± 4
Polyarticular RF-positive1.73 ± 0.5232 ± 448 ± 8
Polyarticular RF-negative1.36 ± 0.3533 ± 752 ± 6
Psoriatic1.05 ± 0.6338 ± 556 ± 4
Systemic1.63 ± 0.6534 ± 656 ± 6

Table 7 shows the levels of education attained by adults with JIA. The General Certificate of Education Advanced levels, which are the UK national examinations, are usually taken at age 18 years, and are used as entry requirements for university enrollment. Students up to age 16 years follow a national curriculum of subjects, and the General Certificate of Secondary Education examinations are usually taken at age 16 years, at the end of high school education.

Table 7. Educational achievement in adults with juvenile arthritis*
 PatientsLocal controls
  • *

    Values are the percentage of patients who attained qualifications. Data for local controls were obtained from the Department of Education, National Office for Statistics, UK. GCE Advanced Level = General Certificate of Education Advanced Level (UK National Examinations usually taken at age 18 years); GCSE = General Certificate of Secondary Education (UK National Examinations usually taken at age 16 years).

University degree1510
GCE Advanced Level21.623
GCSE or equivalent4637
No formal qualification2718.7

In this study, the academic achievement of patients with JIA was comparable, if not better, than that of local controls, and this trend for excellent educational attainment was observed for both secondary and tertiary education (P < 0.05, patients versus controls, by Mann-Whitney U test). There was no significant difference in HAQ scores between patients achieving success on examinations at the secondary or tertiary level of education and those patients who did not. The unemployment rates for patients were 3-fold higher than those for local controls (Table 8); the term “unemployed” excluded persons involved in full-time or part-time education and those who were homemakers (mostly women). There was a trend, albeit not significant, that patients who were unemployed (versus those who were either employed or homemakers) were more likely to be male, to have systemic or RF-positive polyarticular-onset JIA, to have greater disability (median HAQ scores 1.7 for unemployed versus 0.3 for employed/homemakers), and to have lower educational achievement as measured by their examination success (relative risk 1.36). The SF-36 scores were significantly lower for patients who were unemployed (or homemakers) compared with those who were employed or involved in higher education; this trend was significant (P < 0.05) for all physical and mental domains except emotional role and mental health.

Table 8. Employment status in adults with juvenile arthritis*
  • *

    Values are the percent.

  • P < 0.05 versus local controls (7.4%), according to the National Office for Statistics, UK, 2001.

Full-time employment39
Part-time employment9.4
Full-time education10.8
Part-time education4.7


This is the largest study to date in which the SF-36 was used as a validated measure of generic health status and QOL (28–31) in adults with juvenile arthritis in the UK. A major adverse impact of JIA was demonstrated in many patients.

The patient population in this study is representative of the spectrum of subtypes of juvenile arthritis but is biased toward the severe end of the disease spectrum, with a greater proportion of patients with systemic-onset or RF-positive polyarticular-onset JIA, many of whom had undergone joint replacement surgery. Patients with persistent oligoarticular disease, who are generally regarded as having the best prognosis, are underrepresented in our study, based on the expected proportion of subtypes of JIA seen in a regional service (32); we assume that these patients have been discharged from review because they are well and in remission.

Thirty-nine percent of the patients in our study had active disease at the time of the study (according to the PGAS score), and many were taking DMARDs, with or without corticosteroids. This observation is consistent with other outcome studies of patients recruited from hospital outpatients, showing that many patients with JIA have persistent active disease into adulthood (4, 9, 12–16, 22, 23). The current study was started before the core set of outcome variables was published (33), but because the measures proposed by Giannini et al are those commonly used in clinical practice, we are able to report our data using such tools, with the exception that we did not measure acute-phase reactants at the time of the clinical assessments. Despite these limitations, the results of our study are important, because this is the first UK study of long-term outcome in adults with juvenile arthritis in which the SF-36 was used as a validated measure of generic health status and QOL in relation to other measures of functional outcome, and in which the International League of Associations for Rheumatology criteria for JIA were applied (1).

To our knowledge, there is one other study of adults with JIA in which the SF-36 was used (20). This was a postal survey of 43 adults with JRA in the US, who completed the SF-36 and self reported whether or not their arthritis was active. Compared with controls, patients had lower SF-36 scores (i.e., worse generic health status). Interestingly, however, despite the low SF-36 scores, 75% of patients perceived their arthritis to be mild, 20% thought they had moderate symptoms with activity, and 5% reported symptoms at rest. Our adult JIA patients in the UK had worse scores on the SF-36 compared with those of the patients described by Peterson et al (20), which may be explained by the fact that our study had a lower proportion of patients with oligoarticular-onset JIA (23% versus 73% in the US study), which is considered to have a better prognosis.

The generic health scores, measured with the SF-36, demonstrate poor global outcome for patients compared with age- and sex-matched controls. Although not statistically significant, there was a trend showing higher HAQ scores and lower SF-36 scores in the older JIA patients (Table 5), suggesting that as the duration of followup increases, there is functional deterioration, with concomitant worsening in the QOL. Using the HAQ and SF-36 PSS, in which worse functional disability is represented by low PSS and higher HAQ scores, we demonstrated significant differences in outcome between the subtypes of JIA (Table 6). Patients with RF-positive polyarticular-onset and systemic-onset JIA had the worst functional outcome, and the best outcome was observed in patients with oligoarticular-onset disease. However, for the whole group of JIA patients, the HAQ scores were relatively low compared with those of a group of adult patients with RA from our hospital, who had similar disease duration (34).

In 2 recent outcome studies of juvenile arthritis, in which the mean age of patients and the duration of disease were similar to those in our study, similar low levels of functional disability using the HAQ were reported (20, 23). It is difficult to discern if this low level of functional disability in adults with juvenile arthritis reflects better physical outcome than that which is associated with RA of similar disease duration, or that the HAQ is a poor discriminant measure of functional outcome in juvenile arthritis.

Our data show that many adults with JIA have persistent active disease, and that many were taking DMARDs (mostly methotrexate) and had significantly worse HAQ scores than those of patients without active disease. The majority of patients with RF-positive polyarticular-onset JIA had active disease at the time of the study. Many patients had evidence of multiple joint involvement, which can have an impact on physical functioning and QOL that may not be adequately assessed by the HAQ. For example, the high prevalence of TMJ disease and the impact on jaw opening may contribute to the poor oral health observed in JIA patients (35), and the high prevalence of cervical spine disease may limit the ability to drive a vehicle safely, thereby limiting independence and employment opportunities.

The aim of the current study was not to evaluate the JIA criteria, although the subtype of JIA at onset did seem to predict functional outcome, with worse functional disability (i.e., higher HAQ score) being observed in patients with polyarticular-onset and systemic-onset JIA, and the best (i.e., lowest HAQ score) outcome in those with oligoarticular-onset JIA. Furthermore, the pattern of joint involvement at the time of disease onset appeared to be predictive of long-term joint involvement (i.e., those with predominantly symmetric small and large joint involvement had polyarticular or systemic onset, and those with an asymmetric pattern involving large and small joints had oligoarticular or psoriatic onset).

Many of the patients in our study had excellent educational achievement, although for some, the time involved in formal education was extended, presumably as a consequence of periods of ill health resulting in absence from school. The good educational outcome was observed irrespective of the level of functional disability (HAQ score); the reasons for this may include the personality and coping strategies of patients and their families, as well as support from teachers and schools. Many of the patients in this study will have benefited during their school years from the UK 1993 Education Act, which entitles every child with a disability to an individualized program of additional resources through a Statement of Special Educational Need during their school years, to optimize their educational opportunities.

However, it is of concern that despite excellent educational outcome, many patients were unemployed as adults, with higher rates of unemployment compared with those of local controls. There was a trend, although it was not significant, that patients with worse functional disability were more likely to be unemployed. This discrepancy between educational attainment and employment rates has been observed elsewhere in the UK (36, 37) and in Canada (22) but contradicts earlier published studies of employment in JIA (8, 15), which demonstrated employment rates of patients that were comparable to those of siblings and controls. These disparate results may be explained by the different definitions of “employment,” background rates of unemployment, cultural differences, the source of controls, and the numbers of patients in these studies. We believe that our results are significant, because they are compared with those for age- and sex-matched local controls for educational attainment and unemployment rates (information derived from local government sources). Our data suggest that although the UK educational system has worked well for these patients with JIA, there are problems in developing a “readiness for work” that has been described elsewhere (38), as well as more complex issues of attitudes and possible discrimination in the workplace against persons with disabilities (36).

The total SF-36 score reflects self-perceived generic health status and handicap, whereas the HAQ score predominantly reflects functional disability. This study shows that SF-36 (particularly MSS) scores were low for all JIA subtypes compared with controls, and although the HAQ and PSS scores correlated well with each other in patients (not unexpectedly, because they are both measures of functional disability), the MSS scores were low irrespective of the degree of functional disability. This observation warrants further investigation but suggests that QOL from a patient's perspective is dependent on a variety of factors, not just physical impairment and functional disability.

It is well recognized that many young people with chronic disabilities have difficulties with the process of transition into adulthood, including developing relationships, living independently, and finding a vocation (39–41). Despite excellent educational attainment, many of the patients in our study were unemployed, and even though those who were employed often had nonmanual jobs, and many were employed only part-time, their SF-36 scores were significantly higher than the scores of those who were unemployed (or homemakers). These data suggest an important association between employment and QOL in these adults with JIA: the advantages of being employed are many and include financial rewards and feelings of self worth, autonomy, responsibility, and independence. We did not assess directly the mental health status of our patients, but the adverse psychological impact of juvenile arthritis on young people has been observed in an earlier study, in which the impact was more profound in those with onset of disease in adolescent years, and many were unemployed (18). Such factors may well affect their self-perceived QOL scores as compared with those of their healthy peers.

There is emerging evidence that severe physical complications of JIA (including blindness and death) are less common than a decade ago (22, 23, 42, 43). Using traditional methods of measuring functional disability (e.g., Steinbrocker criteria), there does appear to be an improvement in outcome, with a lower proportion of patients having severe disability. This improving outcome is likely to reflect the increasing awareness that JIA is not a benign disease, that joint damage occurs early (44), that early involvement of an experienced multidisciplinary team is important, with a trend toward earlier and more aggressive therapeutic regimens, including early use of intraarticular steroid injections and methotrexate. Based on the early experience using immunotherapies, including biologics and autologous stem cell transplantation (45, 46), the future is even more promising. However, there is no room for complacency, because several studies, including the current one, show that many patients (especially those who are RF positive with polyarticular-onset JIA) have a poor physical outcome. There is still the need for accurate and early prognostic indicators to identify patients whose disease is associated with a poor prognosis and to target more aggressive treatment at an early stage in such patients.

We have shown that regardless of the level of functional disability and the subtype of JIA, the effect of JIA on QOL in young adults is profound, with poor self-perceived health status and, despite excellent educational attainment, high rates of unemployment. Therefore, in order to improve QOL in the long term, the challenge is not only to reduce physical impairment and disability through targeted aggressive treatment but also to address the more complex issues of improving a patient's ability to be an independent adult living in society, with self-worth and self-esteem, regardless of their functional disability. The transition from a pediatric to an adult setting needs to be planned and comprehensive and should include vocational planning, experience and skills for autonomous and independent living, as well as disease-specific and generic health issues (38–41). Currently, there are few adolescent arthritis clinics in the UK and Eire (47), and there is a need to develop transitional care services incorporating input from health, education, and social services, as well as industry, with the aim of improving the long-term global outcome of JIA.