Juvenile idiopathic arthritis (JIA) is a heterogeneous group of disorders that affects more than 1 in 1,000 young persons in the UK (1). Living with JIA is likely to be challenging at any age. However, its impact during adolescence may be particularly difficult (2, 3). Growth and puberty can be affected, and important adolescent tasks (i.e., consolidation of identity, independence from parents, establishing relationships outside the family, finding a vocation) may also be disrupted (2, 4).
Health-related quality of life (HRQOL) is an important outcome measure in understanding the impact of chronic illness (5), including childhood-onset rheumatic diseases (6–8). HRQOL is typically conceptualized as a multidimensional construct that reflects the definition of health proposed by the World Health Organization (9) and, by focusing on the patient's personal perspective of health, supports the view that young people should be consulted about issues and events that directly affect their lives (10–13). It is important to note that HRQOL is conceptually distinct from quality of life (QOL) (8, 14). The latter is a much broader construct in which health is only 1 of many determinants and therefore may not be affected by intervention. In a study of 122 patients (ages 1–18 years) referred for rheumatologic care, Feldman et al (8) found that overall, QOL was higher than HRQOL and that HRQOL explained only a small to moderate amount of the variation in QOL scores.
As yet, the HRQOL of adolescents with JIA has received limited assessment (6–8, 15, 16). Few studies have focused specifically on JIA, and although adolescents are included within the study samples, the authors have not differentiated their findings on the basis of age or developmental level. Knowledge about HRQOL in adolescents is consequently difficult to disentangle from that of younger children. Overall, the findings suggest that children and young persons with JIA have worse HRQOL than healthy controls (15) and that HRQOL decreases with increasing disease severity (8, 16), active joint count (16), pain (6, 7, 16), degree of disability (7, 8), and reduced general health/well- being (7).
One established disease-specific measure of HRQOL in childhood-onset rheumatic disease is the Juvenile Arthritis Quality of Life Questionnaire (JAQQ) (16). Preliminary analyses in Canada have demonstrated the JAQQ to be valid and responsive to change when used with 91 children (ages 1–18 years) with chronic arthritides (16). That study, together with work by Brunner and Giannini in the US (7), has supported construct validity through strong to moderate correlations with functional ability (as measured by the Childhood Health Assessment Questionnaire [CHAQ]) (7); moderate correlations with pain (7, 16), global ratings of health (7), and well-being (7); and weak correlations with active joint count (16) and joint severity (16).
Criterion validity in the context of concurrent validity has also been demonstrated through strong correlations between the JAQQ and the Pediatric Quality of Life Questionnaire Inventory Rheumatology Module (7, 15). However, there are no published studies reporting the validity and reliability of the JAQQ among UK samples and no studies that have specifically explored the HRQOL of adolescents with JIA. Therefore, in addition to describing the HRQOL of adolescents with JIA, the present study aims to examine the usefulness of the JAQQ in measuring HRQOL among adolescents in a UK context and to explore the roles of demographic and disease-related factors in HRQOL. In line with previous studies, it was hypothesized that HRQOL would decrease with worsening disease, disability, and well-being.
PATIENTS AND METHODS
The patients and parents included in this study were part of a national clinical trial to evaluate a program of transitional care (17). Three age cohorts (11, 14, or 17 years, ±1 month), corresponding to early, middle, and late phases of adolescence, were recruited from 10 pediatric rheumatology centers represented in the British Society of Paediatric and Adolescent Rheumatology (BSPAR). To be eligible for the study, patients needed to have a diagnosis of JIA as defined by the revised International League of Associations for Rheumatology criteria (18) and were expected to remain in the pediatric care of a consultant member of the BSPAR for at least 6 months (19).
Clinical data were provided by the senior clinician. Patient and parent data were collected using individual questionnaires designed for self completion, with support from local program coordinators where necessary.
HRQOL was assessed using the JAQQ (16), which consists of 74 items grouped into 4 dimensions: gross motor function (17 items), fine motor function (16 items), psychosocial function (22 items), and systemic symptoms (19 items). The response format for each item uses a 7-point, Likert-type, frequency/difficulty scale (where 1 = none of the time/never and 7 = all of the time/always). Each item also has a “does not apply to me” option. The mean score for the 5 highest scoring items in each of the 4 dimensions is computed as the dimension score; the total JAQQ score is computed as the mean across the 4 dimensions, again with a score range from 1 to 7. Higher scores indicate lower HRQOL. In addition, respondents are also asked to select up to 5 areas of functioning that represent their “biggest problems” within each domain.
Demographic and disease-related data.
Data were collected on age, sex, ethnicity, socioeconomic status based on the reduced 5-class version of National Statistics Socio-Economic Classification (20), family structure, parental marital status/education/employment, and adolescents' educational and prevocational status (including statements of special educational need, a legal requirement in the UK stating that every child with a disability is entitled to an individualized program of additional resources during their school years to optimize their educational opportunities). Other data included JIA onset subtype, age at onset, age at diagnosis, disease duration, and independent health behaviors (self medication, independent consultations).
Core outcome variables.
Core outcome variables included a validated 6-item data set of 1) physician's global assessment (PGA) of overall disease activity, 2) a measure of acute-phase proteins (either erythrocyte sedimentation rate [ESR] or C-reactive protein [CRP]), 3) active joint count, 4) limited joint count, 5) the British version of the CHAQ (scores range between 0 and 3, with higher scores indicating greater disability) as modified for use with adolescents in the UK (including 100-mm visual analog scales for pain [VAS-pain]), and 6) a patient or parent estimate of the patient's overall well-being (VAS-global) (19–22).
Knowledge was assessed using a 16-item, disease-specific, multidimensional measure with multiple-choice response format (23). Final knowledge scores range from 0 to 16, where higher scores indicate greater knowledge.
Satisfaction with rheumatology care.
Satisfaction was measured using a 22-item measure designed for this study. Items included statements about the physical environment, clinic procedures, relationships with health care personnel, information, and support. Respondents were asked to rate the extent to which each item represents “best” and “current” care using a 7-point Likert scale anchored by “strongly disagree” at 1 and “strongly agree” at 7. Satisfaction with each item was conceptualized as the gap between respondents' “best” and “current” score (i.e., their “gap” score). Overall, satisfaction was computed as the mean of all gap scores. Scores ranged between 0 and 7, with higher scores indicating lower satisfaction. The study had Multicentre and Local Research Ethics Committee approval, and all participants gave written informed consent.
SPSS version 11.0 (SPSS, Chicago, IL) was used to perform all analyses on the quantitative data. There was evidence of skewness for some variables, which led to the choice of nonparametric inferential statistics.
The psychometric properties of the JAQQ (16) in a UK context were examined in relation to the following: feasibility, assessed by examining completion rates; face validity, assessed by examining the number of “not applicable” answers; reliability, assessed by examining the JAQQ's internal consistency, using Cronbach's coefficient (results from homogeneous groups are typically expected to reach coefficients >0.8 ); and construct validity. (Based on HRQOL theory and previous results [6–8, 15, 16], it was predicted that adolescents who reported high levels of disease activity, pain, physical disability, and poor general health would have low HRQOL [reflected by high JAQQ scores]. Previous use of the JAQQ [7, 16] suggested a number of a piori predictions: that the strongest correlations would be observed between JAQQ and disability [CHAQ], that pain and well-being would be at least moderately correlated, and that active joint count would be weakly correlated. Therefore, construct validity was examined through Spearman's correlations between the JAQQ and the core outcome variables.) Group differences were analyzed using chi-square, Mann-Whitney, Kruskal-Wallis or Jonckheere-Terpstra, and Friedman's tests, as appropriate. Associations between the study variables were analyzed using Spearman's rho correlations. The strength of statistically significant correlations was defined as very weak (rs < 0.20), weak (rs = 0.20–0.39), moderate (rs = 0.4–0.69), strong (rs = 0.70–0.89), and very strong (rs = 0.90–1) (25).
Stepwise linear regression was used to identify the best predictors of JAQQ. Variables entered into the analysis included only those that displayed significant univariate relationships with the JAQQ. This was done to limit the potential explanatory variables to a moderate number and therefore optimize the stability of the estimated regression lines. Linearity was confirmed through scatterplots and appropriate normality tests on residuals were implemented (i.e., inspection of histograms for the residuals and normal probability plots). The extent to which the model fit the data was indicated by the R2 value, which shows the percentage of variance in JAQQ that can be explained uniquely or jointly by the independent variables. The change in R2 for the regression upon entry of a variable was taken as the proportion of the variation in final JAQQ scores explained by that variable.
In view of the relatively small numbers of certain JIA subtypes in each age group (unpublished observation), patients were divided into 2 main groups, “oligoarthritis persistent” and “other,” which were, by definition, patients who had primarily polyarthritis and/or had systemic involvement (significant extraarticular disease) with arthritis. Ethnicity categories were similarly collapsed into “white/European” and “other.” Statistical significance was set at the 0.05 level.
This study of HRQOL in JIA is the largest to date in the UK and is unique in the international literature for its focus on adolescents. The results indicate that the HRQOL of adolescents with JIA is less than optimal, particularly in the domains of gross motor and systemic functioning, and, as hypothesized, has significant and independent relationships with pain, disease activity, and functional disability. As such, the results are consistent with previous studies of HRQOL (6–8). Improvement of HRQOL is, therefore, inextricably linked with managing these key aspects of care. However, not all of the variation in JAQQ scores could be explained, and understanding the role of other determinants will be crucial in promoting adolescents' HRQOL.
It is of concern that almost one-third of participants reported frustration among their biggest problems. Adolescents with JIA are faced with symptoms that can be difficult to relieve, activities they cannot perform, and the uncertainty of daily fluctuations and long-term prognosis. In this context, frustration is not unexpected. However, establishing realistic expectations and emotional responses (e.g., through cognitive techniques) is an important step in self management. Holman and Lorig (26) specified 7 basic components of self management. These included minimizing or overcoming disability (e.g., through exercise, nutrition, or rest), establishing realistic expectations and emotional responses to the changes of circumstances caused by the disease, interpreting and managing symptoms, learning how to judge the effects of medications and manage their use, learning to become a good problem solver, communicating with health professionals, and using social resources.
Although self-management programs are established in adult rheumatology practice (26) and have been shown to be effective (27–29), similar efforts have not been made in the field of adolescent rheumatology, despite the fact that adolescence represents the time when responsibility for disease management typically shifts from parents and professionals to the adolescent. Self management is just as necessary for adolescents as it is for adults, and it is likely that many adolescents could benefit from such intervention, acknowledging that it would have to be developmentally appropriate.
In addition to frustration, more than one-fifth (23.4%) of adolescents also rated depression among their biggest problems. This is concordant with a recent population-based study by Adam et al (30) who found that 19% of Canadian adolescents with arthritis and rheumatism (n = 213) reported depressive symptoms. Although depression does appear to increase substantially from preadolescence to adolescence (31, 32), Adam et al (30) demonstrated greater prevalence of depressive symptoms among adolescents with rheumatic conditions compared with those without, of which only 5% reported depressive symptoms. In our sample, the 17-year-olds appeared to be at greatest risk, with almost two-thirds reporting depressive symptoms, and although this may not necessarily reflect clinical levels of depression, it clearly warrants further attention. Packham et al (33) found that 21.1% of 246 adults with JIA had experienced significant depression in the past, and of these, 38.5% had their first episode between 15 and 25 years. Similar to our study, they also found that depression was most common in those adults whose JIA had begun between 6 and 12 years, as opposed to those with an early or late onset. However, although many adults had previously experienced depression, at the time of the study only 5.2% were depressed (33). The authors attributed this to more effective coping through increased experience. Another explanation is the particular impact of disease onset on early adolescent development. However, although young people may adapt to their disease in time, adolescent depression still requires attention to prevent it from adversely affecting other aspects of development (e.g., education).
This study has also identified other risk factors for depression including greater disease activity, greater number of active joints, worse pain, worse general health, and greater functional disability. To some extent, this may also explain why the 17-year-olds were more likely to select depression as one of their biggest problems, in that they reported significantly greater pain than younger groups and were less likely to have persistent oligoarthritis JIA. This is concordant with findings of other authors (34, 35), including Adam et al (30), who have demonstrated a strong association between pain and depression. Good pain management and appropriate psychological referral of these patients within rheumatology services is therefore critical. Pain management also has a generic relevance beyond musculoskeletal pain, with 28.2% of adolescents reporting headaches. Palermo and Kiska (34) have also shown complex relationships between chronic pain, depression, and sleep disturbances (an area of health not accessed in the JAQQ) and suggest that the HRQOL of adolescents with chronic pain may be substantially improved through interventions that improve their sleep. Sleep difficulties are well recognized in adolescent health, often with a multifactorial etiology (36). Fatigue is closely interrelated with sleep, and the former was frequently reported to be a problem in the JAQQ. A combination of fatigue management, pacing skills, and sleep hygiene are therefore areas of relevance to the adolescent rheumatology team, and further research into interventions addressing these issues is awaited. The role of psychosocial factors in depression also requires further study. Indeed, physical disease factors have explained only a small amount of the variation in the depression of adults with JIA (33), and this is likely to be the case with adolescents. A useful clinical screening tool for use in adolescent rheumatology clinics is the HEADS (home; education; exercise; activities, affect, ambitions; dental, diet, driving, drugs; sleep, sex) mnemonic (37), which captures affect (for mood/frustration/depression), activities (for functional ability), and sleep (to include fatigue).
In addition to describing the HRQOL of adolescents with JIA and discussing potential interventions, this study also provides evidence for the suitability of the JAQQ in a UK context. Preliminary analysis of the JAQQ's psychometric properties when used with UK adolescents has shown it to be feasible, have face validity in terms of item relevance, have excellent internal consistency, and have good construct validity. The authors have also shown it to be responsive to change (38). However, the JAQQ does appear to have a number of limitations when used with this group. Not all adolescents were able to complete the JAQQ independently, particularly the younger adolescents, and more than one-fifth (21.8%) failed to answer all 78 items. Therefore, although the JAQQ is certainly comprehensive, there is a danger that clinically important information may be lost as a consequence. Perhaps one way to adapt the measure for future use is to exclude some of the items that were seldom selected as adolescents' biggest problems. However, this said, it would be important to retain items that are clinically important (e.g., “blood on stools”).
In terms of face validity, most items were seen as relevant by the majority of adolescents. However, it did appear that at least some adolescents who are unable to perform activities because of their arthritis are mistakenly rating these as not applicable. “Playing a favorite sport” and “riding a bicycle” appeared to be 2 such items and in both cases were rated not applicable by >10% of adolescents who were characterized by significantly greater pain, worse general well-being, and disability. Validity could therefore be improved by defining the not applicable option more explicitly, perhaps in relation to age as in the CHAQ. It is also important to note that both item generation and face/content validity of the JAQQ were determined in relation to professional opinions. However, face validity usually refers to whether a scale reflects the content of the concept in question, usually in the eyes of the respondent (39). The JAQQ does allow young persons to raise additional issues that are problematic for them. However, although many issues were identified, each was raised by <1% of the sample. The failure to volunteer a significant number of new items may reflect that the instrument is sufficiently complete or that young persons are not inclined to offer additional items after completing the many items that make up each dimension (questionnaire fatigue). The low frequency of some items (e.g., bullying, needle phobia) was surprising in light of the authors' clinical experiences and previous research (2) and may actually reflect the fact that the JAQQ has a relative functional bias, and may not encourage adolescents to think outside of the medical context of the JAQQ. Face validity is likely to be enhanced if item generation is undertaken in close reference to adolescents and if it includes developmentally appropriate domains that reflect the impact of JIA on those social, sexual, educational, and vocational aspects of life that have been shown to be salient for adolescents with JIA (2). Response burden may also be minimized by instructing young persons to read all items but only score a specific number (1, 2, or 3 dependent on the dimension) and then to add others if there more pertinent issues.
To interpret the results, a number of caveats should be mentioned. First, this study has used a single measure of HRQOL that has not been validated within a UK context. Although the JAQQ does appear to be suitable for use within the UK, it is possible that the results reflect unique properties of this measure. For instance, at least 3 of the 4 domains pertain to physical symptoms/function, and therefore it is not surprising that JAQQ scores are significantly related to many of the core outcome variables. However, the core variables also correlated with the psychological domain and support the use of the JAQQ as an HRQOL instrument as opposed to purely a functional assessment tool.
A further difficulty in using the JAQQ is the lack of calibration, and although measures such as the JAQQ have been shown to measure statistically significant changes in scores (16, 38), there remains no agreed definition as to what constitutes good and bad HRQOL and what amount of change is deemed clinically meaningful. In part, this is also limited by the lack of a control group, and therefore further research is required to understand how the HRQOL of adolescents with JIA compares with their healthy peers.
The results must also be considered in relation to the sample, which was largely white with a female predominance. As such, it is difficult to establish the extent to which the JAQQ is culturally and ethnically competent, both of which are important in collecting data with adolescents (40).
In conclusion, this study has demonstrated that JIA can have a significant adverse affect on the HRQOL of adolescents, regardless of their age. Unfortunately, there is no evidence to suggest that these issues will spontaneously resolve themselves if left untreated. Indeed, studies show that at least some young adults with JIA are at risk for clinical depression, unemployment, relationship difficulties, and functional problems compared with controls (33, 41–44). However, there is great potential for intervention. Evidence from a recent study has shown significant improvements in the HRQOL of adolescents with JIA who were enrolled in a program of transitional care (38), which, by definition, attended to the “medical, psychosocial, and educational/vocational needs of adolescents as they move from child to adult-centered care” (45).
We would like to thank the young persons and their parents who participated in this study; Bev Thomas, the project secretary who tirelessly chased, retrieved, entered, and collated all the data; and Dr. Paul Davies at the University of Birmingham for his statistical advice. We would also like to thank the following contributors: Eileen Baildam, Jeremy Camelleri, Diane Coulson, Joyce Davidson, Sue Ferguson, Helen Foster, Paul Galea, Janet Gardner-Medwin, Janine Hackett, Ann Hall, Gill Jackson, Jane Kelly, Sue Kemp, Nicki Kennedy, Ruth McGowan, Steph Phillips, Clarissa Pilkington, Alison Swift, Helen Venning, Tracey Whitely, Jane Wilby, Sue Wyatt, and Helena Wythe.