Background/Purpose: Health-related quality of life (HRQOL) is a key outcome in clinical care and research for children with JIA.  Despite excellent clinical control and the availability of biologic treatment, HRQOL remains suboptimal in half of patients with JIA.  The aims of the study were, for children with JIA during the first year post diagnosis with JIA, to:  1) identify distinct patterns of improvement in HRQOL; and 2) identify baseline disease, patient/parental characteristics that predict improvement in HRQOL.

Methods: Patients were enrolled within 6 months of diagnosis (N=230; 2-16 years of age), and followed up semi-annually.  Patient HRQOL (Generic and Rheumatologic Core via self-report if ≥5 years and parental proxy-report) and non-medical patient/parental characteristics were obtained via survey during clinic visits; disease medical characteristics were obtained from clinical database. Primary outcome Generic PedsQL total score has mean 83 in healthy children, with <78.6 considered suboptimal. Longitudinal growth mixture model (GMM) was used to identify patterns (latent class) of improvement in HRQOL.  Multivariable multinomial logistic regression with stepwise selection was used to predict GMM latent class membership using baseline measures; the model predictive ability was evaluated by the receiver operation characteristics (ROC, 1.0 being perfect).

Results: Table 1 presents study results for primary study outcome. Three distinct patterns of improvement were identified for patient self-report: 72% of patients maintained HRQOL at healthy norm (class I:  84.46+0.23×month), 21% improved (class III: 54.46+2.03×month), but 13% of patients with initial suboptimal HRQOL deteriorated over time (class II: 66.63-1.40×month).  Modeling parent proxy-report, four latent classes were identified.  The final logistic regression model for patient self-report found that less pain, adherence barriers, better self-efficacy managing emotion, and social support from parents are most predictive to more favorite pattern of HRQOL improvement (ROC 0.90). The final model for parental proxy-report found that parent- reports of less child pain and adherence barriers, better functional status and lower parental distress at baseline predicts more favorite pattern of improvement in child HRQOL (ROC 0.83).

Conclusion: During the first year on treatment, children newly diagnosed with JIA showed distinct patterns of improvement in HRQOL.  Although a majority of JIA patients maintained or reached healthy norm, 28-43% of patients remained or deteriorated below healthy norm, despite therapy, suggesting at risk factors of poor HRQOL outcomes need to be better considered in the care of children with JIA.

Table 1. Latent Class of Pattern of Improvement (GMM Results) and their predictors (multinomial logistic regression results)

* indicates P<.05; ** indicates P<.01; *** indicates P<.001