Development of the parent version of the juvenile arthritis disease activity score cutoffs for moderate and high disease activity states in juvenile idiopathic arthritis in a large multinational patient sample

Abstract

Background: Measurement of disease activity level is of pivotal importance in the care of patients with juvenile idiopathic arthritis (JIA). According to the most recent requirements, both, parent's and children's perception should be taken into account while evaluating the disease course and assessing effectiveness of therapy. Therefore, a new disease activity evaluation tool, based only on parent assessment of the outcome, is under development and named Parent Juvenile Arthritis Disease Activity Score (parJADAS) [1]. Objectives: The aim of this study is to develop the parJADAS cut-off values of moderate disease activity (MDA) and high disease activity (HDA) in JIA patients. Methods: The parJADAS (score range 0-40) is the sum of 4 values: 1) parent's assessment of disease activity on a 21-numbered circle 0-10 VAS; 2) assessment of pain intensity on a 21-numbered circle 0-10 VAS; 3) proxy assessment of joint disease up to a maximum of 10 joints; 4) assessment of morning stiffness (MS) on a Likert scale, ranging from no MS (0 points) to > 2 hours of MS (10 points). The study dataset is composed of 2,412 patients with JIA, seen in 3389 visits with parJADAS available, enrolled in the the multinational registry PharmaChild, assessing the long-term safety of treatment of children with JIA. At each visit, subjects were subjectively rated as being in inactive disease, low disease activity, MDA, or HDA by the attending physician. For each patient, only one visit per disease state was retained. To identify the cut-offs the following methods were implemented: 1) Mapping: the 25th percentile value of the parJADAS in patients having MDA or HDA, respectively, was calculated; 2) Youden Index: Youden Index (J) identifies the maximum potential effectiveness of the biomarker through the receiver operating characteristic (ROC) curve analysis; 3) Max agreement: The analysis of agreement was based on kappa statistics, which assesses the agreement beyond chance between 2 dichotomous ratings. The first rating was obtained using all possible parJADAS values as hypothetical test criteria; to obtain the second rating, the categorical ratings of each attending physician were dichotomized and were coded as 0 or 1. Results: Preliminary cut-off values for parJADAS with sensitivity and specificity are presented in the table. Conclusion: Tentative cut-off values for classifying the states of MDA and HAD using parJADAS were calculated. The obtained values will be tested in the validation analysis. Once validated the cut-offs are ideally suited to identify subjects at risk of disease flare when remotely monitored with the parJADAS.L6333047692020-11-10 <br />