Background/Purpose: With the advent and implementation of advanced drug therapy clinical inactive disease (CID) has become an attainable target in the treatment of juvenile idiopathic arthritis (JIA).   As a result CID is becoming an endpoint in clinical trials and a goal of quality improvement efforts.  The ACR Provisional Criteria that describe CID include the physician/provider’s global assessment (PGA), which must be scored as zero on a scale of 0 (indicating no disease activity) to 10 (most disease activity) or ≤ 1 when using the Juvenile Arthritis Disease Activity Score.  Scoring of the PGA is not standardized and considerable inter-provider variation is known to exist.  Our goal was to measure the degree of this variation when scoring patients who were known to meet the ACR provisional criteria of CID, (intended for systemic, oligo- and polyarthritis) and determine clinical and/or laboratory parameters that contribute most to the non-uniformity in the assessment of the PGA in clinical practice.

Methods: Twenty patient profiles from the JIA registry at Cincinnati Children’s Hospital Medical Center (CCHMC) who met the ACR provisional criteria for CID were given to a total of 51 providers who were members of the clinical faculty of the Rheumatology division of CCHMC, the Childhood Arthritis and Rheumatology Research Alliance (CARRA), or of the Pediatric Rheumatology –  Care and Outcomes Improvement Network (PR-COIN).  Using Delphi questionnaire and nominal group techniques for consensus building we determined the intra-class correlation (ICC) for the entire provider group, and asked what variables most strongly influenced their scoring of the PGA. 

Results:   The twenty cases included systemic 0, oligoarthritis 6, polyarthritis 7, psoriatic 3, ERA 3, arthritis associated with IBD 1.  Two patients had uveitis.  A total of 51 providers participated in the exercise with a total of 1022 PGA scores.   The ICC was 0.15.  The standard deviation (SD) due to patient differences was 0.127, and SD due to different raters of the patients was 0.697, suggesting that differences in physician-ratings accounted for most of the variability.    Variables that influenced scoring were presence of pain, questionable TMJ involvement, joint loss of motion, morning stiffness, psoriasis, and past history of uveitis.

Conclusion: Combined data from 3 groups of experienced pediatric rheumatologists reveal that substantial variation in scoring the PGA continues to exist when scoring JIA patients with very low disease activity.  JIA is a heterogeneous collection of conditions, with various extra-articular manifestations that providers consider in making assessments.  Due to the complexity of the conceptual framework of the PGA, standardization is likely not feasible.  This implies that other disease states, other than CID, that do not require the PGA to equal zero may be more suitable for clinical trial endpoints and when setting goals of quality improvement efforts.

« Back to 2015 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/inter-provider-reliability-in-scoring-the-physician-global-assessment-of-disease-activity-among-patients-with-juvenile-idiopathic-arthritis-patients-who-met-the-acr-provisional-criteria-for-clinical-i/