Background/Purpose: The LIMIT-JIA trial is the first study of the use of biologic therapy to prevent disease extension in children with newly diagnosed, uncomplicated, oligo-articular course (“early limited”) juvenile idiopathic arthritis (JIA). Study recruitment requires extensive efforts to identify and approach eligible patients. Potential subject identification was initially conducted manually by study staff. This project developed and implemented a computable phenotype to enhance timely identification of potentially eligible subjects for this randomized trial.

Methods: Monthly reports of potentially eligible subjects for the trial were generated by developing, testing, and then applying an “early limited JIA” computable phenotype to The Childhood Arthritis and Rheumatology Research Alliance (CARRA) Registry records throughout 2022. The algorithm inclusion criteria included individuals diagnosed with oligoarticular JIA within the preceding six months; exclusion criteria included extension to polyarticular course, uveitis, treatment with DMARDs, or comorbid diagnoses of psoriasis, inflammatory bowel disease, or sacroiliitis. Study sites received monthly reports and annotated these reports with actual eligibility status of potential subjects from the CARRA Registry. Several sites configured on-demand reports using their electronic health record (EHR) software to identify potential subjects. All sites were surveyed about enrollment using CARRA Registry and their EHR reports. Data analyses were conducted and descriptive statistics were compiled.

Results: There were 15 study sites with 50 potentially eligible subjects identified by the computable phenotype during the study period. The median age was 6.7 [3.6, 12.1] years, and 36 (72%) were female. Thirty-two (64%) were not eligible for the study. The most common reasons that subjects were ineligible were related to temporal factors and included: 13 (41%) had recruitment delayed beyond the six-month enrollment window, 8 (25%) had started systemic therapy, and 5 (16%) no longer met oligoarticular JIA criteria (e.g. developed psoriasis or sacroiliitis) or had a change of diagnosis. Other ineligibility reasons included inability to be recruited virtually, language barriers, and no follow-up scheduled within the enrolment time window. Of the 18 eligible subjects, 9 (50%) declined enrollment. Nine study sites responded to monthly surveys; only three were using their EHR reporting tools to complement CARRA Reports, and most sites reviewed EHR reports monthly. Sixteen subjects were identified using EHR tools, 3 were eligible for the study, but none were approached for the study.

Conclusion: Novel tools to identify subjects for study recruitment are necessary to improve study recruitment, particularly in studies with time-sensitive inclusion criteria and pediatric prevention trials with active treatment arms. Development of accurate computable phenotypes to assist with subject identification requires accurate data and timely recruitment strategies.

ACR Meeting Abstracts - https://acrabstracts.org/abstract/development-of-an-electronic-clinical-phenotype-to-identify-potential-study-subjects-with-juvenile-arthritis/