Abstract
Introduction. Limited evidence guides pediatric rheumatologists on when to withdraw biologic therapy in children with juvenile idiopathic arthritis, resulting in wide variation in clinical practice. This study aimed to develop and evaluate a decision support tool (DST) based on expert opinion to support pediatric rheumatologists in making withdrawal decisions. Methods. A literature review, focus groups, interviews, and prior research informed the design of the prototype DST. Evaluation of the DST's face validity, content validity, acceptance, and feasibility was conducted through user testing interviews and a survey among pediatric rheumatologists from the Netherlands and Canada. Findings were summarized using descriptive and qualitative content analyses. Results. The prototype DST requires input on relevant patient, disease, and treatment characteristics. Its primary output is the predicted likelihood of biologic therapy withdrawal. Pediatric rheumatologists can adjust the importance of characteristics and observe the resulting impact on withdrawal likelihood. Eleven pediatric rheumatologists participated in testing. Key themes identified included the need for 1) clear terminology to ensure consistent interpretation of model inputs, 2) concise instructions on how and when to adjust the relative importance of characteristics, and 3) practice rounds to build trust among pediatric rheumatologists in the DST's output. Participants found the DST feasible for clinical use, with its main value in explaining decisions to patients and engaging them in the decision-making process. Suggested future improvements include tracking the outcomes of withdrawal decisions and integrating predictive models based on clinical data. Conclusions. The DST developed in this study was well-received. Its main value lies in helping pediatric rheumatologists explain their decisions to patients and parents. The top priority for further development is integrating scientific evidence on successful withdrawal decisions. HIGHLIGHTS: Decision support tools that provide structure to decisions based on expert opinion can increase transparency and consistency in medical decision making in the absence of clinical evidence.Data from clinical vignette studies that use an experimental design to elicit treatment preferences can be used to predict treatment decision making.A decision support tool to support biologic therapy withdrawal decisions has the most value in explaining the decision to children with nonsystemic juvenile idiopathic arthritis and their parents.