Patient generated research priorities to improve diagnostic safety: A systematic prioritization exercise

Background: Most people experience a diagnostic error at least once in their lifetime. Patients’ experiences with their diagnosis could provide important insights when setting research priorities to reduce diagnostic error. Objective: Our objective was to engage patients in research agenda setting for improving diagnosis. Patient involvement: Patients were involved in generating, discussing, prioritizing, and ranking of research questions for diagnostic error reduction. Methods: We used the prioritization methodology based on the Child Health and Nutrition Research Initiative (CHNRI). We first solicited research questions important for diagnostic error reduction from a large group of patients. Thirty questions were initially prioritized at an in-person meeting with 8 patients who were supported by 4 researchers. The resulting list was further prioritized by patients who scored questions on five predefined criteria. We then applied previously determined weights to these prioritization criteria to adjust the final prioritization score for each question, resulting in 10 highest priority research questions. Results: Forty-one patients submitted 171 research questions. After prioritization, the highest priority topics included better care coordination across the diagnostic continuum and improving care transitions, improved identification and measurement of diagnostic errors and attention for implicit bias towards patients who are vulnerable to diagnostic errors. Discussion: We systematically identified the top-10 patient generated research priorities for diagnostic error reduction using transparent and objective methods. Patients prioritized different research questions than researchers and therefore complemented an agenda previously generated by researchers. Practical value: Research priorities identified by patients can be used by funders and researchers to conduct future research focused on reducing diagnostic errors.