Abstract
OBJECTIVE: Prognostic enrichment in clinical trials aims to target potential participants at high risk for events of interest, such as disease progression or incident osteoarthritis (OA). The objective of this work is to describe and investigate key parameters that should inform the decision to implement an enrichment strategy. DESIGN: We apply the framework of the Biomarker Prognostic Enrichment Tool (BioPET) to develop the Research-Based Enrichment for Arthritis Clinical Trials (REACT) tool. REACT takes into consideration the cost of screening and enrichment accuracy to determine whether an enrichment strategy is cost- and time-saving as compared with no enrichment. RESULTS: Modestly accurate prognostic enrichment can yield cost-savings if the costs of enrichment are low; we show cost-savings for a scenario with a prognostic enrichment algorithm with an area under the curve (AUC) of 0.6 and a cost of enrichment of $25 per potential participant screened. The increased number of potential participants needed to screen in order to enroll an enriched sample may present logistical and/or feasibility challenges. While the example with a prognostic enrichment algorithm AUC of 0.6 is indeed cost-saving from a monetary perspective, enrichment in this scenario requires screening 1.7 times more potential participants than a strategy with no enrichment. CONCLUSIONS: REACT is a user-friendly tool to assess the value of incorporating an enrichment algorithm into clinical trial design. It could assist investigators in evaluating enrollment strategies to critically evaluate enrichment trade-offs, consider operational feasibility, in addition to qualitative factors such as generalizability and acceptability by participants.