A Statistical Method for Benefit–Risk Assessment in Clinical Trials

In recent years, benefit-risk assessment (BRA) has become a crucial tool in guiding regulatory decisions regarding the safety and efficacy of investigational treatments. Following the FDA's recent guidance on the BRA, we statistically develop a confidence ellipse approach that considers both safety and efficacy indices. This method addresses limitations in existing BRA frameworks by providing a visual and quantitative tool that accounts for uncertainties and heterogeneities in clinical trial data. We introduce three novel indices: the Net Benefit Index (NBI), Relative Benefit Index (RBI), and Success Rate Index (SRI), offering a comprehensive assessment of treatment performance. Through extensive simulations, we compared the confidence ellipse method with established BRA techniques such as global benefit-risk (GBR) assessment (GBR), multi-criteria decision analysis (MCDA), and stochastic multi-criteria acceptability analysis (SMAA). Our results showed the confidence ellipse method's robustness, particularly with larger sample sizes, and demonstrated decreasing sensitivity as sample sizes increased, reflecting enhanced stability. To illustrate the method's practical application, we present a hypothetical Phase 2 clinical trial comparing a new drug combined with chemoradiotherapy against placebo for locally advanced squamous cell carcinoma of the head and neck. The confidence ellipse approach successfully provided a nuanced benefit-risk assessment, incorporating both unweighted and weighted analyses for adverse events. Overall, the confidence ellipse method is a robust, adaptable, and interpretable tool for benefit-risk assessment, offering valuable visual and quantitative insights for decision-making in modern clinical trials and regulatory processes.