Quantitative Benefit-Risk Assessment in Medical Product Decision Making: A Good Practices Report of an ISPOR Task Force.

Benefit-risk assessment is commonly conducted by drug and medical device developers and regulators, to evaluate and communicate issues around benefit-risk balance of medical products. Quantitative benefit-risk assessment (qBRA) is a set of techniques that incorporate explicit outcome weighting within a formal analysis to evaluate the benefit-risk balance. This report describes emerging good practices for the 5 main steps of developing qBRAs based on the multicriteria decision analysis process. First, research question formulation needs to identify the needs of decision makers and requirements for preference data and specify the role of external experts. Second, the formal analysis model should be developed by selecting benefit and safety endpoints while eliminating double counting and considering attribute value dependence. Third, preference elicitation method needs to be chosen, attributes framed appropriately within the elicitation instrument, and quality of the data should be evaluated. Fourth, analysis may need to normalize the preference weights, base-case and sensitivity analyses should be conducted, and the effect of preference heterogeneity analyzed. Finally, results should be communicated efficiently to decision makers and other stakeholders. In addition to detailed recommendations, we provide a checklist for reporting qBRAs developed through a Delphi process conducted with 34 experts. • This report presents current expert consensus in emerging good practice for developing methodologically rigorous and fit-for-purpose quantitative benefit-risk assessments, together with a checklist to support their reporting. • We provide detailed recommendations on each of the 5 main steps of quantitative benefit-risk assessments: formulating the research question, choosing appropriate analysis models, evaluating the attribute sets and diagnosing any potential violations of model assumptions, choosing the appropriate methods to elicit the required preference weights, and communicating the results effectively. • For preference elicitation, we provide an overview of the 3 methods that are appropriate for eliciting outcome weighting: discrete choice experiments, threshold technique, and swing weighting. [ABSTRACT FROM AUTHOR]