Incorporating Prior Data in Quantitative Benefit-Risk Assessments: Case Study of a Bayesian Method

Area of Science:

• Decision Analysis
• Biostatistics
• Pharmacoeconomics

Background:

• Current multiple criteria decision analysis (MCDA) and stochastic multi-criteria acceptability analysis (SMAA) lack methods to integrate prior or external benefit-risk data.
• Quantitative benefit-risk assessment (qBRA) for medical products requires robust methods for incorporating diverse data sources.

Purpose of the Study:

• To demonstrate a novel Bayesian mixture model approach for incorporating prior data into MCDA and SMAA.
• To enhance quantitative benefit-risk assessments (qBRA) for medical products by utilizing external information.

Main Methods:

• Implemented MCDA and SMAA within a Bayesian framework.
• Utilized mixture priors on benefit and risk attributes, blending prior study information with vague priors.
• Varied the degree of data borrowing using a mixing proportion parameter.

Main Results:

• A case study using Rivaroxaban for peripheral artery disease (PAD) demonstrated the method's utility.
• Incorporating 30% of prior data favorably shifted MCDA/SMAA results for Rivaroxaban.
• The approach addressed discrepancies in trial data, such as all-cause mortality.

Conclusions:

• The developed Bayesian method for incorporating prior data into MCDA and SMAA is user-friendly and interpretable.
• Available RShiny App facilitates the application of this Bayesian approach for enhanced qBRA.