Login / Signup

Method for Calculating the Simultaneous Maximum Acceptable Risk Threshold (SMART) from Discrete-Choice Experiment Benefit-Risk Studies.

Angelyn Otteson FairchildShelby D ReedJuan Marcos Gonzalez
Published in: Medical decision making : an international journal of the Society for Medical Decision Making (2022)
Conventional approaches to calculate maximum-acceptable risk (MAR) using discrete-choice experiment data account for 1 adverse-event risk at a time, requiring that decision makers infer the acceptability of treatments when patients are exposed to multiple risks simultaneously.The Simultaneous Maximum Acceptable Risk Threshold (SMART) maps combinations of adverse-event risks that would be jointly acceptable given a specific treatment benefit and provides a transparent and precise portrayal of acceptance of multiple risks.Risk levels that would be accepted using individual MAR estimates might not be acceptable when simultaneous risks are considered, especially when marginal expected disutility of risk is decreasing nonlinearly with risk probabilities.Preference researchers should calculate SMARTs in any discrete-choice study in which 2 or more adverse-event risks are presented, particularly if risk preferences are nonlinear.
Keyphrases
  • machine learning
  • big data
  • electronic health record