The United States Food and Drug Administration and European Medicines Agency are evaluating quantitative methods for assessing the benefits and risks of pharmaceuticals to facilitate regulatory decision-making. Health outcomes modeling, the modeling of a drug's net benefit by quantifying its effects on health outcomes, constitutes one method under consideration by regulators. Little has been published on the feasibility of health outcomes modeling for regulatory use. Rosiglitazone is a type 2 diabetes drug with a controversial benefit-risk tradeoff, making it an intuitive example for examining health outcomes modeling as a regulatory aid. We sought to estimate the incremental net benefit of rosiglitazone monotherapy versus comparators, to evaluate the effect of emergent evidence on a product's net benefit, and to compare model results with past regulatory decisions concerning rosiglitazone. Our model results questioned the initial decision to approve rosiglitazone for monotherapy use since our estimates projected it to be inferior to glyburide. Only postmarketing efficacy data supported such an indication versus glyburide, yet postmarketing data did not support such an indication versus metformin due to sufficient uncertainty in the risk of drug-attributable cardiovascular death. With growing interest in multiple approaches for quantitative benefit-risk assessment, we also sought to compare modeling with other methods. We described three quantitative methods for improving regulatory benefit-risk assessment: health outcomes modeling, stated-preferences for benefit-risk tradeoffs, and multi-criteria decision analysis. We surveyed technical experts to describe current problems in benefit-risk assessment and to identify attributes needed to evaluate whether quantitative methods could address these problems. While each method has been tested using historical examples, they have yet to be evaluated using common measurement criteria. We developed conceptual and operational criteria by which we could prospectively evaluate the potential success of these three methods alongside the existing approach for assessing drug benefits and risks for regulatory decision-making. Since each method was found to possess both strengths and weaknesses for use in regulatory decision-making, feasibility of implementation will dictate which method might be adopted.