Purpose: Comparative effectiveness research (CER) aims to assist patient and population-level decision making in order to improve health care. A gap exists in terms of using a standardized approach to quantitatively weigh interventions with respect to their collective clinical harms and benefits. We present a CER framework for decision makers to integrate clinical evidence by assigning numerical weights to intervention-specific clinical harms, benefits, and uncertainty.
Methods: Building on decision-analytic modeling, we propose a two step approach to CER that explicitly compares interventions in terms of clinical harms and benefits evidence and links this evidence to the quality-adjusted life year (QALY). The first step is a traditional evidence synthesis of intervention-specific harms and benefits. Conflict and subjective judgment may arise in determining which intervention is optimal if one intervention exhibits superior outcomes but the alternative intervention exhibits other superior outcomes. The second step is pursued when clinical equipoise exists. The second step is the development of a decision-analytic model to simulate the population and the progression of disease over an appropriate time horizon. The output of the two steps is the ability to compare and quantitatively link clinical harms and benefits with QALYs. We craft a hypothetical asthma example (intervention A1 vs. A2).
Results: Hypothetical intervention A1 yields better asthma control and trends toward lower severe exacerbation rates with no known difference in mortality. A2 trends toward lower mild exacerbation rates and has higher lung function scores as well as lower dyspnea. All clinical harms and benefits are linked to QALYs based on a validated asthma decision-analytic model. A1 is associated with 0.25 additional average lifetime QALYs compared to A2 with 95% interval of (-0.05, 0.55).
Conclusions: Using decision-analytic models and QALYs in the CER decision-making process gives an explicit, structured, and consistent quantitative approach to weighing all relevant harms and benefits. The use of decision-analytic models in CER may not be the current standard due to the misconception that decision-analytic methods must include cost, because clinical harms and benefits are not commonly displayed as outputs or related to QALYs in cost-effectiveness models, and/or due to fears related to the use of QALYs as an outcome measure. Future research should study effective communication of these added dimensions for payer, research, and clinical stakeholders.
See more of: The 33rd Annual Meeting of the Society for Medical Decision Making