Economic evaluation has become an increasingly important component of NIH-funded intervention research, including studies supported by NIDA and the NIH HEAL Initiative, as investigators and relevant stakeholders seek to understand not only whether interventions work, but whether their benefits justify the resources required for implementation and scale-up.
In clinical trials, these evaluations combine patient-level cost and effectiveness outcomes derived from standardized economic, healthcare utilization, health-related quality-of-life, and other study measures. The joint analysis of these outcomes presents important statistical challenges, particularly in substance use disorder trials where analysts commonly encounter highly skewed and heavy-tailed healthcare costs, small or moderate sample sizes, incomplete follow-up, and complex patterns of missing data. Moreover, cost and effectiveness outcomes associated with a health intervention may exhibit nonlinear dependence relationships that cannot be adequately analyzed by conventional correlation statistics. Under these conditions, commonly used statistical methods can perform poorly, and estimates of incremental cost-effectiveness ratios can be unstable.