Confounding

Cost-effectiveness analyses (CEA) commonly use observational studies, either alongside or instead of data from randomized controlled trials (RCTs). A major concern is that the results suffer from treatment selection bias due to confounding variables that influence both treatment and outcome. Commonly used analytical methods for dealing with selection bias such as regression analysis or propensity score matching can be highly sensitive to model specification. Moreover, these approaches assume no unobserved confounding.

Our team has vast experience on improving methods in CEA for dealing with selection bias and is currently tackling various issues in this area:

  • Assessing the relative performance of a new matching approach (GenMatch) compared to conventional propensity score matching for reducing selection bias in CEA.
  • Evaluating alternative approaches for addressing selection bias when estimating treatment effects in CEA under misspecification.
  • Investigating approaches for evaluation the effectiveness and cost-effectiveness of continuous treatments.
  • Comparing approaches for addressing time-varying confounding when informing sequential decisions
  • Evaluating the relative merits of synthetic control methods to estimate treatment effects in longitudinal settings.

KEY PEOPLE: Richard Grieve, Noemi Kreif, Stephen O’Neill

Comments are closed.