Our CTML Seminar Series continues on February 12th with an exciting talk led by Alissa Gordon on "Bridging the Evidence Gap: Leveraging Historical Control Data to Address Underrepresentation in Health Research." Don't miss this talk taking place at 12:00PM at Berkeley Way West, 5th Floor, Room 5401.
Many subpopulations cannot fully benefit from advances in health research due to lack of representation in trials, partly due to low RCT trial feasibility. To bridge this gap, we propose a non-parametric sensitivity analysis that allows researchers to safely supplement trial data with relevant historical control data in hybrid control trials. This sensitivity analysis leverages omitted variable bias methodologies and estimates the maximum bias introduced under varying severities of violations. We find that with sufficient understanding of the covariate-outcome relationship, this method reliably bounds bias while leaving ample room for efficiency gains from using hybrid control trials.