A0538
Title: A capture-recapture approach to facilitate causal inference for a trial-eligible observational cohort
Authors: Lin Ge - Indiana University Bloomington (United States) [presenting]
Abstract: Recently proposed design-based capture-recapture (CRC) methods are extended for prevalence estimation among registry participants in order to enable causal inference among a trial-eligible target population. The design for CRC analysis integrates an observational study cohort with a randomized trial involving a small representative study sample and enhances the generalizability and transportability of the findings. It is shown that a novel CRC-type estimator derived via multinomial distribution-based maximum likelihood exploits the design to deliver benefits in terms of validity and efficiency for comparing the effects of two treatments on a binary outcome. The design also unlocks a direct standardization-type estimator that allows efficient estimation of general means (e.g., for continuous outcomes such as biomarker levels) under a specific treatment. This provides an avenue to compare treatment effects within the target population in a more comprehensive manner. Simulations demonstrate the validity and efficiency of the proposed estimators under the CRC design. Finally, an illustrative data application is presented, comparing anti-s antibody seropositive response rates for two major Covid-19 vaccines using an observational cohort from Tunisia.
