PhD candidate in Biostatistics, Jincheng Zhou, will present:
PhD Adviser: Haitao Chu
Abstract: Noncompliance occurs in a randomized clinical trial (RCT) when some participants do not follow their assigned treatments. It can introduce bias to the estimation of a treatment effect. The complier average causal effect (CACE) approach provides a useful tool for addressing noncompliance, where CACE measures the effect of an intervention in the latent class of population that complies with its assigned treatment. Meta-analysis of RCTs has become a widely-used statistical technique to combine and contrast results from multiple independent studies. However, there are no existing methods which can effectively deal with heterogeneous noncompliance in a meta-analysis of RCTs.
To address this gap, Bayesian hierarchical random effects models are developed to appropriately account for the inherent heterogeneity in noncompliance between studies and treatment groups. I will first present a Bayesian hierarchical model to estimate CACE in meta-analysis with binary or ordinal outcomes where heterogeneous compliance rates are available for each study. The approach is then extended to deal with incomplete noncompliance when some RCTs do not report noncompliance data. The results are illustrated by a re-analysis of a meta-analysis estimating the effect of epidural analgesia in labor on the outcome cesarean section, where noncompliance vary substantially between studies. Simulations are performed to evaluate the performance of the proposed approach and to illustrate the importance of including appropriate random effects. In addition, a user-friendly R package “BayesCACE is developed to implement the CACE analysis for binary outcomes based on the proposed Bayesian hierarchical models.
Refreshments will be served prior to the presentation.