Assessing Surrogacy in Clinical Trials Using Counterfactual Models

Yun Li
PhD Candidate, Department of Biostatistics
University of Michigan
*Candidate for the Biostatistics/ Cancer Center Faculty Position

Monday, February 4th
3:30pm
Mayo D199
Minneapolis Campus

Abstract:

A surrogate marker (S) is a variable that can be measured earlier and often easier than the true endpoint (T) in a clinical trial. It can be very useful if it can reliably facilitate early prediction of the effect of the treatment (Z) on T. Most previous research has been devoted to developing surrogacy measures to quantify how well S can replace T or examining the use of S in predicting the treatment effect. However, the research often requires one to fit models for the distribution of T given S and Z. It is well known that such models do not have causal interpretations because the models condition on a post-randomization variable S. In this paper, we directly model the relationship among T, S and Z in a causal inference framework, specifically using a potential outcomes framework introduced by Frangakis and Rubin (2002) for surrogate markers. We propose a Bayesian estimation method to evaluate the causal probabilities associated with the cross-classification of the potential outcomes of S and T when S and T are both binary. We use a log-linear model to model the odds ratios of the potential outcomes. The quantities derived from this approach always have causal interpretations. This causal model is not identifiable from data without additional assumptions. To reduce the non-identifiability problem and increase precision for the statistical inferences, we incorporate assumptions that are plausible in the surrogate context by using prior distributions. We also explore the relationship among the surrogacy measures based on traditional models and this counterfactual model. We use the causal probabilities to predict the treatment effect when T is partially observed. Then we extend the method to the multiple trial setting using hierarchical modeling. The methods are applied to data from the collaborative initial glaucoma treatment study.

A social tea will be held at 3:00 P.M. in A434 Mayo. All are Welcome.
For more details contact 612-624-4655 or see http://www.biostat.umn.edu/seminar_academic.html