Abstract
In subgroup analysis of clinical trials and precision medicine, it is important to assess the causal effect of a new treatment against an existing one and classify the new treatment favorable subgroup if it exists. As the original randomization does not apply to comparisons between subgroups, for unbiased estimate the causal inference method will be used, in particular the doubly robust procedure, in which a propensity score model and a regression model need to be specified. As long as one of the models is correctly specified, the causal effect will be estimated unbiased. However, it is known that any subjectively specified model more or less deviates from the true one, and so the doubly robust procedure may still not be robust. To overcome this issue, we apply a recently proposed method to allow the identification of subgroups and causal inference in subgroups. The model is a semiparametric robust and flexible procedure, in which both the propensity score model and the regression model are semiparametric, with monotone constraint on the nonparametric parts. Simulation studies are conducted to evaluate the performance of the proposed method and compare some existing methods. Then the method is applied to analyze a real clinical trial data.
Acknowledgments
There is no conflict of interest among the authors.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Additional information
Notes on contributors
Ao Yuan
Ao Yuan is professor of Biostatistics. His research interests include clinical trial, semiparametric methods and causal inference, and Bayesian inference.
Anqi Yin
Anqi Yin got her Ph.D in Biostatistics in 2022. Her research interests include statistical methods in biomedical studies, semiparametric methods and causal inference.
Ming T. Tan
Ming T. Tan is professor of Biostatistics. His research interests include the design and analysis of clinical trials, population science and drug combination studies.