Causal Inference

Causal Inference 

Causal inference is a rapidly growing interdisciplinary subfield of statistics, computer science, econometrics, epidemiology, psychology, and social sciences. Association obtained from traditional statistical analysis such as regression cannot be interpreted as causality without further assumption. Causal inference focuses on exploring the rigorous assumptions, study designs, and estimation strategies that allow researchers to draw causal conclusions based on a clinical trial or observational data. Directed acyclic graph (DAG) and potential outcome framework have been developed to rigorously represent the causal relationship between variables. One area of causal inference is to learn the causal network structure from observational studies with or without additional experiments. Another area of causal inference is to estimate either the causal effect of a specific exposure on a specific outcome or the causal effect of a specific pathway under a hypothesized DAG framework with potential unmeasured confounding variables. The third area of causal inference is to develop sensitivity analysis of untestable identification assumptions and to propose new study designs that allow the better evaluation of causal effects. Causal inference has wide application in various scientific area including learning gene regulatory network, estimating drug or vaccine efficacy when there is non-compliance and interference, evaluating biomarker surrogacy, learning optimal dynamic treatment regime, understanding whether the treatment effects are through certain pathway. Our department faculty (Drs Dai, R. and Zheng) have developed some innovative methods in this area:

  1. Zheng, C., and Liu, L. (2021). Quantifying direct and indirect effect for longitudinal mediator and survival outcome using joint modeling approach. Biometrics. doi: 10.1111/biom.13475
  2. Liu, L., Zheng, C., and Kang, J. (2018). Exploring causality mechanism in the joint analysis of longitudinal and survival data. Statistics in Medicine. 37: 3733-3744.
  3. Zheng, C. and Zhou, X.H. (2017). Causal mediation analysis on failure time outcome without sequential ignorability. Lifetime Data Analysis. 23: 533-559.
  4. Zheng, C., Dai, R., Hari, P.N. and Zhang, M.J. (2017). Instrumental variables with competing risk model. Statistics in Medicine. 36: 1240-1255.
  5. Zheng, C. and Zhou, X.H. (2015). Causal mediation analysis in multilevel intervention and multicomponent mediator case. Journal of the Royal Statistical Society: Series B, 77: 581-615.
  6. Yin, J., Zhou, Y., Wang, C., He, P., Zheng, C. and Geng, Z. (2008). Partial orientation and local structural learning of causal networks for prediction. Proceedings of Machine Learning Research, 3: 93-105.