Research

1. Robust design and analysis of cluster randomized trials

Cluster randomized trials (CRTs), in which groups rather than individuals are randomized to intervention conditions, play a central role in evaluating healthcare, public health, and policy interventions delivered at the hospital, clinic, nursing home, provider, or community level. As modern health systems increasingly implement complex, multi-component, and implementation-focused interventions, CRTs have become indispensable tools for pragmatic and implementation research. However, these trials also introduce methodological challenges that differ fundamentally from those in individually randomized trials, including complex intracluster correlation, limited numbers of clusters, heterogeneity across sites, and practical constraints in trial implementation.

My research is to develop rigorous and practical statistical methods that address these challenges, improving the design, analysis, and interpretation of CRTs so that they can generate credible, interpretable, and decision-relevant evidence for clinical and public health practice.

Small sample inference in cluster randomized trials under misspecified correlation structures:

Ouyang Y, Taljaard M, Forbes AB, Li F. Maintaining the validity of inference from linear mixed models in stepped-wedge cluster randomized trials under misspecified random-effects structures. Stat Methods Med Res. 2024 Sep;33(9):1497-1516.
Ouyang Y, Kulkarni MA, Protopopoff N, Li F, Taljaard M. Accounting for complex intracluster correlations in longitudinal cluster randomized trials: a case study in malaria vector control. BMC Med Res Methodol. 2023 Mar 17;23(1):64.

Estimating intracluster correlation coefficients (ICCs):

Ouyang Y, Hemming K, Li F, Taljaard M. Estimating intra-cluster correlation coefficients for planning longitudinal cluster randomized trials: a tutorial. Int J Epidemiol. 2023 Oct 5;52(5):1634-1647.
Ouyang Y, Li F, Li X, Bynum J, Mor V, Taljaard M. Estimates of intra-cluster correlation coefficients from 2018 USA Medicare data to inform the design of cluster randomized trials in Alzheimer's and related dementias. Trials. 2024 Oct 30;25(1):732.

2. Bayesian methods in clinical trials

Bayesian methods provide a principled framework for designing and analyzing clinical trials by combining prior information with accumulating trial data. This approach is especially valuable in modern clinical research, where evidence may come from historical studies, expert knowledge, real-world data, or interim trial results. My research focuses on Bayesian clinical trial methodology, including adaptive designs, borrowing of external information, and decision-making under uncertainty. By developing statistically rigorous and clinically interpretable methods, I aim to improve trial efficiency, support more ethical use of patient data, and help translate emerging evidence into reliable medical decisions.

Bayesian methods in cluster randomized trials:

Ouyang Y, Luo B, Dixon SN, Al-Jaishi AA, Devereaux PJ, Walsh M, Wald R, Zwarenstein M, Anderson S, Garg AX. Bayesian Analysis of Time-To-Event Data in a Cluster-Randomized Trial: Major Outcomes With Personalized Dialysate TEMPerature (MyTEMP) Trial. Can J Kidney Health Dis. 2025 Jun 28;12:20543581251341710.
Zhan D, Ouyang Y, Xu L, Wong H. Improving efficiency in the stepped-wedge trial design via Bayesian modeling with an informative prior for the time effects. Clin Trials. 2021 Jun;18(3):295-302.

Bayesian adaptive trial designs:

Robinson CH, Parekh RS, Cuthbertson B, Fan E, Ouyang Y, Heath A. Using Bayesian pre-trial simulations to optimize the design of adaptive clinical trials in childhood nephrotic syndrome. Contemp Clin Trials. 2025 Jun;153:107918.
Zhan D, Ouyang Y, Vila-Rodriguez F, Karim ME, Wong H. Bayesian adaptive enrichment design in multi-arm clinical trials: The BayesAET package for R users. Comput Methods Programs Biomed. 2025 Aug;268:108833.