YSPH biostatistician developing advanced statistical methods for complex clinical trials

Randomized clinical trials are considered the gold standard for researchers seeking to identify effective new treatments for illness and disease. But these trials require rigorous and highly nuanced statistical analysis that can be challenging in the face of large, complex data sets and multiple endpoints. Done properly, such trials can lead to life-saving advances in science and medicine. Done poorly, they can lead to years of misguided research and costly delays.

Dr. Fan Li, PhD, an Associate Professor in the Department of Biostatistics at the Yale School of Public Health (YSPH) and a leading expert in causal inference and clinical trial methodology, was recently awarded a $2.6 million R01 grant from the National Institutes of Health to develop innovative solutions to help address these challenges. Dr. Li is also a faculty member in the Center for Methods in Implementation Science, and Yale Center for Analytical Sciences. He holds a secondary appointment in the Section of Cardiovascular Medicine, Department of Internal Medicine, at the Yale School of Medicine.

Over the next four years, Dr. Li and his team will develop new causal inference methods to strengthen the design and analysis of cluster-randomized trials (CRTs). For these types of trials, treatments are randomized across institutions, such as hospitals or clinics, rather than a pool of individuals. So, a new intervention involving 20 hospitals might see 10 receive the intervention while the other 10 serve as controls. Every qualifying patient within an intervention hospital would receive the new treatment, and their outcomes would then be compared to patients in the 10 control hospitals. These studies tend to be complex and require non-conventional methods for causal inference.

CRTs that evaluate multiple clinical outcomes and complex composite endpoints can be challenging to conduct. These situations arise frequently in cardiovascular clinical studies, which often span multiple institutions and involve multiple outcomes such as stroke, heart attack, and death, along with patient-reported quality-of-life measures. However, current analytic strategies often fail to capture the full clinical picture or provide clear guidance for statistically efficient analysis.

Dr. Li’s team plans to create new statistical theory, tools, and guidance that will help researchers better measure treatment benefits across different health outcomes in CRTs. The new approach will properly take into account the hierarchical data structure arising from cluster randomization and multiple outcomes. By mid-2029, the team plans to release free, regularly updated software that will help clinical researchers reach clearer, more patient-centered conclusions about how treatments affect people with complex health conditions.

“My work aims to develop statistical tools that sharpen the scientific question, improve the interpretability of treatment effects, and ultimately provide estimates that are clinically informative and directly useful to patients, clinicians, and decision makers,” Dr. Li said. “There are thousands of CRTs registered in the public domain. But statistical methods, especially causal inference methods, for these complex trials are significantly lagging. When researchers design a CRT with this level of complexity, they’re often constrained by a lack of statistical methods and tools.”

Without suitable analytical tools, researchers frequently adapt methods intended for individually randomized trials, an approach that can distort results and lead to inaccurate conclusions, Dr. Li said. “This challenge becomes even more pronounced when trials involve multiple outcomes or complex composite endpoints, where naive adaptations can obscure true effects and further compromise scientific validity,” he said.

The work is being conducted in collaboration with Yale’s Cardiovascular Medicine Analytics Center (CMAC), directed by Dr. Guangyu Tong, which provides advanced analytical capabilities for cardiovascular research, and Yale’s Clinical and Translational Research Accelerator (CTRA), directed by Dr. F. Perry Wilson, which unites physician-scientists committed to rigorous clinical investigation and innovative trial designs. Dr. Li and his Yale team also work closely with collaborators at Mississippi State University, the University of Washington, and the University of Pennsylvania. Together, this interdisciplinary team aims to develop novel methods that enable investigators to analyze multiple clinically meaningful endpoints simultaneously under complex cluster-randomized settings, incorporating clinician input to ensure that the treatment effect estimates are most relevant to patients.

The project’s broader goal is to help researchers produce clearer, more reliable evidence for public health decisions. “At the core of my research agenda is a commitment to producing clear, transparent, and methodologically rigorous evidence that strengthens clinical and public health decision making,” Dr. Li said.