University of Michigan biostatistics researcher turns complex health data into better treatment decisions
Honored with the 2026 Excellence in Research Award, Donglin Zeng develops tools that help clinicians identify more effective treatments for individual patients and guide public health decisions
When Donglin Zeng first studied mathematics as an undergraduate student in China, he was drawn to the way numbers could help explain real-world problems—especially how diseases spread through a population.
That early curiosity stayed with him through major public health events, including flu outbreaks in China and the SARS outbreak in 2002. Today, it continues to shape his work as a professor of Biostatistics at the University of Michigan School of Public Health.
Zeng, who joined Michigan Public Health in 2023 after more than 20 years on the faculty at the University of North Carolina at Chapel Hill, has become a leading researcher at the intersection of statistics, machine learning and precision medicine. His work focuses on developing statistical tools that help doctors, scientists and public health leaders make better decisions using complex health data. That includes identifying which treatments may work best for individual patients and improving how researchers understand disease risk, treatment effects and long-term health outcomes.
Zeng was honored recently with the School of Public Health’s 2026 Excellence in Research Award, which recognizes faculty whose research shows strong impact, innovation, productivity, creativity, leadership, interdisciplinarity and recognition by peers. His record reflects that broad reach: He has published more than 350 articles, been cited more than 17,000 times and holds an H-index of 65, a measure of sustained research impact.
Among his most influential contributions is a 2012 paper that helped bring machine learning into the field of individualized treatment decisions. The paper has been cited more than 1,100 times and helped spark major growth in research on precision medicine and dynamic treatment strategies. Zeng’s work also played a role during the COVID-19 pandemic, when his research on the long-term effectiveness of vaccines contributed to findings published in the New England Journal of Medicine, JAMA and The Lancet.
Colleagues say Zeng’s influence comes not only from the strength of his methods, but also from the way he uses them to solve urgent public health problems.
“Dr. Zeng is an outstanding biostatistical and public health data scientist who is generating transformative and highly collaborative research,” said Veera Baladandayuthapani, chair of the Department of Biostatistics. “This, coupled with his dedicated mentorship and extensive professional leadership, significantly advance public health research.”
In this Q&A, Zeng discusses his path from mathematics to public health, the promise of precision medicine, the role of biostatistics in global health crises and why collaboration and public trust are essential to solving today’s most complex health challenges.
Dr. Zeng is an outstanding biostatistical and public health data scientist who is generating transformative and highly collaborative research. This, coupled with his dedicated mentorship and extensive professional leadership, significantly advance public health research.”
— Veera Baladandayuthapani, chair of the Department of Biostatistics
Is there anything in your past that helped lead you to your interest in public health?
During my undergraduate studies in mathematics, I became fascinated by how simple mathematical models could explain patterns of disease transmission in a population. I was also curious about how public health decisions, such as lockdowns and other control measures, were made during epidemic outbreaks, including the H3N2 and influenza B outbreaks in China in 1992 and the SARS outbreak in 2002.
What parts of public health are the most interesting for you?
Mathematical and statistical models, when combined with real-world data, can help explain complex public health phenomena, such as how diseases spread, which populations are most at risk, and how interventions may change the course of an outbreak. These models provide a quantitative framework for understanding uncertainty, evaluating potential scenarios and guiding evidence-based decision-making in public health.
Has there been an obstacle or challenge that you’ve overcome to get where you are today?
Yes. Coming from a mathematics background that emphasized rigor, I initially found it challenging during my PhD program to understand how decisions could be made based on imperfect, random data from empirical studies. Early in my career, I also found it difficult to communicate my work effectively to public health scientists.
What drew you to biostatistics?
I was trained in mathematics and statistics, and I have found it both profound and exciting to apply these tools to biomedical problems, where the science is complex and the data are often challenging. Biostatistics provides the ideal field for bringing these interests together.
What is your main area of research, and what drew you to work in that area?
I have worked on developing statistical models and methods to address a wide range of biomedical and public health problems. For example, my research on semiparametric models focuses on estimating the scientific quantities of primary interest, such as intervention effects and risk scores, while minimizing assumptions about aspects of the data-generating process that are not central to the scientific question.
My work in precision medicine focuses on using complex data from diverse studies to support optimal decision-making, while placing less emphasis on fully modeling all relationships among data variables.
What did it mean to you to be recognized with the Excellence in Research Award this spring?
It was a great honor. I truly appreciate the school’s recognition of my work and its impact. Since I have been at the School of Public Health for only three years, this recognition has also helped introduce my work to a broader community within the school.
Your research sits at the intersection of statistics, machine learning and precision medicine. How do these fields come together in your work to help clinicians identify the most effective treatments for individual patients?
They are all closely integrated. Precision medicine is the goal and provides the clinical motivation. Statistical frameworks provide rigor, making the conclusions justifiable, reproducible and generalizable. Machine learning offers powerful algorithms and computational tools for identifying optimal solutions in large, complex datasets. To help clinicians select the most promising treatments for individual patients, each of these components is essential.
Your 2012 paper helped introduce machine learning techniques into individualized treatment decision-making and has since influenced a large body of follow-up research. Looking back, what made that work especially important, and how has the field evolved since then?
This was the first time we introduced machine learning tools to address the problem of individualized treatment decision-making. Before then, statisticians had primarily focused on traditional regression models, while computer scientists had centered their work on classification problems. Our work showed that these perspectives could be integrated.
Since then, the field has seen explosive development, particularly within statistics. Our paper has been cited more than 1,100 times. Extensions now include multicategory and continuous treatments, survival outcomes, dynamic treatment regimes, statistical policy search, and methods that incorporate deep learning and data integration.
Your research on the long-term effectiveness of COVID-19 vaccines helped inform findings published in journals such as the New England Journal of Medicine, JAMA and The Lancet. What did that work reveal about the role of biostatistics in shaping public health policy during a global health crisis?
Through this work, I found that traditional biostatistical methods can sometimes be extremely useful for addressing urgent public health problems, such as evaluating time-varying vaccine effects. The methods do not always need to be modern or trendy to be impactful. However, data are essential. In this case, we were fortunate to have access to vaccine surveillance data from the state of North Carolina, which made this work possible.
What is something about public health you wish everyone knew?
No one can solve public health problems alone. Meaningful progress requires collaboration across many fields, including epidemiology, biostatistics, public policy and community engagement.
As machine learning and AI tools become increasingly influential in public health, it is especially important to ensure that research is transparent, justifiable and reproducible, while also protecting data privacy. These principles are essential for earning and maintaining public trust.
— Written by Bob Cunningham
