Baiming Zou
University of North Carolina at Chapel Hill, NC·Mathematics & Statistics
Doctors and health systems collect huge amounts of patient data, but most of it comes from everyday care or observational studies rather than carefully controlled trials. In this kind of data many factors are tangled together, and patients respond differently to the same treatment.
- Fields
Mathematics & Statistics
Public Health & Epidemiology
Computer Science
- Topics
machine learning
deep learning
causal inference
- Grants
- 2 known grants, through Nov 2028
What they work on
- Develop machine learning methods that adjust for complex confounding in observational data, to reveal causal links between clinical features and heart, lung, blood and sleep disorder outcomes in the Hispanic Community Health Study/Study of Latinos.
- Build deep learning hidden-subgroup analysis frameworks that find groups of patients who respond differently to treatment, so health management plans can be tailored to individual needs.
- Establish a flexible, scalable feature importance test to identify biomarkers, measurable signs of disease, that improve prediction of health and disease outcomes.
- Create deep learning models that predict acute and chronic postoperative pain from electronic health records and the TEMporal PostOperative Pain Signatures cohort, to support timely pain control and reduce opioid overuse.
- Develop deep learning-based semi-parametric methods for irregularly measured, non-normal pain data, to identify causal mechanisms of postoperative pain and find optimal individualized pain management.
More about the research
Ordinary statistical tools can then give biased or contradictory answers. This research builds machine learning tools that can pull reliable cause-and-effect conclusions and personalized recommendations out of such messy data. It applies them to two health problems: heart, lung, blood and sleep disorders, and pain after surgery.
Why it matters
Patients with heart, lung, blood and sleep disorders, and surgical patients at risk of poorly controlled pain, could benefit from more reliable conclusions drawn from real-world health data. The tools could help clinicians predict who will develop acute or chronic postoperative pain, reduce opioid overuse, identify biomarkers that improve outcome prediction, and choose management strategies tailored to individual patients rather than the average patient.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Enhanced Machine Learning Tools for Complex Data Evaluation and Integration in Advancing Health Outcomes
NIH · Dec 2024 – Nov 2028
Robust Computational and Data Analytic Tools for In-depth Understanding Postoperative Pain Mechanism with Enhanced Pain Management and Clinical Decision Making
NIH · Sep 2024 – Aug 2028
Closest labs
Pablo Martinez Amezcua
Johns Hopkins University, MD·Public Health & Epidemiology
Examine whether the association between hearing loss and cognitive decline over about six years in Hispanic adults differs by language acculturation, using HCHS/SOL data.
Peng Liu
Iowa State University, IA·Mathematics & Statistics
Develop functional regression models for imaging outcomes built on a distributed learning framework, so that estimation and inference stay accurate yet scalable for large imaging datasets.
Caleb Miles
Columbia University, NY·Mathematics & Statistics
Develop simple parametric and semiparametric efficient estimators of the average treatment effect in multi-study, multi-outcome data fusion, using related follow-up measurements as proxies for outcomes missing in some trials.
The school
University of North Carolina at Chapel Hill: 72 labs on PI Pages, $1.6B a year on research, 523 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.