Mingzhao Hu, Ph.D.

Location

Contact

SUMMARY

Mingzhao Hu, Ph.D., is a statistician who develops and applies statistical methods to study cognitive aging and Alzheimer's disease. His work integrates longitudinal cohort data with blood and imaging biomarkers to quantify how cognition and disease risk evolve over time. Dr. Hu develops multistate progression models that translate multimodal biomarker profiles into interpretable estimates of absolute risks of dementia and works on a cascade of biomarkers for disease severity. He also supports research in other chronic diseases — such as liver disease — by adapting artificial intelligence (AI) tools to enhance solutions for difficult clinical problems.

Focus areas

• Longitudinal biomarker trajectories and disease staging. Dr. Hu studies how blood, imaging and cognitive biomarkers shift with aging and throughout the course of disease. Using flexible longitudinal tools — such as mixed‑effects models, changepoint models and subgroup patterns — this work pinpoints when disease signals first emerge and when their progression accelerates, strengthening how biomarkers are interpreted in clinical research.
• Risk prediction and multistate disease progression. Dr. Hu develops hidden Markov models to estimate absolute risk of dementia across age and biomarker profiles. These models infer a person's unobserved disease stage by linking measurable biomarkers to underlying latent states, allowing transitions between states to be estimated over time. By capturing how individuals move from normal aging to preclinical and symptomatic stages, the models generate individualized risk curves that can be accessed through an online calculator for clinical applications.
• Biomarker cutpoints and staging frameworks. Dr. Hu evaluates candidate thresholds and data-driven staging rules, emphasizing interpretability, transportability across cohorts and alignment with clinically meaningful outcomes.
• Statistical methods for complex longitudinal data. Dr. Hu advances methodology for joint models, integrating state‑space and related frameworks to support inference when processes evolve over time and are measured with statistical noise.
• Synthetic data for study emulation and methods evaluation. Dr. Hu builds and evaluates privacy-preserving synthetic cohorts to develop harmonization pipelines, compare screening strategies and enhance data quality under realistic data-generating mechanisms.

Significance to patient care

Alzheimer's disease develops slowly over many years, and people want to understand what their biomarker results mean for their future. Dr. Hu's research helps turn blood tests and brain scans into clearer estimates of a person's short‑ and long‑term risk of developing memory problems. These estimates can guide decisions about who may benefit from closer follow‑up, prevention strategies or participation in research.

His methods also strengthen clinical studies by making the results more reliable, especially in situations where patients miss follow‑up visits or their diagnosis change over time.

Professional highlights

• Guest editor, Journal of Data Science, 2025-present.
• Statistics editor, Journal of Cardiovascular Magnetic Resonance, 2024-present.
• American Statistical Association:
  • Program chair, Symposium on Data Science and Statistics, 2026.
  • Program chair, Joint Statistical Meetings, 2025.
• Program chair, Eastern North American Region of the International Biometric Society Spring Meeting, International Biometric Society, 2026.