Statistical Center for Translational Research in Medicine (SC-TRM)
Advancing translational medicine through innovative biostatistics, real-world data, and clinical research expertise.
SC-TRM Key Ongoing Research Projects
Develop statistical methods and software tools for improving screening/diagnosis for under-screened/under-diagnosed diseases
Summary: Under-diagnosis occurs when an individual living with a medical condition has not been diagnosed. Overlooking of diagnosis prevents patients from opportunities of appropriate treatments and prevention strategies, and often is more serious in some population subgroups. Recent emphasis on electronic health records (EHRs) provides a unique opportunity for addressing under-diagnosis, which to date remains largely unexploited due to lack of suitable analytical methods. The overarching goal of this ongoing project is to develop data-driven methods to accurately identify patients from EHRs who possibly miss their diagnosis and provide insights for reasons of under-diagnosis.
Principal Investigator:
Jinbo Chen
Grant Number:
5R01LM014401
| Selected Publications |
|---|
| Zhang L, Ding X, Ma Y, Muthu N, Ajmal I, Moore JH, Herman DS, Chen J. A maximum likelihood approach to electronic health record phenotyping using positive and unlabeled patients. JAMIA 2020, 27:119-126. |
| Zhang L, Ma Y, Herman D, Chen J. Testing calibration of phenotyping models using positive-only electronic health record data. Biostatistics 2022, 23:844-859. |
| Lee SH, Ma Y, Wei Y, Chen J. Optimal sampling for positive only electronic health record data. Biometrics 2023, 79(4):2974-86. |
| Ma W, Cohen JB, Chen J. A semiparametric method for addressing underdiagnosis using electronic health record data. Biometrics 81, 2025. |
| Sheth S, Andersen L, et al. (2026). Prospective genetic validation of a machine learning algorithm trained to flag patients with undiagnosed familial hypercholesterolemia. (Under review). |
| Peng J, Ma Y, Chen J. Classification based electronic health records with supplemented labels. AOAS (under revision) |
Develop innovative statistical methods for evaluating and understanding algorithmic fairness in healthcare AI, with the goal of advancing ethical and equitable use of AI in clinical practice.
Summary: This emerging research direction addresses a critical challenge in healthcare AI: ensuring that increasingly powerful predictive algorithms are not only accurate, but also fair, ethically responsible, and consistent with the principles of equitable clinical care.
| Selected Publications |
|---|
| Parikh RB, Zhang Y, Kolla L, Chivers C, Courtright KR, Zhu J, Navathe AS, Chen J. Performance drift in a mortality prediction algorithm among patients with cancer during the SARS-CoV-2 pandemic. JAMIA 2023, 30:348-54. |
| Kolla L, Linn K, Navathe AS, Kreisler C, Roberts CB, Park SH, Singh H, Feng J, Chen J, Parikh RB. Performance Drift in a Nationally Deployed Population Health Risk Algorithm in the US Veterans Health Administration. JAMA Health Forum 2025, 6(8):e252717. |
| Hegarty S, Linn KA, Zhang H, Teeple S, Albert P, Parikh RB, Courtright K, Kent DM, Chen J. Assessing algorithm fairness requires adjustment for risk distribution differences: re-considering the equal opportunity criterion. Biostatistics (in press). |
| Zheng XY, Hegarty S, Zhang J, Chen J. Fairness evaluation for risk prediction models under heterogeneous risk distributions. Biometrics (under review) |
| Xiong H, Park SH, Wang R, Hegarty S, Navathe AS, Linn KA, Chen J. Equivalence testing for algorithmic fairness: risk-standardized equal-opportunity assessment in clinical risk prediction (Under review) |
Develop an accurate model for predicting CVD risk among men with prostate cancer
Summary: Cardiovascular disease (CVD) is the leading cause of death in men with prostate cancer (PCa), the most common male cancer in the US. The overall objective of this newly funded proposal is to develop robust, accurate and precise CVD risk prediction tools that account for the competing risk of PCa mortality and incorporate cancer, CVD-specific, and individual patient characteristics, including the SDOH.
Multiple Principal Investigator:
– Jinbo Chen
– Bonnie Ky
Grant Number:
R01HL182015-01
Develop innovative statistical methods for risk model development and validation that address complex study designs and real-world data challenges
Summary: In particular, we have developed methods to evaluate the added predictive value of information that is costly to obtain, while addressing two important challenges: (1) efficient risk model development and validation under two-phase study designs and (2) integration of external data sources that may be systematically biased or non-representative of the target population.
| Selected Publications |
|---|
| Zheng J, Zheng Y, Hsu L. Risk Projection for Time-to-event Outcome Leveraging Summary Statistics With Source Individual-level Data. J Am Stat Assoc. 2022;117(540):2043-2055. |
| Parikh RB, Hasler JS, Zhang Y, Liu M, Chivers C, Ferrell W, Gabriel PE, Lerman C, Bekelman JE, Chen J. Development of Machine Learning Algorithms Incorporating Electronic Health Record Data, Patient-Reported Outcomes, or Both to Predict Mortality for Outpatients With Cancer. JCO Clin Cancer Inform 2022, 6:e2200073. |
| Zheng J, Zheng Y, Hsu L. Re-calibrating pure risk integrating individual data from two-phase studies with external summary statistics. Biometrics. 2022 Dec;78(4):1515-1529. |
| Cao Y, Haneuse S, Zheng Y, Chen J. Two-phase stratified sampling and analysis for predicting binary outcomes. Biostatistics 2023, 24:585-602. |
| Zheng J, Dong X, Newton CC, Hsu L. A Generalized Integration Approach to Association Analysis with Multi-category Outcome: An Application to a Tumor Sequencing Study of Colorectal Cancer and Smoking. J Am Stat Assoc. 2023;118(541):29-42. |
| Cao Y, Ma W, Zhao G, McCarthy AM, Chen J. A constrained maximum likelihood approach to developing well-calibrated models for predicting binary outcomes. Lifetime Data Anal. 2024, 30:624-48. |
| Hasler J, Ma Y, Wei Y, Parikh R, Chen J. A semiparametric method for risk prediction using integrated electronic health record data. Ann Appl Stat. 2024, 18:3318-37. |
| Dai G, Shao L, aChen J. Moving beyond population variable importance: concept, theory and applications of individual variable importance. Journal of the Royal Statistical Society Series B-Statistical Methodology 2025, 87:816–32. |
| Dai G, Carroll RJ, Chen J. Valid and efficient inference for nonparametric variable importance in two-phase studies. Biometrics 2025, 81(3):ujaf095. |
| Zheng J, Steinfelder R, Yin H, Qu C, Thomas M, Thomas S, Andrews C, Augusto B, Corley D, Lee J, Berndt S, Chan A, Chanock S, Gignoux C, Goldberg S, Haiman C, Huyghe J, Iwasaki M, Le Marchand L, Lee S, Melendez J, Mesa I, Ogino S, Sifontes V, Um C, Visvanathan K, White L, Williams A, Willis W, Wolk A, Yamaji T, Vadaparampil S, Jarvik G, Burnett-Hartman A, Milne R, Platz E, Figueiredo J, Zheng W, MacInnis R, Palmer J, Schmit S, Landorp-Vogelaar I, Peters U, Hsu L. MyGeneRisk Colon: A Web-Based Tool for Personalized Colorectal Cancer Risk Prediction Based on Genetics and Lifestyle. 2026 January 01; :2026.04.03.26349669. |
| Shao L, Dai G, Chen J. A general framework for heterogeneous variable importance: Pointwise and uniform inference. Biometrics 2026, 82(1):ujag015. |
Advance statistical methodology for innovative analyses of EHR data
Summary: We have developed novel statistical methods to address case contamination in EHR-based case-control studies and innovative study designs for label validation that facilitate more efficient and reliable model validation.
Principal Investigator:
Jinbo Chen
| Selected Publications |
|---|
| Wang L, Schnall J, Small A, Hubbard RA, Moore JH, Damrauer SM, Chen J. Case contamination in electronic health records based case-control studies. Biometrics. 2021 Mar;77(1):67-77. |
| Schneider C, Li T, Zhang D, et al, Chen J, Rader D. Large-scale identification of undiagnosed hepatic steatosis using natural language processing. Eclinicalmedicine 2023. |
| Dai G, Ma Y, Hasler J, Chen J, Carroll RJ. A robust approach for electronic health record-based case-control studies with contaminated case pools. Biometrics. 2023 Sep;79(3):2023-2035. |
| Lee J, Kolla L, Chen J. Towards optimal model evaluation: enhancing active testing with actively improved estimators. Sci Rep 2024, 14:10690. |
| Lee J, Dai G, Rader DG, Chen J. Model calibration and evaluation via optimal subsampling using electronic health record data. JRSS-A 2026, qnag036. |
SC-TRM has been Empowering Researchers Throughout the K-Award Lifecycle
SC-TRM is dedicated to cultivating the next generation of clinical and translational research leaders. It has been actively supporting K-award applications and continued to provide rigorous methodological support throughout investigators’ award periods.
Notably, a range of clinical investigators have successfully leveraged this support to transition from mentored K-awardees to independent scientific leaders, establishing and leading externally funded research programs that advance clinical and translational science.
Some Ongoing K-Award Collaborations:
| K-Award |
|---|
| K23 |
| Neuro-EM K12 |
| K23 award application (Fundable score) |
| K01 application (ongoing) |
| K23 application (ongoing) |
| Success application of multiple summer research awards. |