Research in Biostatistics and Data Science

Project-specific training opportunities for biostatisticians and data scientists are often not available outside of certain advanced graduate (or occasionally undergraduate) degree programs. Lacking these opportunities, the effectiveness of the biomedical workforce suffers due to shortages in qualified biostatisticians and data scientists who can design efficient and robust studies that can be analyzed to produce reliable and generalizable results and thus applied to the most important problems in health research. For this project, we focus on pairing undergraduates with biostatistics and translational mentors in areas of interest to the NHLBI and NIAID to work on specific research projects. The Summer Institute in Biostatistics and Data Science at the University of Texas Medical Branch provides trainees with these opportunities for direct, hands-on training. The objective of this project is to develop this training opportunity to expose students to statistical design and analysis problems from real research, while providing them with the conceptual and technical tools to address these problems. As such, students should be able 1) to assess study designs for strengths and weaknesses in addressing specific biomedical questions, 2) to use the tools of statistical modeling, data science, and hypothesis testing along with state-of-the-art statistical software to work on problems as well as produce results and conclusions for these problems, and 3) to understand how developing these skills can lead to a wide variety of career opportunities both inside and outside of academia.

UTMB Summer Institute in Biostatistics and Data Science

DMAC has been funded by Cancer Prevention & Research Institute of Texas Since 2021. The Core builds data infrastructure provides training and pilot awards for research along the entire cancer continuum.

See the DMAC Website

NIH Artificial Intelligence/Machine Learning Consortium to Advance Health Equity and Researcher Diversity (AIM-AHEAD)

MPIs: Suresh K. Bhavnani (University of Texas Medical Branch), Rodney Hunter (Texas Southern University)

Numerous studies have shown that patients with complex conditions such as cancer tend to have multiple and overlapping combinations of characteristics ranging from comorbidities to social determinants of health (SDoH). Such heterogeneity makes patients similar and different in complex ways, requiring powerful AI/ML modeling methods to discover significant patterns, enable humans to examine them for disparities in data and methods, and interpret them for clinical translation. This project develops a novel AI/ML pipeline using a Human-Centered Artificial Intelligence (HCAI) approach to identify complex co-occurrences of SDoH forming subtypes in cancer, and which enables expert panels to examine the models and interpret the subtypes.

Trauma Research and Combat Casualty care collaborative (TRC4)

MPIs: Celeste Finnerty, Suresh K. Bhavnani (University of Texas Medical Branch)

While Texas has the largest oil and gas extraction industry in the US, it unfortunately also leads to the number of work-related fatalities due to traumatic injuries. Because burn injuries are one of the most severe forms of trauma with high rates for morbidity and mortality, there is an urgent need to analyze the combination of factors that increase the risk of adverse outcomes among burn victims. However, such research currently faces two major hurdles: (1) the lack of a Texas-wide registry of burn victims critical for enabling researchers to analyze clinical data from across the University of Texas system that serve many patients from the oil and gas industry, and (2) the lack of an understanding of the complex co-occurring factors that increase the risk for morbidity and mortality among subgroups of burn patients.

This project directly addresses the above hurdles by: (1) building a University of Texas Burns (UT-Burns) registry with critical variables identified by stakeholders for analyzing long-term burn trajectories. This registry will initially use data from the University of Texas Medical Branch (UTMB) electronic health record (EHR), but which will be scalable in the future to include data from across the UT system; and (2) analyzing the data in the registry through Human-Centered AI (HCAI) methods for identifying subgroups of burn patients based on the co-occurrence of prior comorbidities. The HCAI analytical method will enable an understanding of how a combination of factors result in high-risk burn subtypes enabling a more targeted design of interventions. The HCAI approach will also enable a stakeholder panel to inspect the data and algorithms for bias against underrepresented minorities, and the clinical interpretability of the results.

MPI: Yong-Fang Kuo, Mukaila A. Raji

This R01 related to opioid prescription in older adults has been funded by National Institute of Drug Abuse since 2016. The current studies focus on effectiveness, toxicity and safety of opioid and benzodiazepine substitutes. More than 50 papers were published.

Read more about the project

See all publications

2) Annual Wellness Visit Policy: Impact on Disparities in Early Dementia Diagnosis and Quality of Healthcare for Medicare Beneficiaries with Alzheimer’s Disease and Its Related Dementia

MPI: Huey-Ming Tzeng, Yong-Fang Kuo, Mukaila A. Raji

The R01 related to annual wellness visit policy impact on disparity in early dementia diagnosis and quality of health care has been funded by the National Institute on Aging since 2023. This is convergent parallel mixed-methods design research. A recent paper analyzing Medicare data was published in JAMA Network Open with media attention. The recruitment of caregivers for qualitative study was called.

Read the article

Read the article

Read the article