Research

My research involves the creation of industrial engineering and operations research (IE/OR) methodologies to answer decision-making problems arising in public health and medicine.

I currently have three main streams of research: medical decision-making, healthcare access, infectious disease prevention and control. My work is motivated by addressing fundamental problems in these areas arising from challenges in maternal health, poliovirus, and chronic disease management with previous work on COVID-19. From a methodological perspective, I leverage data-driven optimization, Markov decision processes, agent-based modeling, simulation, and predictive models.

 

More details about each of these areas are provided below.

Maternal Health

Medical Decision Making

Pregnant woman in black holds her exposed stomach expectantly. Original public domain image from Wikimedia Commons

In this work, we are creating machine learning and decision-analytic models to help clinical decision-making throughout the perinatal period. I am working with Dr. Garcia, collaborators within Emory University’s Department of Obstetrics and Gynecology, and a group of PhD students 

Check out this related paper:

Support for this work came from an Imagine, Innovate, and Impact (I3) Emory School of Medicine/Georgia Tech internal grant through the Georgia CTSA NIH award (UL1-TR002378) and the National Science Foundation through a Graduate Research Fellowship awarded to Ph.D. student, Meghan Meredith.

Health systems improvement

The maternal mortality rate in the United States is higher than that of other industrialized nations and is quickly growing (up 89% from 2018) with staggering racial/ethnic and urban/rural disparities.

The reasons for the maternal mortality crisis are complex and often are due to factors that span across multiple levels of influence, including patient-, provider-, community-, and health system-level factors. I am working with my Ph.D. student to analyze systems-level interventions, such as perinatal regionalization strategies.

Here is a related paper:

Recent work in this area is being supported by supported by a seed grant through the National Center for Advancing Translational Sciences of the National Institutes of Health under Award Number UL1TR002378. 

Chronic Disease Management

Medical Decision-Making

 The management of chronic diseases often involves “sequential decision-making under uncertainty.” For example, a patient visits their doctor roughly every year for a check-up. The doctor has to decide how to best treat the patient considering the benefits of treatment (e.g., reducing the risk of a heart attack) and the potential costs and harms of treatment (e.g., side effects) under uncertainty about how the patient’s disease will progress. Markov Decisions Processes (MDPs) are a mathematical framework that can be used to optimize these type of sequential decisions. My work in this area has focused on designing treatment policies that are more robust to parameter ambiguity and that are interpretable to stakeholders.
Markov decision process model of blood pressure and cholesterol control
Markov decision process model of blood pressure and cholesterol control
 
Treatment recommendations from multiple MDP models
Treatment recommendations from multiple MDP models

Some of my papers on chronic disease management and/or MDPs include:

This work was supported by the National Science Foundation under grant numbers DGE-1256260 (Steimle’s NSF GRFP) and CMMI-1462060 (Denton).

Healthcare Access

With my PhD student Abel Sapirstein, I am using similar mathematical models to study access to tertiary cancer care in Rwanda. Rwandans experience significant delays in the diagnosis of cancer, in part because of challenges in accessing cancer care.

This work is supported by an NDSEG Fellowship awarded to my PhD student Abel Sapirstein.

Here is a related paper:

Infectious Disease Outbreak Prevention & Control

Mathematical modeling and analytics for responding to COVID-19

Covid 19 Corona Coronavirus - Free image on Pixabay

Much of my work in this area was aimed at helping institutions of higher education identify strategies for students to safely return to campus in the face of COVID-19. These decisions involve complex trade-offs and have to be made despite uncertainty about the future. The main goals of this project were to (1) create analytical tools for campus planners to make informed return-to-campus decisions and (2) generate general insights around the impact of these operational decisions on risk to students, faculty, and staff.

Check out the latest tools available on our team’s Github page.

Some papers on COVID-19 include:

This work was supported by a seed grant from the ISYE Thos and Clair Muller Endowment Fund and the Georgia Tech EVPR COVID-19 Rapid Response Seed Grant Program.

Informing Global Polio Eradication Efforts Through Innovative Modeling Approaches

“Perhaps the biggest threat to [polio eradication] now is an explosion of vaccine-derived polio outbreaks in Africa that affected almost two dozen countries last year and paralyzed more than 500 children in 2020 and again in 2021”.

— From Science Magazine, 2022

While the Global Polio Eradication Initiative has successfully eradicated two of the three serotypes of poliovirus, a major challenge in achieving complete eradication is outbreaks of “circulating vaccine-derived poliovirus” (cVDPV). cVDPV outbreaks are caused when the oral poliovirus vaccine used to control outbreaks reverts to a form that has similar properties to wild poliovirus. I am working with Dr. Keskinocak and a great team of students to create mathematical models and analytical techniques that can be used to inform poliovirus outbreak prevention and response.

Check out this related paper:

Support for this work comes from a cooperative agreement between Georgia Tech and the CDC (GH22-2272 U2R GH001919).