Committee:

Iva Miljkovic (thesis advisor/dissertation director), EPISeyoung Kim, EPIJoe Zmuda, EPIAaron Barchowsky, EOHMegan M. Marron, EPIAllison Kuipers, Michigan State University

View on site | Email this event

Advisors and Committee Chairs

Lu TangVictor TalisaAbstract

Sepsis is a serious and potentially life-threatening condition caused by the body's response to an infection. It affects roughly two million individuals in the US each year. Despite decades of research, it remains a leading cause of mortality in critical care. One of the biggest challenges in sepsis management is heterogeneity. Patients can look very different from one another – in terms of their underlying conditions, disease progression, and response to treatment. Because of this variability, clinical decision-making is often difficult. This dissertation focuses on addressing challenges posed by patient heterogeneity in sepsis care using machine learning methods.

In the first project, we develop Federated Learning of Robust Individualized Decision Rules (FLoRI), a flexible machine learning-based individualized treatment framework that can be deployed in complex clinical settings. Through a novel objective function and federated learning, FLoRI provides safe treatment recommendations, improves generalizability across patient populations, and mitigates infrastructure limitations across healthcare systems simultaneously. The performance of FLoRI is demonstrated through an application to University of Pittsburgh Medical Center (UPMC) sepsis patients. FLoRI effectively improves the survival by 2-3 percentage points among sepsis patients and by 10 percentage points among sepsis patients with higher risk of death.

In the second project, we develop Tilted Individualized Decision Rules (TIDE), an individualized treatment framework that is robust to data contamination. Through tilted empirical risk minimization, TIDE down-weights corrupted observations and emphasizes the signal from the majority of the data, leading to more reliable treatment recommendations. TIDE is then extended to TIDE+ for multistage settings, allowing treatment decisions to be dynamically adjusted over time. The performance of TIDE is demonstrated using UPMC sepsis data. TIDE effectively mitigates the impact of contaminated observations and yields treatment recommendations that improve patient outcome compared to standard approaches.
In the third project, we develop Registered Trajectory Clustering for Sepsis Phenotyping (ReTraC), a clustering framework that identifies sepsis subphenotypes using electronic health records (EHRs). ReTraC jointly aligns and clusters patient trajectories through iterative updates. Both simulation studies and real-world applications to UPMC data demonstrate improved clustering accuracy and the identification of clinically interpretable subphenotypes.

Public health significance: This dissertation advances public health by addressing key challenges in sepsis care using real-world EHR data. The proposed methods improve the reliability of individualized treatment decisions and enable more precise identification of patient subgroups, ultimately supporting more effective and personalized care in critical care settings.

View on site | Email this event

Committee:

Caterina Rosano, EPI (committee chair)Zongqi Xia, Neurology (advisor)Sonja Swanson, EPIChung-Chou Chang, BiostatisticsKangho Suh, Pharmacy and TherapeuticsAbstract:

Multiple sclerosis (MS) is a chronic neuroinflammatory and neurodegenerative disease that leads to disability. Improving long-term disability outcomes is a major goal of MS care. In real-world clinical settings, disability trajectories vary across patients and are shaped by both risk profiles and treatment decisions. Important gaps remain in understanding how patient characteristics influence disability trajectories and how treatment strategies can be optimized across clinical contexts. The overall goal of this dissertation is to improve understanding of determinants of disability in MS and to generate evidence for optimizing disability outcomes using longitudinal and causal inference approaches applied to real-world data.

In Aim 1, I applied multi-state Markov models to EHR-linked MS registries to evaluate how comorbidity burden influences disability transitions and trajectories. I found higher psychiatric and cardiometabolic comorbidity burden was associated with greater transition intensity toward worse disability states, lower transition intensity toward improvement, higher 5-year probability of reaching severe disability, and fewer years spent in low disability states. These findings supported a more integrated approach to MS care in which improving long-term outcomes requires attention to mental and vascular health in addition to MS itself.

In Aim 2, using the same modeling framework, I assessed the association between treatment use and disability outcomes and whether these associations varied by age. I found higher-efficacy treatment use was associated with more favorable disability outcomes, with greater benefit observed at younger ages. These findings suggest that the comparative effectiveness of treatment on disability is age-dependent and support a more individualized treatment approach.

In Aim 3, I further investigated treatment initiation timing. Using observational data and a clone-censor-weight framework to emulate a target trial, I found that delayed initiation of high-efficacy therapy was associated with increased risk of disability progression or death. Age-stratified analyses showed a consistent pattern of worse outcomes with delayed initiation. These findings underscore the universal importance of timely initiation of high-efficacy DMTs after diagnosis.

Together, this dissertation provides new evidence on the determinants and optimization of disability outcomes in MS, and helps inform future research and clinical efforts toward more comprehensive management and more individualized treatment strategies.

View on site | Email this event

Dissertation Committee

Brittany Brown-Podgorski, PhD, MPH, Assistant Professor, Department of Health Policy and Management, School of Public Health (Committee Chair)

View on site | Email this event