Thesis Title: Health Analytics for Decision Making in Healthcare Spatial Access

Advisor: Dr. Nicoleta Serban, School of Industrial and Systems Engineering

Committee members:

Dr. Turgay Ayer, School of Industrial and Systems Engineering

Dr. Pinar Keskinocak, School of Industrial and Systems Engineering

Dr. Lindsey Bullinger, School of Public Policy

Dr. Monica Gentili, School of Engineering, University of Louisville

Date and Time: 10 am, Friday Jan 8th, 2021

Meeting URL: https://bluejeans.com/354616660

Meeting ID: 354 616 660 (bluejeans)

Abstract:

Appropriate access to healthcare services is important for preventing the spread of disease, reducing hospitalizations and emergency department use, and increasing quality of life. However, within the United States healthcare system, there exist many disparities in access to care. In this dissertation, we aim to quantify and assess disparities in healthcare access, for informed decision making towards improving access. Compared to existing methods, our approach allows for local-level estimates, is data-rich, and is statistically rigorous.

In Chapter 2 of this dissertation, we focus on access to pediatric primary care services in seven states. We design an optimization model to match primary care need with supply while taking into consideration system constraints such as health insurance acceptance and maximum travel distance. Output of this model enables computation of census tract-level average distance children must travel to reach their primary care providers and average congestion children face to schedule visits with their providers. We perform statistical inference, both between and within states, to determine whether there are significant disparities in travel distance and congestion.

In Chapter 3, we focus on primary care for non-elderly adults in Georgia and how it may be impacted by the Affordable Care Act (ACA). Specifically, we project the supply and need to primary care services starting from year 2013 through year 2025 under two scenarios: ACA implementation without Medicaid expansion and ACA implementation with expansion. Similar to Chapter 2, we use an optimization model to obtain census tract-level estimates of availability and accessibility and test whether they significantly change by year 2025 due to ACA implementation or additional Medicaid expansion. We additionally evaluate the impact of two other policies intended to improve access: increasing the number of residency positions in Georgia and implementing a parity program so that more providers accept Medicaid insurance.

In Chapter 4, we begin analysis of psychosocial services for Medicaid-insured children. Using Medicaid claims data for 34 states, we identify which providers are likely to treat Medicaid-insured children and their practice settings. We estimate per-provider and per-state psychosocial service caseloads and compare across states, urbanicity/rurality, and provider specialties.

Finally, in Chapter 5, we develop a modeling framework for one potential intervention to increase access to psychosocial services: collaboration between mental health providers and primary care providers. Our framework mimics providers’ making individual decisions on who they partner based on their unique preferences. We create this framework by extending congestion games into a setting in which players have their own private cost function for each resource and resources have their own capacities and preferences over the players. We construct a polynomial-time algorithm to find a Nash equilibrium for singleton games with non-decreasing cost functions under this setting and demonstrate our model for services to Medicaid-insured children in New York.