Graduation Year

2024

Document Type

Dissertation

Degree

Ph.D.

Degree Name

Doctor of Philosophy (Ph.D.)

Degree Granting Department

Information Systems and Decision Sciences

Major Professor

Giti Javidi, Ph.D.

Committee Member

Kimberly Johnson, Ph.D.

Committee Member

Nasir Ghani, Ph.D.

Committee Member

Sriram Chellappan, Ph.D.

Committee Member

Doug Rohrer, Ph.D.

Keywords

Big Data, Patient Data, Maternal Outcomes, Substance Use Disorder, Fairness-informed models

Abstract

The following dissertation is an accumulation of applied data analytics research in healthcare. In the first chapter, the focus is on detecting ineffective self-management in diabetic patients. Through the application of Natural Language Processing (NLP), the project investigates links between the diabetic patients' discharge notes and their compliance to exercise, dietary restrictions and medications. A bi-product of this research is the development of an online dashboard that allocates scores to six compliance categories. This tools allows healthcare provider to assess patient progress and compare it to the population averages. The dashboard also offers a patient summary generated using the Chat GPT 3.5 turbo engine. The second chapter is dedicated to counterfactual fairness analysis in forecasting a proxy for quality of prenatal care using the PRAMS dataset. This study sheds lights on the racial disparities in unbalanced datasets that could be exacerbated through classification models and offers a penalty function as a solution. Finally, the third chapter presents a fairness-informed isomorphic approach to disparities in patient carepaths in the treatment process of SUD patients. This study presents a novel fairness-informed VF2 model that traverses treatment carepaths and yields the disparate carepaths for each race. This approach will provide detailed information regarding the exact points of disparities such as geographical location (City and County) as well as age distribution, referral source codes and procedures those patients with disparities experience. This study presents an in-depth discussion on barriers in access to outpatient treatment for SUD patients.

Download

Share

COinS