DOI
https://doi.org/10.25772/559H-Q509
Author ORCID Identifier
https://orcid.org/0000-0002-1581-108X
Defense Date
2020
Document Type
Dissertation
Degree Name
Doctor of Philosophy
Department
Health Related Sciences
First Advisor
Dr. Christian Wernz
Second Advisor
Dr. Lauretta Cathers
Third Advisor
Dr. Ernie Steidle
Fourth Advisor
Dr. Loretta Schlachta-Fairchild
Fifth Advisor
Dr. Angela Duncan
Sixth Advisor
Dr. Sarah Marrs
Abstract
The purpose of the research was to develop a predictive model of the effect of patient characteristics on outcomes and resource utilization in telehealth interventions for chronic disease management. Clinical studies have shown high variability in telehealth across settings and populations. The central hypothesis of this study was that patient characteristics, including disease type, illness severity, demographic information, and socio-technical preferences play a role in determining telehealth intervention outcomes. A non-experimental, retrospective study evaluated the feasibility of creating a Telehealth Needs score using Medicare and Medicaid claims data from 2016 to 2017. Claims were summarized into de-identified cases to train a series of outcome prediction models using parametric and non-parametric Bayesian analytical machine learning methods. Telehealth and non-telehealth cohorts were matched by illness severity and stratified by disease type, as well as the number of comorbidities to control for group differences. The telehealth cohort demonstrated statistically significant differences with higher mean total utilization (representing outpatient and prescription use) and lower mean inpatient utilization as compared to the non-telehealth cohort. This effect demonstrated a shifting of clinical resources from inpatient to outpatient services for the telehealth intervention. The creation of a Telehealth Needs score using machine learning for outcome prediction is feasible. Adoption of the Telehealth Needs score may improve patient outcome prediction by reducing variability in digital health services delivery across populations and settings in the Commonwealth of Virginia.
Rights
© The Author
Is Part Of
VCU University Archives
Is Part Of
VCU Theses and Dissertations
Date of Submission
2-18-2020