DOI
https://doi.org/10.25772/HE3K-H056
Author ORCID Identifier
https://orcid.org/0000-0001-7248-4890
Defense Date
2023
Document Type
Dissertation
Degree Name
Doctor of Philosophy
Department
Pharmaceutical Sciences
First Advisor
Teresa M. Salgado
Abstract
Background: By 2034, the US faces a potential primary care physician shortage, exacerbated by an aging population and population growth. High rates of avoidable hospitalizations for chronic conditions persist, prompting a shift towards value-based care. Despite the documented enhancement of chronic disease management through pharmacist involvement, resource constraints call for targeted services. Despite the rise of machine learning in health care, models have yet to predict patients at risk of not meeting quality measures.
Methods: This study assessed the effect of a multidisciplinary primary care model on quality measure achievement and health care utilization. Quality measure achievement was analyzed using a generalized linear mixed model with random effects, while health care utilization was examined through both a mixed model and negative binomial regression. Machine learning algorithms (Random Forest and XGBoost) built predictive risk models to identify adults at risk of not meeting quality measures, with analyses conducted using SAS and R statistical software.
Results: Patients receiving care from a pharmacist significantly improved HbA1c control, becoming ~4 to 5 times more likely to achieve this control. However, factors like late-year care initiation, higher baseline HbA1c values, and more chronic medications decreased this likelihood. Healthcare utilization patterns showed similar rates of hospital admission and emergency visits between groups, suggesting no significant impact of pharmacist involvement. The machine learning models contributed to identifying key factors, notably baseline HbA1c values, study group, and late-year care initiation. XGBoost showed superior prediction capability, affirming the importance of pharmacist intervention in HbA1c control.
Conclusion: Including pharmacists in primary care clinics significantly improves diabetes-related quality measure achievements. Machine learning models, notably XGBoost, affirmed these results, emphasizing the importance of baseline HbA1c and time of the patient’s index date in HbA1c control. These findings highlight the need for an integrated approach to diabetes management, emphasizing the critical role of pharmacists.
Rights
© The Author
Is Part Of
VCU University Archives
Is Part Of
VCU Theses and Dissertations
Date of Submission
8-10-2023