Date of Award

2024

Document Type

Thesis

Degree Name

Master of Science (MS)

Department

Pathology

First Advisor

Sarah Nowak

Second Advisor

Jill Warrington

Abstract

Why doesn’t a patient show up for their appointment? Is it too far? Are there too many appointments? Or something else? Urine drug testing clinics often observe patient scheduled visit absenteeism, and this can be used as a data source to answer our questions and explore other potential correlations between factors. With a well-developed electronic health data system, a retrospective study was performed on a large data set collected between January 2019 and December 2021 across the U.S. and more than half a million patient encounters; it contained nearly a year of quarantine, and pandemic status was also analyzed as one variable. Patient demographic information (including gender, race, and age), self-reported drug use, participating programs, scheduling method, assigned risk of relapse, and collection site are included in the study. Factor analysis of mixed data was done to give a clean and easy interpretation of the data set. A multiple linear regression formula was established for single case prediction; 9 simple backpropagation neural network structures were also tested. Forty percent of patients had at least one unexplained absence on the record; however, the frequency of missing an appointment increased significantly (p < 0.001) by geographic distance from a clinic for >20 miles (long distance) and between 0-2 miles (walking distance). Automated scheduling methods also demonstrated higher rates of missed appointments than manually scheduled visits. Patient groups who filled their race as unknown provided a significant (p

Language

en

Number of Pages

32 p.

Download

Included in

Pathology Commons

Share

COinS