The relationship between internet adoption and self-reported diabetes prevalence in United States counties
No Thumbnail Available
Authors
Meeting name
Sponsors
Date
Journal Title
Format
Thesis
Subject
Abstract
The Internet connects many people with type 2 diabetes or at-risk for diabetes to needed healthcare services, yet some underserved and vulnerable populations in the United States cannot access them due to disparities in Internet availability, adoption, and digital literacy. To better understand this phenomenon, social factors such as education, socioeconomic position, geographic location, and Internet adoption must be assessed to determine their relationship with diabetes. The aim of this study is to examine the relationship between Internet adoption and diabetes prevalence in United States counties, while controlling for other social determinants of health. Due to the complex, networked relationship of social determinants of health and health outcomes, the secondary aim is to measure the predictive power and reduction in error of non-linear machine learning models compared with linear regression. This study utilized a cross-sectional, ecological methodology using nationwide county-level data obtained from the 2021 American Communities Survey and the 2021 Behavioral Risk Factors Surveillance System. Data were linked by state and county name. Analysis was completed with descriptive statistics, two-stage ordinary least squares linear regression, and a gradient boost algorithm in Python notebooks. Linear regression and gradient boost model were compared to find optimal reduction in model error and explanation of variance.
Table of Contents
DOI
PubMed ID
Degree
Ph. D.