Differential and Combined Effects of Physical Activity Profiles and Prohealth Behaviors on Diabetes Prevalence among Blacks and Whites in the US Population: A Novel Bayesian Belief Network Machine Learning Analysis

摘要

The current study assessed the prevalence of diabetes across four different physical activity lifestyles and infer through machine learning which combinations of physical activity, sleep, stress, and body mass index yield the lowest prevalence of diabetes in Blacks and Whites. Data were extracted from the National Health Interview Survey (NHIS) dataset from 2004–2013 containing demographics, chronic diseases, and sleep duration (?=?288,888). Of the total sample, 9.34% reported diabetes (where the prevalence of diabetes was 12.92% in Blacks/African Americans and 8.68% in Whites). Over half of the sample reported sedentary lifestyles (Blacks were more sedentary than Whites), approximately 20% reported moderately active lifestyles (Whites more than Blacks), approximately 15% reported active lifestyles (Whites more than Blacks), and approximately 6% reported very active lifestyles (Whites more than Blacks). Across four different physical activity lifestyles, Blacks consistently had a higher diabetes prevalence compared to their White counterparts. Physical activity combined with healthy sleep, low stress, and average body weight reduced the prevalence of diabetes, especially in Blacks. Our study highlights the need to provide alternative and personalized behavioral/lifestyle recommendations to generic national physical activity recommendations, specifically among Blacks, to reduce diabetes and narrow diabetes disparities between Blacks and Whites.