Relationship between county-level crime and diabetes: Mediating effect of physical inactivity

This paper assessed the extent to which physical inactivity accounts for the relationship between the crime rate and prevalence of type 2 diabetes in the United States. Using 2018 US county-level data, we compared unadjusted and adjusted prevalence of type 2 diabetes between high and low crime counties for 2,966 US counties. Average causal mediating effects of residents’ reported physical inactivity were estimated for each comparison. Counties with a higher crime rate were more likely to have higher percentages of people with type 2 diabetes than counties with a lower crime rate, even after adjusting for potential confounding factors such as racial distribution, income level, food insecurity, and neighborhood walkability (adjusted coefficient for top 40% vs. bottom 40% of crime rate distribution = 0.36; p < 0.001). Over 60% of the adjusted relationship between county-level rate of crime and type 2 diabetes was found to be mediated by physical inactivity. This study reinforces potentially overlooked public health benefits of effective anti-crime measures via improved physical activity.