Greg Midgette
Greg Midgette is a senior policy researcher at RAND and an associate professor of criminology and criminal justice at the University of Maryland. He specializes in program evaluation, policy analysis, and empirical research methods. His work focuses on alcohol and drugs, illicit markets, community corrections, the impact of public safety policies and practices on public health, and the influence of policy design on disparities.
Session
Introduction. Excessive drinking is a leading source of preventable death in the US. To reduce alcohol consumption among justice-involved individuals, an increasing number of jurisdictions are implementing 24/7 Sobriety programs (hereinafter 24/7). 24/7 requires participants to undergo frequent alcohol testing, predominantly by twice daily breathalyzer or a transdermal alcohol monitoring bracelet. Under the program, those testing positive are subject to an immediate but short stay in jail (typically 24 to 48 hours). This analysis measures the association between implementation of 24/7 and mortality rates in two early-adopting jurisdictions, North Dakota and South Dakota.
Study design: Heterogeneous count difference-in-differences (DID) analysis based on staggered adoption of 24/7 at the county level variation in the timing of 24/7 implementation. County-level mortality rates are generated from the CDC’s Multiple Cause of Death data.
Setting: 119 counties in North Dakota and South Dakota covering every month from 2000-2019 (N=28,560)
Intervention. Implementation of 24/7 in a county.
Main outcome measures. County-month deaths rates for any death as well as those diagnosed as cancer, circulatory, digestive, or external injury.
Results. The association with overall deaths was statistically significant across both states, but the effects are concentrated in South Dakota. The cause-specific analyses found that 24/7 implementation was associated with a ten percent decrease in external injury deaths across both states (p=.029). Sex specific subgroup analyses revealed these findings are driven by declining mortality in men in the pooled sample, but were large amongn women in South Dakota.
Conclusions. Implementation of 24/7 in a county is negatively associated with county-level rates of deaths among men in both states. While our analyses do not allow us to assess the mechanisms driving this association, the findings are consistent with a reduction in alcohol misuse. Individual-level analyses are needed in North Dakota to rule out the ecological fallacy.