Frequent Alcohol Testing Combined with Swift, Certain and Modest Sanctions Is Associated with Drop in Mortality
Feb 9, 2016
Published in: The Lancet Psychiatry, 2016
Posted on RAND.org on February 18, 2016
BACKGROUND: In the UK and USA, various jurisdictions have launched new approaches for managing alcohol-involved offenders that might have public health implications. These programmes require participants to abstain from alcohol and submit to frequent alcohol testing with swift, certain, and modest sanctions for violations, with the aim to reduce crime and keep alcohol-involved offenders in the community. In this study we examine whether the 24/7 Sobriety programme in South Dakota, USA--the largest such programme to date--is associated with reductions in mortality. METHODS: With a differences-in-differences design, we used variation in the timing of 24/7 Sobriety implementation across South Dakota counties between 2005 and 2011 to estimate the association between programme introduction and county-level mortality. We used monthly, county-level, aggregate counts for mortality from January, 2000, to June, 2011. We assessed total deaths, and deaths due to external injuries, circulatory disorders, digestive disorders, and cancer (as a potential placebo). FINDINGS: Between January, 2005, and June, 2011, 16 932 people (about 3% of the adult population) participated in the 24/7 Sobriety programme. The analysis was based on a sample size of 9 108 county-month observations (ie, 66 counties ×?12 months ×?11·5 years). Implementation of 24/7 Sobriety was associated with a 4·2% (95% CI 1·5–6·9) reduction in all-cause adult mortality, with the largest associations among women (8·0%, 95% CI 3·9–11·8) and individuals older than 40 years (4·3%, 95% CI 1·4–7·0). Associations were most evident among circulatory disorders. INTERPRETATION: 24/7 Sobriety might have public health benefits, which could extend beyond individuals directly enrolled in the programme. However, further research, including randomised controlled trials and analyses of individual-level data, is needed to corroborate the finding, reassess the size of these associations, and gain insight into causal mechanisms. Should a negative association be replicated, it might represent a substantial advance in our understanding of how criminal justice interventions could help shape public health.