An ecological study of factors associated with rates of self-inflicted death in prisons in England and Wales.

This study describes ecological associations between self-inflicted death rates and prison-level environmental indicators over the period 2000-2002 in England and Wales. The objective was to assist in the development of interventions for reducing the incidence of self-inflicted deaths in prisons in England and Wales, by identifying risk factors, including overcrowding, positive drug tests, the number of assaults, purposeful activity, offending behaviour programmes, and cost per prisoner. Poisson regression was used to estimate associations between self-inflicted death rates and these potential risk factors, controlling for different categories of prison. The annual rate of self-inflicted death during 2000-2002 was 1.14 per 1000 (95% CI 0.98 to 1.34), with no evidence for a difference in the two years. Highest rates were in the Male Local and Women's prisons at 1.86 (95% CI 1.42 to 2.26) and 2.27 (95% CI 1.35 to 3.84) per 1000 respectively. In a multivariate analysis, overcrowding, assault rate and purposeful activity were significant. In an analysis controlling also for prison category, only purposeful activity remained independently significant, as a protective factor (RR 0.57, 95% CI 0.35 to 0.92, p=0.02) with weaker evidence for a positive association with positive drug tests (RR 1.41, 95% CI 0.96 to 2.05, p=0.08). Despite concerns about the quality of routinely collected data and the interpretation of ecological associations, this study suggests that a higher level of purposeful activity is independently associated with lower rates of self-inflicted death, whatever the prison category. This adds support to other studies conducted at the level of the individual prisoner.