Understanding extreme mortality among prisoners: a national cohort study in Scotland using data linkage

Lesley Graham, Colin M Fischbacher, Diane Stockton, Andrew Fraser, Michael Fleming, Kevin Greig

Research output: Contribution to journalArticlepeer-review

Abstract

BACKGROUND: Mortality is known to be extremely high among people who have been imprisoned, but there is limited information about the factors that explain this increased risk.

METHODS: Standard record linkage methods were used to link Scottish prison records and mortality data for all individuals imprisoned in Scotland for the first time between 1 January 1996 and 31 December 2007.

RESULTS: Among 76 627 individuals there were 4414 deaths (3982 in men). When compared with the general population, the age-standardized mortality rate ratio for those imprisoned was 3.3 (95% CI: 3.2, 3.4) for men and 7.6 (6.9, 8.3) for women. Further adjustment for an area measure of deprivation accounted for part but not all of this excess risk [adjusted rate ratio 2.3 (2.2, 2.4) and 5.7 (5.1, 6.2) for men and women, respectively]. Relative risks were highest for drug and alcohol related causes, suicide and homicide and were markedly higher among women than men. Out of prison deaths were most frequent in the first 2 weeks after release from prison. Mortality rates were lower in those with longer total duration in prison and higher in those with multiple short episodes in prison.

CONCLUSION: People who have been imprisoned in Scotland experience substantial excess mortality from a range of causes that is only partly explained by deprivation. The association of increased mortality with multiple periods in prison and the concentration of deaths in the early period after prison release both have implications for policy and practice.

Original languageEnglish
JournalEuropean Journal of Public Health
DOIs
Publication statusPublished - 11 Feb 2015

Fingerprint Dive into the research topics of 'Understanding extreme mortality among prisoners: a national cohort study in Scotland using data linkage'. Together they form a unique fingerprint.

Cite this