Suicide in prisoners: a systematic review of risk factors.

OBJECTIVE: To examine factors associated with suicide in prisoners. DATA SOURCES: Studies were identified through electronic searches of MEDLINE (1950-February 2007), PsycINFO (1806-February 2007), EMBASE (1974-February 2007), and CINAHL (1982-February 2007) without language restriction using the search terms prison, jail, felon, detainee, penal, and custody combined with suicide. STUDY SELECTION: Included studies were investigations that reported on prisoners dying by suicide who were compared with prisoners in control groups (which were randomly selected or matched, or consisted of the total or average prison population). Subgroup analysis and meta-regression were used to explore sources of heterogeneity. DATA SYNTHESIS: Thirty-four studies (comprising 4780 cases of prison suicide) were identified for inclusion in the review, of which 12 were based in the United States. Demographic factors associated with suicide included white race/ethnicity (OR = 1.9, 95% CI = 1.7 to 2.2), being male (OR = 1.9, 95% CI = 1.4 to 2.5), and being married (OR = 1.5, 95% CI = 1.3 to 1.7). Criminological factors included occupation of a single cell (OR = 9.1, 95% CI = 6.1 to 13.5), detainee/remand status (OR = 4.1, 95% CI = 3.5 to 4.8), and serving a life sentence (OR = 3.9, 95% CI = 1.1 to 13.3). Clinical factors were recent suicidal ideation (OR = 15.2, 95% CI = 8.5 to 27.2), history of attempted suicide (OR = 8.4, 95% CI =6.2 to 11.4), having a current psychiatric diagnosis (OR = 5.9, 95% CI = 2.3 to 15.4), receiving psychotropic medication (OR = 4.2, 95% CI = 2.9 to 6.0), and having a history of alcohol use problems (OR = 3.0, 95% CI = 1.9 to 4.6). Black race/ethnicity was inversely associated with suicide (OR = 0.4, 95% CI = 0.3 to 0.4). Few differences were found in risk estimates when compared by study design or publication type. CONCLUSIONS: Several demographic, criminological, and clinical factors were found to be associated with suicide in prisoners, the most important being occupation of a single cell, recent suicidal ideation, a history of attempted suicide, and having a psychiatric diagnosis or history of alcohol use problems. As some of these associations included potentially modifiable environmental and clinical factors, there is scope for targeting these factors in suicide prevention strategies for individuals in custody