BACKGROUND: Psychiatric hospital in-patients are known to be at high risk of suicide, yet there is little reliable knowledge of risk factors or their predictive power. AIMS: To identify risk factors for suicide in psychiatric hospital in-patients and to evaluate their predictive power in detecting people at risk of suicide. METHOD: Using a case-control design, 112 people who committed suicide while in-patients in psychiatric hospitals were compared with 112 randomly selected controls. Univariate analysis and multivariate analyses were used to estimate odds ratios and adjusted likelihood ratios. RESULTS: The rate of suicide in psychiatric in-patients was 13.7 (95% CI 11.7-16.1) per 10,000 admissions. There were five predictive factors with likelihood ratios > 2, following adjustment: planned suicide attempt, 4.1; actual suicide attempt, 4.9; recent bereavement, 4.0; presence of delusions, 2.3; chronic mental illness, 2.2; and family history of suicide, 4.6. On this basis, only two of the patients who committed suicide had a predicted risk of suicide above 5%. CONCLUSIONS: Although several factors were identified that were strongly associated with suicide, their clinical utility is limited by low sensitivity and specificity, combined with the rarity of suicide, even in this high-risk group.
BACKGROUND:Psychiatric hospital in-patients are known to be at high risk of suicide, yet there is little reliable knowledge of risk factors or their predictive power. AIMS: To identify risk factors for suicide in psychiatric hospital in-patients and to evaluate their predictive power in detecting people at risk of suicide. METHOD: Using a case-control design, 112 people who committed suicide while in-patients in psychiatric hospitals were compared with 112 randomly selected controls. Univariate analysis and multivariate analyses were used to estimate odds ratios and adjusted likelihood ratios. RESULTS: The rate of suicide in psychiatric in-patients was 13.7 (95% CI 11.7-16.1) per 10,000 admissions. There were five predictive factors with likelihood ratios > 2, following adjustment: planned suicide attempt, 4.1; actual suicide attempt, 4.9; recent bereavement, 4.0; presence of delusions, 2.3; chronic mental illness, 2.2; and family history of suicide, 4.6. On this basis, only two of the patients who committed suicide had a predicted risk of suicide above 5%. CONCLUSIONS: Although several factors were identified that were strongly associated with suicide, their clinical utility is limited by low sensitivity and specificity, combined with the rarity of suicide, even in this high-risk group.
Authors: Patricio Molero; Michael F Grunebaum; Hanga C Galfalvy; Mary A Bongiovi; David Lowenthal; M Goretti Almeida; Ainsley K Burke; Ellen Stevenson; J John Mann; Maria A Oquendo Journal: Arch Suicide Res Date: 2014