Citation

Johansson LM, Sundquist J, Johansson SE, Bergman B. Acta Psychiatr. Scand. 1997; 95(2): 125-131.

Affiliation

Department of Psychiatry, Karolinska Institute, Huddinge Hospital, Stockholm, Sweden.

Copyright

(Copyright © 1997, John Wiley and Sons)

DOI

unavailable

PMID

9065677

Abstract

The aim of this longitudinal study was to determine the influence of ethnicity, social factors and self-reported long-term somatic and psychiatric illness on suicide in a random sample of the Swedish population. The study is based on face-to-face interviews conducted between 1979 and 1985 with a random sample of the Swedish population consisting of 47,762 Swedish-born subjects and 4407 individuals born elsewhere. The sample has been followed via register data concerning cases of suicide (suicides and undetermined deaths) until 31 December 1993. In total, 102 males and 46 females committed suicide prior to this date. Living alone and self-reported somatic illness with impaired health status were very strong risk factors for suicide, with risk ratios of 2.15 (CI, 1.51-3.05) and 1.80 (1.19-2.72), respectively. Ethnicity, defined as being born outside Sweden, had a risk ratio of 1.87 (1.18-2.97) in a model controlled for sex and age. However, this risk decreased with increasing age in the final model. Furthermore, an increased risk of suicide was found among female subjects living in rented flats and among male subjects irrespective of form of tenure, as well as among residents of large urban areas. Respondents with a self-reported long-term psychiatric illness with impaired health status also had a high risk of suicide, which decreased with increasing age. The main finding of this study is that somatic diseases and psychiatric disorders, which are known risk factors for suicide, may be revealed in surveys conducted by interviewers without medical training. Thus self-reported psychiatric and somatic illness appear to have a good potential for predicting suicide, even if the prevalence of psychiatric disorders is to some extent underestimated.

Language: en