Suicide prediction revisited

Suicide Life Threat Behav. 1993 Spring;23(1):1-10.


Data from a 1983 prospective study of suicide in a cohort of 4800 psychiatric inpatients were reanalyzed using logistic regression, which is more appropriate for a binary outcome. The results were the same as in the previous study: too few of the subsequent suicides were identified and there were too many false positives to make this procedure useful. Several additional "artificial" logistic regression analyses were done: one series randomly removed increasing numbers of nonsuicide cases to increase the base rates; another series added an increasingly powerful hypothetical "test." Both of these maneuvers helped, but fell well short of perfection.

Publication types

  • Research Support, U.S. Gov't, Non-P.H.S.
  • Research Support, U.S. Gov't, P.H.S.

MeSH terms

  • Humans
  • Personality Assessment / statistics & numerical data*
  • Regression Analysis
  • Risk Factors
  • Suicide / psychology*
  • Suicide / statistics & numerical data
  • Suicide Prevention