We conducted a clinical investigation of 80 suicides who met the DSM-III-R
criteria for schizophrenia. The results using this approach showed no signi
ficant difference with those of preceding studies, and general features reg
arding the phenomenology of suicide among schizophrenics worldwide were rec
ognized. The present study, however, served to underscore the following poi
nts: (i) suicide of schizophrenics must be considered a concern at all stag
es of the disease; (ii) the subjective strength of the will to die may be m
ore important for the committing suicides than the lethality of the methods
employed; and (iii) a change in the environment, for example, a hospital a
dmission or discharge, may trigger suicide. A control group of 80 living sc
hizophrenics with no past attempted suicide was then matched to the suicide
group with respect to sex and illness duration, in order to identify the p
redictors of suicide. In a logistic regression analysis, the presence of su
icidal ideation, degree of anxiety estimated by positive and negative syndr
ome scale, and birth order were revealed as predictors of suicide. As to th
e birth order, the risk is higher in middle children.