Use of case vignettes in suicide risk assessment.

Thirty-three brief case histories of suicidal patients were given to 19 experienced crisis workers for 7-point ratings of short- and long-term suicide risk. The ratings revealed considerable variability, raising questions about the reliability of such global assessments of suicidality. The most consistently rated cases were selected to operationally define "mild," "moderate," and "high" risk. Thus, each level was "anchored" by several vignettes. It was hoped that these anchor points would lead to more uniform future ratings among crisis workers. The correlation between short- and long-term risk ranged from near zero for some vignettes to as high as .82 (median = .46), demonstrating the need to rate both separately. Long-term risk was more difficult to rate, as demonstrated by a larger number of cases judged to be unrateable because of "insufficient information" (94 vs. 53 rater-case combinations). When the anchor vignettes were provided as a guide to the same sample of crisis workers, their ratings of suicide risk, as expected, showed significantly improved consistency.