Objective: To derive and test a series of brief diagnosis-specific scales t
o identify subjects who are at high probability of meeting diagnostic crite
ria and those who may safely be spared more extensive diagnostic inquiry. M
ethod: Secondary data analysis of a large epidemiological data set (n = 1,2
86) produced a series of gate and contingent items for each diagnosis. Find
ings were replicated in a second retrospective analysis from a residential
care sample (n = 884). The DISC Predictive Scales (DPS) were then used pros
pectively as a self-report questionnaire in two studies, in which parents (
n = 128) and/or adolescents (n = 208) had subsequent diagnostic interviewin
g with the Diagnostic Interview Schedule for Children or the Schedule for A
ffective Disorders and Schizophrenia for School-Age Children. Results: All
analyses showed that gate item selection was valid and that any missed case
s were due solely to inconsistent reports on the same questions. Screening
performance of the full scales was shown to be good, and substantial reduct
ions in scale length were not associated with significant changes in discri
minatory power. Conclusions: The DPS can accurately determine subjects who
can safely be spared further diagnostic inquiry in any diagnostic area. Thi
s has the potential to speed up structured diagnostic interviewing consider
ably. The full DPS can be used to screen accurately for cases of specific D
SMIII-R disorders.