Mental disorder among homeless and poor people: a comparison of assessment methods.

This study assessed mental disorders among 144 homeless and poor adults using four different methods: (a) history of psychiatric hospitalization, (b) structured clinical interview, (c) self-report symptom checklist, and (d) interviewer ratings. These four methods yielded divergent estimates of mental illness, ranging from 3-70%. Correlations assessing the degree of overlap among the measures were generally modest in magnitude. The results suggest that the variation in rates of mental illness across existing studies is due to methodological differences and that, with the exception of the structured interview, the various methods fail to adequately distinguish mental disorder from substance abuse.