This study aims to determine the epidemiology (nature, prevalence, risk factors, course, and consequences) of psychiatrist-ascertained mental disorders that co-occur with alcohol and/or drug abuse disorders among representative samples of household-residing and homeless adults. It will also estimate specific psychiatrist-ascertained treatment needs and use of specific health services by persons in these two groups. Characteristics of those with mental morbidity co-occurring with alcohol and/or drug abuse disorders will be compared to those with only one type of disorder. The subjects come from: 1) a multi-stage probability sample of adult community residents selected from the household-residing population; and, 2) an otherwise unselected probability sample of homeless adults drawn from shelters, missions, and the jail. Psychiatrists directly examined both samples with the same standardized method, which included the Present State Examination (PSE-9), and also performed brief physical exams on the community sample. Internists performed physicals on the homeless sample. Morbidity data will be expressed using all 5 DSM-3 axes, as well as the CATEGO-ID system, making the results suitable for direct national and cross-national comparisons. The community sample was re-interviewed after one year with questions detailing mental symptoms, functional status, and use of specific health services. Aggregating data from the household sample, weighting each subject according to the strata and response rates, and adjusting to the 1980 census will produce estimates that can be generalized to the entire population. Community study subjects were also interviewed with a lay-administered NIMH-DIS. Psychiatrist-ascertained data will be used to develop algorithms that will be applied to DIS symptom data for improved identification of persons likely to have mental morbidity requiring treatment. These algorithms will then be applied to the DIS symptom data of persons with and without DIS-DSM-3 alcohol and/or drug use disorders of 5 ECA sites. These data will then be weighted by sociodemographic characteristics so that they resemble the population of the United States.
