Six different models or methods for indirectly asessing the need for mental health, alcoholism, and drug abuse services across a state will be tested for validity, using Colorado as the sample state. The models include: - linear regression equations (e.g., the NIMH-recommended model); - flat rate with a social-indicator index adjustment; - prevalence-variability model using a social-indicator index; - three-category model based on five social indicators; - vertical and horizontal synthetic estimation (national, state, or large-area epidemiologic survey data disaggregated and reaggregated for subareas of Colorado) NIMH Health Demographic Profile System data selected as key indicators of the need for mental health, alcoholism, and drug abuse services, will be fitted to the different needs-assessment models, which will be used to estimate the need for services in 48 randomly sampled subareas of Colorado. Stratification is used to represent high-density Hispanic and Black ethnic minority areas for special attention in modeling needs. Additional local and state social indicators (e.g. crime rate, economic data) will also be used in appropriate models if such data is available and economical to incorporate. Validation criteria for the models consist of three types of needs or problem-prevalence rates across the 48 subareas, as determined by direct survey interviews. They include (1) rate of diagnosable disorders assessed with the Diagnostic Interview Schedule, (2) scores on the CES-D scale used as a national comparison scale for psychological symptoms and distress, and (3) measured impairment in daily functioning, including areas of work, home management, interpersonal relations, legal problems, physical functioning, and subjective well-being.
