Predicting alcoholism service needs from a National Treatment Utilization Survey.

Even when the prevalence of alcohol abuse and alcoholism allows for resource allocation based on differential problem size, further planning methods are necessary in order to determine what specific types of services should be developed to meet the estimated need. Using population-based variables (e.g., total population, White and minority populations) and easily obtainable alcohol indicator variables (consumption, revenue, alcohol-related deaths and arrests), regression models were developed to predict state bed capacity data, obtained from the 1980 National Drug and Alcoholism Treatment Utilization Survey (NDATUS). Prediction models are presented for detoxification, quarterway and halfway house, other residential, hospital, and outpatient. Several variables, particularly total population, deaths, and arrests, are predictive of alcoholism service levels by state, as reflected on NDATUS. Based on this study, estimates of specific service needs within a geographic area can be made by varying the values of the independent variables in the models to reflect the characteristics of the area.