The development of a health risk appraisal instrument for alcoholism.

The purpose of this study was to develop an instrument to identify those at risk for alcoholism. Risk indicators and predictors of alcohol abuse were identified based on a review of the literature. Questionnaire items based on empirical data and theory were constructed. The risks were mapped in a time framework recognizing there are early, intermediate and late antecedents of alcoholism. Reversible or preventable risks as well as fixed risks were identified. The data were collected from two criterion groups; 151 members of Alcoholics Anonymous and 164 non-alcoholics. Determination of items that discriminate the two criterion groups was accomplished using concurrent validity coefficients. Factor analyses on various item clusters were done to reduce the number of variables. For the factor scales used in the composite index the estimates of internal consistency ranged from.53 to.91 with two exceptions. Test-retest for the composite risk index reliability was.85. Analysis of the data was done for item selection and final scale construction. A discriminant analysis was also performed to determine the item weights and the percent correct classification (91.1%). Cross-validation revealed the percent correctly classified (90.8%). A logistic stepwise regression was performed using the same grouping of variables that was employed in the discriminant analysis. The validity coefficient produced was R =.83. Cross-validation showed no shrinkage in the validity coefficient (R =.85). This was a first effort at risk appraisal for alcoholism. The results should serve as a stimulus for other kinds of prospective studies that are necessary for identification of risk factors for alcoholism, prevention of alcohol abuse and reduction of risk for alcoholism. All of the variables employed in the final instrument were early predictors and were not typical of questionnaires related to quantity and frequency of drinking patterns. Thus, the findings suggest those at risk for alcoholism are less likely to deny these early behaviors, attitudes, and life events. The questionnaire has potential application in risk prediction, identification of risk areas, counseling, and education for prevention