A diagnostic formula for alcoholism.

Publisher Website
Google Scholar
Abstract
A sample of 63 alcoholics (mean age 47.2; 45 men) was matched by age, sex, examining physician and socioeconomic status with 63 nonalcoholics. Both groups were drawn from the patients of a family-practice clinic. Because the alcoholics were ambulatory and had not been previously hospitalized for alcoholism (with 1 exception), they were considered to be in an early- or middle-stage of alcoholism. Each patient received a diagnostic score based on 5 variables (medical conditions associated with alcoholism): respiratory conditions, cardiovascular problems, gastrointestinal difficulties, skin conditions, and bone and joint disorders. A stepwise discriminant analysis was performed to test the ability of the variables to discriminate between alcoholics and nonalcoholics. The alcoholics had higher mean scores on all variables, indicating that they had significantly more diseases in each category (P < .0001). The scores on each variable were then multiplied by numerical weights (equivalent to .beta. weights in linear regression) which were assigned to each variable according to its importance in predicting alcoholism. The respective weights for respiratory, skin, gastrointestinal, and bone and joint disorders were 113, 101, 51 and 37; cardiovascular problems had no importance in discriminating between patients. The weighted scores were then summed to obtain a total diagnostic score and a total score of 142 was established as the cutoff value (i.e., a score higher than 142 indicates an alcoholic patient). Using this technique, about 70% of the alcoholics and 78% of the nonalcoholics were correctly categorized.

This publication has 1 reference indexed in Scilit: