The demographic (ever married) and alcohol-related characteristics in Table 3 that differentiated between the AUD Groups 1 and 3 were likely to be correlated, so these six items were entered into a backward elimination logistic regression analysis to determine which variables remained significantly related to a combined AUD plus SUD outcome (Group 1) among individuals in Groups 1 plus 3. The results demonstrated significant odds ratios (ORs) for Group 1 membership for the number of AUD items endorsed (OR 1.20), an alcohol-induced mood diagnosis (OR 6.76), and having received formal AUD treatment (OR 3.89). The two drug-related items and the SRET score that differentiated between Group 1 and Group 2 in Table 3 were entered into a logistic regression predicting Group 1 membership from among Groups 1 plus 2, with only the SRET contributing significantly (OR 2.12). If SRET was removed from the equation, approximately equal contributions were seen from use of tobacco products and the number of SUD items.
