| Characteristic | Univariate Associations* | Multiple Analysis† | ||
|---|---|---|---|---|
| OR (95% CI) | P | OR (95% CI) | P | |
| Age | 1.01 (1.00–1.02) | .07 | 1.02 (1.01–1.03) | <.01 |
| Sex | 1.69 (1.13–2.55) | .01 | 2.00 (1.26–3.19) | <.01 |
| Household composition | 1.06 (0.64–1.77) | .81 | ||
| Education | ||||
| Basic | Reference category | |||
| Intermediate | 0.88 (0.54–1.45) | .62 | ||
| High | 0.80 (0.48–1.36) | .41 | ||
| Smoking status | 0.64 (0.42–0.96) | .03 | ||
| Comorbidity | ||||
| Presence of chronic somatic comorbidity without psychiatric comorbidity | Reference category | |||
| Presence of psychiatric co-morbidity without chronic somatic comorbidity | 0.63 (0.31–1.30) | .21 | ||
| Presence of both chronic somatic and psychiatric comorbidity | 0.57 (0.27–1.21) | .14 | ||
| Attitudes toward alcohol use | ||||
| Less alcohol improves health | 0.55 (0.32–0.96) | .03 | ||
| Less alcohol is not enjoyable | 0.71 (0.48–1.04) | .08 | ||
| Less alcohol use complicates relaxing | 0.53 (0.35–0.81) | <.01 | 0.58 (0.37–0.90) | .02 |
| Importance of changing alcohol use | ||||
| Unimportant | Reference category | |||
| Neutral | 0.82 (0.55–1.22) | .34 | 0.67 (0.43–1.05) | .08 |
| Important | 0.42 (0.22–0.80) | .01 | 0.39 (0.19–0.80) | .01 |
| Practice type | 1.06 (0.73–1.53) | .77 | ||
| Level of urbanization | ||||
| Large city | Reference category | |||
| City | 1.13 (0.66–1.92) | .65 | ||
| Small urban | 1.74 (1.04–2.93) | .04 | ||
↵* Data were calculated using multilevel logistic regression analysis to calculate the odds ratio (OR) and 95% confidence interval (CI) for each patient and practice characteristic independently.
↵† A prediction model was calculated using a backward stepwise regression procedure, starting with all patient and practice characteristics and then eliminating all variables that did not contribute (P > .1) to the model.