Table 2.
Two logical regression models developed to predict drinking outcome at follow-up (n = 861)
| Dependent variable | Low risk drinking at follow-up | Clinically significant reduction in alcohol problems a | ||||||
|---|---|---|---|---|---|---|---|---|
| Predictor variables | B | S.E. | p | OR | B | S.E. | p | OR |
| Woman | −0.46 | 0.15 | 0.002 | 0.63 | −0.41 | 0.15 | 0.005 | 0.66 |
| Drinks last week | - | −0.02 | 0.01 | 0.001 | 0.98 | |||
| Days drinking | −0.11 | 0.04 | 0.015 | 0.90 | - | |||
| Days binge drinking | −0.25 | 0.05 | <0.001 | 0.78 | - | |||
| AUDIT score | - | 0.05 | 0.02 | <0.001 | 1.06 | |||
| Drugs | −0.50 | 0.26 | 0.05 | 0.61 | - | |||
| Depression | 0.04 | 0.02 | 0.045 | 1.04 | - | |||
| Readiness scale VAS | 0.14 | 0.05 | 0.003 | 1.15 | 0.13 | 0.04 | 0.002 | 1.14 |
| Completed program | 0.39 | 0.16 | 0.018 | 1.47 | - | |||
| User diary | −0.35 | 0.17 | 0.036 | 0.70 | - | |||
| All modules at once | 0.38 | 0.16 | 0.015 | 1.46 | 0.47 | 0.15 | 0.001 | 1.60 |
| Spoke to someone | 0.28 | 0.16 | 0.076 | 1.32 | - | |||
| Contact with profess. | - | −0.39 | 0.19 | 0.037 | 0.68 | |||
| Pharmacological treatm. | - | −0.61 | 0.32 | 0.058 | 0.55 | |||
| Constant | −0.79 | 0.48 | 0.101 | 0.45 | −1.49 | 0.44 | 0.001 | 0.23 |
| Nagelkerke R 2 | 0.16 | 0.08 | ||||||
| Percent correct | 65.7 | 60.2 | ||||||
| Hosmer and Lemeshow test χ2 (p) | 12.42 (0.134) | 6.05 (0.641) | ||||||
Predictors that did not contribute significantly to and where excluded from both models: age, anxiety, health index and VAS (EQ-5D), the physical, psychological, social and environmental quality of life (WHOQOL-BREF), action score (RCQ), and having used calendar
aA shift from a more problematic category (hazardous/harmful/probable dependence) to a less problematic category at least one level lower (no problematic use/hazardous/harmful)
OR odds ratio
“-” indicates that predictor was excluded because it did not contribute significantly to the model