# Table 3 Odds Ratios (ORs) for positive change in ordered logistic regression models for the different variables (n = 40)

Bivariate

Model I

Model IIreduced

Independent variables

OR

95% CI

OR

95% CI

OR

95% CI

Gender

Women

ref.

ref.

-

Men

5.69**

1.59-20.33

2.63

0.54-12.84

-

Age group

15-39

40-49

ref.

ref.

-

50-64

.35

0.07-1.87

.10*

0.01-0.86

-

65+

.23

0.05-1.16

.28

0.04-2.22

-

.57

0.10-3.44

1.98

0.16-23.94

-

Category of alcohol use at baseline

Hazardous use1

ref.

ref.

ref.

Harmful use

18.3**

2.41-138.49

21.6*

1.69-276.06

19.1*

2.01-181.37

Dependence

11.4**

1.91-68.63

128***

8.21-999.99

32.0**

3.93-259.77

Other professional help

No

ref.

ref.

-

Yes

.56

0.18-1.77

.80

0.15-4.25

-

Medication

No

ref.

ref.

ref.

Yes

.19*

0.05-0.71

.08**

0.01-.50

.12**

0.03-0.54

Help from friends

No

ref.

ref.

ref.

Yes

.40

0.06-2.87

.12

0.01-1.98

.09*

0.01-0.96

from family

No

ref.

ref.

ref.

Yes

4.29*

1.21-15.16

2.35

0.48-11.47

3.852

0.93-15.85

Treatment

Reduce the no. of times

ref.

ref.

-

Goals

Reduce the amount

.11

0.01-1.06

.14

0.01-3.36

-

To be able to control

.60

0.14-2.55

.72

0.12-4.21

-

Complete or period stop

.71

0.15-3.30

1.00

0.16-6.07

-

Else

.20

0.01-10.55

.01

0 .00-1.71

-

1. *p < .05 **p < .01 ***p < .001
2. 1The cut-off was 8 for men, but 6 for women. The only person with non-hazardous alcohol use at baseline was included in the reference group.
3. 2Help from the family (n = 14) was left in the model as a control, as it was confounded with help from friends.