The aim of this study is to investigate whether motivation-tailored alcohol interventions are more effective when delivered by person or by computer-generated feedback letters. A sample of 920 general hospital inpatients with risky drinking will be recruited through a computerized screening procedure. Patients with more severe alcohol problems will be excluded from the study. Participants will be allocated by time frame randomization to one of three study arms: (1) personal counseling based on Motivational Interviewing, (2) computer-expert system intervention that generates individualized feedback-letters, and (3) control group (treatment-as-usual). The interventions differ in their channel of delivery, but not regarding their content. Both intervention groups receive interventions at three time points: directly after the baseline-assessment at the general hospital, and 1 and 3 months later by mail and phone, respectively. Outcome will be assessed six, 12, 18 and 24 months after baseline.

Inclusion Criteria:

• General hospital inpatients with risky drinking (AUDIT-C >= 4/5 (women/men) and AUDIT < 20)

Exclusion Criteria:

• Patients already recruited during an earlier hospital stay
• Patients physically and mentally not capable of participating in the study
• Patients with a hospital stay of less than 24 hours
• Patients with insufficient language/ reading skills
• Patients employed at one of the departments participating in the study or conducting the study
• Patients with more severe alcohol problems (AUDIT >= 20)