Major Finding: At baseline 83% of the therapist group, 76% of the computer group, and 78% of the control group reported severe peer aggression during the past year. At 6 months the percentages were 45%, 49%, and 49%, respectively. At baseline 53% of the therapist group, 49% of the computer group, and 54% of the control group reported binge drinking. At 6 months the percentages declined to 33%, 33%, and 34%.
Data Source: Randomized, controlled trial of 726 adolescents, aged 14–18 years, who reported both past-year aggression and alcohol consumption.
Disclosures: The study was supported by the National Institute on Alcohol Abuse and Alcoholism. The authors reported no financial conflicts.
A brief intervention in the emergency department resulted in modest reductions in violence and alcohol use, according to a randomized, controlled trial involving 726 adolescents.
The teenagers, all of whom reported violence and alcohol abuse during the past year, were randomized to receive either a brochure (the control condition) or a 35-minute intervention delivered via computer or by a therapist. Both interventions were targeted at alcohol use and violence and were based on motivational interviewing techniques and skills training. The interventions included a review of goals, tailored feedback, a decisional balance exercise, role plays, and referrals, wrote Maureen A. Walton, Ph.D., of the University of Michigan, Ann Arbor, and her colleagues (JAMA 2010;304:527–35).
On every violence-related measure, all three groups, including the control group, showed substantial declines from baseline at 3 months and again at 6 months. For example, at baseline 83% of the therapist group, 76% of the computer group, and 78% of the control group reported severe peer aggression during the past year. At 3 months the percentages reporting aggression were 48%, 54%, and 62%, respectively, and at 6 months they were 45%, 49%, and 49%.
Similarly, all three groups showed substantial declines in every alcohol-related measure at 3 months and additional declines in most alcohol-related measures at 6 months. For example, at baseline 53% of the therapist group, 49% of the computer group, and 54% of the control group reported binge drinking. At 3 months the percentages declined to 34%, 29%, and 35%, respectively, and at 6 months they were 33%, 33%, and 34%.
Included in the study were adolescents aged 14–18 years who were being seen in a level I trauma center for a variety of reasons. Excluded were teens experiencing suicidal ideation, abnormal vital signs, insufficient cognitive orientation, and several other conditions. Of 3,764 patients approached for screening, 446 refused and 2,509 did not meet inclusion criteria, the most important of which was reported alcohol use and reported violence within the past year. More than 100 others refused participation, leaving 726 to be randomized. A total of 626 completed the 6-month assessment.
Several of the between-group differences were statistically significant. Compared with those in the control group, teenagers in the therapist group were significantly less likely to report severe peer aggression, an experience of peer violence, or consequences of violence at 3 months. None of those differences were statistically significant at 6 months. At 6 months, but not at 3 months, those in the therapist group reported significantly fewer alcohol consequences than controls.
The investigators reported several encouraging results from a number-needed-to-treat analysis. For example, only eight at-risk adolescents would need to receive the therapist intervention to prevent severe peer aggression in one adolescent. Ten at-risk adolescents would need to receive the therapist intervention to prevent one from being victimized by a peer. And 17 adolescents would need to receive the therapist intervention to prevent alcohol consequences in 1 teen.
In an accompanying editorial, Dr. Richard Saitz and Dr. Timothy S. Naimi of Boston University criticized several aspects of the study. They noted that the study's trial registration suggested that the investigators measured quite a few additional primary outcomes that they did not mention in their report, including drug use, injury, delinquency, and weapon carrying.
The fact that the investigators measured so many primary outcomes raises the concern of type I experimental error because of multiple comparisons. “If this study had measured more objective outcomes such as physician-documented injury events or school-based reports of violent incidents, rather than self-reported risk behaviors, the findings might have been more convincing,” they wrote.
In addition, Dr. Saitz and Dr. Naimi suggested that participants in face-to-face counseling might be less likely to report unsafe or undesirable behaviors at follow-up (JAMA 2010;304:575–7).
Dr. Saitz reported having been a consultant for online alcohol-related screening and brief intervention education projects supported by National Institutes of Health grants. He also has been compensated by Beth Israel Deaconess Hospital and the National Institute on Alcohol Abuse and Alcoholism for serving on data and safety monitoring boards. He has been or expects to be compensated as a speaker or consultant on alcohol and drug topics by multiple government agencies, academic institutions, professional societies, and private companies.