Though computer-delivered alcohol interventions can have short-term benefits, they are not as effective as face-to-face interventions in reducing college students' alcohol consumption in the long term, according to a recent study led by Kate Carey, professor of behavioral and social sciences at Brown. The study highlights important components of interventions that these computer programs, which have grown increasingly popular on college campuses, could integrate to be more effective.
Carey's study, which will be published in the December issue of the journal Clinical Psychology Review, analyzed the results of 48 previous studies. In most of the studies, heavy drinkers at public universities participated in at least one intervention session, either on the computer or in person. Most of the sessions provided them with information about alcohol and alcohol-related problems, feedback about their own level of consumption and comparisons of their alcohol consumption to other students. In many cases, students also participated in one or two follow-up assessments to test the effectiveness of the initial intervention.
"We have been immersed in this literature for a while and wondered what the big picture was ... what studies in lots of different labs across different universities were revealing about the efficacy of these interventions," Carey said.
Carey's team found that at short-term follow-ups within 13 weeks of the intervention, both students who received the CDIs and the FTFIs showed decreased alcohol consumption when compared to the control group, which received an alcohol assessment but no education or feedback.
But in the intermediate and long-term follow-ups, only the FTFI group differed significantly from the control group.
Neither the number of interventions nor their duration impacted their effectiveness, a result Carey described as "counterintuitive."
Face-to-face interventions are done by trained counselors and educators, which ensures some level of quality standardization, Carey said. "The data did suggest the effects of the CDI are more heterogeneous," Carey said. "It's entirely possible that some of them were just not thoughtfully designed."
In addition to examining the interventions' overall effectiveness, Carey's team examined which components improved their performance. For example, the CDIs that identified high-risk situations actually performed poorer than those that did not. Carey's team also found that the effectiveness of the CDIs was lower for females compared to males.
Designing a better CDI is definitely possible, Carey said. Many of the components that made the FTFI interventions successful, like personalized feedback "can be translated to computer delivery," Carey said.
Brenda Curtis, a health communication research scientist at the Treatment Research Institute in Philadelphia, was not affiliated with Carey's study but has developed web-based intervention programs to help treat people addicted to smoking. She said she was not surprised by the results of the study. The current CDIs on the market are not living up to their full potential, she said.
"The study points out, 'Guys, get on your game,'" Curtis said. "It's time for us to up the ante, do some more rigorous work and figure out what works and what we need to do."
When she examined interventions in her own research, she found that certain kinds of personal tailoring improved their efficacy, she said.
When a person enters information about their drinking behavior and views on alcohol, "computer programs now just say, 'Oh, you drink four drinks. Well, that's more than the college norms,'" Curtis said. "That's basic level stuff, as opposed to getting down to things theoretically and clinically to things that change behavior," Curtis said. CDIs have great potential to deliver personalized and engaging feedback, she said.
A weakness of Carey's study was that it did not examine any comprehensive approaches to alcohol education, Curtis said. She added that the CDIs should have been held to a higher standard by being tested against an active control, rather than against a group that did not receive any type of alcohol education. But, she said, many of the programs may not have fared well in such a comparison.
"We're hoping people can use these findings to work towards even better interventions," Carey said.
At Brown, any student who has been treated by Brown Emergency Medical Services or encounters disciplinary difficulties due to alcohol must set up an appointment with a counselor, said Frances Mantak, director of health education.
Counselors will talk to students about the specific time they ran into trouble and give them feedback about how their drinking compares to their peers, Mantak said. Such interventions are usually only one session, though people in high-risk situations may have follow-up appointments.
Last year, 183 students were referred for appointments from Brown EMS, and 40 were referred from Providence Rescue, Mantak said, adding that 11 were referred from the deans. Forty-eight percent of students who came in for appointments were first-years and about 15 percent were returning for at least a second visit.
CDIs could potentially help free up time so that health educators may focus on other aspects of alcohol education, like prevention, Mantak said. But there are some personal conversations that would be difficult for a computer to replicate. If someone is using marijuana to manage anxiety, for example, or is a low-risk drinker who grew up with an alcoholic, a computer program may not be able to zone in on the important topics to discuss.
"Lots of times, I'm helping them not be too hard on themselves that they got EMS-ed, and it's hard for a computer to do that," Mantak said.