Practice Profile: Alcohol Interventions for College Students

Carey and colleagues (2012) conducted a meta-analysis to examine the efficacy of computer-delivered interventions (CDIs) and face-to-face interventions (FTFIs) on alcohol use for college students. A comprehensive search was conducted through multiple electronic databases, including PsycINFO, PubMed, Dissertation Abstracts, ERIC, CINAHL, and The Cochrane Library. Studies were determined eligible for inclusion if they 1) examined the efficacy of alcohol interventions at an individual level, 2) sampled college students, 3) conducted either a randomized controlled trial (RCT) or quasi-experimental design (QED) with a wait-list or no-treatment control condition, 4) measured alcohol behavior, and 5) provided adequate information necessary to determine an effect size. Studies were excluded if they 1) did not focus on alcohol use only (i.e., combined substance use interventions), 2) sampled non-college students, 3) used an active control condition, or 4) included a structural-level intervention component.

A total of 48 studies (47 RCTs and 1 QED) were identified, which included 37,480 participants. Of the 48 studies, 22 examined FTFIs, and 26 examined CDIs. The average age of participants in the FTFI studies was 19 years, and 56 percent were women. The final sample size included 5,237 students whose race/ethnicity was White (83 percent), Hispanic (11 percent), Black (7 percent), and Asian (5 percent). Of the participants in the FTFI studies, 68 percent were freshmen, 13 percent were sophomores, 10 percent were juniors, and 5 percent were seniors. Nearly all participants in the FTFI studies (97 percent) reported alcohol use prior to the intervention. The average age of participants in the CDI studies was 20 years, and 51 percent were women. The final sample size for these studies comprised 32,243 students who were 81 percent White, 13 percent Hispanic, 10 percent Asian, and 7 percent Black. Of these students, 73 percent were freshmen, 16 percent were sophomores, 12 percent were juniors, and 10 percent were seniors. Most participants (95 percent) had also reported alcohol use prior to the intervention.

Outcomes of interest included alcohol consumption (i.e., a certain quantity consumed over a particular period such as a week or a month, frequency of heavy drinking, or peak blood alcohol concentration) and alcohol-related problems (which were operationalized using multi-item scales). The outcomes were examined at the short-term follow up (less than 13 weeks postintervention), intermediate follow up (14–26 weeks postintervention), and long-term follow up (more than 27 weeks postintervention). The CrimeSolutions review of this meta-analysis focused on the outcomes at the long-term follow up. Effect sizes were calculated using Cohen’s d. If a study reported dichotomous outcomes, an odds ratio was calculated and transformed to d using the Cox transformation. The weighted mean effect sizes were calculated using random-effects procedures.

Samson and Tanner-Smith (2015) conducted a meta-analysis to determine the efficacy of single-session interventions to reduce alcohol use among college students. This meta-analysis consisted of a subset of studies from a larger meta-analysis (Tanner-Smith and Lipsey 2015), which synthesized results from studies examining brief alcohol interventions for all adolescents and young adults. A comprehensive search strategy was used to identify published and non-published studies. First, electronic databases (such as Clinical Trials Register, Dissertation Abstracts, ERIC, PsycINFO, PubMed, and Social Services Abstracts) were searched from 1980 through December 31, 2012. Other sources were then searched to identify grey literature.

Studies were considered eligible for inclusion in the larger meta-analysis if they had 1) a comparison condition of no treatment, a wait-list control, or some form of treatment as usual; 2) participant samples that included adolescents and young adults (defined as individuals ages 11–25) or samples that consisted entirely of undergraduate college students; 3) a study design of a randomized controlled trial (RCT) or controlled quasi-experimental design; 4) provided enough information to determine an effect size, and 5) assessed an intervention effect on at least one outcome variable that measured alcohol use or alcohol-related problems (Tanner-Smith and Lipsey 2015). Samson and Tanner-Smith (2015) then identified studies from this sample that used experimental or controlled quasi-experimental designs to test the effects of a single-session brief alcohol intervention for undergraduate students ages 25 years or younger. Studies were considered eligible if the intervention was delivered in a single session and if the study participants exhibited hazardous levels of alcohol use (defined as a group average or as a majority of participants drinking more than two drinks per day or four drinks in a sitting in the past month).

A total of 73 studies were identified as eligible for inclusion. The majority of studies (79 percent) were conducted in the United States. Studies were also conducted in Australia/New Zealand (7 percent), Great Britain (5 percent), and Scandinavia (3 percent). Approximately one third (37 percent) of comparison groups were active conditions rather than no- treatment or wait-list control groups. With regard to the intervention type, 48 percent used motivational enhancement therapy/motivational interviewing, 41 percent used feedback only, 8 percent used cognitive–behavioral therapy, and 3 percent used psychoeducational therapy. In regard to gender and race, participants were 51 percent female) and 78 percent White.

The outcome of interest was alcohol consumption among heavy-drinking college students (measured as frequency of use, quantity of use, peak consumption, and blood alcohol concentration). Effect sizes were calculated using a standardized mean difference effect size (Cohen’s d).

Scott-Sheldon and colleagues (2012) conducted a meta-analysis to determine the efficacy of alcohol expectancy challenge (EC) interventions for alcohol abuse prevention in college students. A comprehensive search strategy was used to obtain relevant studies from electronic databases (such as PsycINFO, PubMed, Dissertation Abstracts, ERIC, CINAHL, and The Cochrane Library), reference sections of relevant manuscripts, professional journals, databases of alcohol-related interventions for college students from the College Drinking Meta-Analytic Team at Brown University, and responses to listserv requests. Studies were determined eligible for inclusion if the authors 1) examined a behavioral intervention to reduce alcohol expectancies, 2) sampled college students, 3) used a randomized controlled trial (RCT) or quasi-experimental design (QED) with a control condition or a within-group design with pre- and posttest outcomes, 4) assessed alcohol expectancies, and 5) provided information necessary for calculating effect sizes. Studies were excluded if the authors 1) did not focus solely on alcohol use (i.e., combined substance use interventions), 2) sampled non-college students, 3) did not assess alcohol expectancies, or 4) included a structural-level intervention component.

A total of 14 studies (that examined 19 interventions) were identified. The studies included 1,417 participants. Most studies were conducted at large public universities in the United States (71 percent). Of these, 36 percent were in the Southeast, 14 percent were in the Midwest, 7 percent were in the Northeast, 7 percent were in the Southwest, and 7 percent were in multiple regions. Studies were also conducted in the Netherlands (14 percent) and Sweden (7 percent). Most studies were published in journals (86 percent) between 1986 and 2008. The remaining 14 percent were unpublished dissertations. Most participants were White (88 percent) and male (60 percent). The mean age of participants was 20 years, and 49 percent were first-year college students. Nearly all participants (91 percent) reported using alcohol in the past, and none reported a history of or current alcohol treatment.

The outcomes of interest were alcohol-related expectancies and consumption. Alcohol expectancies were measured with a variety of validated measures such as the Alcohol Expectancy Questionnaire. Alcohol consumption was also measured in a variety of ways (e.g., quantity consumed over a period such as a week or a month, or frequency of heavy drinking). Effect sizes were calculated using Cohen’s d. If a study reported dichotomous outcomes, an odds ratio was calculated and transformed to d using the Cox transformation. The weighted mean effect sizes were calculated using random-effects procedures.

Using meta-analytic techniques, Scott-Sheldon and colleagues (2014) analyzed the efficacy of alcohol interventions to prevent alcohol misuse in first-year college students. To identify eligible studies, several electronic databases (including PubMed, PsycINFO, ERIC, Dissertation Abstracts, and The Cochrane Library) were searched. References of relevant papers, abstracts from professional journals, and author responses to requests were also used to retrieve studies. Studies that fulfilled the selection criteria and were available by April 2013 were included in the meta-analysis. To be included, studies had to 1) examine an individual- or group- level intervention intended to reduce or prevent alcohol use, 2) sample first-year college students, 3) use a randomized controlled trial (RCT) with a comparison group, 4) measure alcohol use, and 5) provide sufficient information to calculate an effect size. Studies were excluded if they 1) did not focus solely on alcohol use (i.e., combined substance use interventions), 2) sampled other college students (i.e., upperclassmen), or 3) included a structural-level intervention component.

A total of 41 studies (examining 62 interventions) were identified. The sample consisted of 24,294 participants, most of whom had reported previous alcohol use (85 percent). Of the 62 interventions, 37 percent were FTFIs, 32 percent were CDIs, 12 percent were computer-facilitated FTFIs, 10 percent were delivered through mail or email, 8 percent were a combination of FTFIs and CDIs and mail-based, and 3 percent were written. Of the studies, 31 took place at public universities, 8 took place at private universities, and 1 took place at both public and private universities. Participants were 77 percent White, 9 percent Black, 8 percent Asian, and 6 percent Hispanic. The average age was 19 years, and 57 percent were women.

The outcomes of interest were alcohol consumption and alcohol-related problems. Alcohol consumption was measured in a variety of ways (e.g., quantity consumed over a period such as a week or a month, or frequency of heavy drinking). Alcohol-related problems were operationalized using multi-item scales. Effect sizes were calculated using Cohen’s d.