Brief Alcohol Screening and Intervention for College Students (BASICS)

Study 1

Summary

Marlatt et al. (1998), Baer et al. (2001), and Roberts et al. (2000) screened high school students intending to attend the university and selected 348 students-to-be who were predicted to be at high risk for drinking problems in college. After random assignment, the treatment group but not the control group underwent the brief intervention during the freshman year. Assessments at baseline, 6 months, 2 years, and 4 years measured both drinking rates and harmful consequences. A separate group of normal students not at high risk was followed for comparison.

Marlatt et al. (1998), Baer et al. (2001), and Roberts et al. (2000) found that participants at the University of Washington who received BASICS demonstrated a significantly

• Greater deceleration of drinking rates and problems over time in comparison with control participants through the 2- and 4-year follow-ups.

Evaluation Methodology

Design: In the spring of 1990, questionnaires were mailed to all students who were accepted and had indicated an intention to enroll at the University of Washington the following autumn term (by sending in a $50.00 deposit), who were matriculating from high school, and who were not over 19 years of age. Each student was offered $5.00 and entrance into a prize drawing for return of the questionnaire. Of 4,000 questionnaires sent, 2,179 (54%) completed forms were returned. Of these, 2,041 students (51% of the total screening sample) provided usable questionnaires and indicated a willingness to be contacted for future research. A two-step process was used to acquire participants for the study: sample selection and participant recruitment. From the screening pool, a high-risk sample was selected based on the following criteria: (a) reported drinking at least monthly and consuming at least five to six drinks on one drinking occasion in the past month or (b) reported the experience of three alcohol-related problems on three to five occasions in the past three years on the Rutgers Alcohol Problem Inventory (RAPI). The selection criteria identified the top quartile 25% of the screening sample (n = 508) as high risk. An additional normative comparison sample was randomly selected from the pool of 2,041 responders (n = 151); as this sample was selected to represent normative practices, it included students at all risk levels, including those previously screened as high risk (n = 33, 22%).

On arrival on campus for the fall term, these students were invited by means of a personal letter to participate in a 4-year longitudinal study of alcohol use and other lifestyle behaviors, including a 45 minute interview and completion of questionnaires for a $25.00 payment. Receipt of the letters was confirmed by telephone. Students in the high-risk group also agreed to be randomly assigned to participate in the intervention program or to the no-treatment control condition. All participants agreed to additional paid follow-up questionnaire assessments and to provide the names and addresses of two collateral reporters who could be called to confirm the drinking practices of participants. Of the 508 high-risk students, 366 (72%) were successfully recruited, and 151 (76%) of the normative sample was recruited, and 108 agreed to participate. There were no differences on screening measures between those recruited and those not recruited. Of the 366 high-risk students, 11 were removed because of extreme levels of drinking and given immediate attention, and 7 returned questionnaires too late to be randomized. The remaining 348 high-risk students were randomized to the intervention or assessment control conditions.

Follow-up assessments occurred at six-months post-baseline, and then annually by mail each subsequent fall term. All collateral reporters were telephoned after each participant assessment, and collateral assessments were successfully completed for approximately 50% of subjects at each assessment point. Of the 456 combined high-risk and normative students invited to participate at baseline, 403 (88%) provided data at the 2-year follow-up; of these, 379 (83%) provided complete data at both the 1- and 2-year follow-ups. Complete data sets at baseline and all 4 follow-up periods (baseline through the four-year follow-up) were provided by 328 participants (76%). Missing data were replaced by a multiple imputation method using maximum likelihood estimation. Similarly, no significant differences were observed between prevention and control condition for alcohol consumption, related consequences, or demographic and individual differences. Attrition rates at the four-year assessment did not differ significantly between high-risk and normative comparison groups, or between randomized high-risk group conditions.

Sample: The high-risk sample was 55% female and 84% Caucasian, while the normative comparison sample was 54% female and 78% Caucasian. High-risk drinkers reported drinking about twice a week, consuming almost 11 drinks a week, reaching an estimated blood alcohol concentration (BAC) of .12 weekly, .18 on peak episodes. Normative comparison students reported drinking about 5.5 drinks per week during one drinking occasion, reaching a typical estimated peak blood alcohol level of .08%.

Measures: The initial four-page screening questionnaire asked participants to rate their frequency of alcohol consumption, typical quantity of alcohol consumed on weekend evenings, and the most alcohol consumed on one occasion over the past three months. The Brief Drinker Profile, a one-hour structured interview, was administered at baseline to measure family history of alcohol problems, history of conduct disorder, and personal drinking history. The Daily Drinking Questionnaire (DDQ) was used to measure the actual number of drinks for each day of a typical drinking week, yielding drinking days per average week and average drinks per drinking day. The 23-item Rutgers Alcohol Problem Inventory (RAPI) was included as a measure of alcohol-related problems over the previous three years. In addition, the Alcohol Dependency Scale (ADS) was used to measure the severity of drinking problems. An additional questionnaire packet was administered containing measures of drinking and drug use, problems associated with alcohol abuse and dependence, and a variety of psychosocial measures. Collateral assessments to corroborate self-reported drinking behavior and consequences were conducted by phone.

Analysis: At the two-year follow-up, the effects of treatment on drinking over time were assessed using multivariate repeated measures MANOVAs. Short-term results (baseline to 6-month follow-up) were analyzed only for drinking patterns, while long-term results were also analyzed for all dependent variables at the annual assessments. For both short- and long-term outcomes, separate repeated measures analyses were completed for sets of related dependent measures. Individual factors such as gender, family history of alcoholism, conduct history, and residence type were evaluated in separate MANOVAs on a post-hoc basis. At the four-year follow-up, three distinct mixed model analyses provided hypothesis tests with restricted maximum likelihood parameter estimates for the effects of group and time on the three major outcome factor scores: frequency, quantity, and negative consequences. The mixed models included specifications for a priori contrasts for adjacent time points (baseline vs one-year follow-up, one-year vs two-year follow-up, etc.), repeated measures effects for time, time-by-treatment interactions, and random-effects estimates for subjects.

Outcomes

Short-term: (Marlatt, Baer, Kivlahan, Dimeff, Larimer, Quigley, Somers, and Williams, 1998)

There was a significant short-term treatment effect at the six-month follow-up in each of the dependent measures when analyzed separately. In comparison with those in the control condition, those in the treatment group reported drinking less frequently over time, less quantity over time, and less peak quantity over time. Analysis of responses to the Daily Drinking Questionnaire (DDQ) revealed a similar pattern. Multivariate repeated measures analysis of two scores derived from the DDQ, drinking frequency and average drinks per drinking day, also revealed a significant multivariate effect for time and a significant multivariate treatment X time interaction, which was associated with both average quantity of drinking and drinking frequency estimates.

Long term: In general, drinking and related problems tended to decline over the two year period. Individuals in the treatment condition demonstrated significantly greater reductions in drinking rates and problems than did those in the control condition. Compared to individuals in the control condition, individuals in the treatment condition reported drinking less frequently over time, less quantity over time, and less peak quantity over time. Each of these differences was significant, although the magnitude of the effect sizes was modest. A significant effect for time and a significant treatment X time interaction were noted in the analysis of drinking frequency and average drinks per day as measured by the Daily Drinking Questionnaire (DDQ). A significant interaction among treatment, time, and drinking quantity and frequency indicated that the treatment X time effect was differentially associated with average quantity of drinking. Again, the magnitude of the treatment effects was modest.

Analysis of alcohol related problems revealed similar significant effects for time and treatment X time favoring those receiving the brief intervention. The magnitude of treatment effects for alcohol related problems was larger than those noted among drinking rate measures. Similar effects were noted with the measure of alcohol dependence (ADS), with significantly fewer individuals in the treatment group showing mild alcohol dependence as compared to the no-treatment control group. Despite a developmental trend of fewer problems over time for all participants, analyses of a composite score combining the Rutgers Alcohol Problem Inventory (RAPI) and ADS scores significantly favored the treatment group.

Two-year follow-up of clinical change: (Roberts et al., 2000)

This article examined changes in clinical classifications based on drinking rates (using the Rutgers Alcohol Problem Inventory) and alcohol-related problems (using the Alcohol Dependency Scale) for the 80% of the sample with follow-up data at two years. The clinical classification divided the respondents into categories based on falling either above or below a cut point that separated high-risk drinkers from functional drinkers. The two outcomes were measured in terms of change from the baseline category to the follow-up category, and the change defined five transitions: newly high risk, reliably worse, no change, reliably improved, and resolved. Further, the analysis examined these categories separately for those above or below the cut point. The tables report percentage figures without significance tests. A summary provided by the authors (p. 503) states: "using empirically derived criteria for clinical significance, we found that participants in the high-risk treatment group showed better outcomes than did the high-risk controls." For example, 48% of the high-risk treatment group versus 38% of the high-risk control group reached the resolved category.

Four-year follow-up: (Baer, Kivlahan, Blume, McKnight, and Marlatt, 2001)

Significant prevention group X time interactions over four years were observed with respect to negative drinking consequences and for drinking quantity. In general, drinking problems declined significantly over time, and the treatment produced significant differences in alcohol use and related problems over four years, favoring the treatment group. Individual change analyses suggest that the dependence symptoms of those receiving the treatment are more likely to decrease and less likely to increase when compared to those in the control group.

Study 2

Summary

Borsari and Carey (2000) used a randomized controlled trial to compare a sample of 60 binge drinking college students assigned to one of two conditions: (1) BASICS; or (2) no-treatment control. Assessments at baseline and six-week follow-up included measures of alcohol consumption and alcohol-related problems.

Borsari and Carey (2000) found at six-week follow-up that, relative to the no-treatment control group, the intervention group showed a significantly lower:

• Number of drinks consumed per week
• Number of times consuming alcohol in the past month
• Frequency of binge drinking in the past month.

Evaluation Methodology

Design:

Recruitment: Participants were recruited from an introductory psychology class at a large Northeastern university. Those who reported drinking five or more drinks (four or more for women) on one occasion two or more times in the past month were eligible to participate. Of 109 individuals screened, 63 (58%) met the selection criteria. Of those, 60 (95%) were invited to participate and all agreed.

Assignment: Participants were randomly assigned using a coin flip to intervention (n = 29) or no-treatment control (n = 31) groups.

Assessments/Attrition: Assessments occurred at baseline and six-week follow-up. One control group participant did not complete the follow-up, for an overall attrition rate of 1.7%.

Sample:

The sample had a mean age of 18.58 years and was 57% female. Approximately 12% of participants were minorities. Most (87%) lived in on-campus dormitories.

Measures:

Outcome measures came from surveys completed by students. Four outcomes were computed for analysis: (1) number of drinks consumed per week; (2) number of times consuming alcohol in the past month; (3) frequency of binge drinking in the past month; and (4) alcohol-related problems in the past 30 days. The authors emphasized the confidentiality of responses and the importance of accuracy, but information on the validity and reliability of measures was not provided.

Analysis:

Three-stage regression models were used to test for intervention effects. Demographic variables (i.e., age, residence, and gender) were entered at Step 1, group membership at Step 2, and hypothesized mediators at Step 3. Variables not accounting for significant variance were then dropped from later models. The models did not appear to include baseline outcome controls. The authors noted that although the three alcohol consumption variables were significantly correlated (rs = .29-.59), they were analyzed separately and a Bonferroni adjustment for multiple tests was used.

Missing Data Method: Not discussed, though the one participant who did not complete the follow-up was very likely dropped from the regression analysis.

Intent-to-Treat: Although no details were provided, it appears that all available data were used in analyses, regardless of program participation.

Outcomes

Implementation Fidelity:

Participants reported high levels of satisfaction (M = 3.5 on 4-point scale) with the intervention and agreed that the information provided reflected their actual drinking (M = 3.0), that they would recommend such a session to a student like themselves (M = 3.4), and that they would recommend the session to a friend with a drinking problem (M = 3.2).

Baseline Equivalence:

The authors reported no significant differences for the outcome measures and mediators (Table 1). They also referred to no differences in sociodemographics but provided no details on statistical significance.

Differential Attrition:

Not examined, but overall attrition was only 1.7%.

Posttest:

Students assigned to the treatment condition relative to those in the no-treatment control condition showed significantly better outcomes on three of four tests at the six-week follow-up -- fewer number of drinks consumed per week (d = .21), fewer number of times consuming alcohol in the past month (d = .28), and lower frequency of binge drinking in the past month (d = .12). In response to a request from Blueprints, the lead author confirmed that Table 2 contained several typos and sent the computer output with the correct figures. Consistent with the results reported in the text, the corrected results show negative coefficients for group that indicate lower drinking in the intervention group than the control group for all three significant outcomes.

Mediation analyses revealed that the relationship between intervention condition and follow-up drinking was mediated by perceptions of how many drinks the typical student consumes per week but not by perceptions of how many drinks their friends consume per week, positive expectancies regarding heavy drinking, or negative expectancies regarding heavy drinking.

Long-Term:

Not examined.

Study 3

Summary

Murphy et al. (2001) used a randomized trial to examine 84 Auburn University undergraduates who reported high levels of drinking. Participants were assigned to one of three groups (the BASICS treatment, an educational intervention, or an assessment-only control group), and were followed for 9 months.

Murphy et al. (2001) found that heavier drinking BASICS participants at Auburn University showed significantly

• Greater 3-month decreases in drinking measures that were maintained at 9 months.

Evaluation Methodology

Design: Auburn University undergraduate students (n = 299) completed a series of screening questionnaires in exchange for extra course credit. No specifics regarding which courses were selected, or what criteria were used to select these courses were provided. Participants (n = 99) who were in the upper 33% of the screening sample in terms of reported drinks per week, as measured by the Daily Drinking Questionnaire (DDQ), and who endorsed two or more alcohol-related problems on the Rutgers Alcohol Problem Inventory (RAPI), were randomly assigned to one of two interventions (BASICS or education) or a control group. Randomization was conducted separately by gender and was stratified by drinks per week and RAPI score. See the original write-up for the specifics of the BASICS intervention. Participants assigned to the education condition watched "Eddie Talks," a 30-minute video consisting of a male college student discussing the negative interpersonal and academic consequences resulting from his alcohol abuse. Students in this group then participated in an individual discussion with a clinician that focused on their reaction to the video and to a sheet containing generic information about the risks of heavy alcohol consumption among college students. The 20-minute discussion focused on the student's thoughts about college student drinking in general, rather than on his or her personal alcohol consumption.

Of the 99 selected students, 84 (85%) completed one of the two interventions (BASICS: n = 30, education: n = 29) or participated as a control (n = 25). Of the 84 participants in all three groups, 79 (94%) completed one or both of the follow-up assessments; one control, four education, and 0 BASICS participants did not complete either follow-up. There were no significant differences on drinking variables between completers and those lost to attrition. Similarly, there were no significant differences on drinking variables between randomized participants who completed the intervention phase and those who did not. Finally, there were no significant differences between participants in the BASICS, education, or control groups on any demographic or baseline drinking variables.

Sample: The participants were 54% female, 94% were Caucasian, 83% were freshmen or sophomores with a mean age of 19.60 years. Participants averaged 2.6 binge drinking days per week and 21.9 total drinks per week.

Measures: All drinking-related measures other than the Alcohol Dependence Scale (ADS) were administered at screening and at the three- and nine-month follow-up assessments. The ADS was administered at screening and nine-month assessments. The timeframe covered by drinking measures at each assessment was the previous two months. Alcohol consumption was assessed with the Daily Drinking Questionnaire (DDQ). Alcohol-related problems were measured by the Rutgers Alcohol Problem Inventory (RAPI). In addition, students in the BASICS and education groups completed an intervention evaluation at post-test, and assessed whether or not the program had a "significant impact" on their drinking at the three-month follow-up. Finally, at both the three- and nine-month follow-ups, all participants were asked if their drinking had stayed the same, decreased, or increased relative to their initial assessment level of drinking.

Analysis: Within-group effect sizes were calculated by dividing the within-group difference between the baseline and follow-up mean by the pooled, weighted pre-post standard deviation. The between-group effect sizes were calculated by dividing the difference between adjusted group means by the pooled, weighted between-group standard deviation. Changes in drinking variables from baseline to three- and nine-months were calculated using multivariate analysis of covariance, with the baseline scores as the covariates.

Outcomes

Post-test: At the three-month follow-up, a significant group effect was noted. In addition, for the analyses of covariance that compared group changes in drinks per week and frequency of binge drinking, the tests for homogeneity of pre-post regression were significant. Specifically, among participants who consumed at least 25 drinks per week at baseline (29% of the total sample), BASICS participants showed greater 3-month reductions in drinks per week than did education participants. The BASICS condition also demonstrated a significant advantage over the control condition among participants who consumed at least 26 drinks per week at baseline (26% of the total sample). Among participants with at least 3 binge nights per week (49% of the total sample), the BASICS intervention yielded significant reductions in binge drinking relative to control values. The BASICS intervention was also superior to the educational intervention among participants with 4 or more binge nights per week (19% of the total sample). There were no significant group differences on frequency of drinking or RAPI score, but BASICS participants demonstrated more change than control and education participants on frequency of drinking and more change than control participants on RAPI score. Education participants showed slightly greater RAPI decreases than BASICS participants.

Long-term: At the nine-month follow-up, there was no significant multivariate effect of group, although BASICS participants showed small to moderate advantages over control and education participants on drinks per week, drinking days per week, and binge days per week. BASICS participants also demonstrated slightly more change than control participants on the RAPI and the ADS. Education participants showed a slight advantage over BASICS participants on the ADS, but the two groups showed a similar amount of change on the RAPI.

In general, at the nine-month follow-up, BASICS participants maintained their moderate reductions in drinking variables from the three-month follow-up, but control and education participants had also improved on their three-month outcomes at nine months. Heavier drinking BASICS participants maintained their large reductions across the drinking measures.

Participant evaluations: Participants in the BASICS group generally gave more favorable reviews to the program than did participants in the education group. At the three-month follow-up, more than half of BASICS participants indicated that the intervention they completed had a "significant impact" on their drinking, compared to only 20% of education participants. At three-months, a greater percentage of BASICS participants (69%) than education (24%) or control (17%) participants indicated a decrease in their drinking. Similar results were obtained at nine-months: 63% of BASICS participants indicated a decrease in their drinking compared to 29% of education and 40% of control participants.

Study 4

Summary

Larimer et al. (2001) used a cluster randomized trial of 12 fraternities and 159 participants, with a 1-year follow-up assessment of drinking, alcohol problems, and dependence.

Larimer et al. (2000) found that fraternity pledges in the treatment condition in a West Coast university showed

• Greater decreases in total weekly alcohol consumption and typical peak blood alcohol concentrations than did pledges in the control condition.

Evaluation Methodology

Design: Participants were recruited from the incoming pledge classes of 28 fraternities at a large West Coast university as part of a larger study investigating the effectiveness of alcohol-based intervention programs in Greek-letter organizations. The presidents of all fraternity houses received written and telephone requests from project staff asking if they would be willing to participate in an evaluation of an alcohol education program. Twenty-one fraternities expressed an interest in participating. Each house completed an initial screening of house-wide drinking rates and related consequences. Fraternities with fewer than ten incoming pledge class members were removed from further consideration. Of the remaining fraternities, 12 were randomly selected to participate further in the study. Individual pledge-class members of the selected fraternities were recruited using posters, personal letters, and announcement made during house dinners and meetings. Of the 224 potential pledge class members, 166 (74%) chose to participate in the study and completed a baseline assessment. Of these, seven members were identified by a clinical review committee as having endorsed extensive alcohol dependence or psychiatric symptoms and were removed from participation. Of the 159 eligible fraternity members, 120 (75%) completed a 1-year follow-up assessment.

Each fraternity was randomly assigned to either the intervention condition (n = 6 houses, 77 participants) or the assessment only/treatment-as-usual control group (n = 6 houses, 82 participants). Participants assigned to the intervention condition received the BASICS program, described in detail in the original write-up, administered by either a trained college-aged peer or a professional staff member. Intervention houses also received a one-hour house-wide feedback program. These programs were similar to the individual feedback sessions in terms of their approach and content, but were focused on identifying house drinking norms; the wide variety of drinking patterns; organizational consequences associated with drinking; and encouraging house-wide change. Programs were conducted on site at individual fraternities and attendance by 80% of members was required. Of the 82 participants in the intervention condition, 64 (78%) completed the individual interview. Participants in the assessment only/treatment-as-usual comparison group did not receive any feedback, but each house was required by university policy to receive at least one didactic presentation regarding alcohol use. All houses were paid $100 for participating in the study.

Individual study participants completed a baseline packet of questionnaires assessing current drinking rates, the prevalence of alcohol-related consequences, symptoms of alcohol dependence, and perceptions of drinking norms within their houses during the fall or winter quarter of their first year of fraternity membership. Follow-up assessments were completed one-year later. Individual participants were paid $20 for both the baseline and follow-up assessments. No significant differences were found between the two conditions for the average number of drinking days per week, the average number of drinks per occasion, or estimates of peak blood alcohol content (BACs), based on self-reported drinking. Analyses of those who completed the follow-up with those who did not revealed no significant differences on 17 measures of baseline drinking and alcohol-related consequences.

Sample: The sample reported a mean age of 18.8 years, was 81.1% Caucasian, 1.3% Native American, 1.3% Hispanic, 12.6% Asian, and 3% did not identify an ethnic group. The demographic distribution of the sample was representative of the fraternity population on the campus.

Measures: The Daily Drinking Questionnaire (DDQ) measured average alcohol consumption over the past three months. The quantity-frequency/peak alcohol use indices was used to measure the number of drinks and amount of time spent drinking on a typical and a peak occasion during the past month. This scale produced estimates of typical and peak blood alcohol concentrations (BACs). The 23-item Rutgers Alcohol Problems Index (RAPI) was used to evaluate the frequency and severity of alcohol-related problems. The Alcohol Dependence Scale (ADS) was used to assess symptoms of physical dependence on alcohol, while the Drinking Norms Rating Form (DNRF) was used to measure individual perceived norms of alcohol use. The Short Michigan Alcoholism Screening Test for mother and father (SMAST) was used to assess the lifetime incidence of problematic drinking behavior. The University of Rhode Island Change Assessment (URICA) was used to measure readiness to change alcohol use behavior. Finally, the Alcohol Perceived Risk Assessment (APRA) was used to evaluate perceptions of the risk of experiencing negative consequences as a result of drinking.

Analysis: All 120 participants who completed follow-up assessments were included in all analyses. Outlier analyses identified one participant with substantially elevated average quantity and total average use. Therefore, data were removed for these drinking scales. The effects of treatment on alcohol use and consequences over time were assessed using multiple univariate repeated measures ANCOVA. Individual student perceptions of pledge class drinking norms were entered as a covariate. Univariate F -tests of time and treatment effects were also performed. Chi-square analysis was used to examine differences in follow-up participation among individuals who received peer or professional interviews.

Outcomes

Post-test: No immediate post test data were analyzed in this evaluation.

Long-term: All data collected were analyzed at the one-year follow-up.

Drinking outcomes: A significant treatment effect was observed for total average use. In comparison with students in the control condition, students in the intervention condition reported significantly greater reductions in average drinks per week. No significant treatment interactions emerged for either the quantity per occasion or frequency of alcohol consumption. A significant main treatment effect was observed for typical peak blood alcohol concentrations (BAC). In comparison with students in the control condition, students in the intervention condition reported significantly greater reductions in typical peak BACs. Analysis of alcohol problems produced no significant results for either negative consequences (RAPI) or symptoms of dependence (ADS). Students did not report lower levels of problems over time or differential reductions in problems based upon treatment condition.

Peer versus professional interviewers: No significant difference in follow-up participation was observed between the two groups. A significant group effect was observed, however, for typical peak BAC. Students who received their individual interviews from college-aged peers reported significantly greater reductions in typical peak BACs than did students who received their interviews from professional staff members. Baseline differences between interview conditions were substantial, although not statistically significant.

Risk factors associated with treatment: The predicted mediators of the treatment, family history of alcohol problems, motivation to change drinking behaviors, and perceived risk for personal problems, were unrelated to the outcome.

Study 5

Summary

Turrisi et al. (2009) randomly assigned a sample of 1,275 athletes enrolled in a public northeastern and a public northwestern university to BASICS, Parent Based Intervention (PBI), BASICS+PBI, or control. Assessments were conducted at baseline (prior to matriculation) and 10 months after baseline.

Turrisi et al. (2009) found that, among a sample of athletes enrolled in a public northeastern and a public northwestern university, BASICS significantly lowered the

• Levels of peak blood alcohol concentration
• Number of drinks consumed on a typical weekend during the first year of college.

Evaluation Methodology

Design:
Recruitment/Sample size/Attrition:
Participants were incoming freshmen at a large public northeastern (site A) and a northwestern (site B) university. Participants were randomly selected from the registrar's database of incoming freshmen. At site A, 2,328 students were selected, and at site B, 1,672, resulting in a total of 4,000 students. These students were contacted by mail and 1,796 consented to participate and completed the Web-based screening assessment, yielding a 45% response rate. Of those who completed the screening survey, 1,419 met inclusion criteria (participation in high school or club team athletics). Of these individuals, 1,275 completed the baseline assessment and 1,090 (85.5%) completed the follow-up assessment. Following completion of baseline assessment, the parents of all students were invited to participate, but did not provide assessment data on their child's drinking.

Study type/Randomization/Intervention:
The study employed a randomized controlled trial design. Pre-screened high school athletes who had completed the baseline assessment (n = 1,275) were randomized to one of four study conditions: (1) BASICS-only (n = 277), (2) Parent Based Intervention-only (PBI-only, n = 316), (3) combined BASICS and PBI (PBI+BASICS, n = 342), and (4) control group (n = 340). BASICS interventions were 45-60 minutes and were conducted one-on-one with a trained peer facilitator. PBI is a handbook-based intervention designed to raise parental awareness of alcohol abuse and consequences among college students and to increase parental effort to address this issue with their teen. The 35-page handbook was mailed to parents in the summer (May-August) prior to college matriculation. Parents were requested to work through the material and to discuss the topics with their child. Students in the control group did not receive any intervention but completed the assessments.

Assessment:
All study participants completed the baseline assessment during the summer of 2006 before college matriculation. The post-test assessment was conducted approximately 10 months after baseline (spring semester 2007).

Sample characteristics:
Participants (mean = 17.92 years) were 44.4% male (n = 566) and 55.6% female (n = 709); 4.5% identified as Hispanic, 79.8% as white, 10.1% as Asian, 3.7% as multiracial, 2.0% as black or African American, 0.5% as Native Hawaiian or other Pacific Islander, 0.2% as American Indian/Alaskan Native, and 3.2% as other; 0.4% did not identify race/ethnicity.

Measures:
Validity of measurements:
All items and scales have been used in prior published research and validity and reliability has been established.

Primary outcomes:

• Alcohol use: Peak blood alcohol content (peak BAC) was calculated using participant's responses to a question regarding maximum drinks consumed on an occasion within the past 30 days and the number of hours they spent drinking on that occasion. Using the Daily Drinking Questionnaire participants were also asked to indicate the number of drinks consumed on a typical weekday and typical weekend.
• Alcohol-related consequences: The 23-item Rutgers Alcohol Problem Index (RAPI) was used to assess alcohol-related consequences within the past 3 months (alpha = .848).

Secondary outcomes:

• Descriptive drinking norms: 2 items from the Core Institute's Campus Assessment of Alcohol and Other Drug Norms were summed to create a composite descriptive norms variable (alpha = .858).
• Injunctive norms: Injunctive norms were assessed with respect to participants' closest friends and parents (alpha = .710; alpha = .897).
• Beliefs about alcohol: Positive vs. negative beliefs about alcohol were assessed with 4 items (alpha = .837).
• Attitudes toward drinking: 2 items assessing attitudes toward drinking (alpha = .871)

Analysis:
The study employed analysis of covariance (ANCOVA) to examine drinking outcome mean differences at post-test by treatment condition, controlling for baseline drinking and gender.

Intention-to-treat: The study followed the intent-to-treat principle. Missing data for attritors were imputed using a maximum likelihood approach.

Outcomes

Implementation fidelity:
Facilitators of the BASICS intervention were trained (10 weekly meetings) by clinical psychologists and counselors specializing in interventions for college student drinking. In addition, peer facilitators were monitored through coding random 20-minute segments of every session, using the Motivational Interviewing Treatment Integrity coding system. Facilitators demonstrated sufficient proficiency levels (4.49).

For the PBI intervention, attempts were made to ensure that parents read the materials. Parents were asked to evaluate the handbook through notes and a brief questionnaire and return it to the researchers. In the PBI-only group, 63% (n = 199) of the parents returned the handbook, the evaluation, or both. For 21 of the 26 topics covered on drinking, more than 85% of the parents indicated that they discussed the material with their teens. In the BASICS-only group, 53.8% (n = 149) of the teens attended BASICS, and 68.2% (n = 189) of the parents returned the consent form or the brief survey, yielding 112 (40.4%) participants in this group who met both criteria. In the combined group, 53.8% (n = 184) of the teens attended BASICS, and 59.9% (n = 205) of the parents returned the handbook, the evaluation, or both (59.9%, n = 205), yielding 112 (32.7%) participants in this group who met both criteria. In the control group, 70% (n = 238) of the parents returned the brief survey.

Baseline Equivalence:
A test for baseline equivalence revealed no significant differences between groups on all outcome measures and the majority of socio-demographic controls. Only one significant difference was observed between conditions for gender (F = 9.170, p < .05). Therefore, gender was included as a covariate in all models.

Differential attrition:
A statistical test for differential attrition indicated no significant differences between groups on demographic characteristics (gender, ethnicity), as well as on outcome measures. However, the results revealed a small but significant difference in the participant dropout rate among the four intervention conditions (Chi-square = 14.00, p < .01). The control group had the highest number of follow-up survey completers (89.7%), compared with those in the parent condition (88.3%), the BASICS condition (82.3%), and the combined condition (81.3%).

Post-test:
PBI-only: At post-test, ANCOVAs showed no significant difference between the Parent Based Intervention (PBI) and the control group on any of the four outcome measures (peak blood alcohol concentration, drinks per weekend, drinks per week, Rutgers Alcohol Problem Index).

BASICS-only: Out of 4 tests, 2 (50%) were significant, demonstrating beneficial program effects. Compared to the control group, BASICS-only significantly lowered the levels of peak blood alcohol concentration (.107 vs.122, p<.05) as well as the numbers of drinks consumed on a typical weekend (5.8 vs. 6.6, p<.05). However, no effects were observed for drinks per week and alcohol related problems.

PBI+BASICS: The strongest results, however, were revealed for the combined PBI+BASICS intervention (4 out of 4 tests were significant). The combined intervention led to a significant decrease in peak blood alcohol concentration (.096 vs.122, p<.05, d=.26), significantly lower numbers of drinks consumed on a typical week (7.3 vs. 8.4, p<.05, d=.16) or weekend (5.6 vs. 6.6, p<.05, d=.20), as well as a significant reduction in the self-reported number of alcohol-related problems (2.8 vs. 3.5, p<.05, d=.20).

Mediator analysis:
A mediator analysis conducted by Turrisi et al. (2009) revealed that descriptive and injunctive peer norms were significant mediators between the combined intervention (BASICS+PBI) and all drinking outcomes. Relative to participants in the control group, those who received the combined intervention perceived that typical college students drink less and perceived their peers to be less favorable toward their drinking behaviors, and, in turn, they reported lower peak blood alcohol concentration, fewer drinks per week and weekend, as well as lower numbers of alcohol related consequences. In addition, alcohol beliefs mediated the relationship between intervention and peak blood alcohol concentration and alcohol-related negative consequences.

Long-term effects:
Long-term effects were not assessed in this study.

Study 6

This study lacked a treatment-as-usual control group and compared the original face-to-face BASICS program with similar but less costly web-based and group-based alcohol interventions. The hypothesis was that all delivery forms would be equally effective.

Summary

Alfonso et al. (2013) used a randomized controlled trial to examine 173 students at a state university who had been mandated for alcohol treatment. The students were randomly assigned to three conditions: BASICS, a group-based version of BASICS, or an electronically delivered version of BASICS. The students were followed for three months after the intervention to assess the levels and harms of alcohol use.

Alfonso et al. (2013) found no significant differences between the three versions of the program (the study lacked a no treatment or usual-treatment control group).

Evaluation Methodology

Design:

Recruitment: The sample consisted of 173 undergraduate students at a state university in the southeastern United States. The students had been mandated for treatment by the university because of a violation of alcohol consumption rules (e.g., alcohol possession under age 21, visible intoxication, DUI), and they were recruited from violation referrals and a drug counseling center. Participants were deemed ineligible if at the initial screening they reported no alcohol use or met criteria for alcohol dependence.

Assignment: The eligible students were assigned randomly to one of the three intervention conditions using a computer-generated randomized number list. The three conditions included BASICS, a group-based alcohol intervention modeled after BASICS called CHOICES, and an electronically delivered alcohol intervention called e-CHUG that, like BASICS, utilized social norms feedback and motivational enhancement. The study did not report the randomized sample sizes of the conditions.

Assessments/Attrition: Participants completed an anonymous online survey at baseline, when they received a unique identification number, plus an anonymous follow-up questionnaire at three months post-intervention. The anonymity, designed to decrease socially desirable responses, meant that only the participants providing their unique code at follow-up could be included. It thus appears that the 173 students cited in the study were those whose baseline and posttest data could be linked. The study did not report figures on attrition or missing data, but the unequal sizes of the three conditions (53, 72, and 48) indicate a problem. Assuming approximately equal assignment, the gaps between the reported condition numbers suggest high attrition.

Sample:

The sample consisted of 57% men and ranged in age from 18 to 25 years, with a mean age of 18.77. Most students were white (83%), with the remainder split among Asians, Blacks, and others.

Measures:

The outcome measures came from self-reports over a period of the previous four weeks.

• Peak alcohol consumption was measured using the Alcohol Timeline Followback procedure, a measure with evidence of reliability and validity, to indicate the largest number of drinks consumed over all drinking occasions.
• Average blood alcohol content was computed from reported alcohol consumption, weight, and length of drinking sessions.
• Peak blood alcohol content was computed the same way but measured the single highest level reached over the four-week period.
• Alcohol-related harm was measured with the 23-item Rutgers Alcohol Problem Index and eight items from the Drinker Inventory of Consequences, with a total score obtained by adding the number of times each harm was experienced.

Analysis:

The analyses used mixed-model ANOVAs with repeated measures, and the tests focused on condition-by-time interactions.

Intent-to-Treat: The authors stated that intent-to-treat analyses could not be conducted because of the inability to match anonymous identification numbers for some subjects.

Outcomes

Implementation Fidelity:

Results from coding the videotapes of 45% of the individual and group sessions indicated that counselors met the criteria for 'competence' on all six global motivational interviewing scales. Specifically, they had mean scores above four on the five-point Likert scale.

Baseline Equivalence:

Chi-square and univariate ANOVA tests found no significant differences between conditions for sex, race, ethnicity, class standing, type of residence, and alcohol-related harms. A multivariate analysis of variance for the three dependent variables measuring alcohol use found a significant difference for average blood alcohol content.

Differential Attrition:

The anonymity of respondents made it impossible to test for differential attrition. However, the differences in the reported sizes of the three conditions (53, 72, and 48) suggest high differential attrition (assuming approximately equal random assignment).

Posttest:

All four outcomes - negative alcohol-related harms, average blood alcohol content, peak blood alcohol content, and peak number of drinks consumed in one setting - showed significant time effects, but none showed significant differences across conditions in the change over time. The authors concluded that the less costly electronic version of the program was equally effective as the face-to-face version.

Long-Term:

Not examined.

Study 7

Summary

Kulesza et al. (2010) used a randomized controlled trial to examine 114 college students who were at high risk for alcohol problems. They were randomized to a 10-minute intervention (n = 39), a 50-minute intervention (n = 35), or an assessment-only, waitlisted control group (n = 40) and assessed for alcohol use and alcohol-related problems at four weeks after the intervention.

Kulesza et al. (2010) found that, compared to the control group, the 10-minute intervention session, but not the 50-minute intervention session, led to significantly lower

• self-reported alcohol consumption at the four-week follow-up.

Evaluation Methodology

Design:

Recruitment: The sample of high-risk college drinkers was recruited through the Psychology Subject Pool and received course credit for participation. Students were at high risk if they 1) drank at least monthly and had one occasion of binge drinking in the last month, or 2) reported three or more alcohol-related problems. Of the 536 participants completing the screening, 134 met inclusion criteria and 114 agreed to participate.

Assignment: Students were randomized to a 10-minute intervention (n = 39), a 50-minute intervention (n = 35), or an assessment-only, waitlisted control group (n = 40).

Assessments/Attrition: Assessments came at baseline and at a four-week follow-up. Because of a two-week gap between baseline and the intervention, the time span for the follow-up differed across conditions. For the intervention groups, the follow-up came four weeks after the intervention and about six weeks after baseline; for the control group, it came four weeks after baseline.

Sample:

The sample of college students came from psychology classes at a single university. It was predominately Caucasian (84%) and female (72%), with an average age of 20 years.

Measures:

The study used three outcomes measures. First, the Rutgers Alcohol Problem Inventory assessed personal, social, and academic problems from alcohol use. Second, the Daily Drinking Questionnaire assessed the typical number of drinks consumed and hours spent drinking on each day of the week over the past month. Third, the Protective Behavioral Strategies Survey assessed students' use of various cognitive-behavioral strategies to reduce the harm associated with alcohol consumption.

Analysis:

The study used analysis of covariance for each outcome, with baseline scores as a covariate.

Intent-to-Treat: The analysis used all subjects.

Outcomes

Implementation Fidelity:

Not presented.

Baseline Equivalence:

The authors stated that "There were no differences across conditions in age, gender, or ethnicity." The baseline outcomes listed in Table 1 appear similar, but no tests were reported for these measures.

Differential Attrition:

No attrition.

Posttest:

Overall ANOVA tests reached significance for one of the three outcomes - the Daily Drinking Questionnaire measure of alcohol consumption (eta-squared = .06). Pairwise posthoc comparisons with Bonferroni corrections indicated that participants in the 10-minute intervention - but not the 50-minute intervention - had significantly fewer drinks per week at follow-up compared to the control condition (d = .53). The conditions did not differ significantly on alcohol-related negative consequences or on use of protective behaviors.

Long-Term:

Not examined.

Study 8

Summary

Kulesza et al. (2013) used a randomized controlled trial to examine 278 college students who were at high risk for alcohol problems. They were randomized to a 10-minute intervention (n = 95), a 50-minute intervention (n = 93), or an assessment-only control group (n = 90) and assessed for alcohol use and alcohol-related problems at four weeks after the intervention.

Kulesza et al. (2013) found that, compared to the control group, both the 10- and 50-minute intervention session led to significantly lower

• Self-reported alcohol consumption at the four-week follow-up.

Evaluation Methodology

Design:

Recruitment: Undergraduate students at a large Southern university who were enrolled in Psychology courses joined the study to earn extra course compensation. The sample was limited to high-risk students, who were defined as having reported 1) drinking at least monthly and binge drinking on at least one occasion in the past month or 2) three or more alcohol-related problems on three to five occasions in the past three years. Of 672 screened students, 289 (43%) met the inclusion criteria and 278 (41%) participated in the study.

Assignment: Students were randomized to a 10-minute brief intervention session (n = 95), a 50-minute brief intervention session (n = 93), or an attention-control group (n =90). Control participants were asked to come to the clinic and spend about 15 minutes with the therapist discussing topics unrelated to their alcohol consumption (i.e., football, academics). The text referred to the counselors as being "both graduate students." With two interventions and two counselors, there was potential for confounding of the interventions with the counselor delivering the intervention.

Assessments/Attrition: Follow-up measures were completed four weeks after the intervention or control visit. According to the authors, only 10 students (3.6%) dropped out. However, comparing Tables 1 and 2 shows that the 50-minute group fell from 93 to 81, the 10-minute group fell from 95 to 90 - a combined loss of 17 subjects - while the control group rose from 90 to 97.

Sample:

The average age of the participants was 20.1 years, and they consumed an average of 16.2 drinks per week. The majority were Caucasian (87%) and female (71%), which is consistent with ethnic/racial and gender representation of the campus Psychology student population.

Measures:

The study examined two self-reported outcomes of alcohol consumption and alcohol problems that, based on other studies, had good psychometric properties. The authors took special steps to ensure confidentiality. First, the Rutgers Alcohol Problem Inventory measured the number of alcohol-related problems in personal, social, or academic functioning. Second, the Daily Drinking Questionnaire assessed the average number of drinks per week.

The study also examined four student self-reported mediating outcomes. They were measured at the same follow-up assessment as the outcome rather than between the baseline and follow-up. First, the Protective Behavioral Strategies Survey measured the number of cognitive-behavioral coping strategies used to reduce the harm associated with alcohol consumption (i.e., limiting/stopping drinking, manner of drinking, and serious harm reduction). Second, the Drinking Norms Rating Form measured student perceptions of alcohol use among their peers. Third, the Comprehensive Effects of Alcohol measured positive and negative alcohol outcome expectancies. Fourth, the Situational Confidence Questionnaire measured self-efficacy to abstain from alcohol in high-risk drinking situations.

Analysis:

The authors conducted a multivariate analysis of covariance (MANCOVA) with covariates for baseline values of the two key outcomes. For the individual outcomes, the authors performed ANCOVAs and utilized a Bonferroni correction.

Intent-to-Treat: The analyses used all available data.

Outcomes

Implementation Fidelity:

The study noted only that the intervention sessions were carefully monitored to ensure adherence to the prescribed length of time.

Baseline Equivalence:

Table 1 shows no significant condition differences on five measures, including the two outcomes plus age, race, and sex. However, Table 3 lists several other measures obtained at baseline that were not included in the tests.

Differential Attrition:

The authors noted only minimal attrition of 3.6% but the reported Ns suggest that attrition differed substantially across conditions.

Posttest:

The MANOVA found significant differences between the two key outcomes, and follow-up ANCOVAs found significant differences between conditions for alcohol consumption but not alcohol problems. Participants in both the 10- and 50-minute interventions reported significantly fewer drinks consumed per week than the control condition, with the 10-minute condition having a stronger effect (d = .51) than the 50-minute condition (d =.31). However, the two active treatment conditions did not differ significantly from one another.

Four mediation tests for the significant outcome of alcohol consumption found two significant mediators. Changes in descriptive norms and coping skills, but not refusal self-efficacy or positive alcohol expectancies, were significantly related to the intervention and significantly mediated intervention effects on alcohol consumption.

Long-Term:

Not examined.

Study 9

This study had a special emphasis on the effectiveness of the program for a sample of students who had been mandated to receive alcohol treatment rather than for a sample of student volunteers.

Summary

Logan et al. (2015) used a randomized controlled trial to examine 61 college students who were mandated for a university alcohol program. They were randomized to BASICS (n = 18), an Alcohol Skills Training intervention (n = 22), or a treatment-as-usual Alcohol Diversion intervention (n = 16). The students were followed for six months and assessed for alcohol use and alcohol problems.

Logan et al. (2015) found that the two interventions groups combined (BASICS and alcohol skills training), when compared to the control group, had significantly lower

• Estimated blood alcohol concentration levels.

Evaluation Methodology

Design:

Recruitment: The undergraduates (ages 18+) recruited for the study came from a university in the southern United States. The students had been referred to Judicial Affairs after violating campus alcohol policy and mandated to take part in the university Alcohol Diversion Program. Joining the study would meet the university requirement. Of the 90 students contacted, 61 (67.8%) consented and completed baseline measures. The recruitment process took place during the academic year from October 2009 to April 2010.

Assignment: Participants were randomized to three conditions: a BASICS personalized feedback intervention (n = 18), an Alcohol Skills Training Program intervention (n = 22), or a treatment-as-usual Alcohol Diversion Program intervention (n = 16). These sample sizes do not include five participants who did not complete an intervention. The Alcohol Skills Training Program used motivational interviewing in a group session and a harm-reduction focus, while BASICS used motivational interviewing in a one-on-one counseling session and a personalized feedback focus. The Alcohol Diversion Program focused on education, was delivered by the campus police department, and did not include motivational interviewing techniques.

Assessments/Attrition: Participants completed web assessments at baseline and follow-ups at two, four, and six months after baseline. About 58% completed all three follow-ups, 21% completed some follow-ups, and 21% did not complete any follow-up. The outcome results in Table 2 listed Ns of 45-56 (74-91%).

Sample:

The sample had an average age of 19.2 years and was 42.6% female, 96.7% White, 59.0% freshmen, and 27.9% sophomores.

Measures:

Alcohol use. Two self-reported outcome measures came from the Daily Drinking Questionnaire: 1) estimated blood alcohol concentration based on number of drinks, hours drinking, sex, and weight, and 2) weekly drink totals calculated by adding the number of typical drinks on each of the seven days.

Consequences. One self-reported outcome measure came from the Rutgers Alcohol Problem Index, which assessed alcohol-related consequences (e.g., hangovers, missing classes, getting into fights) within the past two months (alphas ranged from .85 to .94 over the multiple follow-ups).

Analysis:

The analysis used hierarchical linear models to estimate the trajectories over time in the outcomes. Interactions between interventions and time tested for program effects. Two notably inconsistent outliers at four months had extremely high consequences but low blood concentration levels and were treated as missing. Square root transformations were performed for nonnormal distributions of the outcome variables.

Intent-to-Treat: It was unclear how the study treated the five participants (8.2%) who did not complete the intervention, but otherwise the multilevel models allowed the use of subjects who did not complete all four surveys.

Outcomes

Implementation Fidelity:

Most participants (62.5%) completed the intervention before the two-month follow-up, with 25.0% completing before the four-month follow-up and 6.3% completing before the six-month follow-up. The authors noted (p. 35) that occurrence of the intervention at various time points raised some concerns with time delay. Based on independent coding of session tapes, the interventions did more to follow motivational interviewing requirements than the control group, and BASICS used one key motivational interviewing requirement, complex reflections, more than the Alcohol Skills Training Program did.

Baseline Equivalence:

The authors stated that "There were no baseline differences between the three conditions on demographics, alcohol use, or consequences" but did not present figures.

Differential Attrition:

Completers and non-completers did not differ by condition or baseline measures with two exceptions: Men were more likely not to complete the follow-ups and Alcohol Diversion Program participants were more likely not to complete the follow-ups if they had high baseline drinking scores.

Posttest:

For one of three outcomes, estimated blood alcohol concentration, the combined BASICS and Alcohol Skills Training groups scored significantly lower than the treatment-as-usual control group. The two interventions did not differ from one another. The study did not report tests for BASICS alone relative to the control group.

Long-Term:

Not examined.

Study 10

Summary

Eggleston (2007) used a randomized controlled trial that assigned 299 students at one university to three conditions: BASICS with risk and normative feedback, revised BASICS with normative feedback alone, or an assessment-only control group. However, consent followed assignment, reducing the sample to 115. An assessment six months after baseline measured alcohol use and alcohol-related problems.

Eggleston (2007) found no significant differences between the intervention groups and the control group at the six-month follow-up for any of the nine outcomes.

Evaluation Methodology

Design:

Recruitment: College students (n = 1,197) enrolled in an introductory psychology class at Ohio State University in 2005-2006 were screened for alcohol consumption and alcohol-related problems. Eligible students (n = 299) were 18 years or older and fell in the top quartile of the screening sample in terms of alcohol use severity.

Assignment: The 299 eligible students were randomly assigned within strata (gender and alcohol composite score) to one of three study conditions: BASICS with risk and normative feedback, revised BASICS with normative feedback alone, or an assessment-only control group. However, as consent followed assignment, only 115 of the 299 randomized students joined the study by completing the baseline assessment. Analyses revealed no significant differences between those agreeing and not agreeing to participate on gender, age, and five alcohol-related measures.

Assessments/Attrition: Participants were re-assessed six months after the baseline assessment. Only 38 of the 115 participants (33%) completed the follow-up assessment, and individual measures had additional missing data, up to 16% for one.

Sample:

The sample came from students at one university. Of the 115 students who agreed to participate, 61% were female, and the average age was 19.0 years. The study offered no other information on the sociodemographic makeup of the sample.

Measures:

The study examined four measures of alcohol use, two measures of alcohol-related problems, two measures of alcohol abuse or dependence, and one measure of perceived risk of alcohol use. All measures came from commonly used instruments and most had evidence of good reliability.

Analysis:

The analysis used regression models that entered predictors in the following order: demographic characteristics, initial value of the dependent variable, and study condition. Missing follow-up data were imputed with the first-observation-carried-forward approach.

Intent-to-Treat: The intent-to-treat analysis imputed missing follow-up data to include all participants with baseline data.

Outcomes

Implementation Fidelity:

Ratings of the session content indicated that 100% of the risk-normative condition sessions and 92% of normative-feedback condition sessions had full compliance

Baseline Equivalence:

Tests for baseline equivalence in Table 3.2 that used the 113 participants (of 115 total) with baseline data showed no significant differences across the three conditions for 21 measures.

Differential Attrition:

Tests for differential attrition in Table 3.3 found no significant differences between dropouts and completers for 22 measures.

Posttest:

For the ITT sample of 115 with imputed data, there were no differences between the intervention groups and the control group at follow-up for any of the nine outcomes. There was some evidence of moderation, "with individuals who consumed more alcohol or experienced more alcohol-related problems" at baseline "being more likely to have a negative response to treatment."

Long-Term:

Not examined.

Study 11

Summary

Simão et al. (2008) used a randomized controlled trial that assigned 334 Brazilian university students at-risk for alcohol problems to a BASICS intervention group or a no-treatment control group but only 266 students remained at baseline for study. Alcohol use and alcohol-related problems were assessed at 12- and 24-month follow-ups.

Simão et al. (2008) found significantly greater declines in the intervention group than the control group for measures of

• The number of drinks per occasion
• Frequency of drinking
• At-risk drinking.

Evaluation Methodology

Design:

Recruitment: From 2000 to 2004, all 5,052 first-year students from seven campuses of Sao Paulo State University were invited to participate in a screening for alcohol consumption. From 4,100 responders, 1,057 met the criteria for at-risk drinkers (i.e., high scores on an alcohol use disorders screening measure and five or more harmful consequences related to alcohol use according to an alcohol problems measure). Using only freshman students from the Biological Sciences campus, the study invited 334 to take part in this project and all consented.

Assignment: The 334 students were randomly assigned to a BASICS intervention group or a no-treatment control group. However, the loss of subjects reduced the sample size to 266, with 145 in the intervention group and 121 in the control group. The loss came from students who dropped out of college, refused to participate, could not be contacted, or reported alcohol dependence or abstinence.

Assessments/Attrition: Follow-up assessments came 12 and 24 months after baseline. Figures on attrition were reported for the reduced sample of 266 students. At 12 months, 263 (99%) completed the assessment, and at 24 months, 216 (81%) completed the assessment. Relative to the randomized sample of 334, 79% completed the 12-month assessment, and 65% completed the 24-month assessment.

Sample:

The sample students had a mean age of 19.6 years and a gender breakdown of 56% male and 44% female. All were single, most lived with roommates (67.3%), and nearly all did not work (90%). About 26% had a history of alcohol abuse.

Measures:

The study used self-reported measures that came from validated instruments, though few details on validity and reliability for the sample were presented.

Measures from the Brief Drinking Profile included:

• Number of drinks per occasion
• Frequency
• Number of drinks in the past 30 days
• Number of drinks per weekend

Other measures included the AUDIT score for at-risk drinking and the RAPI score for harmful alcohol consequences.

Analysis:

The repeated-measures analysis of covariance models included covariates for gender and the baseline outcome plus an interaction term for treatment-by-time.

Intent-to-Treat: The study excluded 5.6% of participants who after consenting refused to participate, and it did not follow the 10.0% who dropped out of college. The analysis otherwise used all available data with listwise deletion.

Outcomes

Implementation Fidelity:

Not examined.

Baseline Equivalence:

Multivariate analyses of variance for six variables at baseline showed a significant overall difference between the treatment group and the control group. The study did not report individual tests for the six baseline measures, but the control group had consistently higher scores on the measures of alcohol use and alcohol-related problems.

Differential Attrition:

The study did not present tests for differential attrition, but the high attrition, difference across conditions in the size of the analysis sample at baseline (55% versus 45%), and difference in attrition rates across conditions (29% versus 7%) indicate a problem.

Posttest and Long-Term:

The analysis did not distinguish between tests at 12 and 24 months. The treatment-by-time interactions showed significantly greater declines in the intervention group than the control group for five of six outcomes (number of drinks per occasion, frequency, number of drinks per weekend, at-risk drinking, and harmful alcohol consequences but not number of drinks in the past 30 days). For two outcomes (number of drinks per occasion and at-risk drinking), however, the means at 24 months appear higher for the intervention group than the control group, suggesting possible iatrogenic effects.

Study 12

Terlecki (2008) presented preliminary results, while Terlecki et al. (2010) presented the final results for the full sample.

Summary

Terlecki (2008) and Terlecki et al. (2010) used a randomized controlled trial that assigned 92 university students to four conditions: mandated treatment, mandated wait-list control, voluntary treatment, or voluntary assessment-only control. A posttest assessment four weeks post-intervention measured alcohol use and alcohol-related problems.

Terlecki (2008) and Terlecki et al. (2010) found that the intervention group had lower scores than the control group for

• Drinking quantity and peak drinks per occasion at four weeks.

Evaluation Methodology

Design:

Recruitment: Over several consecutive semesters, the study recruited two separate groups of students, one mandated and one volunteer. First, the mandated sample came from 59 students who had violated a campus alcohol policy (e.g., public drunkenness) and were screened for heavy drinking. Of those screened, 49 were eligible and 47 enrolled in the study. Second, the volunteer sample came from 98 students who were in psychology classes or responded to ads. Of those screened for heavy drinking, 53 were eligible and 45 enrolled. The total enrolled sample thus equaled 92.

Eligible students (a) reported high levels of drinking, (b) endorsed at least three alcohol-related problems in the past three years, and (c) received high scores on a measure of at-risk alcohol use.

Exclusion criteria included self-reported treatment for alcohol use or alcohol-related problems, endorsement of severe and persistent alcohol-use patterns (e.g., physiological alcohol dependence), request for more intensive treatment (e.g., intensive outpatient program), and self-reported history of an alcohol-related disciplinary event for student volunteers.

Assignment: Random assignment occurred within the groups of mandated and voluntary students to define four conditions: mandated treatment (n = 19), mandated wait-list control (n = 24), voluntary treatment (n = 22), or voluntary assessment-only control (n = 19). The sample sizes refer to the numbers who completed the study, as the randomized numbers were not reported.

Assessments/Attrition: A follow-up assessment came four weeks after the program for the intervention groups and six weeks after baseline for the control groups. The extra two weeks for the control groups accounted for a two-week period of alcohol self-monitoring between the assessment and feedback interviews for the intervention groups. Of the 92 enrolled and randomized, 84 (91%) completed the posttest survey.

Sample:

The sample was 62% male, about 84% white, and about 20 years old on average (range 18-24).

Measures:

The measures came from self-report surveys obtained via the Internet. The study used standard measures but provided few details on reliability or validity for the sample.

Measures from the Daily Drinking Questionnaire, the Quantity/Frequency Peak Index, and the Rutgers Alcohol Problems Index included: 1) quantity of typical weekly alcohol consumption; 2) frequency of consumption; 3) typical quantity per drinking occasion; 4) peak drinks per occasion; and 5) negative consequences related to alcohol.

Analysis:

The analysis of covariance models for the posttest outcomes included the baseline outcomes and intervention condition as covariates. Significance tests adjusted for multiple tests with a .0125 alpha level.

Intent-to-Treat: According to Table 2 in Terlecki et al. (2010), participants who did not attend the baseline assessment or feedback session were excluded. The latter exclusion appears to violate the intent-to-treat criterion, but with both exclusions making up only 5% of the randomized sample, the violation may not be important.

Outcomes

Implementation Fidelity:

Not examined.

Baseline Equivalence:

Of six sociodemographic measures and one alcohol risk measure, two differed significantly across conditions at baseline. For example, the mandated intervention and control groups were 79% and 92% male, respectively, while the volunteer intervention and control groups 46% and 26% male, respectively. Fraternity/sorority membership also differed significantly across conditions. Of the tests for baseline differences on the five outcomes, one was significant (typical alcohol consumption).

Differential Attrition:

Not examined.

Posttest:

The results in Table 3 of Terlecki et al. (2010) showed two significant posttest effects in five tests of the intervention (with both mandated and volunteer students combined): typical drinking quantity (d = .36) and peak drinks per occasion (d = .42). The pairwise tests gave less clear-cut results. For quantity of drinking, the intervention group scored significantly lower than the control group among volunteer students but not mandated students. For peak drinks, the comparisons showed no significant differences between the intervention and control groups among either mandated or volunteer students.

Long-Term:

Not examined.

Study 13

The dissertation of Terlecki (2011) presented the same tables and much of the same text as the article of Terlecki et al. (2015). The more compact version of Terlecki et al. (2015) served as the basis for the details below, with some unique results on readiness to change taken from Terlecki (2011). Terlecki et al. (2021) examined a subset of the data from the two earlier articles.

Summary

Terlecki (2011) and Terlecki et al. (2015, 2021) used a randomized controlled trial that assigned 255 university students to four conditions: mandated treatment, mandated wait-list control, voluntary treatment, or voluntary assessment-only control. Posttest assessments ranging from four weeks to 12 months post-intervention measured alcohol use and alcohol-related problems.

Terlecki (2011) and Terlecki et al. (2015, 2021) found that the intervention group had lower scores than the control group for

• Drinks per week, typical drinks, and peak drinks at four weeks
• Typical drinks, peak drinks, and alcohol problems at 12 months.

Evaluation Methodology

Design:

Recruitment: The study recruited two groups of students between ages 18-24, one mandated and one volunteer. First, the mandated sample came from students who had violated a campus alcohol policy (e.g., public drunkenness) and were screened for heavy drinking. Second, the volunteer sample came from students who were in psychology classes or responded to ads. Of the 550 who expressed interest in the study, 520 completed the eligibility screening, 309 met eligibility criteria, and 255 (83% of those eligible) enrolled in the trial.

Inclusion criteria for mandated and voluntary participants were (a) drinking at least monthly and endorsing past month binge drinking in the past year; (b) reporting at least three alcohol-related problems on three to five occasions in the past year; and (c) scoring high on the Alcohol Use Disorder Identification Test of risky drinking practices. Exclusion criteria included serious alcohol use disorder symptoms, such as physiological dependence, and previous alcohol-related disciplinary referrals.

Assignment: Computer-based urn randomization was used to assign the 255 students to a BASICS intervention group (n = 131) or a control group (n = 124). The volunteer control group served as an assessment-only condition, while the mandated control group received BASICS after a six-week waitlist period. Completion of the baseline assessment came after randomization, but the completion rates, 88% in the intervention group and 89% in the control group, were nearly identical.

Assessments/Attrition: Assessments came at baseline and four weeks, three months, six months, and 12 months after the program session. Because intervention participants attended the BASICS session about two weeks after the baseline assessment, the control group students actually completed the first follow-up assessment six weeks after baseline.

Of the 255 randomized students, 211 (83%) completed the four-week survey. The subsequent follow-ups excluded an unreported number of students who were assigned to the waitlisted mandatory control group and received the program. Considering only the 131 students randomized to the intervention group, 74% completed the three-month survey, 64% completed the six-month survey, and 56% completed the 12-month survey.

Sample:

The sample was 60% male, about 88% white, and about 20 years old on average (range 18-24).

Measures:

The outcomes came from self-report surveys obtained via the Internet. The study used standard measures but provided few details on reliability or validity for the sample.

Measures from the Daily Drinking Questionnaire, the Quantity/Frequency Peak Index, and the Rutgers Alcohol Problems Index included: 1) quantity of drinks per week; 2) frequency of drinking; 3) typical drinks per occasion; 4) peak drinks per occasion; and 5) alcohol-related problems.

To measure readiness to change, Terlecki (2011) summed questions about contemplation and action to reduce alcohol consumption.

Analysis:

The analysis presented separate results from baseline to four weeks and from four weeks to 12 months because the long-term follow-up excluded the waitlisted control group students who started the program after four weeks. The long-term model examined if the four-week change was sustained. For both periods, multilevel longitudinal models adjusted for over-time correlations for each subject. The predictors included condition, referral group (mandatory, volunteer), time, gender, Greek-system membership, and a time-by-condition term labeled as the intervention slope in the table. Full information maximum likelihood estimation and imputation using the last observation carried forward allowed the analysis to include all participants with baseline data.

Intent-to-Treat: The analysis included all participants who completed the baseline survey regardless of their attendance at the intervention session. Only the 12% with no baseline data had to be excluded.

Outcomes

Implementation Fidelity:

A checklist review provided by the counselors found that 100% of three core BASICS components were addressed in the intervention sessions and that 97% of three other core components were addressed in the sessions. About 85% of the intervention participants attended the intervention session. 

Baseline Equivalence:

In eight tests across four conditions for the sample completing the baseline assessment (Table 1), one significant difference emerged for gender, but the difference stemmed from the comparison of mandated and voluntary students rather than from the comparison of intervention and control students. However, a large but non-significant difference appeared in the voluntary sample for Greek membership - 30% in the intervention group versus 41% in the control group. For the five outcomes, the authors reported that, with a control for gender, the control condition had a significantly higher score on the baseline measure of alcohol problems than the intervention group, but they also reported that the multilevel models in Table 3 showed no intercept (i.e., baseline) condition differences for the five outcomes.

Differential Attrition:

Attrition analyses indicated that, relative to completers, participants who did not complete the 12-month follow-up assessment were significantly more likely to be former Greek-system members and mandated students and to have reported fewer baseline peak drinks. No significant differences were found between completers and non-completers for condition, age, gender, race, ethnicity, year in school, living situation, or other baseline intervention outcomes.

Posttest:

The four-week results in Table 3 (Terlecki et al., 2015) show that of the five outcomes, the intervention students reported significantly fewer drinks per week (d = .59), typical drinks (d = .76), and peak drinks (d = .92) than control students. The effect sizes are calculated relative to the assessment-only control group rather than both the assessment-only and waitlist control groups. The effects were similar for both the mandated and volunteer samples.

Terlecki (2011) reported that the intervention did not significantly affect readiness to change and that readiness to change did not moderate the intervention effect.

Terlecki et al. (2021) examined a subset of the original sample of 255 students. The subset included 120 students who scored either high or low on social anxiety measures (n = 63 for the intervention group and n = 57 for the control group). The study added measures of protective behavior strategies used to prevent the harm of heavy drinking (alpha = .79). Tests for baseline equivalence among the subset showed no significant differences on five sociodemographic measures (Table 1) or eight outcome measures (Table 2). However, the conditions appeared to differ substantially on anxiety: The intervention group had 52% with high anxiety versus 67% in the control group.

In analyzing the four-week posttest data, the study used multivariate analysis of covariance for five drinking outcomes (typical drinks, peak drinks, weekly drinking quantity, drinking frequency, and alcohol problems) and three protective strategies outcomes (stopping/limiting, manner of drinking, and serious harm reduction). As reported in the text (p. 14), four of the five drinking outcomes and one of three protective strategies outcomes were significant and favored the intervention group. Moderation tests showed stronger intervention effects for the low anxiety group, suggesting that high anxiety obstructed the efficacy of the program.

Long-Term:

The results in Table 3 (Terlecki et al., 2015) for the four-week through the 12-month assessments are difficult to follow given the labels used in the tables and the complexities created by exclusion of the waitlisted mandatory control students. However, the row labeled intervention slope appears to show the time-by-condition interactions. These coefficients indicated beneficial effects of the program through 12 months for three of five outcomes: typical drinks (d = .11), peak drinks (d = .42), and alcohol problems (d = .56). The effects appear to be similar whether the mandatory and voluntary intervention students were combined or kept separate.

Study 14

Summary

Terlecki et al. (2011) used a randomized controlled trial that assigned 110 university students to intervention (n = 55) and control (n = 55) conditions. Posttest assessments at four weeks post-intervention measured alcohol use and alcohol-related problems.

Terlecki et al. (2014) found that the intervention group had significantly lower scores at the four-week posttest than the control group for

• Weekly drinking quantity
• Alcohol problems.

Evaluation Methodology

Design:

Recruitment: The study recruited undergraduate students aged 18 to 24 at a large southern public university. Recruits came from the psychology research participation pool for students meeting course requirements and from the Office of Judicial Affairs for students satisfying a requirement for an alcohol-related violation. Of 158 students who were screened, 110 met the inclusion and exclusion criteria.

The inclusion criteria consisted of (a) consuming five to six drinks per occasion, (b) endorsing three or more alcohol-related problems on three to five occasions in the past three years; and (c) scoring in the harmful/hazardous drinking range on the Alcohol Use Disorder Identification Test. The exclusion criteria consisted of (a) chronic alcohol use (e.g., alcohol dependence) that warranted more intensive treatment, (b) requesting more intensive alcohol use treatment, and (c) reporting a previous campus alcohol policy violation among volunteers recruited via the psychology department research pool.

Assignment: The study randomized the 110 participants to an intervention group (n = 55) or an assessment-only control group (n = 55). The baseline assessment came after randomization, with 105 of 110 (95%) completing the baseline survey.

Assessments/Attrition: The follow-up came six weeks after baseline or four weeks after the program session for the intervention group. The timing adjusted for the two-week period of alcohol self-monitoring that occurred between the baseline assessment and the intervention session. Figure 1 lists 91 participants (83%) as having completed the follow-up assessment but only 70 (64%) as being included in the analysis.

Sample:

The final sample of 70 students was 85.7% white, 20.55 years old on average, 68.6% male, 64.3% junior or senior, 74.3% living off-campus, and 37.1% Greek system members.

Measures:

The study used standard measures but provided few details on reliability or validity for the sample. The self-reported alcohol consumption measures, obtained via an internet survey, came from the Daily Drinking Questionnaire, the Quantity/Frequency/Peak Index, and the Rutgers Alcohol Problem Index. The authors computed five drinking outcome measures from the instruments: weekly quantity, weekly frequency, typical drinks per occasion, peak drinks per occasion, and alcohol-related problems.

Analysis:

The tests for main effects examined mean condition differences at posttest without covariates. The tests for moderation used analyses of covariance with covariates for referral group, trait anxiety scores, and pretest scores of the outcome variables. A Bonferroni correction adjusted for multiple comparisons.

Intent-to-Treat: According to Figure 1, three intervention participants (3%) who did not complete the session were dropped but only because they could not be contacted. In addition, the study did not explain the discrepancy between the 91 students completing the follow-up and the 70 used in the analysis.

Outcomes

Implementation Fidelity:

A review of the individual sessions showed that three core BASICS components were covered in 100% of intervention sessions and that three others were covered in 94% to 98% of sessions. For participant adherence, 95.4% completed assessment self-report measures and attended the assessment interview.

Baseline Equivalence:

Table 1 shows no significant condition differences in seven tests for baseline drinking outcomes and social anxiety measures. The tests did not include sociodemographic measures.

Differential Attrition:

The study did not present tests for baseline differences between dropouts and completers, but the CONSORT diagram indicates a higher completion rate in the intervention group (69%) than in the control group (58%).

Posttest:

For the full sample, univariate tests (without covariates) indicated that the intervention group reported significantly lower scores on two of the five outcomes: weekly drinking quantity (d = .63) and alcohol problems (d = .80). Moderation tests found that social anxiety interacted significantly with condition for one of the five outcomes (typical drink consumption). The moderation revealed a stronger intervention effect among those with low baseline social anxiety.

Long-Term:

Not examined.

Study 15

The sample in this study differed from other studies in selecting heavy drinking students with emotional problems.

Summary

Whiteside (2010) used a randomized controlled trial that assigned 145 university students to a BASICS group, a revised BASICS group with an emotional control component, and a control group. Posttest assessments at one month and three months measured alcohol use, alcohol-related problems, and emotional problems.

Whiteside (2010) found that the BASICS only intervention group had significantly higher scores than the control group for

• Emotional regulation at the one-month follow-up.

Evaluation Methodology

Design:

Recruitment: Participants were voluntary members of the University of Washington Psychology Department Subject Pool who met the screening criteria of heavy drinking and symptoms of depression or anxiety. A total of 2,336 students were screened from autumn 2007 to autumn 2008, and 202 (8.6%) met the screening criteria. Participants were considered eligible if they binge drank at least once in the past month, drank at least weekly, reported that at least "some of the time" they drank to deal with negative emotions, and had a high score on the Beck Depression Inventory or the Beck Anxiety Inventory.

Assignment: The 202 participants were randomly assigned to BASICS (n = 67), BASICS with Dialectic Behavior Therapy that added content on emotional control (n = 67), or a relaxation control condition (n = 66). However, consent and the baseline assessment came after randomization, leaving 145 (71.78%) of the 202 randomized participants. The author stated that "There were no differences between those eligible and enrolled versus those eligible but not enrolled on any demographic characteristics or screening criteria." By condition, participation rates varied from 64% for the BASICS with Dialectic Behavior Therapy group to 79% in the control group, but the authors stated that "There were no statistically significant differences between conditions in baseline completion."

Assessments/Attrition: Assessment occurred at baseline and at one-month and three-months post-baseline. According to the CONSORT diagram, of the 145 participants with baseline data, 113 (78%) and 101 (70%) completed the one-month and three-month follow-up assessments, respectively.

Sample:

The sample included 40% men and 60% women, and it ranged in age from 17 to 26 years. Participants were 66% Caucasian, 26% Asian, 5% Multi-Racial, and 5% Hispanic.

Measures:

The study used standard measures with good alpha reliabilities (excepting a value of .55 for weekly and binge drinking). Drinking measures came from the Drinking Motives Questionnaire, the Harvard Alcohol Survey, and the Rutgers Alcohol Problem Index. The depression and anxiety measures came from the Beck Depression Inventory, Beck Anxiety Inventory, and Difficulties in Emotion Regulation Scale. The eight outcomes included depression, anxiety, emotion regulation, alcohol-related problems, coping drinking motives, binge drinking, drinks per week, and peak blood alcohol content.

Analysis:

The analyses used ANCOVA models with baseline outcomes as covariates.

Intent-to-Treat: Outside the loss of subjects between randomization and the baseline assessment, the study appeared to use all available data, regardless of completion of the intervention session.

Outcomes

Implementation Fidelity:

All sessions were reviewed for adherence and therapists received written feedback, but the study presented no quantitative information. The CONSORT diagram shows that 94% of those completing the baseline assessment attended the intervention session.      

Baseline Equivalence:

Table 1 presents the baseline mean outcomes across the three conditions but no tests. The text points to "a failure of randomization for the drinks per week outcome at baseline, such that those in the BASICS condition consumed significantly more drinks per week" than the control condition. No other mention of significant baseline differences suggests that this was the only non-equivalent measure, but the study was not explicit.

Differential Attrition:

Attrition rates did not differ significantly across conditions at either the one-month or three-month follow-ups. No other tests were presented.

Posttest:

In overall tests for differences across all three conditions, four reached significance at one month but only one (depression) reached significance at three months. In pairwise tests comparing BASICS to the control group, one (emotional regulation, d = .36) reached significance at one month and one (coping drinking motives, d = .40) reached significance at three months. However, the overall test across the three conditions was not significant at three months for coping drinking motives. At three months, then, none of the outcomes showed both a significant difference across all three conditions and a significant difference between BASICS and the control group. The effects of the revised BASICS with Dialectic Behavior Therapy to address emotional problems showed more consistent effects relative to the control group than BASICS alone.

Long-Term:

Not examined.

Study 16

This study consists of two parts, the first a general survey of drinking behavior and the second an experimental evaluation of BASICS. The write-up examines only the second part.

Summary

McPherson (2012) used a randomized controlled trial that assigned 37 Australian university students to a BASICS group, a brief online alcohol intervention group, and a waitlisted control group. Posttest assessments at one month and three months measured alcohol use, alcohol-related problems, and readiness to change.

McPherson (2012) found no significant posttest differences that favored the intervention group over the control group for any of the seven outcomes.

Evaluation Methodology

Design:

Recruitment: Based on responses to a general screening survey of 1,046 students at the Royal Melbourne Institute of Technology, 182 registered interest in participating in the study, and 90 were eligible because of self-reported moderate or high risk of hazardous drinking (but not alcohol dependence). Of the 90, only 37 consented to join the study.

Assignment: The study randomly assigned the 90 students to one of three conditions: the e-CHUG online intervention (n = 30), the BASICS brief motivational feedback intervention (n = 30), or a waitlist control condition (n = 30). One therapist delivered the intervention to all those assigned to BASICS.

Assessments/Attrition: Assessments were conducted at baseline and at one month and three months after baseline. Of the 90 randomized students, only 29 (32%) completed the one-month survey, and only 24 (27%) completed the three-month survey.

Sample:

The sample was about 30% males, had an average age of about 27 years, and had completed about 3.2 years of tertiary schooling.

Measures:

The study used standard measures from student self-reports obtained via the Internet and reported minimal information on reliability and validity. For the four main outcomes of alcohol risk, modified alcohol risk, readiness to change, and alcohol-related problems, the participants completed the Alcohol Use Disorder Identification Test, Readiness to Change Questionnaire, and Rutgers Alcohol Problems Inventory. In addition, the two intervention groups only completed measures of peak alcohol consumption, weekly alcohol consumption, and estimated BAC.

Analysis:

The study conducted an analysis of variance, with a time measure that included the baseline outcome and the two follow-up outcomes. The key time-by-condition interaction coefficients were not consistently reported, however. Multiple imputation replaced missing data for those who completed the follow-up surveys but not for dropouts.

Intent-to-Treat: The analyses could not include participants who failed to complete the baseline survey and attend the intervention session.

Outcomes

Implementation Fidelity:

Not examined.

Baseline Equivalence:

The author reported that there were no significant differences between the conditions on measures of age, gender, readiness to change, completion of tertiary school, and alcohol use disorders. However, the tests used the analysis sample, and even non-significant differences were large: 14% of BASICS versus 45% of the control group were male (Table 22), and 29% of BASICS versus 40% of the control group reported hazardous or harmful drinking (Table 23). Also, the text noted significant differences between the BASICS and e-CHUG groups on baseline measures of peak alcohol consumption and average weekly alcohol consumption.

Differential Attrition:

Little's MCAR test showed the data to be missing completely at random. Comparison of those "who volunteered for intervention but did not commence, those who commenced intervention but did not complete and those that completed the intervention" showed no significant differences on sociodemographic and alcohol screening measures. Yet, attrition differed substantially across conditions. It was 17% in BASICS, 23% in e-CHUG, and 40% in the control group. Also, differences at baseline between conditions for the analysis sample suggest the presence of differential attrition

Posttest:

In tests for four outcomes using all three conditions, only one showed a significant difference, but the control group reported lower scores at follow-up on the Alcohol Use Disorder Identification Test than either intervention group.

Analyses on three drinking behavior outcomes excluded the control group, which was not surveyed on these items. Although the BASICS group appeared to do better than the e-CHUG group, none of the differences reached statistical significance.

Long-Term:

Not examined.

Study 17

Summary

Borsari (2003) and Borsari and Carey (2005) used a randomized controlled trial that assigned 72 students who had violated university alcohol policies to a BASICS intervention group or an alcohol-education control group. Posttest assessments three months and six months after baseline measured alcohol use and alcohol-related problems.

Borsari (2003) and Borsari and Carey (2005) found that, relative to the control group, the BASICS intervention group reported significantly

• Fewer alcohol-related problems at six months.

Evaluation Methodology

Design:

Recruitment: Recruitment occurred at two universities in one metropolitan area in the northeastern United States. Students who had violated school alcohol policy were recruited from Fall 2000 to Spring 2002 and screened for high-risk drinking. Those eligible for the study scored high on the Alcohol Use Disorders Identification Test or reported two or more binge-drinking episodes in the past 30 days. Those who had been referred more than once for excessive drinking, had requested more intensive treatment, or primarily used other substances (e.g., marijuana) were excluded. A total of 72 students were invited to participate in the project.

Assignment: The eligible students who agreed to participate were randomly assigned to the intervention or control group and then completed the baseline assessment (Borsari, 2003, p. 38). The control group students received alcohol education but no attempts were made to elicit personal information or facilitate problem recognition. One person delivered both interventions, possibly confounding the design. However, eight students (11% of the 72 randomized) were dropped after randomization, leaving 34 in the intervention group and 30 in the control group. Of the lost students, one declined to participate, two were referred for more intensive treatment following the baseline assessment, two students did not return for the intervention, and three did not return for a follow-up assessment.

Assessments/Attrition: The study included baseline, three-month, and six-month assessments. Of the 72 eligible students, 64 (89%) completed the baseline assessment, 60 (83%) completed the three-month assessment, and 57 (79%) completed the six-month assessment.

Sample:

The sample was about 18% female, 92% Caucasian, and averaged 19.1 years of age. The participants reported an average of about 3.2 binge drinking events per month and 20 drinks per week.

Measures:

The study used standard measures of five drinking outcomes: number of drinks consumed per week, the frequency of binge drinking in the past 30 days, typical blood alcohol content, peak blood alcohol content, and alcohol-related problems. Based on data collected from close friends, "it did not appear that students systematically misrepresented their alcohol use."

Borsari (2003) included an additional scale of drinking knowledge.

Analysis:

The analysis used hierarchical linear modeling in which time was nested within individuals and time-by-treatment terms included the baseline values of the outcomes and measured intervention impact. The models controlled for gender, campus, and days between the infraction and intervention. Tests of significance used the false discovery rate to adjust for multiple comparisons.

Intent-to-Treat: The study used all available data, including those who completed only one follow-up. However, the 11% who did not complete the baseline and intervention were not followed.

Outcomes

Implementation Fidelity:

The data in Table 1, which came from the coding of audiotapes, demonstrated that the intervention sessions covered a high percentage of the content (median 93%).

Baseline Equivalence:

In the 16 tests for condition differences in sociodemographics and baseline drinking, two were significant. The intervention group had higher scores on the screening measure of high-risk drinking and on typical blood alcohol level at baseline. Other figures in the table did not include significance tests but revealed substantial differences. For example, 44% of the intervention group versus 33% of the control group reported being drunk in public.

Differential Attrition:

The authors reported that "Attrition analyses revealed no baseline differences between participants who completed the study and those who did not and between participants who completed one versus two follow-ups."

Posttest:

For four drinking variables, none of the time-by-group effects were significant. For alcohol-related problems, there was a significant time-by-group effect, and contrasts between baseline and the six-month follow-up indicated a significant between-group difference (d = .39). The contrasts over five outcomes and two follow-ups thus indicated that only one of 10 tests reached significance.

For the measure of alcohol knowledge, the time-by-group interaction was not significant, and the intervention group did not demonstrate greater knowledge at either follow-up.

Tests in Borsari et al. (2003) found no evidence of either mediation or moderation of intervention effects.

Long-Term:

Not examined.

Study 18

Summary

Butler (2007) and Butler and Correia (2009) used a randomized controlled trial that assigned 114 students who were high-risk drinkers to three conditions: BASICS in-person, computerized BASICS, or an assessment-only control group. A posttest assessment four weeks after baseline measured alcohol use and alcohol-related problems.

Butler (2007) and Butler and Correia (2009) found that, relative to the control group, the BASICS in-person intervention group reported significantly fewer

• Drinking occasions, binge episodes, and drinks per week at four weeks.

Evaluation Methodology

Design:

Recruitment: The study screened 300 students enrolled in undergraduate courses at one university. Eligible students reported high-risk drinking defined as having at least two binge episodes and two alcohol-related problems in the past 28 days. Of the 300 screened students, 114 met the eligibility criteria and were invited to participate, but 30 (26%) were not included in the study. Of the 30, 10 did not reply to the invitation, four did not attend the session, and 16 did not attend the follow-up.

Assignment: The 114 students invited to participate were randomized within gender into three conditions: in-person BASICS (n = 28), computerized BASICS (n = 30), and an assessment only control group (n = 26). The reported sample sizes refer to the analysis sample, after dropping those who did not attend the sessions. All of the face-to-face interventions were conducted by a single graduate clinician, thus confounding the intervention with the person delivering it.

Assessments/Attrition: Assessments came at baseline and four weeks after the intervention. Attrition consisted of those lost before intervention (12%) and those lost after (14%).

Sample:

The sample was about 65% female and about 91% white. The students on average were about 20 years old.

Measures:

The study used standard measures from the Daily Drinking Questionnaire and the Rutgers Alcohol Problem Index, though it provided little information on validity. The four outcome measures included alcohol use days, binge drinking days per month, standard drinks consumed during a typical week, and alcohol-related consequences.

Analysis:

Butler & Correia (2009) used analysis of covariance models with the baseline outcome included as a covariate, while Butler (2007) used one-way analysis of variance with no baseline control.

Intent-to-Treat: The study did not employ "an intent-to-treat approach, as only individuals who completed the intervention were included in the follow-up analyses."

Outcomes

Implementation Fidelity:

Not examined.

Baseline Equivalence:

The tests used the analysis sample of 84 and showed one significant difference for nine measures. The face-to-face group reported a significantly lower level of mean years of education than the other groups.

Differential Attrition:

The tests for baseline equivalence offer some evidence that differential attrition was modest. In addition, the authors stated that "the 84 participants who completed the study did not differ from the 30 eligible participants who did not complete the study, or from the subset of 20 participants who were randomly assigned and scheduled for a feedback and/or a follow-up session but did not attend, on any of the measured variables."

Posttest:

Of the four outcomes examined in the models controlling for baseline outcomes in Butler & Correia (2009), three (drinking occasions, binge episodes, and drinks per week) had lower scores for the BASICS in-person condition than the control group. The in-person and computerized conditions did not differ significantly from one another on any of the four outcomes.

Long-Term:

Not examined.

Study 19

Summary

DiFulvio et al. (2012) used a quasi-experimental design that compared 2,065 undergraduate students who participated in the program with 648 comparison students who were similar on sociodemographic measures and high-risk drinking. Assessments of alcohol use and alcohol-related problems came at baseline and six months later.

DiFulvio et al. (2012) found that the intervention group scored significantly lower than the comparison group on

• Typical, peak, and binge drinking at six months.

Evaluation Methodology

Design:

Recruitment: The study began with undergraduate students from one university who had been mandated to receive the program due to campus alcohol or other drug policy violations. Of the 2,672 students participating in the program from February 2006 through June 2008, 2,490 (93%) consented to participate in the study. After removal of 378 because they were repeat offenders of the campus alcohol policy and 47 because they self-referred to the program, 2,065 remained eligible.

Assignment: The quasi-experimental design selected a comparison group from a random sample of campus undergraduate students after stratifying by gender, class year, ethnicity, and residential status (on or off campus). Students were not matched directly, but the stratified sampling sought to obtain the same demographic percentages as the program group. A total of 1,500 students were invited to participate and 908 (61%) agreed. The comparison students were restricted to those at least 18 years of age who scored high on the Alcohol Use Disorders Identification Test and had not previously completed the BASICS program. The remaining sample of 648 (71% of those consenting) served as the comparison group.

The timing of assessments appeared to differ across the conditions. The intervention group participated in the program from February 2006 through June 2008, while the control group completed baseline and follow-up surveys in December 2006 and May 2007, respectively.

Assessments/Attrition: Along with a baseline assessment, participants completed a six-month follow-up survey. Of those eligible, 1,390 (67%) in the intervention group and 506 (78%) in the comparison group completed the follow-up.

Sample:

The sample was about 40% female, 90% white, and 19.5 years of age on average.

Measures:

The study used standard student self-report measures, though with little or no information on validity, to examine nine outcomes: single-episode drinking concentrations (typical and peak blood alcohol content), alcohol drinking (typical number of drinks and peak number of drinks), weekly cumulative alcohol consumption (number of drinks in a typical week and number of drinks in a peak-drinking week), high-risk drinking behaviors (binge and frequent binge drinking), and negative consequences associated with alcohol use.

Analysis:

The analyses used generalized linear modeling techniques with time, group, and a time-by-group interaction that included the baseline outcome.

Intent-to-Treat: The study used all available data, but the intervention group was limited in the quasi-experimental design to students who had already completed the program.

Outcomes

Implementation Fidelity:

Not examined.

Baseline Equivalence:

In Table 1, two of five sociodemographic measures differed significantly across conditions, but the differences were small in size (intervention students were older by 0.9 years and less likely to be white by 1.1%). In Table 2, four of eight outcomes differed significantly, but again the differences appeared to be relatively small.

Differential Attrition:

Attrition rates differed across conditions - 33% in the intervention group versus 22% in the comparison group. The study compared differences for dropouts and completers separately for males and females and for the intervention and comparison groups. For the intervention group, there were no significant differences for men or women. For the comparison group, men showed significant differences on all eight baseline outcomes, but women showed no significant differences.

Posttest:

The intervention group scored significantly lower than the control group on five of nine drinking outcomes (Table 5), including measures of typical, peak, and binge drinking. Tests for moderation suggested that the intervention worked better for males than females and for moderate- to high-risk drinkers than low-risk drinkers.

Long-Term:

Not examined.

Study 20

Summary

Horner (2010) used a randomized controlled trial that assigned 150 undergraduate students who were high-risk drinkers to three conditions: BASICS (n = 40), a group-based values clarification program (n = 68), or an assessment-only control group (n = 42). Posttest assessments of alcohol use and alcohol-related problems came immediately after the session, one month later, and three months later.

Horner (2010) found no significant posttest differences across conditions on any of the drinking or readiness to change outcomes.

Evaluation Methodology

Design:

Recruitment: The sample was gathered over six consecutive semesters and came from 150 undergraduate students at a Northeastern State University who were referred for violating campus drinking policies.

Assignment: After giving consent and completing the pre-assessment questionnaire, the 150 students were randomly assigned to one of three interventions: BASICS (n = 40), a group-based values clarification condition that served as the University's treatment-as-usual program (n = 68), and an assessment-only control group (n = 42). The second group was oversampled to adjust for expected high rates of attrition.

Assessments/Attrition: Assessment occurred at baseline, posttest (immediately after the intervention session), one-month follow-up, and three-month follow-up. The gap between baseline and the posttest was 7-10 days for the BASICS and control conditions. However, the gap for the values-clarification condition ranged from a few days to more than three months. By the time of the three-month follow-up, 43% of the participants had been lost.

Sample:

The sample averaged 19 years of age and was 69% male and 77% white.

Measures:

The seven outcomes came from standard self-report measures (i.e., Daily Drinking Questionnaire, Frequency-Quantity Questionnaire, Rutgers Alcohol Problem Index, Readiness Ruler): total number of drinking days per week, total number of drinks consumed per week, peak alcohol use in the last month, alcohol dependence, personal alcohol-related consequences, social alcohol-related consequences, and readiness to change alcohol use. The study reported reliability and validity information that came primarily from other studies.

Analysis:

The study used repeated measures multivariate analysis of variance for the three outcomes with time-by-condition interactions that included the baseline outcomes.

Intent-to-Treat: Although attrition was high, the study used all available data.

Outcomes

Implementation Fidelity:

Not examined.

Baseline Equivalence:

The text stated that there were no significant differences for the four sociodemographic measures or the seven outcomes. However, some of the group means appeared to vary widely (80% versus 64% with a 3.0 GPA, or 62% versus 74% male).

Differential Attrition:

Attrition rates at three months were 33% for the BMI condition, 66% for the values-clarification condition, and 31% for the control condition. The differences were significant, due to the loss of participants in the values-clarification condition.

At the posttest, tests for differences of completers and dropouts in the values-clarification condition found no significant differences for the baseline outcome measures.

At the one-month follow-up, tests of completers and dropouts across all conditions found one significant difference (personal alcohol-related consequences), and at three months found no significant differences.

Posttest:

MANOVA tests across all drinking outcomes combined found that the time by condition interaction was non-significant. The results were similar when examining the results through only the one-month follow-up. Tests also found no significant effect on readiness to change.

Long-Term:

Not examined.

Study 21

Murphy et al. (2010) presented the results of two separate trials. The one labeled Study 1 is described here (and the other is included as Study 22). Teeters et al. (2015) examined a supplemental outcome from Study 1.

Summary

Murphy et al. (2010) and Teeters et al. (2015) used a randomized controlled trial that assigned 74 undergraduate students who were screened for heavy drinking episodes into two conditions: BASICS (n = 39) and a control group (n = 35). The control group used the Alcohol 101 Plus CD-ROM program. Posttest assessments of alcohol use, alcohol-impaired driving, and motivation to change came immediately after the session, one month later, and six months later.

Murphy et al. (2010) and Teeters et al. (2015) found that, relative to the control group, the intervention group reported significantly

• Lower alcohol-impaired driving at six months
• Greater subjectively rated changes in drinking at one month
• Greater self-ideal and normative discrepancy after the intervention session
• Greater motivation to change after the intervention session.

Evaluation Methodology

Design:

Recruitment: Participants were students from a large metropolitan public university in the southern United States who were at least 18 years old and reported one or more heavy drinking episodes in the past month. To recruit an ethnically diverse sample, a lower eligibility threshold was used to screen for heavy drinking among minority students. Of 130 eligible students, 74 (57%) enrolled.

Assignment: After completing the baseline assessment, participants were randomly assigned within gender and ethnicity categories to two conditions: BASICS (n = 39) and a control group (n = 35). The control group used the Alcohol 101 Plus CD-ROM program.

Assessments/Attrition: Murphy et al. (2010) examined assessments at baseline, immediately after the session, and one month after the session. Teeters et al. (2015) examined a six-month follow-up. Of 74 participants, 69 (93%) completed the one-month follow-up.

Sample:

The sample was 73% Caucasian, 23% African American, 2.7% Hispanic/Latino, 2.7% Asian, 1.4% American Indian, and 41% male.

Measures:

Murphy et al. (2010) examined two alcohol use measures from the Daily Drinking Questionnaire: the total number of standard drinks that they consumed each day during a typical week and the frequency of heavy drinking in the past month. Evidence of reliability and validity came primarily from citations to other studies.

The study also examined three risk and protective factors. Measures of normative and self-ideal discrepancy used the Discrepancy Ratings Questionnaire. Normative discrepancy assessed students' perceptions of their drinking compared to the average college student; self-ideal discrepancy assessed how alcohol affected their relationships with others, health, and appearance. Higher scores on both measures indicate a greater perceived discrepancy and worse outcomes. A measure of motivation to change drinking came from the Readiness Ladder, which involves circling an image on the rung that most closely corresponded to their thoughts of changing their drinking. An additional measure of subjective change in drinking asked if their drinking increased, decreased, or stayed the same during the past month.

Teeters et al. (2015) used a single-item measure of alcohol-impaired driving that referred to a positive response to the statement, "I have driven a car when I knew I had too much to drink to drive safely."

Analysis:

Murphy et al. (2010) used analysis of covariance models to examine post-session outcomes (drinking perception discrepancy and motivation to change drinking) and follow-up drinking outcomes (drinks per week and past-month frequency of heavy drinking), with the baseline values as a covariate. Teeters et al. (2015) used logistic regression for the binary measure of alcohol-impaired driving, again with the baseline outcome as a covariate.

Intent-to-Treat: The analyses used all available data, dropping only the 7% who did not complete the follow-up.

Outcomes

Implementation Fidelity:

A review of a subsample of the sessions by an independent rater assigned an average rating of 1.98 across 26 program components (where 2.0 meant the intervention component was delivered in a manner that was consistent with the protocol in terms of both content and motivational interviewing style). The rating on the same scale averaged 2.04 across the 10 specific motivational interviewing skills.

Baseline Equivalence:

Murphy et al. (2010) reported that "There were no significant group differences in drinks per week, heavy drinking, motivation, or discrepancy at baseline."

Differential Attrition:

Murphy et al. (2010) stated that attrition rates did not differ significantly across conditions, and "there were no demographic or baseline drinking differences between completers and non-completers." Also, a footnote added that, after replacing missing follow-up data for non-completers with the last observation carried forward, the results did not change. Teeters et al. (2015) did not report on attrition, but the lead author provided information to Blueprints. Tests for baseline equivalence using the analysis sample of n = 65 (n = 35 for the intervention group and = 30 for the control group) found no significant condition differences (p < .05) for six sociodemographic or three baseline outcome measures.

Posttest:

Murphy et al. (2010) found no significant differences across conditions in the two drinking outcomes (drinks per week and past month frequency of heavy drinking). For the risk and protective factors, the intervention group relative to the control group reported significantly greater self-ideal and normative discrepancy, motivation to change immediately after the session, as well as subjectively rated changes in drinking at one month. Eta-squared values ranging from .053-.178 indicated medium to large effect sizes. Additional tests showed little mediation and moderation.

Teeters et al. (2015) found at the six-month follow-up that the intervention group was significantly less likely to report alcohol-impaired driving (OR = 3.71).

Long-Term:

Not examined.

Study 22

The article presented the results of two separate trials. The one labeled Study 2 is described here (the other is listed under Study 21).

Summary

Murphy et al. (2010) used a randomized controlled trial that assigned 133 undergraduate students who were screened for heavy drinking episodes into three conditions: BASICS (n = 46), a web-based program called e-CHUG (n = 45), or an assessment-only control condition (n = 42). Posttest assessments of alcohol use and motivation to change came immediately after the session and one month later.

Murphy et al. (2010) found that, relative to the control group, the intervention group reported significantly lower

• Typical drinking per week and heavy drinking per month at one month
• Self-ideal discrepancy immediately after the intervention session.

Evaluation Methodology

Design:

Recruitment: Participants were students from a large metropolitan public university in the southern United States who were at least 18 years old and reported one or more heavy drinking episodes in the past month. To recruit an ethnically diverse sample, a lower eligibility threshold was used to screen for heavy drinking among minority students. Of the 219 eligible students, 133 (61%) enrolled.

Assignment: After completing the baseline assessment, participants were randomly assigned within gender and ethnicity categories to three conditions: BASICS (n = 46), a web-based program called e-CHUG (n = 45), or an assessment-only control condition (n = 42).

Assessments/Attrition: Assessments occurred at baseline, immediately after the session, and one month after the session. Of the 133 participants, 118 (89%) completed the one-month follow-up.

Sample:

The sample was 65.4% Caucasian, 30.1% African American, 2.3% Hispanic/Latino, 2.3% Native American, 0.8% Hawaiian, 0.8% Asian, and 50% male.

Measures:

The study examined two alcohol use measures from the Daily Drinking Questionnaire: the total number of standard drinks that they consumed each day during a typical week and the frequency of heavy drinking in the past month. Evidence of reliability and validity came primarily from citations to other studies.

The study also examined risk and protective factors. Measures of normative and self-ideal discrepancy used the Discrepancy Ratings Questionnaire. Normative discrepancy assessed students' perceptions of their drinking compared to the average college student; self-ideal discrepancy assessed how alcohol affected their relationships with others, health, and appearance. Higher scores on both measures indicate a greater perceived discrepancy and worse outcomes. A measure of motivation to change drinking came from the Readiness Ladder, which involves circling an image on the rung that most closely corresponded to their thoughts of changing their drinking. An additional measure of subjective change in drinking asked if their drinking increased, decreased, or stayed the same during the past month.

Analysis:

Analysis of covariance models examined post-session outcomes (drinking perception discrepancy and motivation to change drinking) and follow-up drinking outcomes (drinks per week and past-month frequency of heavy drinking), with the baseline value as the covariate.

Intent-to-Treat: The study used all available data, dropping only the 11% who did not complete the follow-up.

Outcomes

Implementation Fidelity:

Review of a subsample of the sessions by an independent rater assigned an average rating of 2.00 across 26 program components (where 2.0 meant the intervention component was delivered in a manner that was consistent with the protocol in terms of both content and motivational interviewing style). The rating on the same scale averaged 2.02 across the 10 specific motivational interviewing skills.

Baseline Equivalence:

The authors reported that "There were no significant group differences in drinks per week, heavy drinking, motivation, or discrepancy at baseline."

Differential Attrition:

Attrition rates did not differ significantly across conditions. One baseline outcome, drinks per week, was significantly higher for non-completers than completers, but "there were no other demographic or drinking differences between completers and non-completers." Because the model controlled for drinks per week, the difference in loss for completers and non-completers likely did not bias the results. Also, a footnote added that, after replacing missing follow-up data for non-completers with the last observation carried forward, one result changed: The contrast between BMI and e-CHUG on heavy drinking became significant.

Posttest:

The results found significant differences across conditions for drinks per week (d = .42) and past month frequency of heavy drinking (d = .52). The intervention group reported lower drinking relative to the control group but not the web-based program. For the risk and protective factors, the intervention group reported significantly greater post-session self-ideal discrepancy than the web-based program and greater subjectively rated changes in drinking at one month than the web-based program and control group. Additional tests showed little mediation and moderation.

Long-Term:

Not examined.

Study 23

The study aimed to demonstrate that BASICS presented via a video-conferencing session between client and therapist (i.e., telehealth) would work as well as a face-to-face session. As such, it did not have a no-treatment control group and did not predict the superiority of BASICS to the alternative.

Summary

King et al. (2020) used a randomized controlled trial that assigned 51 undergraduate students who were screened for heavy drinking episodes into two conditions: BASICS and a videoconferencing or telehealth version of BASICS. Posttest assessments of alcohol use and problems came at baseline, and at one-, two-, and three-months post-intervention.

King et al. (2020) found no effects for BASICS relative to the videoconferencing version of BASICS.

Evaluation Methodology

Design:

Recruitment: The sample consisted of 51 undergraduates from a large midwestern University who received extra course credit for participation. Eligible participants indicated that they had engaged in heavy episodic drinking within the last two weeks.

Assignment: It appears that the three therapists randomly assigned each client to receive either face-to-face BASICS or the telehealth video-conferencing version. Each therapist thus delivered both types of sessions and performed the random assignment.

Assessments/Attrition: Assessments occurred at baseline and at one-, two-, and three-months post-intervention. The randomized sample of 51 dropped to 15 by the three-month follow-up (29.4% completion).

Sample:

The sample of undergraduates averaged 19 years of age and was 61% female.

Measures:

The two outcomes included 1) the Alcohol Use Disorders Identification Test (AUDIT), which consisted of a 10-item brief assessment to screen for problematic and excessive drinking (alpha = .734) and 2) the Rutgers Alcohol Problem Index (RAPI), which consisted of a 23-item assessment evaluating the frequency of problems associated with alcohol consumption over a specific period (alpha = .873).  

Analysis:

The analysis used a piecewise multilevel regression to model the linear slope for the change from baseline to one-month post-treatment, and from one- to three-month post-treatment. The conditional model controlled for covariates, and the time variable included the baseline outcomes. Therapist competency was included as a control variable.

Intent-to-Treat: The analysis used full information maximum likelihood "to address missing variables." The authors further stated that they applied an intent-to-treat analysis and that "all participants randomly allocated to their treatment groups were analyzed in their groups regardless of program dosage, adherence, or assessment retention."

Outcomes

Implementation Fidelity:

Coding of a sample of the sessions utilizing the Motivational Interviewing Treatment Integrity Code assigned an average score of 3.94, where 3.5 indicated beginning proficiency and 4.0 indicated full proficiency.

Baseline Equivalence:

The authors stated that "No significant treatment condition difference was observed for gender, age, year in college, Greek membership, ethnicity, mental health diagnosis, AUDIT, and RAPI baseline scores." Appendices A and B present the condition means. Although not significant, some differences were relatively large (e.g., 46% versus 26% Greek membership in BASICS and telehealth BASICS, respectively).

Differential Attrition:

The authors stated that "post-hoc analyses found no significant dropout differences between groups across baseline and all follow up time points (p-values ranging from .243 to .879)."

Posttest:

Tests found no significant condition difference in either outcome measure (excessive drinking or alcohol problems) for the one-month post-treatment means, the change from baseline to one-month post-treatment, or the change from one month to three months post-treatment. The authors noted, however, that both conditions led to significant reductions in the drinking and drinking problems outcomes, particularly among those with a mental health diagnosis.

Long-Term:

Not examined.

Study 24

This study of event-specific prevention focused on reducing alcohol use during 21st birthday celebrations.

Summary

Neighbors et al. (2012) used a randomized controlled trial that assigned 92 university students to six conditions: BASICS, four interventions focused on 21st-birthday drinking, and an assessment-only control group. A posttest assessment one week after the 21st birthday measured drinking and its consequences during the birthday celebration.

Neighbors et al. (2012) found that the BASICS-only intervention group relative to the control group reported significantly

• Lower blood alcohol content (d = -.25)
• Fewer alcohol-related problems (d = -.27).

Evaluation Methodology

Design:

Recruitment: Participants included college students at a large public northwestern university who were turning age 21 and reported an intent to engage in heavy drinking on their birthday. From a list of 3,045 students who were invited to participate between December 2008 and December 2009, 1,558 (51.2%) completed the screening assessment, and 642 (41.2%) met screening criteria and were invited to participate in the study. Of those invited, 599 (93.3%) completed the baseline assessment.

Inclusion criteria included (a) intending to consume four (for women) or five (for men) drinks during their 21st birthday; (b) listing the e-mail address of at least one friend, 18 years or older, with whom they planned to celebrate their birthday; and (c) having not previously participated in the study as a friend. The study thus utilized data from the participants (i.e., those who were turning 21) and friends (i.e., those who were listed by a participant as being a supportive friend and who would be present at the 21st birthday celebration).

Assignment: The study randomized 642 participants to one of six conditions: 1) BASICS, 2) 21 BASICS, 3) 21 BASICS with friend, 4) 21 web BASICS, 5) 21 web BASICS with friend, and 6) control. The condition sample sizes ranged from 106-110.

• The BASICS condition did not review any content specific to 21st birthday drinking, while the four other interventions specifically targeted 21st birthday drinking.
• Participants in the two web interventions received an e-mail link to personalized feedback based on their prior survey responses rather than meeting with a counselor.
• Participants in the two friend interventions named a friend joining in the celebration, who received web-based personalized feedback.
• Participants in the control condition only completed the assessments.

The baseline assessment came after consent, with 599 of 642 (93%) completing the survey. Completion rates varied from 92% to 95% across the six conditions.

Assessments/Attrition: The baseline assessment came one week before the 21st birthday, and the posttest assessment came one week after the birthday. Of the 642 randomized participants, 585 (91%) completed the baseline and follow-up assessments.

Sample:

The sample of students was 46% male and 54% female. Most students were white (68.1%), with 15.9% Asian, 7.7% multiethnic, and 8.3% other. On average, participants intended to consume about 10 drinks on their 21st birthday.

Measures:

The study examined three student self-reported outcomes: (a) total number of alcoholic drinks consumed on the 21st birthday, (b) estimated peak blood alcohol content on the 21st birthday, and (c) total number of drinking-related negative consequences on the 21st birthday (e.g., vomiting, hangover, arrests). The measures came from standard instruments used in previous research but were modified to apply to the 21st birthday. The last measure showed good reliability.

Analysis:

The analyses used a negative binomial regression model for total drinks and total consequences and ordinary least squares regression for blood alcohol content. Given the focus on 21st birthday drinking, the outcome did not exist at baseline. The models included gender and drinking intentions assessed at baseline as covariates but not measures of usual alcohol consumption.

Intent-to-Treat: The study used all available data regardless of the treatment received by participants.

Outcomes

Implementation Fidelity:

The scores from coding of session tapes for the in-person interventions exceeded competency criteria. For BASICS, 71 (66%) of the 107 randomized students received the intervention.

Baseline Equivalence:

The outcomes, which were specific to the 21st birthday celebration, could not be measured at baseline, and tests for sociodemographic differences between conditions were not presented. The authors stated only that "There were no overall differences in [baseline] intentions across conditions."

Differential Attrition:

The study did not examine differential attrition, perhaps because total attrition was only 9%. Calculations from the CONSORT diagram show that attrition ranged from 7% to 13% across the six conditions.

Posttest:

The students in the BASICS intervention reported significantly lower blood alcohol content (d = -.25) and significantly fewer alcohol-related problems (d = -.27) than the control students but no difference for the number of drinks. When combined into a single group, the four interventions that focused directly on the 21st birthday celebration did significantly worse than BASICS for blood alcohol content and did no better than BASICS for the other two outcomes.

Long-Term:

Not examined.