Use of the Internet for Prevention of Binge Drinking
Results across the fourteen articles are summarized on Table 1 and Table 2. The following sections summarize data presented in both tables. Out of 14 articles only 2 studies were conducted outside of the USA, with one study being conducted in New Zealand (Voogt et al., 2013) and the other in the Netherlands (Kypri et al., 2014). Most of the studies utilized a randomized control trial (RCT) study design. Ten studies had control groups and the remaining four studies had randomized experimental groups with no control group. Sample sizes ranged from a low of 77 to a high of 7815 total participants (Doumas et al., 2009; Paschall et al., 2011). The summary table suggests that average age of all studies range between 19 and 22 years old (Neighbors et al., 2009; Moore et al., 2005). Similarly most of the intervention was targeted towards freshmen (Walters et al., 2007; Neighbors et al., 2009), even though there were few studies that was targeted towards junior and senior only (Moore et al., 2005). Out of 14 articles, only 3 articles utilized a theory-based intervention into the study design and implementation of the intervention. Theories that were identified included Social Cognitive Theory and Health Belief Model (Moore et al., 2005), Social Norms Feedback Theory (Alfonso et al., 2013), and Social Norming Theory (Doumas et al., 2009). Most of the studies (n = 11) used pre-existing web-based programs such as mystudentbody.com (Chiauzzi et al., 2005), e-CHUG (Walters et al., 2007; Alfonso et al., 2013), Web-BASICS (Neighbors et al., 2012; Alfonso et al., 2013; LaBrie et al., 2013), CHOICES (Alfonso et al., 2013), AlcoholEdu (Paschall et al., 2011). The remaining studies (n = 3) used unique web-based interventions. Online surveys, emails, and websites were methods for intervention delivery for personalized or generic feedback, and group specific or alcohol related educational sessions that were integrated in all of the studies. In three studies, web-based interventions were compared with face-to-face interventions (Neighbors et al., 2012; Alfonso et al., 2013; LaBrie et al., 2013). Another study compared a web-based intervention to a print-based intervention (Moore et al., 2005). Most of the web-based interventions were limited to evaluation of program through online assessment and computer generated feedback while other interventions also included an educational component. The duration of the studies ranged from 2 days (Neighbors et al., 2012) to 2 years (Neighbors et al., 2009). The average length of study was 12 weeks. Brief interventions with short durations were targeted towards specific events such as a students' 21st birthday. The 2-year intervention was actually a brief intervention with periodic follow-up to evaluate the efficacy of the Internet-based intervention over a longer time. All studies but one reported significant reductions in overall drinking quantity and frequency, as well as predicted changes in other outcome measures such the Rutgers Alcohol Problem Index (RAPI) score, peak Blood Alcohol Content (BAC), and perceived norms.
Outcome measures included antecedents of problematic drinking behavior (i.e. RAPI Score, perceived drinking norms) and drinking behaviors (i.e. number of typical drinks, frequency of alcohol use). The RAPI scale was used to assess problematic drinking behaviors in six studies, and in all the other studies, a modified RAPI scale was used. Other instruments used in studies included the Academic Role Expectations and Alcohol Scale score (AREAS), College Students Alcohol Problems score, and peak and average Blood Alcohol Content (BAC). Similarly, descriptive norms, perceived alcohol use, alcohol-related consequences, perceived gender non-specific drinking norms, 30-day frequency of drinking and driving, 30-day frequency of unprotected sex, and attitude of personal alcohol use also were used as measures of antecedents of problematic drinking behavior. Alcohol Use Disorders Identification Test (AUDIT) or a modified version of AUDIT was a commonly found instrument used to measure drinking behavior. Three studies used the AUDIT instrument and three studies used the Daily Drinking Questionnaire (DDQ) as a measuring tool for drinking behavior. Other instruments used included a 7-day drinking calendar, the Brief Drinker's Profile, and the Alcohol Consumption Index. Studies also assessed frequency and quantity of drinks consumed by asking questions such as: 30 day and 14 day frequency of alcohol use, 30 day and 14 day frequency of binge drinking, number of typical drinks per occasion, weekly drinking, and 21st birthday drinking.
Among the studies reviewed in this article, nine provide justification for sample size (Walters et al., 2007; Hedman and Akagi, 2008; Neighbors et al., 2009; Alfonso et al., 2013). Eight of the studies had pre- and post-assessments (Moore et al., 2005; Hedman and Akagi, 2008; Doumas et al., 2009; Neighbors et al., 2012; Alfonso et al., 2013), and seven studies had three or more assessment points (Walters et al., 2007; Neighbors et al., 2009; Paschall et al., 2011). Only three studies included process evaluations.
Results
Results across the fourteen articles are summarized on Table 1 and Table 2. The following sections summarize data presented in both tables. Out of 14 articles only 2 studies were conducted outside of the USA, with one study being conducted in New Zealand (Voogt et al., 2013) and the other in the Netherlands (Kypri et al., 2014). Most of the studies utilized a randomized control trial (RCT) study design. Ten studies had control groups and the remaining four studies had randomized experimental groups with no control group. Sample sizes ranged from a low of 77 to a high of 7815 total participants (Doumas et al., 2009; Paschall et al., 2011). The summary table suggests that average age of all studies range between 19 and 22 years old (Neighbors et al., 2009; Moore et al., 2005). Similarly most of the intervention was targeted towards freshmen (Walters et al., 2007; Neighbors et al., 2009), even though there were few studies that was targeted towards junior and senior only (Moore et al., 2005). Out of 14 articles, only 3 articles utilized a theory-based intervention into the study design and implementation of the intervention. Theories that were identified included Social Cognitive Theory and Health Belief Model (Moore et al., 2005), Social Norms Feedback Theory (Alfonso et al., 2013), and Social Norming Theory (Doumas et al., 2009). Most of the studies (n = 11) used pre-existing web-based programs such as mystudentbody.com (Chiauzzi et al., 2005), e-CHUG (Walters et al., 2007; Alfonso et al., 2013), Web-BASICS (Neighbors et al., 2012; Alfonso et al., 2013; LaBrie et al., 2013), CHOICES (Alfonso et al., 2013), AlcoholEdu (Paschall et al., 2011). The remaining studies (n = 3) used unique web-based interventions. Online surveys, emails, and websites were methods for intervention delivery for personalized or generic feedback, and group specific or alcohol related educational sessions that were integrated in all of the studies. In three studies, web-based interventions were compared with face-to-face interventions (Neighbors et al., 2012; Alfonso et al., 2013; LaBrie et al., 2013). Another study compared a web-based intervention to a print-based intervention (Moore et al., 2005). Most of the web-based interventions were limited to evaluation of program through online assessment and computer generated feedback while other interventions also included an educational component. The duration of the studies ranged from 2 days (Neighbors et al., 2012) to 2 years (Neighbors et al., 2009). The average length of study was 12 weeks. Brief interventions with short durations were targeted towards specific events such as a students' 21st birthday. The 2-year intervention was actually a brief intervention with periodic follow-up to evaluate the efficacy of the Internet-based intervention over a longer time. All studies but one reported significant reductions in overall drinking quantity and frequency, as well as predicted changes in other outcome measures such the Rutgers Alcohol Problem Index (RAPI) score, peak Blood Alcohol Content (BAC), and perceived norms.
Outcome measures included antecedents of problematic drinking behavior (i.e. RAPI Score, perceived drinking norms) and drinking behaviors (i.e. number of typical drinks, frequency of alcohol use). The RAPI scale was used to assess problematic drinking behaviors in six studies, and in all the other studies, a modified RAPI scale was used. Other instruments used in studies included the Academic Role Expectations and Alcohol Scale score (AREAS), College Students Alcohol Problems score, and peak and average Blood Alcohol Content (BAC). Similarly, descriptive norms, perceived alcohol use, alcohol-related consequences, perceived gender non-specific drinking norms, 30-day frequency of drinking and driving, 30-day frequency of unprotected sex, and attitude of personal alcohol use also were used as measures of antecedents of problematic drinking behavior. Alcohol Use Disorders Identification Test (AUDIT) or a modified version of AUDIT was a commonly found instrument used to measure drinking behavior. Three studies used the AUDIT instrument and three studies used the Daily Drinking Questionnaire (DDQ) as a measuring tool for drinking behavior. Other instruments used included a 7-day drinking calendar, the Brief Drinker's Profile, and the Alcohol Consumption Index. Studies also assessed frequency and quantity of drinks consumed by asking questions such as: 30 day and 14 day frequency of alcohol use, 30 day and 14 day frequency of binge drinking, number of typical drinks per occasion, weekly drinking, and 21st birthday drinking.
Among the studies reviewed in this article, nine provide justification for sample size (Walters et al., 2007; Hedman and Akagi, 2008; Neighbors et al., 2009; Alfonso et al., 2013). Eight of the studies had pre- and post-assessments (Moore et al., 2005; Hedman and Akagi, 2008; Doumas et al., 2009; Neighbors et al., 2012; Alfonso et al., 2013), and seven studies had three or more assessment points (Walters et al., 2007; Neighbors et al., 2009; Paschall et al., 2011). Only three studies included process evaluations.