Skip to main content

Table 1 Summary of included studies: salient characteristics and outcomes

From: Building the first step: a review of low-intensity interventions for stepped care

Intervention modality Investigators Study design Sample Elements of intervention Magnitude of alcohol consumption change & dependence measure Attract individuals* Positive outcomes** Inspire help-seeking†
Bibliotherapy Apodaca & Miller (2003) Meta-analytic review of 22 bibliotherapy studies Included studies broadly targeting “problem drinkers” Included studies invited participants to read and implement self-help materials Bibliotherapy versus control (self-referred drinkers), d = 0.31 X XX X
Dependence measure: varied across studies (e.g., abstinence rates, alcohol consumption frequency, liver enzyme levels)
Bibliotherapy Sobell et al (2002) Randomized trial Individuals who consumed >12 drinks per week or ≥5 drinks on ≥5 days in the year prior to assessment Participants received 1 of 2 written interventions by mail: 1) Motivational enhancement-based feedback and advice (personalized based on participants’ drinking history and patterns); 2) General information on the effects of alcohol, guidelines for safe use, and information on self -monitoring No significant differences between groups in alcohol consumption change X XX X
Motivational enhancement/personalized feedback (n = 414);
Bibliotherapy/drinking guidelines (n = 411)
Bibliotherapy Cunningham et al (2002) Randomized trial Individuals who expressed interest in self-help and scored ≥8 on the AUDIT Drink Wise, which uses CBT principles, served as the self-help text. Personalized assessment/feedback was based on the “Drinker’s Check-up” and included personalized assessment, normative feedback, and information regarding the potential negative consequences of alcohol use Self-help book & personalized assessment/feedback versus single treatment or control, d = 0.21   X  
Self-help book (n = 22); personalized assessment/feedback (n = 21); self-help book & personalized assessment/feedback (n = 17); no materials (n = 26)
Dependence measure: Mean days per week of ≥5 drinks per occasion during a typical week in the 6-month follow-up period
Bibliotherapy Bamford et al (2005) Randomized trial Individuals presenting for clinic-based alcohol treatment Participants received a 6-page leaflet based on FRAMES Leaflet versus no leaflet, d = 0.20   X X
Leaflet condition (n = 180); no leaflet condition (n = 181)
Dependence measure: Self-rated categorical (yes/no) reduction in alcohol use at follow-up
Bibliotherapy Wild et al (2007) Randomized trial Current drinkers (used alcohol in the 12 months prior to assessment) with no previous participation in alcohol treatment who had an interest in self-help materials The self-help pamphlet gave normative feedback regarding drinking habits and included information on the hazards of alcohol use at various consumption levels, guidelines for reducing alcohol use, and referral to a local treatment hotline Unable to calculate   X  
Pamphlet (n = 877); no pamphlet (n = 850)
Bibliotherapy Kavanagh & Connolly (2009) Randomized trial Individuals who met DSM-IV criteria for an AUD Participants were enrolled in primary care and received information regarding the effects of alcohol, self-monitoring forms, and a self-help booklet. The mailed correspondence treatment further included personalized progress letters and 8 CBT-based newsletters Dependence measure: Mean drinks per week = 0.39   XX  
Immediate mailed intervention (n = 103); delayed mailed intervention (n = 101)
Telephone-delivered intervention Brown et al (2007) Randomized trial Individuals who met DSM-IV criteria for alcohol abuse or dependence who were drawn from primary care The telephone-based intervention consisted of 6 sessions based on MI principles and the stages of change model. Behavioral techniques were used to monitor sobriety goals. Participants were also sent individualized letters summarizing progress after sessions Telephone-based intervention versus control (male participants) = 0.16   XX  
Telephone-based intervention (n = 445); control (n = 452)
The intervention was not superior to control for women
Dependence measure: Total standard drinks consumed in the month prior to 3-month follow-up interview
Telephone-delivered intervention Mello et al (2008) Randomized trial Noncritically injured emergency department patients who used alcohol at risky levels per NIAAA quantity/frequency guidelines Targeted individuals with a recent alcohol-related injury. Counselors initiated 2 brief calls (30 minutes & 15 minutes) based on MI principles No significant differences between groups in alcohol consumption change   X  
Telephone-based intervention (n = 140); control (n = 145)
Computer-based intervention Hester et al (2005) Randomized trial Individuals who scored ≥8 on the AUDIT Participants were given access to computer-based MI. Components include assessment and normative feedback, decisional balance, and negotiating sobriety goals Pre versus post for computer-based intervention, = 1.05   X X
Computer-based intervention (n = 35); control (n = 26)
Dependence measure: Mean of 3 dependent variables (average drinks per day, drinks per drinking day, and average peak BAC level) during 12-month follow-up period
Computer-based intervention Neumann et al (2006) Randomized trial Emergency department patients who scored ≥5 on the AUDIT Patients in the active treatment condition were given access to computer-delivered personalized advice and normative feedback. Feedback was based on the MI principles and FRAMES guidelines. Other components included information about alcohol and provider referral information Computer-based intervention versus control = 0.20   X  
Computer-based intervention (n = 561); control (n = 575)
Dependence measure: Proportion of participants who met British Medical Association criteria for at-risk drinking at 6 months post-treatment
Computer-based intervention Boon et al (2011) Randomized trial Drinkers with alcohol consumption levels exceeding the limits set by the Dutch guidelines for low-risk drinking Participants in the treatment group received normative feedback and information regarding the negative consequences of alcohol use. Personalized advice was informed by participant drinking patterns, self-efficacy and attitude, and stage of change Intervention versus control = 0.25   X  
Computer-based advice (n = 230); control (n = 220)
Dependence measure: Meeting or failing to meet Dutch guidelines for low-risk drinking at 1-month follow-up
Internet-based intervention Cunningham et al (2009) Randomized trial Drinkers who scored ≥4 on the 3 consumption items on the AUDIT-C & expressed an interest in self-help Participants in the active condition were mailed a URL that allowed them to access the screening program, which provided a personalized assessment and normative feedback Mailed intervention URL versus control, ηp2 = 0.08   X  
Sent URL by mail to participate in intervention (n = 92); control (n = 93)
Dependence measure: Mean drinks consumed per week during follow-up period
Internet-based intervention Pemberton et al (2011) Quasi-randomized trial Active duty military personnel Participants assigned to active treatment received either AS or DCU. Controls received no intervention. Both interventions were adapted for a military population via expert consensus AS versus control = no significant differences in alcohol consumption change   X  
Internet-based intervention based on social learning theory (AS) (n = 686);Internet-based intervention based on MI principles (DCU) (n = 1470); control (n = 914)
DCU versus control = 0.10
Dependence measure: Average drinks per drinking occasion during 1-month follow-up
Internet-based intervention Riper et al (2007) Randomized trial Drinkers whose consumption levels exceeded Dutch guidelines for low-risk drinking Online self-help protocol was consistent with CBT and self-control principles. The intervention was accessed via the study website and included goal setting and analysis of alcohol behavior. Participants also had access to a peer-to-peer chat room Online self-help versus control = 0.40 X X  
Online self-help, (n = 130); control (n = 131)
Dependence measure: Mean weekly alcohol consumption during 6-month follow-up period
Internet-based intervention Postel et al (2010) Pre-post design Individuals concerned about their drinking Therapy was delivered online by a therapist who communicated with the patient asynchronously. Treatment was a blend of CBT and motivational enhancement, along with elements from the stages-of-change model. Therapy assignments were given in 2 stages; patients could choose (with therapist approval) to move to the second stage of treatment Pre- versus post-intervention = 1.11 X XX  
(N = 527)
Dependence measure: Mean weekly alcohol consumption immediately post-treatment
Internet-based intervention Blankers et al (2011) Randomized trial Score >8 on the AUDIT and consumption of >14 standard drinks in a week SAO: An automated, fully self-guided Internet intervention based on elements of MI and CBT. TAO versus WL = 0.59 SAO versus WL = 0.36   XX  
Internet-based self-help, (SAO) (n = 68); Internet-based therapy (TAO) (n = 68); control (WL) (n = 69)
Dependence measure: Number of drinks in the 7 days prior to 3-month follow-up
TAO: A therapist-led online intervention (elaborated version of SAO’s MI/CBT protocol plus 7 synchronous chat-based therapy sessions)
Control: waitlist (WL) condition
Internet-based intervention Wallace et al (2011) Randomized trial Individuals who accessed the DYD website and scored ≥5 on the AUDIT-C Participants in the active condition received access to the DYD interactive online alcohol intervention based on CBT, MI, and relapse prevention techniques DYD versus control = no significant differences in alcohol consumption change   X  
“Down Your Drink” Internet-based intervention (DYD) n = 3972); control (n = 3963)
Internet-based intervention (television-supported) Kramer et al (2009) Randomized trial Drinkers whose consumption levels exceeded Dutch guidelines for low-risk drinking Participants in active treatment were asked to use a CBT-based, television-supported online self-help intervention. The 5-part series depicts a coach guiding 2 individuals with alcohol problems through an intervention. Participants also received written self-help materials and access to the website Television-based intervention versus control = 0.90   X  
Television-supported intervention (n = 90); control (n = 91)
Dependence measure: Mean weekly alcohol consumption at follow-up
  1. *Attract Individuals = X indicates the intervention appeared to attract individuals who might otherwise not seek help (defined as those who had not previously sought treatment or who expressed disinterest in formal treatment).
  2. **Positive Outcomes = X indicates the intervention significantly reduced alcohol use; XX indicates the intervention significantly reduced alcohol use in more severe drinkers (alcohol dependence diagnosis or AUDIT >19).
  3. †Inspire help-seeking = X indicates the intervention appeared to be associated with future help-seeking.
  4. Abbreviations: MI motivational interviewing, CBT cognitive-behavioral therapy, AUDIT Alcohol Use Disorders Identification Test, NIAAA National Institute on Alcohol Abuse and Alcoholism.
  5. Note: For 4 of the studies, Cohen’s was calculated by the authors using the Effect Size Determination Program (Lipsey & Wilson, 1996) based on either given study chi-square statistics (Bamford et al., 2005; Boon et al., 2011); alcohol consumption outcome summary statistics (Brown et al., 2007); or by comparing the proportion of participants in each condition who no longer met criteria for at-risk drinking as defined by that study (Neumann et al., 2006). Apodaca and Miller (2003) limited effect size calculations to bibliotherapy studies where patients received ≤1 meeting with a clinician. The Cohen’s calculation for Bamford et al. (2005) was based on participants’ reduction in alcohol use. The Cohen’s calculation for Brown et al. (2007) compared alcohol consumption at 3-month follow-up for the experimental group versus control (as opposed to change from baseline). Although the Mello et al. (2008) treatment was not associated with a change in alcohol consumption, impaired driving scores did improve with controls (= 0.31). Pemberton et al. (2011) was designated as quasi-randomized because assignment to active treatment was only available at select study sites. The Cohen’s calculation for Postel et al. (2010) only included participants who completed the treatment (Parts 1 & 2).