Volume 8 Supplement 1

International Network on Brief Interventions for Alcohol and Other Drugs (INEBRIA) Meeting 2013

Open Access

Self-selection in a randomized trial of web-based primary and secondary prevention alcohol brief intervention

  • Nicolas Bertholet1Email author,
  • Joseph Studer1,
  • John A Cunningham2,
  • Jean-Bernard Daeppen1,
  • Gerhard Gmel1 and
  • Bernard Burnand3
Addiction Science & Clinical Practice20138(Suppl 1):A10

DOI: 10.1186/1940-0640-8-S1-A10

Published: 4 September 2013

How much a randomized controlled trial (RCT) sample is representative/differs from its declared source population is a challenging question, with major implications with regard to generalizibility of results. The question is crucial for freely available web-based interventions tested in RCTs.

We compared participants in a primary/secondary prevention web-based alcohol brief intervention RCT to its source population. There is a mandatory army recruitment process in Switzerland at age 19 for men. Between 8.2010 and 7.2011, 12,564 men attended two recruitment centers (source population) and were asked to answer a screening questionnaire on substance use; of the 11,819 (94%) who completed it, 7,034 (56%) agreed to participate in a cohort with regular assessments. In 2012, irrespective of their drinking, cohort participants were invited to the RCT; 1,549 agreed to participate. Using chi-square and t-test, we compared screening data of RCT participants to other members of the source population with respect to weekly alcohol use, maximum number of drinks/occasion, Alcohol Use Disorders Identification Test (AUDIT), prevalence of binge drinking (≥6 drinks per occasion at least monthly), unhealthy alcohol use (>210g of ethanol/week and/or ≥1 binge episode/month), and abstinence.

RCT participants (n=1,549) drank less than other members of the source population (n=10,270): they reported a mean (SD) of 6.2(8.1) vs 7.4(10.7) drinks/week, t=5.2, p<.0001; 9.7(7.7) vs 10.3(9.1) maximum number of drinks/occasion, t=2.6, p=.009; and 6.6(4.4) vs 7.2(4.9) AUDIT scores, t=4.7, p<.0001. The prevalence of binge drinking was lower among RCT participants (36.7% vs 43.9%, χ2=28.8, p<.0001), as was unhealthy alcohol use (37.1% vs 44.1%, χ2=27.5, p<.0001) and abstinence (7.8% vs 9.7%, χ2=5.8, p=0.016). RCT participants differed from other members of the source population: they reported less drinking and lower levels of consequences or risk, but were also less likely to be abstainers, indicating that the self-selection applies to both ends of the drinking spectrum.

Authors’ Affiliations

Alcohol Treatment Center, Department of Community Medicine and Health, Lausanne University Hospital
Centre for Mental Health Research, Australian National University
Clinical Epidemiology Center, Department of Community Medicine and Health, Lausanne University Hospital


© Bertholet et al; licensee BioMed Central Ltd. 2013

This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.