Aims: The direct effects of screening on drinking behaviour have not previously been evaluated experimentally. We tested whether screening reduces self-reported hazardous drinking in comparison with a non-screened control group.
Design: Two-arm randomized controlled trial (RCT), with both groups blinded to the true nature of the study.
Setting and participants: A total of 421 university students aged 18-24 years, recruited in five London student unions.
Interventions: Both groups completed a brief pen-and-paper general health and socio-demographic questionnaire, which for the experimental group also included the 10-item Alcohol Use Disorders Identification Test (AUDIT) screening questionnaire.
Measurements: The primary outcome was the between-group difference in AUDIT score at 2-3-month follow-up. Eight secondary outcomes comprised other aspects of hazardous drinking, including dedicated measures of alcohol consumption, problems and dependence.
Findings: A statistically significant effect size of 0.23 (0.01-0.45) was detected on the designated primary outcome. The marginal nature of the statistical significance of this effect was apparent in additional analyses with covariates. Statistically significant differences were also obtained in three of eight secondary outcomes, and the observed effect sizes were not dissimilar to the known effects of brief interventions.
Conclusions: It is unclear to what extent these findings represent the effects of screening alone, a Hawthorne effect in which drinking behaviour has changed in response to monitoring, or whether they indicate reporting bias. These possibilities have important implications both for the dissemination of screening as an intervention in its own right and for behavioural intervention trials methodology.