Screening and intervention for excessive drinking produce small results

Author and Disclosure Information

  • BACKGROUND: The World Health Organization has recommended screening for excessive alcohol consumption. However, the effectiveness of screening, as a precursor to brief intervention, has not been systematically evaluated. This meta-analysis aims to measure the effectiveness of screening followed by a brief intervention in a primary care setting as a significant factor in the reduction of alcohol consumption.
  • POPULATION STUDIED: The 8 US, UK, and Australian studies conducted in a primary care setting were combined to include a screened group of 134,693 adults aged 17 to 84 years. Of these, 12,345 screened positive for excessive alcohol consumption (variously defined as consuming from >11 to >29 drinks per week); 3317 were randomized to receive a brief intervention.
  • STUDY DESIGN AND VALIDITY: This meta-analysis of 8 randomized controlled trials was designed to evaluate the effect a 2-step process—screening for excessive alcohol consumption followed by a brief intervention—had on decreasing alcohol consumption in a clinically meaningful manner. The authors used robust methodology outlined by the Cochrane collaboration to locate and select relevant studies; they performed a solid and clearly outlined internal and external validity assessment on each trial before inclusion.
  • OUTCOMES MEASURED: The primary outcome was number needed to screen (NNS); that is, the number of people identified as excessive drinkers who would need to receive the intervention for 1 person to benefit. Secondary outcomes were the proportion of patients positive for excessive drinking on screening, the proportion given brief interventions, and the effect of screening.
  • RESULTS: Nine percent of patients (range, 3.3%–18%) screened positive. Further assessment of drinking history identified 2.5% (range, 0.9%–5.4%) who qualified for a brief intervention. Of these patients, about 10% decreased their drinking as a result of the intervention (absolute risk reduction [ARR]=10%; 95% confidence interval [CI], 7.1%–13.9%). The pooled screening effect was 2.6 (95% CI, 1.7–3.4) patients who achieved sensible drinking out of the 1000 screened.



It is reasonable to consider screening for excessive alcohol consumption if time and circumstances permit, realizing the ultimate benefit will be extremely small.

Overall, if a practitioner screens 1000 patients, carries out further assessment in 90 (9%) who screen positive, and gives feedback, information, and advice to the 25 (2.5%) who qualify for brief intervention, 2 or 3 patients can be expected to have reduced their alcohol consumption to below recommended maximum levels after 12 months. This results in a number needed to screen with outcome measured at 1 year (NNS1) of 500. To put this in perspective, the NNS5 (to prevent 1 death in 5 years) for dyslipidemia is 418; for hypertension, 274–1307; for hemoccult testing, 1374; for mammography in those aged 50–59 years, 2451.

Next Article: