John: Hello, I'm John Hilton, editor in the Cochrane Editorial and Methods department. The Cochrane Drugs and Alcohol Group has produced several reviews of the effects of interventions intended to reduce heavy drinking. The collection was extended in September 2017 with an investigation of the use a computer to provide personalised advice. We asked one of the authors, Fiona Beyer from Newcastle University in the UK, to tell us what they found.

Fiona: Hazardous or harmful drinking can involve regular consumption of alcohol over the recommended limits, or occasional consumption of high volumes, known as binging. It causes over 60 diseases, as well as many accidents, injuries and early deaths. One way that helps people reduce their drinking by around 2 to 3 units a week is a structured conversation with a doctor or nurse who provide feedback and advice to the drinker. However, with people using mobile devices more and more, we wanted to find out whether similar interventions, delivered digitally through computers or mobile phones might also be effective and whether there are specific ingredients of the intervention that are most effective. Our findings were promising and suggest that these personalised digital interventions do help people to reduce their heavy drinking.
We looked for randomised trials that recruited people who lived at home, had answered a questionnaire suggesting that they drank over recommended limits, and would be willing to test a digital intervention via a computer or mobile device designed to reduce their drinking. We carried out one analysis where the drinking of people given the intervention was compared to that of similar people who received either nothing or some brief written information about their health or drinking, and a second analysis where the comparison was with the use of a conversation-based intervention. 
We found 57 eligible studies and were able to pool the results of 41 of these in our primary meta-analysis. This showed that, on average, people receiving a digital intervention drank up to 23 grams less alcohol per week compared to those who did not. That’s about one and a half pints of beer or a third of a bottle of wine each week. Although this effect is small compared to the amount some people were drinking in the first place, if digital interventions were widely used and this benefit was seen in a large population, it could have a large impact on alcohol-related diseases and harms. 
Five small studies compared the drinking of people who received a digital intervention to those receiving a conversation-based intervention, but there was no evidence that either intervention was better than the other. 
We found that the interventions used a median of nine behaviour change techniques or BCTs, which means that many BCTs have not been used at all in these interventions. Of the ones that have been included, three were associated with reduced drinking. These were behaviour substitution or specifying something less harmful to do when the person feels like drinking; problem solving or prompting the person to think about their drinking and create strategies to achieve a more healthy drinking pattern; and the presentation of verbal or visual information from a credible source, such as the logo of a well-known and trusted organisation. 
In summary, although we found no evidence that digital interventions have more or less of an impact on drinking than brief conversation-based interventions, we have shown that there was a small but significant reduction in drinking in people who used a digital alcohol intervention compared to those who received either nothing or just brief written information.

John: If you would like to read more about these digital alcohol interventions, how they were delivered and their impact on other aspects of heavy drinking, the full review is available online. Just go to Cochrane Library dot com and search 'digital and harmful alcohol use'.