Do Digital Interventions Work?

Welcome to the drinkbreak blog. We are going to using this blog to provide a "behind the scenes" look at the history of our intervention, the theory behind it and the evidence to support it. We are also going to describe the development as we go along and some of the challenges we face.

We really welcome your comments and observations so we can learn from them.

So we start with a key question - "do digital interventions work"? Clearly, we think they do - that's why we are putting all this effort in. But we have to qualify that answer - after all we are serious researchers. It depends quite a lot on what is meant by work. It depends on the implicit "who" in the question. Do they work for everyone and all the time? That depends on whether people use the intervention as planned. Do they work better than other, usually face to face types of intervention? Often yes, but again it depends on what alternatives are actually available to people in the circumstances they are in.

So here comes some of the published evidence.

The original version of drinkbreak was a website called "DownYourDrink". We did many studies and found that people who used the site reduced their alcohol consumption. We went on to do a fully online randomised controlled trial (the first ever) and found that the intervention group reduced their self-reported consumption (always self-reported as we can't do breathalysers or take blood remotely - yet). But so did the control group. In fact this is a common problem with RCTs in this field.

Other researchers have also been active and an objective approach was needed to garner the information from all the studies together. We have one - a major systematic review. The question was whether digital interventions are as effective as the same kind in intervention when delivered face to face.

Kaner, Beyer, Garnett et al, (2017) have published their Cochrane review in which they identified 57 studies with a total of 34,900 participants.

This Cochrane review described digital interventions as those that used computers, mobile devices or smartphones to address problematic alcohol consumption; are responsive to user input to generate personalised content and use some of the same intervention content as standard face-to-face versions. However, the reviewers also pointed out key differences between the two types of intervention. Digital ones can be used to provide access to hard-to-reach groups, whereas standard ones are used opportunistically. Digital interventions deliver and record information automatically and users may either miss or appreciate the lack of an interpersonal element. Face-to-face is generally a one-off event, whereas digital content can be repeated which may have an impact on long term outcomes. Finally, enthusiasts for technological approaches may be a different population from those attending clinics etc.

The main conclusion was that participants using a digital intervention drank approximately 23g alcohol weekly (about 3 UK units) less than participants who received no or minimal interventions at end of follow up. Participants who engaged with digital interventions had less than one drinking day per month fewer than no intervention controls and about one binge drinking session less per month.

This might not seem like a big reduction overall, but it is reporting an average reduction so some people may have made big changes and other not much at all. Anyone who does research understands that you cannot predict individual changes from group data.