A randomised controlled trial assessing the effectiveness of a smartphone app intervention for alcohol-dependent adults

Background
Digital interventions, including mobile phones and smartphone applications (apps), have become more common in recent years, offering the opportunity to enhance alcohol dependency treatment acceptability and accessibility. Still, studies focussing on the possible effects of using apps in treatment for alcohol dependence are lacking.

The aim of this study was to estimate the possible consumption-reducing effects of using two apps as complement to treatment as usual (TAU) among alcohol-dependent adults.
 
Methods
Three-armed, parallel, randomised controlled trial (n=162). Eligible participants were randomised to A) TAU + a self-monitoring app (Glasklart), B) TAU + a self-monitoring app including a portable breathalyser (iBAC Pro), or C) TAU only. Alcohol-dependent (ICD-10) adults who had had at least 5 heavy drinking days the past 8 weeks, who had not received other treatment for alcohol problems the past 2 years, and who had no severe somatic diseases or mental disorders, were recruited from a dependency clinic in Stockholm, Sweden, and from online advertisements. TAU included 4 sessions of psychological treatment based on motivational interviewing and cognitive behavioural therapy, including eventual pharmacotherapy, for 12 weeks.

Data was collected between Sep 2020 and Jun 2023, and participants were assessed at baseline, at 12 weeks (post-intervention), and at 26 weeks (follow-up).

The primary outcome was past month mean number of heavy drinking days (HDD) at 12 and 26 weeks, defined as days consuming at least 4 standard drinks (12 grams of pure ethanol/drink), measured by the Timeline Follow Back instrument (TLFB). Secondary outcomes included weekly alcohol consumption, typical number of units consumed, hazardous and harmful drinking, assessed by TLFB, alcohol use disorder identification test (AUDIT-C), and phosphatidyl ethanol (PETh) in blood, as well as alcohol dependence and alcohol-related consequences, assessed by ICD-10 and Short Alcohol Dependence Data Questionnaire (SADD). 

Linear mixed-effects models were used to measure intervention effects. Data was analysed according to intention-to-treat, and per protocol analyses were performed on those completing at least 80 % of the instructed treatment. 

Results
Preliminary analyses showed significant within-group decreases in both mean HDD and PETh-values at each timepoint compared to baseline, and a significant between-group difference in mean PETh, where participants randomised to TAU+Glasklart had a significantly lower mean compared to participants randomised to TAU only (p=0.0245), and TAU+iBAC Pro (p=0.0446). The per protocol analysis confirmed the results from the main analysis. 

Additional analyses will be computed and presented at the conference.

Conclusions
Although preliminary, our analyses show promising results for participants randomised to using the app Glasklart as complement to TAU, but this needs to be further evaluated.