Effectiveness of Machine Learning Based Adjustments to an eHealth Intervention Targeting at Risk Alcohol Use
Background
This study aimed to evaluate effects of three machine learning based adjustments made to an eHealth intervention for hazardous alcohol use or mild alcohol use disorder, regarding a) early dropout, b) participation duration, and c) success in reaching personal alcohol use goals. Additionally, we aimed to replicate earlier machine learning analyses that were published by our group (Ramos et al., 2021).
Methods
We used three cohorts of observational log data from the Jellinek Digital Self-help intervention. First, a cohort before implementation of adjustments based on the machine learning findings (T0; n=320); second, a cohort after implementing two adjustments (i.e., sending daily emails in the first week and nudging participants towards a ‘no alcohol use’ goal; T1; n=146); third, a cohort comprising the prior adjustments complemented with eliminated time constraints to reaching further modules in the intervention (T2; n=236).
Results
We found that more participants aimed for a quit goal, whilst participation duration declined at T2. We found an increase in participants reaching further in the intervention (increased use of modules), yet an increase in early dropout (in time) after implementing the elimination of time constraints at T2. Intervention success increased, yet not significantly. Lastly, application of machine learning demonstrated reliable, as the same predictors were present for intervention drop-out and treatment succes in these smaller datasets of an eHealth intervention for hazardous alcohol use, as in our previously published study (Ramos et al., Frontiers in Psychology, 2021 https://www.frontiersin.org/articles/10.3389/fpsyg.2021.734633).
Conclusions
Strong correlates as indicated by machine learning analyses were found to affect goal setting for alcohol use and use patterns of an eHealth CBT based program for alcohol use problems.