Background: Globally, the number of illegal drug users is rising, posing mental and physical health challenges and increasing societal burdens. Despite a significant need for treatment, only about 10% of these individuals receive it worldwide, often with poor adherence. Traditional treatments, while effective, suffer from high dropout rates due to limitations. The COVID-19 pandemic has spurred the growth of digital interventions like apps and online platforms, offering flexibility and cost-effectiveness that better meet patient needs and improve engagement. However, addressing the persistently high dropout rates in these online treatments is crucial and necessitates further research. Objective: This study aimed to estimate dropout rates among adults with illicit drug use participating in digital psychosocial intervention trials, and to identify factors associated with attrition. Methods: We conducted a systematic search of five major databases for English-language randomized trials published up to January 27, 2025. A total of 40 studies (80 arms; 9,563 participants) reporting 46 dropout rate estimates were included. A random-effects model was used to calculate pooled dropout rates, with meta-regression and subgroup analyses exploring potential moderators. The study was registered on PROSPERO (CRD42024534389). Results: At post-test, the pooled dropout rate in the intervention group across 17 studies was 22.4% (95% CI: 12.4%–37.2%). Dropout was significantly associated with education level, employment status, baseline clinical diagnosis, intervention frequency, and initial medication use. During the longest follow-up (29 studies), the dropout rate was 27.9% (95% CI: 18.8%–39.3%), with marital status, recruitment source, medication frequency, and intervention modality as significant predictors. Control group dropout rates were 25.9% and 28.3%, both higher than those in the intervention group. Conclusions: This meta-analysis revealed substantial dropout among adults with illicit drug use receiving digital psychosocial interventions. Targeted modifications to intervention design may improve engagement and long-term retention. Clinical Trial: The study was registered on PROSPERO (CRD42024534389).

npj Digital Medicine, Published online: 10 October 2025; doi:10.1038/s41746-025-01993-5

Automated AI based identification of autism spectrum disorder from home videos

DNA methylation is an epigenetic modification that regulates gene expression by adding methyl groups to DNA, affecting cellular function and disease development. Machine learning, a subset of artificial intell...