Introduction
The global burden of illicit drug use remains substantial, with hundreds of millions affected by various substances such as cannabis, opioids, and cocaine. Digital psychosocial interventions have emerged to overcome barriers like time, location, and stigma, offering flexible, scalable approaches delivered via apps, websites, and messaging platforms. Yet, dropout—participants who fail to complete courses or withdraw before outcome assessment—remains a persistent challenge in digital treatment settings. This article synthesizes evidence on dropout prevalence in digital psychosocial interventions for adult illicit drug users and identifies factors that predict retention or attrition, aiming to inform the design of more effective digital treatments.
Methods: How the evidence was gathered
Following Cochrane and PRISMA guidelines, researchers conducted a systematic review and meta-analysis of randomized controlled trials evaluating digital psychosocial interventions for adults with illicit substance use. Eligible studies reported dropout data for intervention and control groups at posttreatment and/or the longest follow-up. A broad set of moderators was extracted, including demographic, clinical, therapist, and intervention characteristics, to explore what drives dropout at different treatment stages. Analyses used random-effects models to account for heterogeneity across studies and included meta-regression and subgroup analyses to examine potential predictors of dropout.
Key findings: What influences retention?
Overall, dropout in digital interventions showed substantial variability across studies, with a pooled posttreatment dropout around 22% in the intervention arms and 26% in control groups, indicating a potential retention advantage for digital formats, albeit with notable heterogeneity (I² often exceeding 90%). At the longest follow-up, dropout rates averaged about 28% in the intervention groups versus 28% in controls, again with high heterogeneity. Several patterns emerged from moderator analyses:
- Demographics: Among posttreatment data, higher employment was unexpectedly associated with greater dropout, suggesting work-related time constraints may complicate consistent participation. In long-term follow-up, being single appeared linked to lower dropout, potentially reflecting differences in social support or compensatory online engagement strategies.
- Baseline clinical characteristics: Participants with an existing clinical diagnosis or higher baseline drug use frequency tended to drop out more, underscoring the need for tailored support for higher-risk individuals. Cocaine use at baseline showed particularly elevated dropout in some subgroups, highlighting substance-specific retention challenges.
- Intervention characteristics: More frequent contact and real-time feedback generally supported retention, reinforcing the importance of ongoing engagement and therapeutic alliance in digital formats. Recruitment sources also mattered: online-recruited samples showed higher dropout than campus-based samples, suggesting context and setting influence adherence. The degree of digitalization (fully digital vs. partially digital) produced complex results, with inconsistencies likely driven by reporting gaps rather than true effects.
Publication bias was detected in some analyses, but sensitivity checks and trim-and-fill adjustments indicated that the core conclusions remained robust. Overall, digital psychosocial interventions show promise for improving retention relative to traditional approaches, but substantial heterogeneity signals the need for standardized reporting and more nuanced designs to optimize engagement across diverse populations and substances.
Discussion: Interpreting the patterns
These findings underscore that dropout in digital contexts is not uniform. In the short term, work status, initial clinical diagnosis, specific substances used at baseline, and how often contact occurs with participants were key predictors. In longer follow-ups, relationship status, ongoing drug-use frequency, and recruitment context gained prominence. This suggests a multi-layered approach to enhance retention: increase supportive touches and coaching for higher-risk individuals, tailor modules to substance-specific needs, and design recruitment and onboarding processes that align with participants’ environments.
Practical implications for researchers and clinicians
To reduce attrition and improve data quality, investigators should standardize reporting of digital features (degree of digitalization, level of human support) and clearly document reasons for dropout. Integrating adaptive, multimodal engagement strategies—such as reminders, interactive modules, crisis support, and peer or family involvement—may bolster adherence. Given the diversity of illicit substances and comorbidities, modular treatments that can be customized to individual risk profiles are likely to yield better retention and, ultimately, better outcomes.
Limitations and future directions
Limitations include high between-study heterogeneity and incomplete reporting of digital intervention components. Future work should leverage individual participant data meta-analyses, preregistration, and open data practices to unpack subgroup effects and temporal dropout patterns. Embracing innovations such as gamification, real-time analytics, and patient-centered feedback loops could further reduce dropout, while machine learning approaches may help predict and preempt attrition risks.
Conclusion
Digital psychosocial interventions offer a viable path to engaging adults with illicit drug use, with dropout rates that may be lower than traditional approaches in some settings. However, retention is influenced by an interplay of demographic, clinical, and intervention-specific factors that vary across treatment stages. By standardizing reporting, personalizing content, and enhancing engagement strategies, researchers and clinicians can strengthen retention and maximize the therapeutic potential of digital interventions for adult illicit drug users.
