Preventing Non-Medical Opioid Use for Young Populations through Ehealth Interventions: An AI-Assisted Scoping Review

Background & Purpose: eHealth interventions, delivered through wearable devices, mobile apps, and web-based platforms, have been increasingly used to support opioid misuse prevention and education. However, there remains limited evidence on effective ways to develop and assess eHealth intervention features and outcomes. This review addresses this gap through three aims: identifying key features of eHealth interventions targeting opioid misuse prevention, examining the effectiveness associated with specific intervention features, and comparing effectiveness across different intervention modalities.

Methods: A comprehensive search was conducted across nine databases (PubMed, MEDLINE, CINAHL, PsycINFO, Cochrane, Web of Science, Scopus, ResearchRabbit, Google Scholar), using MeSH terms and Boolean combinations targeting opioid misuse and eHealth modalities. AI-assisted tools, including ResearchRabbit for initial research article extract, and ChatGPT for the data screening with unified prompt to review multiple articles were utilized. Inclusion criteria were set using the PICO framework: no population restriction; digital/eHealth opioid misuse prevention and education interventions, comparison groups (control, treatment-as-usual, or pre-post); behavioral or knowledge-based opioid-related outcomes. Of 11,701 records screened, nine studies met inclusion criteria. Studies were excluded if they were non-eHealth interventions, review/protocol papers, irrelevant topics, lacked control/comparison group, or involved complex chronic conditions.

Results: The nine included studies collectively sampled over 5,319 participants (mean age ≈ 27.7 years). Samples varied in gender composition, including one study with 100% adolescent girls and others with mixed or parent-child dyads. Recruitment methods included social media, traditional media, schools, and survey pool through Qualtrics. Retention rates were generally high (Mean = 80%) in follow-ups. Intervention modalities included web-based skill-building programs (e.g., RealTeen) and media campaigns. Control groups included treatment-as-usual or inactive comparators with most studies using randomized or quasi-experimental designs. Six out of nine studies (N=2,922) reported significant intervention effects; effective interventions incorporated behavior change frameworks (e.g., Social Learning Theory, Cognitive Behavioral Therapy, Information-Motivation-Behavioral Skills Model) and interactive intervention formats (e.g., web modules, serious games, and social media campaigns). The outcomes included reduced substance use (e.g., IRR = 0.519 for marijuana use), increased self-efficacy (data?), and improved opioid safety knowledge (P<.001). Non-significant findings were often linked to low intervention dosage, short duration, or small sample sizes.

Conclusions and Implications: eHealth interventions show promise to prevent opioid misuse by improving knowledge, self-efficacy, and behavioral outcomes (e.g., mental health support). Their scalability and cost-effectiveness make them suitable for schools, clinics, and communities, reaching large groups effectively while promoting family conversations about opioid safety. For social work practice, theory-driven and accessible digital tools are valuable for prevention strategies. Future research should prioritize long-term outcomes, culturally tailored interventions, and real-world implementation.