Effectiveness, Acceptability and Feasibility of Technology-Enabled Health Interventions for Adolescents Living with HIV in Low-and Middle-Income Countries: A Systematic Review

Sustained adherence to antiretroviral therapy (ART) and engagement in care remain challenging for adolescents living with HIV (ALHIV) due to developmental, psychosocial, and contextual factors, particularly in low- and middle-income countries (LMIC). With most ALHIV residing in sub-Saharan Africa, differentiated, youth-friendly, and scalable approaches are needed. Technology-enabled health interventions—using devices such as mobile phones and computers over mobile networks or the internet—offer privacy, support, feedback, and potential for self-management support (autonomy, adherence, goal-setting, problem-solving). Despite promising evidence in high-income settings, data on effectiveness, acceptability, feasibility, and fidelity in LMIC are limited. This review aims to identify technology-enabled interventions implemented for ALHIV in LMIC and assess their effectiveness on health and treatment outcomes, as well as feasibility, acceptability, and fidelity.

Prior reviews show a growing but uneven evidence base. Among adults living with HIV, technology-enabled interventions can improve treatment access, self-management, adherence, retention, risk reduction, and quality of life; eHealth in sub-Saharan Africa appears low-cost and can improve adherence and retention. However, among ALHIV, early reviews found very few interventions outside the USA (2015 review identified only one study outside the USA). A 2016–2018 review of ART adherence interventions identified only one mHealth/SMS intervention for adolescents. A 2000–2019 review of self-management interventions found four technology-enabled adolescent interventions, none in LMIC. A recent LMIC-focused review (2000–2021) identified nine mHealth interventions for ALHIV care and treatment but emphasized quantitative outcomes and not feasibility, acceptability, or fidelity. Overall, applicability across varied LMIC contexts and marginalized adolescent groups remains insufficiently understood, underscoring the need to synthesize both outcomes and implementation/process evidence.

Design: Systematic review following seven steps (question formulation, inclusion/exclusion, search, selection, quality appraisal, data extraction, synthesis) per Egger et al. Conducted and reported per PRISMA-P; registered in PROSPERO (CRD42022336330). Inclusion criteria: Studies from LMIC (2010–2022), English-language, involving ALHIV aged 10–19 years as primary population (allowing broader ages if adolescents predominated), reporting technology-enabled interventions delivering/supporting healthcare directly to adolescents. Study types: RCTs, non-randomized controlled trials, before–after studies, and qualitative studies reporting feasibility/acceptability. Comparators: No intervention, standard of care, waitlist, or non-tech comparators; studies without comparators were also considered. Outcomes: Health-related individual outcomes (e.g., adherence, retention, viral suppression, self-efficacy, mental health, knowledge, quality of life, behaviors); secondary process outcomes (acceptability, feasibility, fidelity). Search strategy: Broad search across Ebscohost (PsycArticles, Academic Search Premier), CINAHL, ERIC, MEDLINE, PubMed, SCOPUS, ScienceDirect, Sabinet, plus Google Scholar; trial registries (ClinicalTrials.gov; WHO ICTRP) for unpublished/ongoing studies; grey literature (theses/dissertations, conference abstracts). Search string combined adolescent terms with ICT/eHealth terms, HIV/AIDS, and LMIC. Study selection: Citations imported into Covidence; duplicates removed; two reviewers independently screened titles/abstracts and full texts; discrepancies resolved by discussion with the team. Data extraction: Two reviewers independently extracted study design, participant characteristics, intervention details, outcomes, and settings, guided by an adapted TIDieR 12-item checklist (name, goal, theoretical foundation, end-user involvement, duration, materials/content, procedures and personnel, device type, technology design, delivery platform/mode). Quality appraisal: Mixed Methods Appraisal Tool (MMAT) used to appraise qualitative, RCT, non-randomized, quantitative descriptive, and mixed methods studies. Scoring converted to grades: ≤0.50 weak; 0.51–0.65 moderately weak; 0.66–0.79 moderately strong; ≥0.80 strong. Synthesis: Narrative synthesis with tabulation due to few, heterogeneous studies.

Study selection and characteristics: Of 267 records identified, 120 titles screened; 11 studies met inclusion. Settings: Mostly Africa—Nigeria (3), Uganda (2), South Africa (2), Kenya (2); others: Guatemala (1), Argentina (1). Designs: 5 RCTs; 4 non-RCT pre–post; 1 non-RCT matched controls; 1 qualitative. Participants: Total N=1544 (reported 846 females, 387 males; two studies did not report sex distribution). Age range across studies: 6–25 years (adolescents predominated by inclusion design). Inclusion often required basic digital literacy (internet/SMS/web) and access to a mobile phone; one study provided smartphones, WhatsApp, SIM, and airtime. Quality: RCTs graded strong (2) to moderately strong (3). Non-RCTs: three moderately strong, one weak; matched-control strong; qualitative strong. Intervention modalities and duration: Technology designs included interactive groups (majority), interactive individual, and non-interactive individual (SMS). Delivery platforms: social media (WhatsApp/Facebook/Mxit) and SMS predominated; others included web-based platforms and mixed modes (SMS/WhatsApp/calls). Duration ranged 6–18 months (most 12 months). Theory: None reported an underpinning theoretical framework; all involved end-users. Primary outcomes: - ART adherence (8 studies measured): Significant improvement found in 2 studies. • Argentina (Stankievich et al., 2018): After 32 weeks of generic messages, 65% (13/20) undetectable VL (<40 copies/mL) and 70% (14/20) VL < 1000 copies/mL; 30% had no VL change. • Guatemala (Sánchez et al., 2021): Text-message group improved adherence by 4 percentage points (p<0.01) vs. non-significant 0.85-point improvement in controls (p=0.64). Other adherence studies reported no significant differences. Measures varied: self-report (VAS, ACTG, CAMP, recall), objective (pill count, VL, MEMS, Wisepill). - Viral load: Measured in 3 studies; only Stankievich et al. showed significant VL improvement (VL <1000 copies/mL). Others (Abiodun: <20 copies/mL; Hacking: <400 copies/mL) showed no significant VL differences. - Linkage to care: Virtual Mentors Program (South Africa) improved ART initiation (mentees 80% [28/35] vs. matched controls 42% [30/70]). - Retention in care: No significant effects where measured. - HIV knowledge: Mixed findings—improved in a Facebook group intervention (Dulli 2020; p=0.003), not significant in a web-based peer support program (Ivanova 2019). - Other psychosocial/behavioral outcomes (social support, self-efficacy, mental health, stigma, behavioral health): Generally no significant effects. Secondary/process outcomes: - Acceptability: High in 10/11 studies; factors affecting acceptability included literacy, personal phone ownership, reliable network, prior phone function use, and message privacy. Preference for commonly used platforms (WhatsApp/SMS) over Facebook/web in some contexts. - Feasibility: High in nearly all studies reporting it (10/11); one did not report feasibility. - Fidelity: Mixed—high in three studies; low in one (password issues, connectivity, data costs, device capabilities affecting participation and content formatting). Facilitator characteristics (training, reliability, engagement) and participant comfort improved fidelity in online groups. Implementation insights and challenges: - Privacy/anonymity concerns in group settings; limited control over content sharing by peers. - Engagement barriers: scheduling conflicts with school/household duties, variable response rates to interactive SMS (e.g., 20.5%–67.5% response among messages successfully delivered of 83%–86%); caregiver restrictions on phone use. - Digital and general literacy, device capability, and connectivity/data costs influenced feasibility and fidelity. - Qualitative reports highlighted perceived benefits: increased hope, morale, community, and peer support; occasional discouragement due to technical errors or misinformation.

Across LMICs, technology-enabled interventions for ALHIV span from simple one-way SMS reminders to interactive group-based social media support. While feasibility and acceptability are generally high, evidence of effectiveness on clinical and behavioral outcomes is weak and inconsistent, with only two adherence-focused studies showing significant improvements and mixed effects on HIV knowledge. No consistent benefits were demonstrated for retention, viral suppression (beyond one study), or psychosocial outcomes. Notably, none of the interventions reported a guiding theoretical framework, and integration with routine services and sustainability considerations were unclear, despite international guidance emphasizing theory-driven, youth-centered design and integration with existing systems. Implementation context matters: familiar, low-data platforms (SMS/WhatsApp) and strong facilitation may enhance engagement and fidelity; however, privacy, literacy, device capability, connectivity, and competing priorities can limit participation. Current evidence is dominated by small-scale, heterogeneous, and often pilot studies, precluding firm conclusions and subgroup analyses (e.g., by adolescent developmental stage or degree of human interaction in the intervention). Overall, the findings underscore the promise of digital approaches for supportive functions (e.g., peer support, knowledge) but highlight the need for rigorous, theory-informed, and contextually adapted designs to realize measurable improvements in core HIV outcomes.

Technology-enabled interventions for ALHIV in LMIC are largely acceptable and feasible but show weak and inconsistent effectiveness on key health outcomes. Future work should develop and test theory-driven interventions that are integrated with routine care and community services, leverage widely used, low-barrier platforms, and address privacy, literacy, device capability, and connectivity constraints. Larger, well-powered studies with rigorous designs and standardized outcome measures are needed to determine effectiveness, identify the most impactful components, and clarify which technology designs and platforms are optimal for ALHIV and health workers, enabling sustainable scale-up.

- Evidence base limitations: Few eligible studies with heterogeneous designs, interventions, measures, and populations limited comparability and precluded meta-analysis. Many were pilot or short-term studies with small samples. - Reporting gaps: No interventions reported an underpinning theory; variable and often self-reported adherence measures were used. - Population and age ranges: Some studies included broader ages (up to 24–25 years) due to limited adolescent-specific data, potentially diluting adolescent-specific effects. - Implementation variability: Differences in device capability, connectivity, and data costs, and facilitator quality affected fidelity and may limit generalizability across LMIC contexts. - Review scope: English-language restriction and LMIC focus may have excluded relevant evidence; reliance on narrative synthesis may introduce subjectivity despite structured appraisal.