Systematic review and meta-analysis comparing educational and reminder digital interventions for promoting HPV vaccination uptake

Cervical cancer is the fourth most common cancer among women globally, with most cases in developing countries where preventive access is limited. Preventive measures include HPV vaccination, screening, and safe sexual practices. Scaling HPV vaccination among adolescent girls could significantly reduce cervical cancer incidence and mortality. WHO targets 90% coverage by age 15, which could avert tens of millions of HPV-related cancer deaths; however, vaccination coverage varies widely, with low uptake in less developed regions. Multiple factors influence HPV vaccination uptake (HVU), including vaccine access, cost, policies, and cultural attitudes. Existing strategies face barriers such as undervaluation, misinformation, attitudinal and structural barriers, and the high cost of incentives. Digital interventions offer scalable solutions for clients (education, myth dispelling, reminders, tracking, scheduling) and providers (education via webinars/websites, clinical decision support, monitoring vaccination status). Prior reviews identify common interventions—client education, client reminders, client education plus reminders, provider education, and provider reminders—implemented through digital tools (SMS, apps, websites, social media), but a direct meta-analytic comparison across these five categories has been lacking. This study aims to estimate pooled effects of these digital intervention types on HVU, determine which increase HVU, compare their magnitudes, and identify variables associated with intervention efficacy.

Prior systematic reviews have documented the effectiveness of multiple strategies to increase vaccination uptake, including client education, reminders, combined education plus reminders, provider education, and provider reminders, often delivered through digital channels (mobile messaging, apps, websites, social media). Evidence from RCTs has been promising, yet a direct meta-analytic comparison across these five intervention types specific to HPV vaccination had not been undertaken. The broader literature on immunizations suggests reminder systems are efficacious for improving uptake, while pure educational approaches often show mixed or limited effects, particularly where baseline knowledge is not the primary barrier.

Protocol and reporting: The review was registered on PROSPERO (CRD42023389004) and followed PRISMA and Cochrane Handbook guidance. Eligibility (PICOS): Population included children, adolescents, and young adults aged 9–26 eligible for HPV vaccination, and/or their parents or healthcare providers. Interventions used digital technologies (e.g., SMS, email, DVD, website, webinar, application) to educate or remind clients or providers about HPV vaccination, with intent to promote HVU. Comparators included usual care or alternative control conditions; non-inferiority trials without a control group were excluded. Outcomes were HPV vaccination initiation (≥1 dose) or completion (all recommended doses), assessed via self-report or provider-verified records/registries; studies measuring only knowledge, attitudes, or intentions were excluded. Study design was randomized controlled trials (including cluster and multi-arm designs). Search strategy: Comprehensive searches of PubMed, PsycInfo, Web of Science, and Cochrane Central were completed in January 2023 using terms for digital interventions, HPV vaccine, and uptake outcomes. Reference lists of relevant reviews and included studies were screened. Study selection: Titles/abstracts were screened independently by two authors using Rayyan (90% overlap; discrepancies resolved by discussion). Full texts were independently assessed (85% overlap; disagreements resolved by discussion). PRISMA: 1,929 records identified; after removing 842 duplicates, 1,087 records screened; 210 full texts assessed; 34 studies included. Data extraction: Extracted study characteristics (year), intervention type (client education, client reminder, client education plus reminder, provider education, provider reminder), sample characteristics (age, gender), and outcome data (sample sizes, uptake in groups). Intention-to-treat data were preferred. For cluster RCTs, effective sample sizes were adjusted using reported design effects and ICCs. Dual coding achieved 84–100% agreement (mean 95%). Authors were contacted for missing data. Statistical analysis: For each comparison, odds ratios (ORs) with 95% CIs were computed for post-intervention HVU (initiation or completion), comparing digital intervention vs control. Random-effects models were used due to anticipated heterogeneity. For multi-arm trials comparing multiple interventions to one control, the control group sample size was split to avoid double-counting. ORs were log-transformed, pooled via random-effects meta-analysis, and exponentiated. Heterogeneity was quantified with I2. Subgroup analyses compared the five intervention categories using mixed-effects models (random within subgroups, fixed between subgroups). Analyses were conducted in Comprehensive Meta-Analysis (CMA) 3.0, which accommodates dependency from multiple effect sizes per study. Risk of bias and publication bias: Two authors independently assessed RCTs with Cochrane ROB 2.0 across five domains; disagreements were resolved by discussion. Publication bias was evaluated by funnel plot and Egger’s test, with Duval and Tweedie’s trim-and-fill to estimate adjusted pooled effects accounting for potentially missing studies.

- Studies included: 34 RCTs yielding 41 effect sizes; total N = 281,280 participants. Most studies were from high-income countries (primarily USA), with two from China (upper-middle income). - Overall effect: Digital interventions increased HPV vaccination uptake (pooled OR 1.25; 95% CI 1.16–1.34; P < 0.001). - By intervention type: - Client reminder (CR): OR 1.41; 95% CI 1.23–1.63; P < 0.001; I2 ≈ 42%. - Provider reminder (PR): OR 1.39; 95% CI 1.11–1.75; P = 0.005; I2 ≈ 59%. - Client education + reminder (CRCE): OR 1.29; 95% CI 1.04–1.59; P = 0.007; I2 ≈ 50%. - Provider education (PE): OR 1.18; 95% CI 1.05–1.34; P = 0.007; I2 ≈ 0%. - Client education (CE): OR 1.08; 95% CI 0.92–1.28; P = 0.35; I2 ≈ 15% (not significant). - Comparative effectiveness: CR, PR, PE, and CRCE produced significantly greater improvements than CE alone. - Moderators of effectiveness: - Gender: Greater benefits in studies targeting only males (OR 2.30; 95% CI 1.52–3.48) and mixed-gender samples (OR 1.30; 95% CI 1.21–1.41) than only females (OR 1.06; 95% CI 1.01–1.12). - Platform: Reminder delivery modes were effective—SMS (OR 1.39; 95% CI 1.23–1.55), preference reminders (OR 1.33; 95% CI 1.07–1.65), and electronic health record alerts (OR 1.39; 95% CI 1.24–1.56). Education platforms showed weaker or non-significant pooled effects (e.g., website OR 1.07; 95% CI 0.99–1.15; video OR 1.55; 95% CI 0.92–2.60; application OR 1.53; 95% CI 0.97–2.42; webinar OR 1.27; 95% CI 0.96–2.69; Facebook OR 1.01; 95% CI 0.95–1.08). - Non-significant moderators: age group, intervention target (patients, parents, providers), intervention site (clinic vs non-clinic), outcome type (initiation vs completion), minority status, and control condition. - Publication bias: Funnel plot appeared symmetrical; Egger’s test was significant (intercept 1.35; 95% CI 0.78–1.92; p < 0.001), suggesting potential bias. Trim-and-fill imputed 14 studies; pooled OR reduced from 1.14 (95% CI 1.09–1.18) to 1.08 (95% CI 1.03–1.12) but remained significant. No evidence of small-study effects.

This study provides a comprehensive meta-analytic comparison of five common digital intervention categories for promoting HPV vaccination uptake. Findings show that digital interventions, particularly client and provider reminders, effectively increase uptake, with effect sizes comparable to those reported for face-to-face clinical or school-based interventions. Client education alone did not significantly improve uptake, aligning with prior evidence that educational efforts without action-oriented prompts may not translate into behavior change. Combining client education with reminders did not outperform reminders alone, suggesting potential redundancy or increased complexity without added benefit. Moderation analyses highlight that interventions targeting male or mixed-gender populations and those delivered via reminder-oriented platforms (SMS, preference-based modalities, EHR alerts) are more effective than education-focused platforms. The results support integrating reminder systems into routine care and leveraging provider-facing tools such as EHR alerts to prompt vaccination, with potential cost-effectiveness advantages and scalability. These findings address the research questions by identifying which digital strategies work best, quantifying their effects, and clarifying contextual factors that enhance efficacy.

Digital interventions increase HPV vaccination uptake, with client and provider reminder strategies showing the strongest and most consistent effects, followed by provider education and combined client education plus reminders. Client education alone was not effective. Health systems should prioritize integrating digital reminder systems for both patients and providers to improve uptake at scale. Future research should: (1) examine implementation and cost-effectiveness in diverse settings, particularly low- and middle-income countries; (2) standardize outcome measures and employ longitudinal designs to assess durability; (3) explore interaction effects between family-level factors (e.g., gender) and program-level features (e.g., platform) to optimize targeted intervention design; and (4) conduct large pragmatic trials assessing real-world engagement and scalability.

Control conditions varied across studies (usual care vs attention/education controls), complicating direct comparisons. There was heterogeneity in intervention content, delivery modalities, and outcome measurement methods, though heterogeneity statistics were generally moderate. Most studies were conducted in high-income countries, limiting generalizability to low- and middle-income settings. Few studies assessed outcomes beyond two years post-intervention, limiting understanding of long-term effects. Some risk-of-bias concerns were noted, particularly around randomization reporting and blinding feasibility in behavioral interventions. Publication bias analyses indicated potential bias, though adjusted pooled effects remained significant.