Estimating the effects of interventions on increasing vaccination: systematic review and meta-analysis

Abstract

As global vaccination rates have reached their lowest point in nearly 15 years, effective interventions are being required globally to promote vaccination; however, there is a lack of rigorous evaluation of the effect of various interventions. Through a global synthesis, we analysed data from approximately 6 125 795 participants across 319 studies in 41 countries to reveal the global landscape of four intervention themes and to assess their effectiveness in increasing vaccination rates. We found an overall positive effect of the interventions across four main themes on improving vaccination. Specifically, dialogue-based interventions increased vaccination rates by 43.1% (95% CI: 29.8 to 57.9%, with effect sizes measured as relative risks (RRs)), though they may not always be effective in adolescents or in the sample with a higher percentage of male participants. Incentive-based interventions, whether implemented alone or combined with other intervention themes, failed to demonstrate a significant effect in children. Reminder/recall-based interventions were also effective for promoting vaccination (38.5% increase, 95% CI: 28.9 to 48.9%), particularly for completing vaccine series. Multi-component interventions exhibited excellent effectiveness in vaccination (54.3% increase, 95% CI: 40.5 to 69.6%), with the combination of dialogue, incentive and reminder/recall proving more effective than other multi-component interventions, but showing no significant effects in populations with high initial vaccination rates. However, we found that in most cases combining additional interventions with a single intervention may not significantly improve their effectiveness, especially for incentive-based interventions, but dialogue-based and reminder/recall-based interventions appear to be beneficial in some specific combinations. These findings underscore the importance of governments, public health officials and advocacy groups implementing appropriate vaccine interventions by selecting interventions tailored to specific populations, strategically promoting the completion of vaccine series and effectively combining interventions to promote global vaccination and save more lives.

Keywords: Public Health.

Figure 1. Effects of interventions on vaccination. Effects of four main interventions (upper panel) and specific interventions under four themes (lower panel) on vaccination. A small number of RR values are not displayed because they fell outside the display range, including two highly positive RRs for dialogue-based intervention, one highly positive RR for reminder/recall-based intervention and six highly positive RRs for multi-component intervention.

Figure 2. Effects of specific intervention type on vaccination split by participants’ age. Only intervention types with more than seven RRs were displayed in the forest plot. Specific effect sizes and other statistics applying to this forest plot are presented in online supplemental table S4.

Figure 3. Effects of specific intervention type on vaccination split by participants’ sex. Only intervention types with more than seven RRs were displayed in the forest plot. Specific effect sizes and other statistics applying to this forest plot are presented in online supplemental table S5.

Figure 4. Comparison of vaccination outcomes under a single strategy in relation to implementation of different multi-component interventions. Results are shown for subcategories of studies that indicated whether a single intervention was combined with other interventions as a specific multi-component intervention (interventions containing dialogue, incentive and reminder/recall, respectively): DI (dialogue-based intervention), II (incentive-based intervention), RI (reminder/recall-based intervention), DI+DI (multidialogue intervention), RI+RI (multi-reminder/recall intervention), DI+II (multi-component intervention combining dialogue and incentive), DI+RI (multi-component intervention combining dialogue and reminder/recall intervention), II+RI (multi-component intervention combining incentive and reminder/recall) or DI+II+RI (multi-component intervention combining dialogue, incentive). and reminder/recall). 95% CIs are depicted in heavy black lines (may be hidden by the mean data point), prediction intervals in thin black lines. The size of each data point is proportional to the precision of the study (1/vi (variance)). Significant differences between interventions are indicated by *, and comparisons are made with dialogue-based, incentive-based, reminder/recall-based strategies, respectively.

Figure 5. Factors explaining the effects of interventions on vaccination. Best models selected based on AICc scores identified the following factors as explaining RRs: (A) age for reminder/recall-based interventions; (B) intervention time for reminder/recall-based interventions, (C) specific vaccination behaviour for reminder/recall-based interventions, (D) region for multi-component interventions and (E) specific intervention type for multi-component interventions. Scattered dots represent RR values from primary studies, with dot size proportional to the weight of each RR in the meta-regressions. Fitted curve (black lines) and 95% confidence band [dark blue band in (A) and (B)] are generated from meta-regression. A small number of RR values are not displayed because they fell outside the display range, including three high negative RRs for North America, and three, four and one high positive RRs for Europe and Central Asia, North America and other regions, respectively.