Practice- and Community-Based Interventions to Increase Human Papillomavirus Vaccine Coverage: A Systematic Review

Abstract

Importance: Vaccines against human papillomavirus (HPV) are recommended for routine use in adolescents aged 11 to 12 years in the United States, but uptake remains suboptimal. Educational interventions focused on parents and patients to increase coverage have not generally demonstrated effectiveness.

Objective: To systematically review the literature on effectiveness of interventions conducted at the practice or community level to increase uptake of HPV vaccines in the United States.

Evidence review: Keyword searches of the PubMed, Web of Science, and MEDLINE databases identified studies of adolescents that included the outcome of HPV vaccination published through July 2014. References of identified articles were also reviewed. A total of 366 records were screened, 38 full-text articles were reviewed, and 14 published studies were included. Results were summarized by different intervention approaches.

Findings: Practice- and community-based intervention approaches included reminder and recall (n = 7), physician-focused interventions (eg, audit and feedback) (n = 6), school-based programs (n = 2), and social marketing (n = 2) (2 interventions tested multiple approaches). Seven studies used a randomized design, and 8 used quasiexperimental approaches (one used both). Thirteen studies included girls, and 2 studies included boys. Studies were conducted in a variety of populations and geographic locations. Twelve studies reported significant increases in at least one HPV vaccination outcome, one reported a nonsignificant increase, and one reported mixed effects.

Conclusions and relevance: Most practice- and community-based interventions significantly increased HPV vaccination rates using varied approaches across diverse populations. This finding is in stark contrast to a recent review that did not find effects to warrant widespread implementation for any educational intervention. To address the current suboptimal rates of HPV vaccination in the United States, future efforts should focus on programs that can be implemented within health care settings, such as reminder and recall strategies and physician-focused efforts, as well as the use of alternative community-based locations, such as schools.

Figure 1

Flow Diagram (Preferred Reporting Items for Systematic Review and Meta-Analyses) of Articles Considered for Inclusion HPV indicates human papillomavirus.

Figure 2

Forest Plot of Intervention Results The text, eTable 1 and eTable 2 in the Supplement, and original sources provide additional study details. Blue squares are estimates presented in the original source article; orange squares are estimates calculated by review authors based on data presented in the original source article. Size of squares is proportional to the study sample size: small squares, fewer than 250 patients; medium squares, 250 through 999 patients; and large squares, 1000 or more patients. Multiple estimates presented from single studies are noted by author name. Error bars indicate 95% CIs. CDS indicates clinical decision support. ^aSample size is not reported. Estimates for nonrandomized clinical trials are not reported but are described as “did not improve” for 13-year-olds and “rate of increase slowed” among 14-year-olds, and both were nonsignificant. ^bEstimate is significant at P < .05 when no 95% CIs were reported. ^cEffect estimate is reported as 6.6 and is not represented on the graph because it is beyond the range used for graphic presentation (n = 1000). ^dEstimates for combined intervention (physician and family) compared with the control condition. ^eEstimates in the article are presented as difference (not ratio) measures and are not included on the plot. Differences for human papillomavirus 1 (HPV-1) were significant at P < .05 for the 11- to 12-year-old group but not the 13- to 18-year-old group, and differences in HPV-3 were significant at P < .05 for the 13- to 18-year-old group but not the 11- to 12-year-old group. ^fEffect estimates of 9.4 (95% CI, 2.6–33.1) for HPV-1 and 22.5 (95% CI, 4.3–118.0) are not represented on the graph because they are beyond the range used for graphic presentation (n ≤250). ^gComparison group is regional counties (not state). ^hComparison group is opt-out condition. ⁱComparison group is standard of care condition. ^jEffect estimate is reported as 6.56 (95% CI, 3.99–10.78) and is not represented on the graph because it is beyond the range used for graphic presentation (n ≥1000).