The Impact of Theory in HPV Vaccination Promotion Research: A Systematic Review and Meta-Analysis

Meta-Analyses of HPV Vaccination Literature

Scholars dedicated to facilitating health promotion are in agreement that meta-analyses and systematic reviews are crucial to the development of the field. ^35,40,41 In the context of vaccination promotion, only a handful of meta-analyses and systematic reviews have endeavored to synthesize theoretical and strategic intervention effects. For instance, O’Keefe and Nan examined the effectiveness of gain- and loss-framing in promoting vaccination. ⁴² The results found that loss-framing was potentially more persuasive than gain-framing. A similar systematic review was conducted by Penţa and Băban, in which they synthesized gain- and loss-framing techniques utilized in vaccination promotion with specific attention on moderators and mediators that may influence the main effects. ⁴³ MacDonald et al focused on vaccination promotion in European countries and identified strategies and delivery methods that were prevalent and effective among targeted populations. ⁴⁴ In addition, Newman and Logie conducted a systematic review and meta-analysis and found that evidence-informed promotional strategies were more effective in boosting the acceptance of vaccines among target populations. ⁴⁵ Key findings and patterns have emerged from these meta-analyses and systematic reviews to inform future research in the field.

In the case of HPV, only 4 systematic reviews have specifically analyzed HPV vaccination promotion. ^15,18,19,22 After analyzing 33 educational interventions, Fu et al found that educational interventions were more influential in increasing HPV vaccination intentions among adolescents and young adults. ¹⁹ Walling et al’s review of 51 national and international intervention studies showed that informational interventions with educational campaigns and environmental interventions that decreased social barriers showed the greatest potential to improve HPV vaccine uptake. ¹⁸ In another review, Francis et al examined 11 intervention studies and highlighted the benefit of using communication technologies such as interactive computer videos in increasing vaccine uptake. ²² More recently, Holloway focused on specific intervention delivery strategies in HPV vaccination promotion. ¹⁵ Results suggested that multi-method strategies, including reminder system and provider recommendations, are more promising for increasing HPV vaccination rates.

Interestingly, none of the recently published reviews examined the influence of theory use in HPV vaccination promotion, even though previous research indicated the importance of theories in developing message strategies and producing behavioral changes. ^46,47 Specifically, Nan employed framing theory in studying HPV vaccination promotion messages. ¹¹ Gain-framed messages are defined as those that highlight the advantages of engaging in certain behaviors, and loss-framed messages as those that stress the consequences of not adopting the behavior. Results from the study indicated that the loss-framed HPV vaccination promotion messages were more persuasive.

Furthermore, previous reviews of HPV vaccinations focused only on intervention studies, ignoring the experimental studies examining the persuasive effects of message manipulations. ^18,19 This inevitably results in an incomprehensive overview of the promotional strategies on HPV vaccination-related outcomes. Although these reviews examined some sort of HPV vaccination-related outcomes, they did not establish existing connections between theory use and outcomes. Thus, the limitations of the recently published reviews warrant a more comprehensive examination of the literature. To address these limitations, this study consists of both a systematic review as well as a meta-analysis of experimental studies to examine the influence of theory on HPV vaccination promotion. We propose the following questions to guide our study:

RQ1: What are the theories used in previous HPV vaccination promotion experimental research?

Method

Search Strategy

Similar to previous reviews, we located the sample of studies in the following databases: Academic Search Complete, Business Source Complete, PubMed, PsycINFO, Web of Science, Communication and Mass Media Complete (CMMC), Cumulative Index of Nursing and Allied Health Literature (CINAHL), and MEDLINE. ^15,19,48 We used the following search queries: “fram*” OR “messag*” OR “strateg*” OR “persuas*” OR “persuad*” OR “narrat*” OR “promot*” OR “appeal” OR “intervention” OR “campaign”; AND “Human papillomavirus vaccin*” OR “HPV vaccin*” OR “Human papilloma virus vaccin*”; AND “behavior*” OR “behaviour*” OR “intent*” OR “attitud*.” Our initial search yielded a total of 2247 studies. All 2247 were screened based on our inclusion and exclusion criteria (detailed in the next section). The detailed screening process is listed in Figure 1.

OPEN IN VIEWER

Figure 1.

Systematic literature review screening process.

Inclusion and Exclusion Criteria

Based on previous research, eligible studies had to meet the following criteria: 1) published in English; 2) published in scholarly and peer-reviewed journals; 3) available online and published between 2006 to 2018 since FDA approved the vaccination in 2006 ⁴⁹ ; 4) utilized experimental or quasi-experimental designs; and 5) measured HPV vaccination-related attitude, intention, or behavior as outcomes. ^{15,18,19,22,48} After applying the inclusion and exclusion criteria to the 2247 studies, 70 studies remained for the systematic review.

In addition, to be included for meta-analysis, studies had to include a control condition. This criterion was included based on prior, which suggests that a comparison condition has to be present to examine the mean effect size of treatment. ^42,50 The comparison group served as the control group without any manipulations, allowing us to examine the effectiveness of promotion strategies in the experimental condition. All studies had to report the relevant statistics regarding outcomes such as correlation coefficients to allow for the calculation of effect sizes (e.g., mean, standard deviation, and sample size for both experimental and control conditions). ^50,51 Thirty studies were included for the final meta-analysis.

Results

The results section is organized around the 4 research questions this study seeks to answer. Research questions 1 and 2 are answered with results from the systematic review portion of the study, and research questions 3 and 4 are answered with results from the meta-analysis and the subsequent moderator analysis.

Study Characteristics

Seventy studies were eligible for our systematic review. In general, most studies were published after 2010 (Figure 2). HPV vaccination studies the most frequently published in Health Communication (11%, k = 8), followed by Health Psychology and Vaccine (7%, k = 5 respectively; for more details, see Figure 3). The majority of the studies were conducted in the United States (77%, k = 54), followed by China including Hong Kong, Macau, and Taiwan (7%, k = 5), Canada (4%, k = 3), Australia (3%, k = 2), Sweden (3%, k = 2), United Kingdom (1%, k = 1), Germany (1%, k = 1), Hungary (1%, k = 1), and the Netherlands (1%, k = 1). Forty-seven percent of the studies (k = 33) had a control group present and 30% (k = 21) used a factorial between-subject design.

OPEN IN VIEWER

Figure 2.

Percentage of publication by year. Note: one study is coded 2019 since it is scheduled in a journal’s 2019 issue but is published online in 2018.

OPEN IN VIEWER

Figure 3.

Percentage of publication by journal.

With regard to participants, 80% (k = 56) of the studies used a convenience sample. Additionally, 47% (k = 33) of the studies used both female and male; 46% (k = 32) used female-only participants; and 7% (k = 5) solely focused on male participants. We also coded whether the participants of each study were parents, adolescents, young adults, or others. More than half the studies target young adults (54%; k = 38), 20% (k = 14) targeted parents, 13% (k = 9) targeted adolescents, 9% (k = 6) targeted adults in general, and 4% (k = 3) used both parents and adolescents as participants.

Research Question 1: Theory Use

To answer research question 1, we examined the use of theory in the sample. Our findings show that 22% (k = 19) of the total sample did not use any theoretical guidance; 26 theories and theoretical constructs were utilized in the rest of the studies (see Figure 4). Framing theory was the most frequently used (24%, k = 23). Within these studies, the gain vs. loss framing was the most prevalent strategy (56%, k = 14). Other major theories included the Health Belief Model (HBM, 13%, k = 12), which identified perceived susceptibility, perceived severity, perceived barriers, perceived benefits, cues to action, and self-efficacy as crucial factors for predicting healthcare behaviors, ⁵⁶ and Narrative (7%, k = 6). ⁵⁷ which emphasized the importance of storytelling in persuasion (see Figure 4 for the complete list).

OPEN IN VIEWER

Figure 4.

Theory Use.

Research Question 2: Theory and Main Outcomes

To answer research question 2, we studied the main outcomes in the published literature. The main outcomes were attitude about HPV vaccination (23%, k = 16), intention to get the HPV vaccine (76%, k = 53), and behavior (27%, k = 19). These were not mutually exclusive. We also examined the relationship between the theory used and the outcomes.

Attitude

Of the total sample, 23% (k = 16) of the studies measured attitude as an outcome or one of the outcomes. Of the studies that measured and reported attitude toward the HPV vaccine, 75% (k = 12) used at least 1 theoretical framework.

The majority of studies (58%, k = 7) used Framing theory as theoretical guidance. Loss framing appeared to be more effective than gain framing, ^11,29 while some studies revealed no main effect of gain vs. loss framing or equivalent effectiveness of gain and loss framing. ^24,28 Positive framing of the vaccine elicited more positive attitudes than negative framing. ^27,58 Further, temporal framing had no main effect on attitude. ⁵⁹

There are a few takeaways from studies using other theories. Specifically, using inoculated messages, facilitating interpersonal discussion of HPV vaccination, attributing the threat of HPV to the virus rather than a human agency, and increasing visual vividness of HPV vaccination messages were likely to have a positive impact on individuals’ attitude (for details, see Table 1).

OPEN IN VIEWER

Table 1.

Summary of Findings on Theory Use and Outcomes.*

Author(s), year	Theory	Attitude	Intention	Behavior
Avery & Park, 2018	11	N/a	Fear appeal > non-fear appeal	N/a
Bell et al, 2014	4	Assigning threat agency to virus > to humans	N/a	N/a
Bennett et al, 2015	24	N/a	Tailoring: no main effect	Tailoring: no main effect
Bigman et al, 2010	12	Positive framing > negative framing	N/a	N/a
Carcioppolo et al, 2013	12	N/a	STI- vs. Cancer-prevention framing: no main effect	N/a
Carcioppolo et al, 2017	11; 13	N/a	Fear appeal > non-fear appeal; guilt appeal: conditional and inconclusive	N/a
Cox et al, 2010	3	N/a	Foot-in-the-door strategies > non-foot-in-the-door strategies	N/a
Cox et al, 2014	1	N/a	Anticipated regret: no main effect	N/a
Diclemente et al, 2011	12	N/a	STI- vs. Cancer-prevention framing: no main effect	N/a
Diclemente et al, 2015	15	N/a	N/a	IMB-based intervention: effective
Dunlop et al, 2010	7; 18	Conversation > non-conversation; narrative: no main effect	Narrative: indirect effect; conversation: no main effect	N/a
Gainforth & Latimer, 2011	12	N/a	Gain vs. Loss framing: no main effect	N/a
Gainforth et al, 2012	12	Gain vs. Loss framing: no main effect	Gain framing > mixed framing (in mothers for sons)	N/a
Gerend & Shepherd, 2007	12	N/a	Loss framing > gain framing	N/a
Gerend & Shepherd, 2012	12	Gain framing = loss framing	N/a	Framing: no main effect
Gerend & Sias, 2009	12	N/a	Loss framing > gain framing	N/a
Gerend et al, 2008	12	N/a	Loss framing > gain framing	N/a
Gerend et al, 2013	14; 24	N/a	HBM-based manipulation: effective; tailoring > non-tailoring	N/a
Gottvall et al, 2010	14	HBM-based intervention: no main effect	N/a	N/a
Grandahl et al, 2016	14	N/a	N/a	HBM-based intervention: effective
Joseph et al, 2016	2	N/a	N/a	Initiation and completion: Not different between groups
Juraskova et al, 2011	12	N/a	STI- vs. Cancer-prevention framing: no main effect	Framing: no main effect
Kennedy et al, 2011	18	N/a	Narrative: indirect effect	N/a
Kim & Hmielowski, 2017	23	N/a	SCT-based manipulations (self-efficacy): effective	N/a
Kim & Nan, 2016	12	Temporal framing: no main effect	Temporal framing: no main effect	N/a
Krawczyk et al, 2012	18	N/a	HBM-based manipulation: effective	N/a
Lai et al, 2015	18	N/a	HBM-based manipulation: effective	N/a
Leader et al, 2009	12	N/a	STI- vs. Cancer-prevention framing: no main effect	N/a
Lechuga et al, 2011	12	N/a	Loss framing > gain framing	N/a
Lee & Cho, 2017	12	N/a	Loss framing > gain framing	N/a
Liu et al, 2019	12; 18	N/a	Gain vs. Loss framing: no main effect; narrative: indirect effect	N/a
Mcree et al, 2010	12	N/a	STI + cancer prevention framing > STI prevention framing	N/a
Mehta et al, 2013	14	N/a	HBM-based intervention: effective	N/a
Nan & Madden, 2012	12	Positive framing > negative framing	N/a	N/a
Nan, 2012a	12	Loss framing > gain framing	Loss framing > gain framing	N/a
Nan, 2012b	12	N/a	Loss framing > gain framing	N/a
Nan et al, 2015	18	N/a	Narrative: indirect effect	N/a
Nan et al, 2017	18	N/a	Narrative: indirect effect	N/a
Nan et al, 2016	12	N/a	Gain vs. Loss framing: no main effect	N/a
Park, 2012	12; 14	Loss framing > gain framing; HBM-based manipulation: no main effect	Gain vs. Loss framing: no main effect; HBM-based manipulation: effective	N/a
Pepper et al, 2012	15; 6	N/a	IMB-led intervention: conditional and inconclusive; higher credibility > lower credibility	N/a
Perez et al, 2016	15	IMB-led intervention > control	IMB-led intervention > control	IMB-led intervention: no main effect
Sanderson et al, 2017	22	N/a	N/a	SEM-guided intervention: effective
Tiro et al, 2015	19	N/a	N/a	Precaution adoption process model-based intervention: no main effect
Vanderpool et al, 2013	15	N/a	N/a	IMB-led intervention: effective
Walter et al, 2017	18	N/a	Narrative: no main effect	N/a
Wen & Shen, 2016	12	N/a	Gain vs. Loss framing: no main effect; temporal framing: no main effect	N/a
Wong, 2015	16	Inoculated > non-inoculated	Inoculated > non-inoculated	N/a
Yang & Guo, 2015	9; 17	Greater visual vividness > lower visual vividness (in imagery ability limited people)	N/a	N/a
Yang & Pittman, 2017	12	N/a	Non-STI framing > STI-framing (in women only); attribution framing: no main effect	N/a
Zimmerman et al, 2017	8	N/a	N/a	Intervention: effective for vaccine innitiation but not completion

Note: *Studies using theories to guide experiment development were reported. Theory codes: Construct: Anticipated regret = 1, Construct: Brief negotiated interviewing = 2, Construct: Foot in the door = 3, Construct: Linguistic agency = 4, Construct: Regulatory fit principle = 5, Constructs: Credibility = 6, Construct: Role of conversation = 7, Diffusion of Innovations = 8, Dual coding model = 9, EPPM = 10, Fear appeal = 11, Framing = 12, Guilt appeal = 13, HBM = 14, IMB = 15, Inoculation theory = 16, LC4MP = 17, Narrative = 18, Precaution adoption process model = 19, Priming = 20, Protection motivation theory = 21, SEM = 22, Social cognitive theory = 23, Tailoring = 24, TPB = 25, TRA = 26.

Intention

Of the total sample, 76% (k = 53) of the studies measured intention as an outcome or one of the outcomes. Of the studies that measured and reported the intention of getting the HPV vaccine, 77% (k = 41) used at least 1 theoretical framework.

The majority of studies (51%, k = 21) used Framing theory as their theoretical guidance for the study. Results indicated that loss-framing may have an edge over gain-framing in increasing one’s intention to get the vaccine. ^{9,11,28,31,32,60,61} However, notably, multiple studies found no main effect of gain vs. loss framing on intention. ^{62 -64,29,30,33} Moreover, framing HPV vaccination as sexually transmitted infection (STI)-prevention did not differ significantly from framing it as cervical cancer prevention in increasing intention. ^{23,65 -67} Instead, a mixed framing of cancer and STI prevention may be the most. ⁶⁸ However, caution is advised in framing the vaccine as a prevention method for STI; the main effect was minimal when using temporal framing and attribution framing. ^30,59,69

Takeaways from studies using other theories are as follows: credible messages and inoculated messages can lead to the positive intention of getting HPV vaccines; interventions that are HBM-based, SCT-based, or guided by the foot-in-the-door strategy (involving acquiring compliance with a small request before making another more demanding request may help increase HPV vaccination intention (for details, see Table 1). ⁷⁰

Behavior

Of the total sample, 26% (k = 18) of the studies measured behavior as an outcome or one of the outcomes. Of the studies that measured and reported HPV vaccine uptake, 61% (k = 11) used at least 1 theoretical framework.

Most studies used Information Motivation Behavioral Skills (IMB, k = 3) and HBM as theoretical frameworks (k = 3). Results indicated that both models merit further investigation. Two IMB-led interventions found no main effect, and one intervention significantly increased uptake. ^12,25,26 HBM was used in 2 studies without guiding manipulation development. ^23,24 Only 1 HBM-based intervention demonstrated success. ⁷¹

Takeaways from studies using other theories suggested that interventions led by the Social Ecological Model (SEM) and Diffusion of innovations may help increase vaccine uptake (for details, see Table 1).

Research Question 3: Mean Effect Size—Meta-analysis

Of the 70 studies included in the systematic review, 30 studies met the inclusion criteria for the meta-analysis. This sample of 30 studies yielded 31 independent effect sizes, with a total of 29,036 participants. The data was first screened for outliers in SPSS. As no outliers were detected, all 31 effect sizes were retained for subsequent analyses. Positive weighted mean effect size, r+, indicates enhanced HPV vaccination-related outcomes for the experimental condition compared to the non-manipulated control condition. Table 2 provides the overall weighted mean effect size of receiving promotional strategies on outcomes. According to Cohen’s guidelines, results indicate a statistically significant moderate effect of the use of promotional strategies compared to the control condition that did not use any promotional strategies, r+ = .25, p < .001.

OPEN IN VIEWER

Table 2.

Summary of Results.

	N	k	r+	95% CI	Q(df)	p
All cases (Outcomes)	29026	31	0.25	[0.15, 0.35]	–	<.001
Use of Theory
No	13018	15**	0.25	[0.19, 0.48]
Yes	16018	16*	0.17	[0.01, 0.32]
Between Levels					2.56(1)	.11
Specific Theory
Framing	1371	3	-0.29	[-0.62, 0.13]
HBM	1217	3	0.04	[-0.37, 0.44]
IMB	622	3**	0.75	[0.49, 0.89]
Others	12808	7	0.11	[-0.17, 0.38]
Between Levels					17.98(3)	<.001

Note: *p < .05, **p < .001.

Discussion

The importance of HPV vaccination in preventing severe diseases in men and women has been repeatedly proven and well recognized. ^1,2 However, low-to-middle income countries often fall short of the benchmark for vaccination coverage. Surprisingly, high-income countries such as the U.S. also fall behind the goal for the HPV vaccine uptake. ⁷² Undoubtedly, there is a need to improve promotional strategies related to HPV vaccination. Therefore, utilizing a systematic review and meta-analysis of prior literature, we aimed to understand the strategies that have been employed in HPV vaccination promotion thus far, and outline a blueprint for future research.

Central to our inquiry is a synthesis of the theories used in previous studies. Notably, 22% of the sample did not use any theoretical guidance. Among the theories used, Framing, HBM, and Narrative were relatively more prevalent, while other behavioral change theories (e.g., the Reasoned Action Approach) and media effects theories (e.g., Priming) remained scarce. The current study also systematically examined outcomes and their relationships with the theories used in previous studies. From the results, intention appeared to be the major outcome investigated in the past, followed by behavior and attitude. However, as for the relationships between outcomes and theories used, no conclusive finding was yielded due to the lack of research replication and the mixed nature of results.

Directions for Future Research

Nevertheless, several observations from our systematic review can inform the design of future research and practical strategies. First, although a few studies did not find significant main effects, the overall finding suggests that loss-framing appeared to be more competent than gain-framing in improving attitude and intention toward HPV vaccination. Second, messages designed based on inoculation theory are more likely to enhance attitude and intention compared to the control. Third, messages that are positively framed, virus-attributed (as opposed to human agency-attributed), and visually vivid or interventions that can facilitate interpersonal discussions could elicit positive attitudes. Fourth, messages and strategies perceived as credible, HBM-based, SCT-based, or guided by the foot-in-the-door strategy appeared to strengthen intention. Fifth, messages or interventions designed based on HBM, IMB, SEM, and diffusion of innovations appeared to be capable of promoting vaccine uptake. Lastly, future research should consider validating the relationship between theories and vaccination behaviors. Our review of the literature reveals that there is currently a shortage of such research, thus limiting the extent of our understanding of vaccination promotion strategies.

The implications of these observations are 3-fold. First, these findings are crucial for health promotion practices. Previous studies suggest that promotional messages not guided by theories may be detrimental or counterproductive. ¹³ Thus, using these empirically tested frameworks as guidance may help health practitioners formulate accurate and effective messages. Second, more research should be devoted to testing theories related to media effects and behavioral change. For example, even though scholars have clearly articulated the effectiveness of social norms (i.e., the component in the Reasoned Action Approach) in HPV vaccination promotion, empirical research testing its effectiveness in increasing the vaccine uptake remains scant. ^73,74 Besides, Fishbein and Yzer suggest that the Theory of Priming, which is underused in previous studies, could strengthen the correlation between individuals’ beliefs and attitudes about HPV vaccination. ⁷⁵ Therefore, we call upon health scholars to further examine the behavioral change and media effects theories underutilized as indicated by our findings (for details, see Figure 4 and Table 1). Third, the complexity of HPV vaccination (e.g., it requires one to complete 3 doses) may result in complications in the process from expressing intentions to enacting the behavior. Thus, we encourage health scholars to collaborate with clinics and nursing programs to measure the effectiveness of messages on behavior in the future.

There are several additional findings that emerged from the meta-analysis results that merit in-depth discussion. Although the present study found that the use of theory was not a significant moderator, researchers should nevertheless continue to examine the role of theory use. It is possible that the insignificant finding obtained in this study was due to the limited sample size as over half of the studies were excluded from meta-analysis due to the absence of control groups. ^5,32 Indeed, some theories (e.g., fear appeal) were demonstrated to be informative and helpful for crafting effective messages. ⁷⁶ Furthermore, previous studies have also demonstrated the importance of theory use in empirical studies. ^46,47 Thus, future studies are encouraged to continue to explore various theoretical frameworks and the mechanisms involved in the processing of these various messages. This can help the field to validate the effectiveness of different theories in this regard.

The field also calls for more controlled experiments investigating theory-guided promotional strategies. Among the 70 eligible studies, half of them are un-controlled, which resulted in a reduced sample for meta-analysis. This not only hinders a comprehensive synthesis of the effectiveness of promotional strategies, but also deters researchers from making scientific and irrefutable conclusions regarding which strategy is the most efficient in promoting HPV vaccination. With the current sample, results suggest that promotional strategies were significantly more effective compared to the control. This finding aligns with previous research and is quite promising for health practitioners and organizations. ^13,48 It indicates that the effort to promote vaccination does not go in vain. Promotional strategies in any format did help increase HPV vaccination-related outcomes. Thus, health organizations should pay close attention to theory-guided strategies in promoting HPV vaccination on social media platforms to reach targeted audience. More importantly, the information, motivation, behavioral skills model (IMB) appeared to be a promising theoretical framework for designing promotional strategies. The IMB framework puts forth that people need to have relevant information, social and personal motivation, and behavioral skills to engage and perform the target behavior fully and correctly. ⁷⁷ Indeed, previous studies also demonstrated that interventions that address these IMB components were more effective than control conditions. ²⁶ Future studies are, therefore, highly encouraged to examine the effectiveness of this theoretical framework further.

Overall, most of the studies were conducted in the U.S. This is worrying since increasing HPV vaccine uptake is more urgent in low-to-middle income countries. ⁷⁸ Although this may be due to the constraint that we only coded articles in English, empirical research in English-speaking low-to-middle income countries such as India, Philippines, and Nigeria were virtually absent. Thus, future research promoting HPV vaccine should also put more focus on the non-U.S. context. Lastly, research solely focusing on males has been scarce. Previous studies showed a glaring discrepancy in the uptake rate of HPV vaccination between male and female individuals. ⁷⁹ More importantly, male individuals have lower levels of awareness and acceptance of the vaccine. ^80,81 Therefore, future research is encouraged to develop strategies that exclusively target males to increase their knowledge levels, perceived urgency, and motivation.

Limitations

Our study is not without limitations. First, there is a possibility that our conclusions were limited by the body of literature included. Only 30 articles of our sample were included for meta-analysis, and subsequent subgroup analyses had a smaller number of independent samples (k < 10). These may contribute to imprecise point estimates. ⁸² This imprecision is mainly attributed to the lack of a large base of original studies that used a controlled experimental design. As a result, only a few studies employed similar theories to guide promotional strategy development among controlled studies. Moreover, to ensure the quality of studies in our meta-analysis, we only included peer-reviewed journal articles, potentially leading to publication. ⁸³ However, our fail-safe N test results adopted to estimate publication bias indicated that the bias was negligent in the current study (Supplementary 1). We also excluded non-English studies by screening them out at the initial stages of the search because of the language constraints. Therefore, while we employed an inclusive data gathering method, we might have also missed some relevant papers in the process of study retrieval due to our inclusion and exclusion criteria.

Third, ethnicity and race could contribute to predicting vaccine uptake. ³⁶ However, most of the studies in the current meta-analysis were carried out with Caucasians as the majority, and only a few studies recruited participants of other races, such as Asian. ^9,30,60,65 Therefore, we did not include the demographic factor as a potential moderator in our analysis because of the small distribution. Lastly, this study was not able to determine which HPV promotional message strategy is the most effective due to the data limitation. A sufficient number of students for each strategy utilized is required for comparison of effect sizes of each strategy. However, the current sample we collected used a variety of strategies. Within each strategy, the sample size was too small to support a meta-analysis of that scale. Moreover, to conduct the current meta-analysis, we only included controlled studies, resulting in a reduced sample size. This is due to the fact that both a control and a treatment group are required for an examination of effect sizes. In our study, the treatment group is the promotional strategy while the comparison group is the control condition with no manipulation. That being said, we recommend future research to replicate our analyses when the original studies of each strategy become adequate.

Conclusion

Despite these limitations, the current study contributes to the HPV vaccination literature aimed at informing the effective design of promotional strategies. Our study is one of the first to use both a systematic review and a meta-analysis to examine the literature. Specifically, we examined the use of theory and its relationship with a range of HPV vaccination-related outcomes and synthesized the respective effect sizes. More importantly, our study provides rigorous findings regarding strategy design that can enhance attitude, intention, and behaviors. Lastly, by offering and summarizing practical suggestions based on prior studies’ results, this examination of the published literature is crucial for efforts of future research and strategy design.

Supplemental Material