Messages to Motivate Human Papillomavirus Vaccination: National Studies of Parents and Physicians Free

This article is featured in Highlights of This Issue, p. 1367

Low uptake of human papillomavirus (HPV) vaccine is a serious public health problem that is thwarting progress in cancer prevention (1). Each year, an estimated 21,300 cancers in the United States are attributable to HPV types 16 and 18 (2). These types are preventable through vaccination, yet only 40% of females and 22% of males ages 13 to 17 have received all three doses in the vaccine series (3). This low HPV vaccination coverage is despite national guidelines for routine administration (4, 5) and relatively high levels of coverage for other adolescent vaccines (6). Missed clinical opportunities for concomitant vaccination, defined as clinical visits at which a child received another vaccine but not HPV vaccine, are an important reason HPV vaccine initiation is not higher (7).

Healthcare provider recommendations are uniquely persuasive in motivating HPV vaccination. However, many physicians offer weak (8) or late recommendations (9), if they offer a recommendation at all. Physicians' HPV vaccine recommendations can be inconsistent, with many physicians selectively recommending the vaccine to patients they perceive are at higher risk for infection rather than all patients (9). This practice is especially worrisome as receiving a physician recommendation is a robust predictor of HPV vaccine uptake (10–15). Given the central role physicians' recommendations play in vaccine uptake, low-quality or absent physician communication about HPV vaccine undermines the vaccine's tremendous potential to prevent anogenital cancers.

Recognizing the need for improved physician communication about HPV vaccination, the Centers for Disease Control and Prevention (CDC) developed and widely disseminated resources for healthcare providers through the national “You Are the Key” campaign (16). To help strengthen physicians' communication, CDC offers nine messages physicians could use to recommend HPV vaccine to parents (6, 17). These messages were designed to address common parental barriers to HPV vaccination (e.g., vaccine safety, efficacy, necessity) and resulted from several rounds of testing with parents and physicians. Beginning with a larger pool of messages, CDC surveyed mothers to assess their perceptions of messages that would help them make the decision to get their children vaccinated. CDC then interviewed pediatricians to assess their understanding and usability of the messages, and conducted focus groups with parents to refine the messages (J. Roark, MPH, written communication, June 2016). However, parents' perceptions of these messages remain unknown in the absence of published evaluation data. Moreover, at an average length of 52 words, these messages are fairly long, which may make them challenging to remember and thus deter physicians' use. Our study sought to identify messages physicians would use to recommend HPV vaccine and that would motivate parental acceptance. The study team also developed shorter messages and examined whether they were as effective as the longer CDC-developed messages.

We conducted two national, online surveys of parents and physicians in the U.S. who were members of standing panels maintained by a survey research company (18, 19). Parents were part of a panel of about 55,000 adults identified through a probability-based sampling of addresses from the U.S. Postal Service's Delivery Sequence File. This sample frame covers about 97% of U.S. households. The survey research company recruited its physician panelists through lists compiled by the American Medical Association (20). The University of North Carolina Institutional Review Board approved both study protocols. We described the participants, procedures, and survey instruments previously (9, 21, 22) and summarize them here briefly.

Parents completed our online survey between November 2014 and January 2015 (22). Eligible respondents were parents of an 11- to 17-year-old child who lived in the household at least half of the time. For parents who had more than one child in this age range, the survey asked parents to respond with regard to the child who had the most recent birthday. This selection method is quick, easy, and relatively noninvasive for participants (23), compared with a potentially more burdensome enumeration approach in which we obtain an identifier and request responses for every eligible child. Parents provided a name or nickname for the referent child, which was used in item stems throughout the survey. To support inclusion of lower-income respondents, the survey company provided a computer and free Internet access to panel members without these resources for the duration of their participation in the survey panel (19). Parents who already had a computer and Internet access received points for completing the survey, which could be redeemed for cash, products, or sweepstakes entries.

The survey company emailed invitations to 2,845 randomly selected parents. Of those, 1,760 accessed the survey site, provided informed consent, and completed an eligibility screener. A total of 1,518 parents met eligibility criteria and completed the survey. After excluding 14 parents who had incomplete data, the final analytic sample included 1,504 parents. We estimated that 13.75% of our total sample was not eligible for the survey, resulting in 2,454 eligible parents. Using American Association for Public Opinion Research (AAPOR) definition 5 (24), the overall response rate was 61% (1,504/2,454).

Physicians completed our online survey between April and June 2014 (9, 21). Eligible respondents were pediatricians and family physicians providing preventive care, including vaccinations, to patients ages 11 to 12. We were particularly interested in patients in this age group because they represent the target ages for routine HPV vaccination (2). The survey company emailed invitations to all 2,368 panel members with pediatric or family medicine specialties, and 1,022 (43%) clicked on the link provided to access the survey site. Of these, 776 (76%) met eligibility criteria and completed the survey. The survey company was unable to provide the information required to calculate a response rate using AAPOR definitions (24). Although data on the percentage of ineligible respondents were unavailable, overall, 33% of physicians in the panel completed the survey. Physicians received $25 to $45 for participating, with higher incentives used to encourage participation later in the fielding process.

The brief and longer motivational messages and their sources appear in Supplementary Table S1. In consultation with pediatricians, family physicians, and other experts on HPV vaccination, the study team developed six brief messages (defined as 25 words or fewer; messages B1–B6) designed to motivate HPV vaccination. Nine longer messages (defined as 26 words or longer; messages L1–L9) came from materials that the CDC developed to aid physician communication with parents and adolescents about HPV vaccination (6, 17). Although relatively little is known about the relationship between health-related print literacy and health-related oral literacy, the cognitive processes necessary to understand information presented orally and in writing are closely connected (25). Thus, we calculated message readability using five standard approaches (Flesch–Kincaid Grade Level, Gunning-Fog Score, Coleman–Liau Index, Simple Measure of Gobbledygook (SMOG) Index, Automated Readability Index) and calculated an average readability grade level for each message; grade 8 or lower is a minimum standard for readability for patient communication materials, and grade 6 or lower is desirable (26). On average, the brief messages were 17 words long, and the reading grade level was 6.3 (Supplementary Table S1). The longer messages averaged 52 words and a reading grade level of 9.8.

Study surveys are available online at www.unc.edu/~ntbrewer/hpv.htm.

The parent survey included the six brief motivational messages, plus nine longer messages (17), for a total of fifteen messages. Parents were randomly assigned to panels (Panel A, Panel B, Panel C), each of which included two brief messages and three longer messages in a random order. The survey asked parents, “Which of these statements made by a doctor would persuade you to get the HPV vaccine for [child's name]?” Before viewing the messages, parents indicated the number of HPV vaccine doses their children had received (coded as 0 or ≥1 doses). For children who had not initiated HPV vaccination, the survey assessed parents' HPV vaccination intention by asking, “How likely are you to get [child's name] the HPV vaccine in the next year? Would you say you … [definitely won't, probably won't, probably will, definitely will].” The survey also evaluated attitudes toward adolescent vaccines, including importance of HPV vaccine: “I feel that the HPV vaccine for [child's name] is … [not important, slightly important, moderately important, very important, extremely important].” On demographic and household characteristics, the survey assessed respondents' sex, age, race/ethnicity, education, marital status, household income, and state of residence. It also assessed the index child's sex and age.

The physician survey included the six brief motivational messages presented to parents and asked, “Which of these statements would you use to persuade parents to get HPV vaccine for their 11–12 year olds?” Physicians could choose as many messages that applied or select none of these. On demographic and professional characteristics, our survey assessed respondents' medical specialty, sex, and number of years in practice since residency. Respondents also indicated their typical adolescent patient volume per week and the percentage of vaccine doses they administer that are provided by the federal Vaccines for Children (VFC) program. On clinical practice characteristics, respondents reported their practice type and the total number of physicians in the clinic. Respondents also reported the state in which their clinic is located, which we categorized these locations into one of four U.S. census regions.

We calculated frequencies and percentages of respondents who indicated that a given message would persuade them (parents) or that they would use a given message to persuade parents (physicians) to get HPV vaccine for their child. For parent data, we used bivariate logistic regression models to determine if message endorsement differed by child's HPV vaccination status. Among parents whose child was not previously vaccinated, we conducted bivariate logistic regression models to assess whether message endorsement differed by HPV vaccine intention. For all parents, we used bivariate logistic regression to identify correlates of endorsement of each motivational message. Candidate correlates were child's HPV vaccination status, parent's intention to vaccinate their child, parent's race/ethnicity, parent's education level, child's age, child's sex, and parents' perception of HPV vaccine importance. The parent Panels A, B, and C were similar on 10 of 11 demographic variables, indicating that randomization successfully created largely equivalent groups; as the panels differed on income (P = 0.05), analyses controlled for income. Given that message endorsement differed by child's vaccination status, analyses also controlled for this variable. For physician data, we used χ² tests to identify correlates of endorsement of each motivational message. Candidate correlates were medical specialty and years in practice. We conducted analyses using SAS Version 9.4. Statistical tests were two-tailed with a critical alpha of 0.05.

About half of parents were female (56%; Table 1). About 9% of parents were non-Hispanic black and 14% were Hispanic. Nearly two-thirds (62%) had at least some college education. Almost half (47%) reported an annual household income of $75,000 or more, and the majority (84%) lived in a metropolitan area. About one-quarter (28%) responded with regard to an 11- or 12-year-old child. About half (51%) of parents reported their index child was male and just over half (54%) reported their child had not received any doses of HPV vaccine.

Just over half of respondents were pediatricians (53%), roughly two-thirds (68%) were male, and over half (55%) had been practicing for 20 or more years. The largest proportion saw 10 to 24 adolescent patients weekly (45%). Most respondents worked in private practices (85%), and about half (51%) reported that their practice had four or fewer physicians.

Parents' endorsement of message persuasiveness averaged 41% (range, 9%–65%) for brief messages and 60% (range, 42%–70%) for longer messages (Table 2). At least half of parents supported eight of the nine longer messages and three of six brief messages. About 18% of parents endorsed all messages they read, whereas 22% did not endorse any.

Among the brief messages, the highest proportion of parents endorsed a message that explicitly expressed a strong recommendation for HPV vaccine (B5 important cancer prevention vaccine; 65%; Fig. 1). Many parents also supported messages that speak directly to their role in preventing their child from getting anal/cervical cancer through HPV vaccination (B3 prevent cancer; 59%) and messages that emphasize their control over whether their child becomes infected with HPV (B4 control getting HPV; 58%). For the longer, CDC-developed messages, the highest proportion of parents supported a message with information about HPV vaccine effectiveness shown in clinical trials and studies in the U.S. and other countries (L8 effective vaccine; 70%; Table 2). About two-thirds of parents also endorsed messages about the prevalence of HPV infection and the cancers and precancerous conditions HPV vaccine protects against (L7 HPV causes cancers; 67%), messages about vaccine safety and an emphasis on minor, short-term side effects (L2 safe vaccine; 65%), and a longer version of the brief message in which a physician clearly gives a strong recommendation for HPV vaccine (L6 important cancer prevention vaccine; 65%).

Endorsement of each of the 15 messages was higher among parents whose children had received HPV vaccine (all P < 0.05; Table 2). Parents whose children had not received any doses of HPV vaccine favored five longer messages (L2 safe vaccine, L5 HPV is common, L6 important cancer prevention vaccine, L7 HPV causes cancers, and L8 effective vaccine; 53%–56%) and three short messages (B3 prevent cancer, B4 control getting HPV, and B5 important cancer prevention vaccine; 48%–49%). Few parents of unvaccinated children endorsed brief messages B6 (5%), which uses a car seatbelt analogy, and B1 (16%), in which a physician draws on the experience of vaccinating his or her own child. About two-thirds (67%) of parents of unvaccinated children supported at least one message, as did 91% of parents whose child received ≥1 dose of HPV vaccine.

In an exploratory analysis, we examined sociodemographic correlates of not endorsing any messages among parents of an unvaccinated child. In bivariate logistic regression analyses, we identified parent race and marital status as significant predictors of not endorsing any messages. In a multivariable model containing these two predictor variables, parents who reported black, non-Hispanic race had lower odds of not endorsing any messages compared with white, non-Hispanic parents [odds ratio (OR), 0.47; 95% confidence interval (CI), 0.24–0.92]. Parents who reported they were not married had lower odds of not endorsing any messages compared with parents who were married (OR, 0.58; 95% CI, 0.37–0.89).

For 14 of the 15 messages, parents with higher prior intentions to vaccinate were more likely to say the messages would be persuasive (all P < 0.001; Supplementary Table S2; Fig. 2). Of parents who probably would not get their child vaccinated, more than 60% endorsed three longer messages (L5 HPV is common, L7 HPV causes cancers, L8 effective vaccine; 61%–64%). About half endorsed three brief messages (B3 prevent cancer, B4 control getting HPV, and B5 important cancer prevention vaccine; 48%–51%). Of parents who definitely would not get their child vaccinated, about one-quarter indicated three brief messages would persuade them to vaccinate their child (B3 prevent cancer, B4 control getting HPV, and B5 important cancer prevention vaccine; 23%–25%). Two longer messages would be effective with about one-quarter of these parents (L5 HPV is common and L8 effective vaccine; 23%–24%). About one-third (33%) of parents who definitely would not get their child vaccinated and 75% of parents who probably would not get their child vaccinated endorsed at least one message (Fig. 3).

Endorsement was not associated with parent race/ethnicity, parent education, child age, or child sex (all P > 0.05; Supplementary Table S3).

Physicians' endorsement of the brief messages averaged 39%; about 4% of physicians endorsed all messages, whereas 7% did not endorse any messages. Of the six brief messages presented to physicians, the highest proportion of physicians endorsed B5 (important cancer prevention vaccine, 69%), closely followed by B3 (prevent cancer, 64%; Fig. 1). The lowest proportion of physicians (9%) reported they would use message B6 (seatbelt analogy) to persuade parents to agree to HPV vaccine.

Compared with pediatricians, a higher proportion of family physicians endorsed message B3 (prevent cancer, 59% vs. 69%, P = 0.002), but a lower proportion of family physicians endorsed messages B1 (physician vaccinates own child, 35% vs. 25%, P = 0.003) and B5 (important cancer prevention vaccine, 74% vs. 63%, P < 0.001). Relative to clinicians who have been in practice for ≤19 years, a higher proportion of those practicing longer endorsed messages B1 (physician vaccinates own child, 24% vs. 36%, P < 0.001) and B4 (control getting HPV, 32% vs. 41%, P = 0.019). No other associations were statistically significant (all P > 0.05).

Through our national survey of parents, we identified three brief (B3–B5) and eight longer (L1–L3, L5–L9) messages that at least half of parents reported would persuade them to agree to HPV vaccination for their children. Physicians also endorsed two of these brief messages touching on the strong belief “in the importance of this cancer-preventing vaccine” (B5) and that the child could get cancer as an adult “but you can stop that right now” (B3). Given the importance of provider recommendation in HPV vaccine uptake (7, 27), perhaps it is not surprising that a high proportion of parents endorsed a straightforward and strong recommendation for HPV vaccine. Equally high proportions of parents endorsed the brief and longer versions of this message, suggesting that such a recommendation need not be lengthy so long as the physician clearly states that he or she strongly recommends HPV vaccine.

Many parents also endorsed messages that included information about vaccine safety and effectiveness (L2, L8), the importance of vaccinating prior to sexual activity (L3), HPV infection being common (L5), and cancer prevention (L1, L6, L7, L9). These messages address the primary parental barriers to HPV vaccination, including concerns about vaccine safety, lack of knowledge, and belief that HPV vaccine is not needed or not necessary (7). Moreover, the messages align with communication best practices that suggest effective messages highlight HPV vaccine's role in cancer prevention, explain risks and prevalence of HPV infection, and provide evidence about HPV vaccine safety (28).

These messages may be especially useful as part of a larger HPV vaccine recommendation strategy. The survey used in the current study informed an intervention in which we developed and led a training that taught clinicians an HPV vaccine recommendation strategy. As part of the strategy, we trained clinicians on an approach for eliciting and easing parents' primary concern about HPV vaccine, and to end with a strong recommendation. In the context of this strategy, we foresee physicians using the messages from the current study during these steps. These messages may be particularly helpful once physicians have identified a focal point for discussion by eliciting parents' primary concern about HPV vaccine. To support their communication with parents, physicians can use the motivational message that best addresses the parent's specific concern. Also, physicians can use a message to end the discussion with a strong recommendation for HPV vaccine (e.g., B5 important cancer prevention vaccine). Given that we did not identify a difference in parental endorsement of messages across demographic subgroups, physicians may be able to effectively use these messages across demographic subgroups.

Of particular interest, some parents without prior intentions to vaccinate their child identified messages that would be persuasive. Two messages that were the most highly endorsed by these parents placed the onus of protection on the parent (B3) and emphasized the control they possess over their child's health (B4). This finding is aligned with the tenets of various theories of fear appeals, which posit that fear messages inspire action if the receiver believes he or she has some control over the situation (29). Parents without prior intentions to vaccinate their child may be particularly receptive to messages that arouse fear while fostering a sense of efficacy. Our findings offer promise for appealing to parents with a range of prior intentions to vaccinate.

Concordance between parents' and physicians' perceptions of the brief messages was fairly high. One exception was message B4 (control getting HPV), where many parents (58%) reported the message was persuasive, yet only 37% of physicians would use this message to persuade parents. It is especially concerning that so few physicians would use this message given that it was one of the top messages that would persuade parents without prior intentions to vaccinate their child. It is unclear whether physicians were not amenable to using this message because they personally do not believe in its content, or because they do not think the message would motivate parents.

We also identified some differences in message endorsement by physician specialty and length of time in practice. Pediatricians typically are stronger proponents of vaccination compared with other physician specialties, including family physicians (30, 31). Thus, it may be unsurprising that pediatricians preferred messages B1 (physician vaccinates own child) and B5 (important cancer prevention vaccine), both of which reflected personal endorsement and experience, whereas family physicians preferred message B3 (prevent cancer), which does not use language that reflects their personal beliefs. Given that so many parents reported message B5 (important cancer prevention vaccine) would persuade them to pursue HPV vaccination, it would be worthwhile to encourage family physicians to use this message or further explore their hesitancy to use it. Using this message could be easily incorporated into routine clinical practice and included in training materials targeted to helping family physicians make effective HPV vaccine recommendations.

To facilitate comprehension, experts recommend that health information be written at a sixth-grade or lower reading level (32, 33). We found that the brief messages were generally around this recommended grade level, whereas the longer messages were around four grades above the recommended level. CDC has since disseminated a new set of messages for physicians to use to address parents' questions about HPV vaccine that are somewhat shorter than the previous CDC messages but still lengthy (38 vs. 52 words; ref. 34). Grade level of the new messages remained high and unchanged at around tenth grade. Given that the messages are designed for physicians to deliver to parents orally, it is worthwhile for future research to explore parents' comprehension of the messages when spoken by a physician. Also of note, the highest proportion of parents endorsed the brief statement (B5 important cancer prevention vaccine) with the highest reading grade level (10), and over half of parents endorsed most of the longer statements, with grade levels ranging from 7 to 12. Future research also is needed to understand the trade-offs between message reading grade level, persuasiveness, message length, and ability for providers and patients to remember the messages.

Study strengths include data from large, national samples of parents, and primary care physicians, including physicians who practice in specialties that deliver the majority of HPV vaccine doses in the U.S. (10). To our knowledge, our study is among the first to evaluate the messages developed and disseminated by the CDC. Study weaknesses include that we achieved a modest cooperation rate for the physician survey, which is a common challenge for physician surveys (35). The study relied on parents' and physicians' self-report in response to hypothetical scenarios; thus, we were unable to account for clinical context or other factors that may affect physicians' use of and parents' receptivity to the messages. For instance, physicians' nonverbal cues may influence parents' perceptions of message persuasiveness. Also, it is possible that physicians will modify the messages rather than say them verbatim; these modifications may enhance or undermine the messages' persuasiveness. Moreover, we did not clarify with parents whether they were motivated by the message in its entirety or if there was a specific component of the message that was particularly stimulating. These limitations highlight that this research should be augmented with future studies that further test these messages. Future research also should extend this work by examining whether these messages lead to HPV vaccine uptake. Furthermore, future work should distinguish between messages designed to address parent concerns and messages intended to direct parents to vaccinate.

Our national study of parents and physicians identified messages perceived as persuasive for motivating HPV vaccination, even among parents disinclined to vaccinate their children. The lack of difference across demographic subgroups in parental endorsement of messages may suggest that these messages can be used across demographic subgroups, and warrants further investigation. Our findings provide physicians with evidence that using these messages may help them to communicate effectively with parents about this important cancer-preventing vaccine.

N.T. Brewer reports receiving commercial research grants from Merck and Pfizer and is a consultant/advisory board member for Merck. No potential conflicts of interest were disclosed by the other authors.

Conception and design: T.L. Malo, M.B. Gilkey, M.E. Hall, P.D. Shah, N.T. Brewer

Development of methodology: M.E. Hall, N.T. Brewer

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): M.E. Hall, P.D. Shah, N.T. Brewer

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): T.L. Malo

Writing, review, and/or revision of the manuscript: T.L. Malo, M.B. Gilkey, M.E. Hall, P.D. Shah, N.T. Brewer

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): T.L. Malo, M.E. Hall, N.T. Brewer

Study supervision: N.T. Brewer

This work was funded via an unrestricted educational grant from Pfizer (to N.T. Brewer). M.B. Gilkey and T.L. Malo were supported by the Cancer Control Education Program at UNC Lineberger Comprehensive Cancer Center (R25 CA57726). M.B. Gilkey was also supported by a career development award from the NCI (K22 CA186979).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.