Missing the Target for Routine Human Papillomavirus Vaccination: Consistent and Strong Physician Recommendations Are Lacking for 11- to 12-Year-Old Males

In 2011, the Advisory Committee on Immunization Practices (ACIP) recommended routine human papillomavirus (HPV) vaccination of all males ages 11 to 12 years, catch-up vaccination for males ages 13 to 21 years, and vaccination for men who have sex with men ages 22 to 26 (1). The ACIP updated this recommendation in 2015 to include the recently licensed 9-valent HPV vaccine (2). Yet HPV vaccination uptake among adolescent males nationally and in Florida is modest. In 2014, HPV vaccine initiation rate among males both in Florida and nationally was approximately 41%; coverage with ≥3 dose among males ages 13 to 17 was 22% in the United States and 17.5% in Florida (range: 9% in Alabama to 43% in Rhode Island; ref. 3). Available data suggest rates are even lower for 11- to 12-year-olds, who constitute the target age group for vaccination (4). Florida has among the nation's highest rates of HPV-related diseases in males, including the fifth highest rate of anal cancer (5). Given the established efficacy of preventing HPV infections that cause anal and penile cancer (6, 7), HPV vaccination has tremendous potential for primary prevention (8, 9), particularly for males from racial/ethnic and sexual minority groups disproportionately affected by HPV-related diseases (10–12).

The Centers for Disease Control and Prevention (CDC; ref. 13) and the President's Cancer Panel (14) advocate consistent and strong physician recommendations, particularly for the target age group of 11- to 12-year-old adolescents, as a primary approach to improving HPV vaccine dissemination. Yet relatively little research has examined the consistency and strength of physicians' HPV vaccine recommendations specific to males following the ACIP's recommendation for routine vaccination. Regarding consistency, results from a statewide survey of pediatricians, family physicians, and nurse practitioners in Minnesota showed that about 46% of providers routinely recommended HPV vaccine for boys ages 11 to 12 years (15). A national survey found that 61% of pediatricians and family physicians started routinely recommending HPV vaccine for males at or before age 12. For strength, a national survey of pediatricians and family physicians found that 52% of pediatricians and 41% of family physicians strongly recommend HPV vaccine for 11- to 12-year-old males (16), compared with 64% of pediatricians and 50% of family physicians in Hawaii (17). To our knowledge, only one study has examined both consistency and strength of physicians' HPV vaccine recommendations as part of a larger recommendation quality measure but did not focus solely on male patients (16). Because available literature demonstrates a disparity in recommendation practices such that fewer physicians report consistent and strong HPV vaccine recommendations for males compared with females (15–18), it is important to further explore factors associated with physicians' HPV vaccine recommendations for males.

Understanding recommendation practices and factors associated with those practices in a state with low vaccine uptake is essential to designing and implementing interventions to improve physicians' HPV vaccine recommendations. The medical encounter presents competing demands that pull physicians in many directions and influence how they recommend/provide preventive services during patient encounters. These ideas serve as the basis for the Competing Demands Model (CDM; ref. 19), originally developed to understand delivery of clinical preventive services in the primary care setting. The CDM proposes that physician (e.g., specialty, attitudes), patient (e.g., knowledge, attitudes), and practice environment (e.g., setting, location) factors influence physicians' delivery of preventive health services. Although several of these CDM factors have been explored by our group (20–22) and others (23), few have delved into vaccine-specific factors within the practice environment, such as the presence of a vaccine coordinator and use of specific types of reminder systems for series initiation and completion.

This study examines the prevalence of physician recommendation of HPV vaccination to vaccine-eligible males ages 11 to 12, 13 to 17, and 18 to 21 in Florida. Given ACIP recommendations emphasize routine vaccination targeting 11- to 12-year-olds and an emphasis on multilevel approaches to improving HPV vaccination rates (24), we assessed examined physician-reported factors relevant to the physician and practice (both general and vaccine specific) domains of the CDM as correlates of HPV vaccine recommendation for this age group.

Physicians were recruited from the American Medical Association (AMA) Physician Masterfile, a database of all licensed U.S. physicians (25). Our initial sampling frame excluded those who (i) were trainees; (ii) locum tenens; (iii) reported their major professional activity as nonpatient care; (iv) were ≥65 years of age, as the AMA Masterfile has been shown to have a significant lag in updating retired physicians (26); and (v) listed a post office box for their address (precluding use of FedEx mailing). Florida pediatric and family medicine (FM) physicians were sampled on the basis of their proportional representation in the Florida physician primary care workforce and randomly selected from the AMA Masterfile (n = 770). We selected only 1 physician per group practice. Informed by Dillman tailored design method (27), prenotice postcards were mailed in May 2014, followed by the first survey mailing in June 2014. Although the survey itself was anonymous, we limited subsequent mailings by using a postage paid postcard that was included with every survey mailing where physicians were asked to provide their name, address, and select from options including: (i) I have completed and sent the survey back in the prepaid envelope; or (ii) I do not wish to participate and have returned the blank survey and the $25 incentive. Follow-up mailings were sent to all those from whom we did not receive a completed postcard through August 2014. Participating physicians received a $25 cash incentive.

Of the 770 surveys mailed, 367 were received. After accounting for undeliverable surveys (n = 36) and ineligible respondents (n = 10), the overall response rate was 51% (367/724). We excluded 12 participants who reported not seeing male patients ages 9 to 26, for an analytic sample size of 355.

Where possible, we used previous survey items to assess HPV vaccination recommendation (20, 28–33); new items that were created to measure constructs not assessed in previous studies are noted below. The final 49-item survey assessed three domains relevant to the CDM: physician characteristics, physician-reported general and vaccine-specific practice characteristics.

Physician-reported characteristics were perceived personal and parental barriers related to HPV vaccination, HPV-related knowledge, and demographic characteristics. Sixteen items assessed physician-reported barriers to immunizing male patients against HPV. Perceived parental HPV vaccination barriers for 9- to 17-year-old male patients were measured using 14 items. Response options for physician-reported and perceived parental barriers were on a 4-point Likert scale (1 = not a barrier at all to 4 = a major barrier). Items were summed to create scores for perceived personal (range: 16–64; Cronbach α = 0.88) and parental barriers (range: 14–56; Cronbach α = 0.90). Lower scores indicated lower perceived barriers. Knowledge was measured using 9 items regarding HPV infection, disease, and vaccine guidelines for males. We built on previous surveys by creating new items to assess knowledge specific to guidelines and financial coverage for vaccinating males. One point was awarded for each correct response, and correct responses were summed to create a knowledge score (range: 0–9).

Physician-reported general practice characteristics included number of physicians in the practice, practice situation (single specialty, multispecialty, other), practice type (private, other), race/ethnic category of the majority of patients seen, whether the practice serves Medicaid patients, typical daily patient load, and practice location (urban, suburban, rural, other). Vaccine-specific practice characteristics were administration of HPV vaccine, Vaccines for Children (VFC) provider, strategies for remembering to discuss HPV vaccine with male patients, specific strategies to get patients into the office for the first and subsequent dose(s) of vaccine, total number of strategies used to get patients into the office for the first and subsequent dose(s) of vaccine, and presence of an office vaccine coordinator. We created new survey items to assess whether other health care professionals (i.e., medical assistant, nurse, nurse practitioner, physician assistant) in the practice discuss and recommend HPV vaccine. The final survey can be obtained by emailing the corresponding author.

In addition to items evaluating domains relevant to CDM, our survey included new and previous items that assessed the context and content of vaccination recommendation (34). New items also assessed provider's acceptance and use of the CDC messages and materials to support HPV vaccine recommendation for adolescent males (35) and intervention preferences.

The primary outcome variables were HPV vaccine recommendations to vaccine-eligible males ages 11 to 12, 13 to 17, and 18 to 21. Given that national guidelines call for consistent and strong recommendations, we assessed recommendation consistency and strength using two questions from prior studies (33). The first question asked providers to indicate how often they recommended HPV vaccination by age group: never/almost never (<10%), occasionally (10%–39%), about half the time (40%–59%), usually (60%–90%), and always/almost always (>90%). Providers were also asked how strongly they recommended vaccination by age group: I recommend against; I make no recommendation for or against; I recommend, but not strongly; and I strongly recommend. In addition to reviewing these items individually, we created a composite variable to reflect that national guidelines call for both consistent and strong recommendations. We combined these two questions into one variable to compare those who reported recommending vaccination consistently (always/almost always) and strongly to physicians who reported any other combination of responses.

Frequencies and percentages were calculated for independent and outcome variables. To reflect national guidelines for consistent and strong HPV vaccine recommendations targeted to adolescent males ages 11 to 12, we focused analyses on assessing correlates of recommendation for this age group. This was done in three steps. First, simple logistic regression models examined each correlate. Second, within each domain, significant univariate correlates were entered into a multivariable model. A backward elimination approach (significance level of stay = 0.05) was used to determine those correlates, making relatively independent contributions to consistent and strong HPV vaccine recommendations for each domain. Significant correlates in the final model for each domain were then entered into a multivariable model using a backward elimination approach to generate a model of the relatively independent correlates across domains. ORs and their 95% confidence intervals (CI) were estimated in each model. All analyses used two-tailed tests with the significance level set at P < 0.05 and were performed using the SAS 9.3 statistical software package (SAS Institute Inc.).

The sample was almost equally comprised of female (51.0%) and male (49.0%) physicians (Table 1), with an average age of 48.7 years (SD = 9.0). Most were white (67.7%) and non-Hispanic (75.1%). About half specialized in FM (49.9%) and had been practicing for ≥16 years (51.9%). HPV vaccine knowledge scores averaged 5.7 (SD = 2.1; range: 0–9).

Physicians most often reported their practice had two physicians (49.9%), was single specialty (66.8%), private practice (67.2%), and in a suburban location (52.6%). More than one third (36.6%) reported seeing mostly patients from minority groups and about one fifth (20.2%) reported no definable majority. The majority reported seeing Medicaid patients either along with other types of insurance (67.6%) or solely (1.9%). Almost half (44.5%) saw 20 to 29 patients daily. The majority reported administering HPV vaccine in their practice (68.8%) and nearly half were VFC providers (45.9%). More than half did not use a strategy to get patients in for the first dose of HPV vaccine (52.5%), but most used ≥1 strategies to get patients vaccinated with subsequent doses (84.3%). Physicians reported using flagging charts (16.6%), automatic prompts (22.2%), and electronic queries (13.6%) to remind themselves to discuss HPV vaccine. About three fourths reported having a vaccine coordinator (72.2%).

The anonymous nature of the survey precludes examining the degree to which survey responders and nonresponders were similar on demographic and practice characteristics. However, we were able to compare responding physicians to the population of physicians in Florida, meeting our study eligibility criteria on characteristics, including age, sex, and clinical specialty. We found no statistically significant difference between responding physicians and the larger population of Florida physicians for age, sex, and clinical specialty (all P > 0.05).

Compared with older adolescent groups, a lower proportion of physicians reported consistently (34.8%), strongly (42.9%), and both consistently and strongly (30.6%) recommending HPV vaccine to the 11- to 12-year-old group (Fig. 1). As shown in Figs. 2 and 3, physicians' mean perceived personal barriers score was 30.3 (SD = 9.6; range: 16–61), and the mean perceived parental barriers score was 37.0 (SD = 8.9; range: 14–56). As can be seen in Table 2, consistent and strong HPV vaccine recommendation for 11 to 12 year olds was associated with the following variables in univariate analyses: physician gender, specialty, personal barriers, HPV knowledge, patient race, Medicaid patients seen, practice location, and all vaccine-specific practice characteristics.

Within each domain (physician characteristics, general practice characteristics, and vaccine-specific practice characteristics), a multivariable analysis using a backward stepwise approach was applied, starting with the significant univariate correlates in each group (Table 2). Two physician characteristics were significantly associated with consistent and strong recommendations: pediatric specialty [adjusted OR (AOR) = 2.55; 95% CI, 1.38–4.71] and fewer (or lower) physician-reported barriers (AOR = 0.94; 95% CI, 0.90–0.97). The lone general practice characteristic was serving Medicaid patients (OR = 3.11; 95% CI, 1.48–6.53). The lone vaccine-specific characteristic was being a VFC provider (OR = 6.48; 95% CI, 3.47–12.1).

A multivariable analysis was performed using backward stepwise regression, starting with the significant correlates in the domain-specific multivariable models (Table 2). In the final multivariable model, fewer (or lower) physician barriers (AOR = 0.91; 95% CI, 0.88–0.94) and being a VFC provider (AOR = 5.43; 95% CI, 2.80–10.55) were statistically significant correlates of HPV vaccine recommendation in 11- to 12-year-olds.

Given that that the total score for physician barriers was inversely associated with strong and consistent HPV vaccine recommendations, we performed a post hoc analysis to explore each of the 16 physician barriers as a correlate. The results of these analyses are presented in Table 3. Alpha was adjusted to 0.0031 (0.05/16) for univariate analyses and α = 0.05 for multivariable analyses. Ten of the 16 individual barriers were significant correlates (left column). Only one of these 10 barriers was significant in a multivariable model (middle column), highlighting the positive correlation among the individual barriers which is also exhibited in the relative high Cronbach α (0.88). The barriers that remained statistically significant following a backward stepwise approach (right column) were physician concern about vaccine safety, concern about adding another vaccine to the schedule, and remembering to discuss the vaccine.

Recent efforts to improve physician recommendation for HPV vaccination have focused on two critical components. First, the recommendation should be consistent for all adolescent males and females, particularly the 11- to 12-year-old age group, for which routine vaccination is recommended. This is clearly demonstrated by the 2011 (1) and 2015 (2) ACIP guidelines endorsed by all professional medical societies that provide preventive care to adolescent males. Second, as noted in the President's Cancer Panel Report focused on HPV vaccination, the recommendation must be strong (24). Our study demonstrates that, despite available guidelines, a minority of physicians indicated that they consistently (35%), strongly (43%), and both consistently and strongly (31%) recommend HPV vaccination to the target 11- to 12-year-old adolescent male age group. Although our study focused on recommendation for males, studies that have examined differences in recommendation by gender highlight a marked disparity in vaccine recommendations between boys and girls (15–18). Without immediate and targeted intervention, the nation is unlikely to achieve Healthy People 2020 goal of 80% of 13- to 15-year-old adolescents receiving the entire 3-dose HPV vaccine series (36).

Relative to physicians reporting high HPV vaccination barriers, those reporting low barriers were more likely to consistently and strongly recommend vaccination. This finding aligns with previous research reporting that barriers to vaccination were associated with physicians not strongly recommending HPV vaccination to females ages 11 to 12 years (23) or offering the vaccine at all (37). Thus, one approach to improving vaccination recommendation may be to reduce provider barriers.

Barriers significantly associated with recommendation in the current study included concerns about vaccine safety, adding another vaccine to the schedule, and remembering to discuss vaccination. Interestingly, a systematic review of provider communication about HPV vaccination identified no studies demonstrating an association between providers' perceptions of HPV vaccine safety and their recommendation behaviors (38). However, our survey took place more recently than those included in the systematic review and may reflect the cumulative impact of media coverage that presents HPV vaccination in a controversial political context and as a vaccine for females (39). Thus, providers responding to our survey may be less aware of the safety of HPV vaccine for males. The ACIP recommendations for universal male HPV vaccination also coincided with political debates surrounding the vaccine during the 2012 presidential campaign (40). In addition, physician concerns about safety may be influenced by events such as the Japanese government's decision in June 2013 to stop proactive recommendation of HPV vaccine due to safety concerns (41, 42).

Physicians in our study also expressed concerns about adding another vaccine to the schedule. This concern has been documented in the context of adding a new vaccine to the early childhood immunization schedule (43) and HPV vaccination (20, 23). This finding is particularly concerning and should be addressed when suggesting that providers deliver a bundled recommendation that includes HPV alongside tetanus, diphtheria, and acellular pertussis (Tdap) and meningococcal vaccines, which are also recommended at ages 11 to 12 for boys and girls (44).

Difficulty in remembering to discuss HPV vaccine with males was the third barrier associated with recommendation for males. Given the strong correlation between provider recommendation and HPV vaccine uptake in males (45), providers forgetting to discuss the vaccine offers one possible explanation for lower series initiation rates in males compared with females, despite the 2011 ACIP recommendation for routine vaccination in age-eligible males (1). The President's Cancer Panel recommends use of provider reminders to reduce missed opportunities for HPV vaccination (24). Almost two thirds of respondents (67%) did not use any reminders to discuss HPV vaccination with patients and approximately half (44%) used no strategy to get patients into the office for the first dose. In our study, we specifically examined the manual and electronic strategies for remembering to discuss HPV vaccine with male patients and specific strategies to get patients into the office for the first and subsequent dose(s) of vaccine. Although our study found no association between use of reminders and recommendation, it is possible that the relatively small number of those using these approaches precluded detection of a statistically significant association.

One study found that clinics using an electronic health record–based point-of-care reminder system targeting both clinicians and patients resulted in significantly more young women initiating and completing the HPV vaccine series in a timely fashion. The authors suggested that this combined strategy may have supported the creation of a common agenda that facilitates HPV vaccine series initiation (46). A recent literature review of interventions to increase HPV vaccination rates suggested that there was sufficient evidence to support the use of Community Preventive Task Force recommendations related to client reminder and recall systems as well as provider assessment and feedback (47). Thus, future interventions may benefit from considering point-of-care reminders that simultaneously prompt physicians and patients to discuss HPV vaccination (48).

In our study, physicians who participated in the VFC program were more likely to recommend vaccination than those not participating. VFC-enrolled providers administer to Medicaid-eligible, uninsured, or American Indian or Alaska Native children younger than age 19 (49, 50). VFC also serves underinsured children in limited public settings (e.g., federally qualified health centers). Davis and colleagues (51) found that physician recommendation for pneumococcal conjugate vaccine (PCV7) was associated with the child's insurance coverage for the vaccine and state VFC policy. Another study demonstrated that patients of VFC providers were 59% more likely to be up-to-date with early-childhood vaccines than patients of non-VFC providers (52). However, it is unclear why these providers are more likely to recommend vaccination. It is possible that participating in the VFC program reduces physicians' up-front costs by offering free vaccines for VFC-eligible children (53). Providers also do not have to send patients elsewhere to receive vaccines. It is also possible that VFC providers also represent those who are more supportive of vaccines overall and willing to offer in-office vaccination in their practice setting. Future qualitative interviews with VFC and non-VFC providers may help to further explore these possible explanations.

To our knowledge, our study is among the first to focus specifically on physicians' HPV vaccine recommendations for males after the ACIP guidelines for routine vaccination of adolescent males. In addition, this is among the first observational studies to describe the current use of electronic and manual reminders for both physicians and patients related to HPV vaccination. This study has several notable strengths, including a statewide sample of primary care providers and an examination of physicians' recommendation with respect to both consistency and strength. There are also limitations. Our cross-sectional survey design precluded our ability to make causal inferences about variables significantly associated with recommendation of HPV vaccination. In addition, we surveyed physicians from a single state; although this design limits our ability to generalize findings to physicians practicing in our states, it allowed us to focus on HPV vaccine recommendations in a state with relatively high rates of HPV-related disease in males. We provided quantitative “anchors” for our questions regarding consistency but not strength of vaccine recommendation. Physicians may vary in their interpretation of a strong recommendation. Physicians may have reported socially desirable responses regarding practice behaviors; however, the anonymity of the survey likely reduced this bias. Physicians most in favor of HPV vaccination may have completed the survey, possibly providing an overestimate of the proportion that consistently and strongly recommend HPV vaccination. In our sample, 31% of provider-reported HPV vaccine currently is not administered to males in their clinical setting. However, we did not include any follow-up questions as to why they currently do not administer. It is possible that they simply do not vaccinate in their office. Although less likely, they may specifically not offer HPV vaccination for males. Finally, our study was limited to physicians, although other health care providers may recommend HPV vaccination. Thus, study of groups delivering care to adolescent males, such as nurses, medical assistants, and physician assistants, is warranted.

Physician recommendation is key to increasing HPV vaccine coverage; yet physicians' current HPV vaccine recommendations do not align with national guidelines and U.S. health organization recommendations. Interventions are needed to support HPV vaccine recommendation consistency and strength. Our research and others' suggest these interventions also should include education about and strategies as well as policy level interventions to address financial barriers to HPV vaccination. Current efforts to bolster physician recommendation have largely been focused on communication skills and office-based strategies to increase HPV vaccination rates. For example, the CDC's You Are the Key Campaign (54) largely provides patient and provider educational materials and provider communication strategies. More recently, the American Academy of Pediatrics HPV Champion Toolkit (55) has extended resources provided in the You Are the Key Campaign to include tools to facilitate change at the practice level (e.g., electronic health record–based reminder recalls). Our survey found that Florida physicians' barriers to consistently and strongly recommending HPV vaccine were related to concerns about vaccine safety, concerns about adding vaccines to the vaccine schedule, and difficulty in remembering to discuss HPV vaccination. These findings suggest that future interventions should include components to address these issues at the physician and practice level. By addressing physicians' challenges and supporting their HPV vaccine recommendations to their male patients, we can increase HPV vaccine coverage in Florida and reduce HPV-related disease.

J.A. Kahn reports receiving a commercial research grant from Merck, Inc. A. Casler has received speakers bureau honoraria from and is a consultant/advisory board member for Merck (Vaccine Division). D. Salmon reports receiving commercial research grants from Pfizer and Crucell and is a consultant/advisory board member for Merck. G.D. Zimet reports receiving commercial research grants from Merck and Roche and is a consultant/advisory board member for Merck. A.R. Giuliano reports receiving a commercial research grant from and is a consultant/advisory board member for Merck & Co. No potential conflicts of interest were disclosed by the other authors.

Conception and design: S.T. Vadaparampil, J.A. Kahn, D. Salmon, R.G. Roetzheim, G.D. Zimet, A.R. Giuliano

Development of methodology: T.L. Malo, J.A. Kahn, R.G. Roetzheim, G.D. Zimet

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): S.T. Vadaparampil, T.L. Malo

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): S.T. Vadaparampil, S.K. Sutton, J.A. Kahn, D. Salmon, G.D. Zimet

Writing, review, and/or revision of the manuscript: S.T. Vadaparampil, T.L. Malo, S.K. Sutton, K.N. Ali, J.A. Kahn, A. Casler, D. Salmon, B. Walkosz, R.G. Roetzheim, G.D. Zimet, A.R. Giuliano

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): S.K. Sutton, K.N. Ali

Study supervision: S.T. Vadaparampil, T.L. Malo

We thank Janine Cory and Allison Kennedy Fisher for their survey feedback and Dr. Ji-Hyun Lee for assistance with sample selection.

This research was supported by funding from the Bankhead-Coley Cancer Research Program (4BB10). T.L. Malo is supported by the UNC Lineberger Cancer Control Education Program (R25CA057726). This work was also supported in part by the Biostatistics Core at the H. Lee Moffitt Cancer Center & Research Institute, an NCI-designated Comprehensive Cancer Center (P30-CA076292).

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.