National guidelines recommend that 11- and 12-year-old adolescents receive 3 doses of human papillomavirus (HPV) vaccine.1 However, HPV vaccine uptake remains low, with coverage levels falling far short of public health goals and uptake of other recommended adolescent vaccines.2,3 In addition, considerable variation in HPV vaccination levels between states exists: receipt of all 3 doses of HPV vaccine ranges from 12.1% of adolescent girls in Mississippi to 57.7% of adolescent girls in Rhode Island.3
Preliminary studies demonstrate the effect of HPV vaccination on reducing population levels of HPV infection and genital warts,4–6 but the potential impact on cancer outcomes remains to be seen. However, examining the relationship between current vaccination patterns and cancer rates could foreshadow the future benefits of widespread vaccination. Bach7 investigated 2 state characteristics that correlated with levels of HPV vaccine initiation among girls (using 2008 vaccination data), highlighting the negative relationship with cervical cancer mortality and the positive relationship with median household income. The author found that HPV vaccine initiation was more common in states in which girls were at lower risk for cervical cancer mortality.
To update and extend these findings and to better understand the ecological pattern of HPV vaccination in states across the United States, we analyzed the relationships between states’ HPV vaccine coverage and (a) the incidence and mortality rates of HPV-related cancers, (b) demographic characteristics, and (c) measures of contact with the health care system. We hypothesized that states with higher HPV-related cancer rates, including cervical cancer, would have lower HPV vaccine coverage.
MATERIALS AND METHODS
We examined data for 50 states and Washington, DC. For the sake of simplicity, we refer to all 51 jurisdictions as states hereafter.
Vaccination data came from the 2012 National Immunization Survey–Teen (NIS-Teen; the most recent year of available data at the time of analyses).3 We examined HPV vaccine initiation (i.e., received ≥1 dose) among adolescents aged 13 to 17 years, stratified by sex. We examined HPV vaccine follow-through (i.e., received all 3 recommended doses, among those who initiated HPV vaccination) only among girls aged 13 to 17 years. Follow-through differs from 3-dose completion in that it examines completion only among vaccine initiators rather than the entire population; NIS-Teen reports these outcomes separately and has noted the different pattern of sociodemographic correlates for initiation and completion of 3 doses versus for follow-through.3 We also gathered data on coverage with tetanus, diphtheria, and pertussis booster (Tdap) and meningococcal vaccine (receipt of first dose) for adolescents aged 13 to 17 years.
Rates of HPV-Related Cancer
We obtained incidence and mortality rates per 100,000 population from the 2001–2010 United States Cancer Statistics database8 for cancers that HPV vaccine is currently approved to prevent9,10: anal cancer (both sexes) and cervical, vaginal, and vulvar cancers (females only). We created a measure of cancer deaths attributable to HPV types protected against by the bivalent and quadrivalent HPV vaccines (i.e., types 16 and 18).11 Thus, we discounted the annual death rate per 100,000 population by the attributable fraction (87% of anal, 76% of cervical, 56% of vaginal, and 44% of vulvar cancer), and then we summed the discounted attributions. Any state with missing data on at least 1 cancer mortality rate (because of fewer than 16 observations during the study period) was coded to missing on the summary measure.
We used data from the US Census12,13 to examine the racial/ethnic composition (2012 data) and median household income (2011–2012 data) of each state. We calculated physicians per 100,000 population for 3 specialties (pediatrics, family practice, and obstetrics/gynecology) using data from the Bureau of Labor Statistics14 and standardized with population estimates from the US Census.12
Contact With Health Care System
We captured adolescent health care adequacy, as reported by parents, using 2 constructs derived from the National Survey of Children’s Health15: the proportion of adolescents in each state (a) with a medical home and (b) who had a health care provider whom their parents considered their “personal” doctor or nurse. Because around 6% of US schools have health centers that may provide vaccines,16 we used data from the School-based Health Alliance17 and the US Census12 to measure the number of school-based health centers per 100,000 population in each state.
We calculated the proportion of adult women in each state (ages 18 years and older) who had received a Papanicolaou (Pap) test in the previous 3 years using data from the Behavioral Risk Factor Surveillance System.18 To account for errors in self-report of Pap testing that vary by race,19–21 we adjusted state-level Behavioral Risk Factor Surveillance System estimates based on the methodology described by Rauscher and colleagues.22
We examined the relationships between HPV vaccination outcomes and other variables using Pearson product-moment correlations yielding r statistics. All statistical tests were 2 tailed, and we describe in text all correlations with a P value of less than 0.10, given the ecological sample size of k = 51 states. We graphed scatterplots for notable relationships. We analyzed data using Stata version 13 (College Station, TX).
According to the 2012 NIS-Teen results,3 53.8% of female adolescents had initiated HPV vaccine and 66.7% of those initiators had followed through with receipt of 3 doses of HPV vaccine. Among adolescent boys, 20.8% had initiated HPV vaccine.
HPV Vaccine Initiation (Girls)
Girls’ initiation was lower in states with higher levels of incidence and mortality rates of HPV-related cancers: cervical cancer incidence (r = −0.29), cervical cancer mortality (r = −0.46; Fig. 1A), and the summary measure of mortality rates for cancers related to HPV 16/18 (r = −0.58; Table 1). Human papillomavirus vaccine initiation among girls was higher in states with higher median household incomes (r = 0.32), lower proportions of non-Hispanic black residents (r = −0.28), higher proportions of residents of “other” races/ethnicities (r = 0.24), and greater concentrations of pediatricians (r = 0.47; Fig. 1B) and OB/GYNs (r = 0.37). Initiation among girls was also higher in states with higher levels of other vaccination outcomes: boys’ HPV vaccine initiation (r = 0.47; Fig. 1C), Tdap vaccination (r = 0.43), and meningococcal vaccination (r = 0.48).
HPV Vaccine Follow-Through (girls)
Follow-through was lower in states with higher rates of vaginal cancer incidence (r = −0.25), cervical cancer mortality (r = −0.30), and the summary measure of mortality rates for cancers related to HPV 16/18 (r = −0.29; Table 1). Interestingly, follow-through was higher in states with higher rates of vulvar cancer incidence (r = 0.30). In addition, follow-through was higher in states with higher levels of other measures of health care access: other vaccination outcomes (r = 0.27–0.51), adolescent health care adequacy (r = 0.26–0.37), concentration of school health centers (r = 0.25), and proportion of adult women with a recent Pap test (r = 0.36; Fig. 2).
HPV Vaccine Initiation (Boys)
Human papillomavirus vaccine initiation among boys was higher in states with lower proportions of residents that were non-Hispanic white (r = −0.31) and higher proportions of residents of other races/ethnicities (r = 0.32; Table 1). Boys’ initiation was also higher in states with higher levels of meningococcal vaccination (r = 0.39) and proportions of adolescents with a personal doctor or nurse (r = 0.26).
In line with our hypothesis, states with higher rates of HPV-related cancers, including cervical cancer, had lower HPV vaccine coverage (both initiation and follow-through) among girls. Girls’ initiation was also associated with demographic composition. However, girls’ follow-through demonstrated more consistent associations with measures related to contact with the health care system. For boys, initiation showed some associations with demographic composition and contact with the health care system, but not HPV-related cancer rates. As initiation among boys and girls was highly correlated, it is possible that the pattern seen among girls could develop as vaccination among boys becomes more widely accepted and coverage increases.
Lower rates of HPV vaccination in areas with higher cancer rates could exacerbate current disparities in cancer incidence and mortality across states. One potential mechanism is that HPV is sexually transmitted,23 and infections spread through sexual networks that are largely geographically bounded.24–26 In addition, HPV vaccine initiation among girls was lower in states with a higher proportion of non-Hispanic black residents, an especially worrisome finding because black women have among the highest risk for cervical cancer.27 Girls’ follow-through was associated with lower incidence rates for vaginal cancer and higher incidence rates for vulvar cancer, although the reasons for the different directions of these relationships are unclear. Promotional efforts should concentrate on increasing HPV vaccination, including follow-through, for both boys and girls in areas with high rates of HPV-related cancer incidence and mortality. Our findings also suggest that cost-effectiveness analyses may overestimate the benefits of current vaccination coverage and underestimate the benefits of increasing coverage, as girls living in areas with the highest rates of HPV-related cancers were the least likely to initiate the HPV vaccine series. Cost-effectiveness models should consider the impact of disproportionate vaccination by cancer risk on the benefits of vaccine promotional efforts.28–31
Ecological measures of contact with the health care system demonstrated consistent, positive associations with HPV vaccine follow-through among girls. That is, girls were more likely to follow-through with vaccination if they lived in states with greater use of other preventive health care services. This pattern is not surprising, but it does underscore the need for health care systems and services that meet the needs of adolescents, who have contact with primary care less often than younger children.32 It will be important to test these associations with follow-through among boys, as well, but follow-through estimates were only available for 11 states in the 2012 NIS-Teen data.3
Some differences emerged between our ecological state-level findings and the individual-level findings in the report of the 2012 NIS-Teen results.3 Human papillomavirus vaccine initiation among non-Hispanic black girls did not differ from non-Hispanic whites in analyses of individual-level data, but initiation among girls was higher in states with lower proportions of non-Hispanic black residents in our ecological analyses. Follow-through was lower among black and Hispanic girls than non-Hispanic whites in the individual-level analyses, but no relationship between rates of follow-through and demographic composition emerged in our ecological analyses. However, for both the individual and ecological analyses, follow-through was positively associated with white race. Initiation was lower among non-Hispanic white boys than other racial/ethnic groups in individual-level analyses, a finding that held true in our ecological analyses. Finally, in individual analyses, all vaccination outcomes were more common among adolescents living below the federal poverty level than those living at or above the poverty level, but we found that median household income was positively associated with girls’ initiation and no relationship for boys’ initiation or girls’ follow-through. These patterns could reflect multilevel influences on vaccination that vary across people and states, including health insurance status and personal beliefs versus public health funding and social norms. To the extent that HPV-related cancer rates vary systematically by individual race as well as state racial composition, additional research should examine the potential causes of these differences in relationships and interventions to address them, with an ultimate goal of reducing disparities in vaccination and cancer.
Our analyses have several limitations, including that the findings from the ecological analyses we report may not generalize to variability among individuals. The small number of observations (k = 51) limits the options for statistical analysis of these ecological phenomena. Notably, the correlations reported here may vary when controlling for additional factors, but adding variables to the models may lead to unstable results. To examine the potential impact of controlling for sociodemographic characteristics on the relationship between HPV vaccination and cervical cancer mortality, we ran exploratory linear regression models that controlled for 2 state-level variables: median household income and demographic composition (percent of residents who are non-Hispanic white, non-Hispanic black, and Hispanic). The coefficients of the associations between mortality and HPV vaccine initiation and follow-through remained statistically significant in these analyses (P = 0.04 and P = 0.02, respectively), increasing our confidence in the findings.
In addition, we gathered data that were generally cross sectional. Such data limit our ability to make conclusions about causal mechanisms driving vaccine coverage, but they instead permitted us to characterize the context in which vaccination took place. Additional studies are needed to investigate these causal relationships. Strengths include our use of data from high-quality sources, including provider-verified NIS-Teen vaccination data, to get a more complete picture of ecological correlations with vaccination than any one data source could provide. Our analyses are also among the first to address geographic differences in HPV vaccination, an understudied and important topic.
In summary, HPV vaccine coverage in 2012 continued to be correlated with HPV-related cancer rates and several other state characteristics. For girls, initiation was associated with state-level demographics and cancer rates, whereas follow-through was related to health care access. For boys, initiation was associated with state-level demographics and health care access. Interventions to increase HPV vaccine coverage in areas with high HPV-related cancer incidence and mortality rates could result in meaningful public health benefits.
3. Centers for Disease Control and Prevention (CDC). National and state vaccination coverage among adolescents aged 13–17 years—United States, 2012. MMWR Morb Mortal Wkly Rep 2013; 62: 685–693.
4. Markowitz LE, Hariri S, Lin C, et al. Reduction in human papillomavirus (HPV) prevalence among young women following HPV vaccine introduction in the United States, National Health and Nutrition Examination Surveys, 2003–2010. J Infect Dis 2013; 208: 385–393 doi: 10.1093/infdis/jit192.
5. Flagg EW, Schwartz R, Weinstock H. Prevalence of anogenital warts among participants in private health plans in the United States, 2003–2010: Potential impact of human papillomavirus vaccination. Am J Public Health 2013; 103: 1428–1435. doi: 10.2105/AJPH.2012.301182.
6. Tabrizi SN, Brotherton JM, Kaldor JM, et al. Fall in human papillomavirus prevalence following a national vaccination program. J Infect Dis 2012; 206: 1645–1651. doi: 10.1093/infdis/jis590.
7. Bach PB. Gardasil: From bench, to bedside, to blunder. Lancet 2010; 375: 963–964. doi: 10.1016/S0140-6736(09)62029-8.
8. Centers for Disease Control and Prevention (CDC). United States Cancer Statistics: Public information data. Available at: http://wonder.cdc.gov/cancer.HTML
. Updated 2014. Accessed June 9, 2014.
9. Centers for Disease Control and Prevention (CDC). FDA licensure of bivalent human papillomavirus vaccine (HPV2, cervarix) for use in females and updated HPV vaccination recommendations from the Advisory Committee on Immunization Practices (ACIP). MMWR Morb Mortal Wkly Rep 2010; 59: 626–629.
10. Centers for Disease Control and Prevention (CDC). FDA licensure of quadrivalent human papillomavirus vaccine (HPV4, gardasil) for use in males and guidance from the Advisory Committee on Immunization Practices (ACIP). MMWR Morb Mortal Wkly Rep 2010; 59: 630–632.
11. Gillison ML, Chaturvedi AK, Lowy DR. HPV prophylactic vaccines and the potential prevention of noncervical cancers in both men and women. Cancer 2008; 113 (10 suppl): 3036–3046. doi: 10.1002/cncr.23764.
14. U.S. Bureau of Labor Statistics. Occupational employment statistics. http://data.bls.gov/oes/
. Accessed June 9, 2014.
15. Child and Adolescent Health Measurement Initiative. The National Survey of Children’s Health. Available at: http://childhealthdata.org/learn/NSCH
. Updated 2012. Accessed June 9, 2014.
16. Brener ND, Wheeler L, Wolfe LC, et al. Health services: Results from the School Health Policies and Programs Study 2006. J Sch Health 2007; 77: 464–485. doi: 10.1111/j.1746-1561.2007.00230.x.
17. School-based Health Alliance. 2010–2011 Census report of school-based health centers. 2012.
18. Centers for Disease Control and Prevention (CDC). Behavioral Risk Factor Surveillance System. Available at: http://www.cdc.gov/brfss/
. Updated 2014. Accessed June 9, 2014.
19. Caplan LS, McQueen DV, Qualters JR, et al. Validity of women’s self-reports of cancer screening test utilization in a managed care population. Cancer Epidemiol Biomarkers Prev 2003; 12 (11 Pt 1): 1182–1187.
20. Pizarro J, Schneider TR, Salovey P. A source of error in self-reports of Pap test utilization. J Community Health 2002; 27: 351–356.
21. Johnson TP, O’Rourke DP, Burris JE, et al. An investigation of the effects of social desirability on the validity of self-reports of cancer screening behaviors. Med Care 2005; 43: 565–573. doi: 00005650-200506000-00007.
22. Rauscher GH, Johnson TP, Cho YI, et al. Accuracy of self-reported cancer-screening histories: A meta-analysis. Cancer Epidemiol Biomarkers Prev 2008; 17: 748–757. doi: 10.1158/1055-9965.EPI-07-2629.
24. Adimora AA, Schoenbach VJ. Social context, sexual networks, and racial disparities in rates of sexually transmitted infections. J Infect Dis 2005; 191 (suppl 1): S115–S122. doi: JID32102.
25. Doherty IA, Padian NS, Marlow C, et al. Determinants and consequences of sexual networks as they affect the spread of sexually transmitted infections. J Infect Dis 2005; 191 (suppl 1): S42–S54. doi: JID32098.
26. Laumann EO, Youm Y. Racial/ethnic group differences in the prevalence of sexually transmitted diseases in the United States: A network explanation. Sex Transm Dis 1999; 26: 250–261.
27. Watson M, Saraiya M, Benard V, et al. Burden of cervical cancer in the United States, 1998–2003. Cancer 2008; 113 (10 suppl): 2855–2864. doi: 10.1002/cncr.23756.
28. Goldie SJ, Kohli M, Grima D, et al. Projected clinical benefits and cost-effectiveness of a human papillomavirus 16/18 vaccine. J Natl Cancer Inst 2004; 96 (8): 604–615.
29. Chesson HW, Ekwueme DU, Saraiya M, et al. The cost-effectiveness of male HPV vaccination in the United States. Vaccine 2011; 29: 8443–8450. doi: 10.1016/j.vaccine.2011.07.096.
30. Elbasha EH, Dasbach EJ, Insinga RP. Model for assessing human papillomavirus vaccination strategies. Emerg Infect Dis 2007; 13: 28–41. doi: 10.3201/eid1301.060438.
31. Armstrong EP. Prophylaxis of cervical cancer and related cervical disease: A review of the cost-effectiveness of vaccination against oncogenic HPV types. J Manag Care Pharm 2010; 16: 217–230. doi: 2010(16)3: 217–230.
© Copyright 2015 American Sexually Transmitted Diseases Association
32. Rand CM, Shone LP, Albertin C, et al. National health care visit patterns of adolescents: Implications for delivery of new adolescent vaccines. Arch Pediatr Adolesc Med 2007; 161: 252–259. doi: 10.1001/archpedi.161.3.252.