Human papillomavirus vaccination among adolescents in Georgia

Abstract

Human papillomavirus (HPV) vaccination coverage for adolescent females and males remains low in the United States. We conducted a 3-arm randomized controlled trial (RCT) conducted in middle and high schools in eastern Georgia from 2011–2013 to determine the effect of 2 educational interventions used to increase adolescent vaccination coverage for the 4 recommended adolescent vaccines: Tdap, MCV4, HPV and influenza. As part of this RCT, this article focuses on: 1) describing initiation and completion of HPV vaccine series among a diverse population of male and female adolescents; 2) assessing parental attitudes toward HPV vaccine; and 3) examining correlates of HPV vaccine series initiation and completion. Parental attitude score was the strongest predictor of HPV vaccine initiation among adolescents (adjusted odds ratio (aOR): 2.08; 95% confidence interval (CI): 1.80, 2.39). Other correlates that significantly predicted HPV series initiation were gender, study year, and intervention arm. Parental attitudes remained a significant predictor of receipt of 3 doses of HPV vaccine along with gender, race, school type and insurance type. This study demonstrates that positive parental attitudes are important predictors of HPV vaccination and critical to increasing coverage rates. Our findings suggest that more research is needed to understand how parental attitudes are developed and evolve over time.

Keywords:

• attitudes
• adolescent
• HPV vaccine
• Health Belief Model
• Theory of Reasoned Action

Introduction

In 2006, the Advisory Committee on Immunization Practices (ACIP) recommended routine human papillomavirus (HPV) vaccination for females ages 9–26 after licensure of the quadrivalent vaccine in 2006.¹ The quadrivalent vaccine was not licensed for males ages 9–26 until 2009, and ACIP did not provide a routine recommendation for HPV vaccination for males until 2011.² Despite routine recommendations for males and females as well as numerous marketing and educational campaigns, vaccination rates remain low in the US.

HPV vaccination rates are suboptimal compared to other adolescent vaccination rates and are well below the Healthy People 2020 goal of 80% coverage for all adolescent vaccines, both nationally and in Georgia where rates lag even further behind.³ In 2013, coverage of 13–17 y old females in the U.S initiating HPV vaccine series was 57.3% and 34.6% among males.⁴ Coverage of all 3 doses among US adolescent females was 37.6% and 13.9% among males.⁴ In 2013, coverage of 13–17 y old females in Georgia initiating the HPV series was 53.7% and 40.5% among males.⁵ Coverage of all 3 doses among Georgia adolescent females was 33.2% and 15.3% among males.⁵

Provider recommendations and positive parental attitudes have been shown as important predictors of parents' decisions to vaccinate their adolescent against HPV.^6-8 Studies have found that among parents, low knowledge of HPV, low perceived susceptibility of adolescents to HPV/sexually transmitted infections, concerns about vaccine safety, and views that HPV vaccination leads to sexual activity were associated with reluctance to allow HPV vaccination.^9-11

The objectives of this study were to 1) describe receipt of one and 3 HPV vaccine doses among a diverse population of male and female adolescents; 2) assess parental attitudes toward HPV vaccine; and 3) examine correlates of HPV vaccine series initiation and series completion among a sample of middle and high school adolescents. This study is unique in that it focuses on a racially and economically diverse sample of adolescents coming from a mix of urban and rural schools, whereas previous HPV studies with parents were conducted prior to ACIP recommendations for HPV vaccination of males, focused on female adolescents only, or relied primarily on health care facilities for recruitment.^12-15

Results

Demographic characteristics and vaccination uptake

The majority of adolescents were African American (74.9%, n = 514) and female (53.1%, n = 364). Over half were covered by Medicaid (59.8%, n = 410) and were enrolled in middle school (63.6%, n = 436) (Table 1). Only 48.5% (n = 333) of adolescents had received one dose of the HPV vaccine (Table 1). Among adolescents who received one dose, 61.9% (n = 206) completed 3 doses. Female adolescents were significantly more likely to have received the initial HPV dose than male adolescents (59.8% vs 36.1% respectively, P < 0.001) (data not shown). Among adolescents receiving one HPV dose (n = 333), females were significantly more likely to receive 3 doses compared to males (67.6% v 51.3% respectively, p = 0.003) (data not shown).

Table 1. Participant demographics and characteristics data

	Responses N=686 n (%)
Race
Caucasian	111 (16.2)
African-American	514 (74.9)
Other	61 (8.9)
Gender
Female	364 (53.1)
Male	322 (46.9)
Mean age in years (SD) (at time of survey)	14.2 (2.4)
Child's insurance
Private insurance	233 (34.0)
Medicaid	410 (59.8)
No insurance	42 (6.1)
Middle school	436 (63.6)
High school	250 (36.4)
Prior receipt of vaccine
1 HPV dose	333 (48.7)
3 HPV doses among those who received 1 (n=333 )	206 (61.9)
If not received any HPV vaccine, intent to receive (n=351 )	216 (61.5)
Study Characteristics
Intervention Arms
Control	210 (30.6)
Parent only	251 (36.6)
Parent and adolescent	225 (32.8)
Study Year
Baseline	117 (17.1)
Year 1	209 (30.5)
Year 2	360 (52.5)
Survey Method
Phone	358 (52)
Online	328 (48)

Table 2. Application of the Health Belief Model (HBM) and Theory of Reasoned Action (TRA)

Theoretical Construct	Item (True/False)
Perceived Susceptibility*	-My child is very likely to get HPV
Perceived Severity*	-The HPV infection can cause a serious disease
Perceived Benefits*	-The HPV vaccine is very effective at preventing cervical cancer
	-Children should be vaccinated against HPV
Social Norms^+	-Most people important to me think I should give my child a HPV vaccine
	-Most of the parents I know take their children for HPV vaccine.

* = Health Belief Model; ⁺ = Theory of Reasoned Action

Parental attitudes toward HPV vaccine

The average HPV attitude score was 3.8 out of a possible 6 (SD = 1.5) (Table 3) with higher attitude scores among parents who reported their adolescent received one HPV dose (mean = 4.5, SD = 1.1) compared to parents who reported their child did not receive one HPV dose (mean = 3.2, SD = 1.6) (95% CI [−1.5, −1.1], t(669) = −11.99) (Table 3). Among parents who reported their adolescent received one HPV dose, there was no significant difference in parental HPV attitude score among adolescents who received 3 HPV doses and those that did not receive 3 doses (95% CI [−0.4, 0.0], t(328) = −1.6) (Table 3).

Table 3. Parental HPV attitude and belief score and differences by receipt of one HPV dose, 3 HPV doses and gender

Characteristic	N	Mean (SD)	T-test value (p-value)
Parental attitude score	686	3.8 (1.5)
1 HPV dose
Yes	330	4.5 (1.0)	−11.9 (P < 0.001)
No	341	3.2 (1.6)
3 HPV doses
Yes	204	4.6 (1.0)	−1.621 (p = 0 .10)
No	126	4.4 (1.1)
Gender
Female	357	3.9 (1.5)	0.718 (p = 0 .47)
Male	315	3.8 (1.5)

Correlates of HPV vaccine initiation and series completion

In unadjusted analyses, correlates of HPV vaccine initiation included: child gender, intervention arm, study year, parental HPV attitude and belief score, and insurance type. In the adjusted model, parents in the 2 intervention groups were less likely to report their child receiving one HPV dose compared to those parents in the control group. Compared to parents in intervention arm 3 (parent and adolescent arm), parents in arm 1 (control arm) had twice the odds of reporting their child received an initial dose of HPV (aOR = 2.14, 95% CI [1.3, 3.4]) (Table 4).

Table 4. Correlates of receiving one HPV dose and 3 HPV vaccine doses

	1 HPV Dose				3 HPV Doses
Variables	Unadjusted OR [95% CI]	p-value	Adjusted OR [95% CI]	p-value	Unadjusted OR [95% CI]	p-value	Adjusted OR [95% CI]	p-value
Gender
Male	Ref		Ref		Ref		Ref
Female	2.55 [1.87, 3.49]	<0.001	3.06[2.13, 4.39]	<0.001	1.98[1.24, 3.14]	0.004	2.11[1.30, 3.43]	0.002
Race
Caucasian	Ref		Ref		Ref		Ref
African American	0.91 [0.60, 1.38]	0.679	0.78 [0.48, 1.28]	0.336	0.59 [0.31, 1.12]	0.10	0.50 [0.25, 0.99]	0.048
Other	0.766 [0.408, 1.436]	0.405	0.932 [0.438, 1.984]	0.855	0.866 [0.324, 2.316]	0.775	0.794 [0.238, 2.229]	0.662
School type
Middle school	Ref		Ref		Ref		Ref
High school	0.93 [ 0.68, 1.27]	0.663	1.02 [0.70, 1.49]	0.891	1.48 [0.92, 2.38]	0.099	1.71[1.02, 2.86]	0.040
Insurance Type
Medicaid	Ref		Ref		Ref		Ref
Private	0.67 [0.48, 0.92]	0.016	0.77 [0.52, 1.14]	0.197	1.56 [0.94, 2.60]	0.085	1.71 [1.00, 2.91]	0.048
No Insurance	1.01 [0.53, 1.92]	0.959	1.06 [0.50, 2.28]	0.866	1.24 [0.50, 3.09]	0.636	1.14 [0.43, 3.00]	0.789
Study Year
Baseline	Ref		Ref		Ref		Ref
Year 1	1.16[0.73, 1.83]	0.520	1.34[0.78, 2.29]	0.283	0.92 [0.44, 1.90]	0.826	0.81 [0.38, 1.74]	0.605
Year 2	1.65[1.08, 2.52]	0.020	1.76[1.07, 2.89]	0.025	0.80 [0.41, 1.54]	0.514	0.65 [0.32, 1.30]	0.224
Intervention Arm
Control	1.65[1.13, 2.42]	<0.001	2.14 [1.33, 3.42]	<0.001	0.839[0.48, 1.44]	0.529	1.13[0.630, 2.04]	0.671
Parent-only	.996[0.69, 1.43]	0.984	.865[1.33, 3.42]	<0.001	1.04 [0.59, 1.82]	0.414	1.092 [0.60, 1.96]	0.768
Parent and adolescent	Ref		Ref		Ref		Ref
HPV attitude and belief score	1.95[1.71, 2.22]	<0.001	2.08[1.80, 2.39]	<0.001	1.18 [0.96, 1.45]	0.107	1.2 [1.00, 1.54]	0.045

* adjusted for gender, study year, intervention arm, and HPV attitude and belief score, age, race and insurance type

Across all 3 study arms, the odds of parents reporting receipt of one HPV dose at baseline increased to 1.76 during the final follow-up (Year 2) (aOR = 1.76, 95% CI [1.07, 2.89]). Parents with more positive attitudes about HPV vaccine had higher odds of reporting receipt of an initial dose of HPV vaccine compared to parents with lower HPV attitudinal scores (aOR = 2.08, 95% CI [1.80, 2.39]). The odds of female adolescents receiving one HPV dose was 3 times greater than male adolescents (aOR = 3.0, 95% CI [2.1, 4.3]) (Table 4).

When assessing correlates associated with receiving 3 doses of HPV, gender, race, HPV attitude and belief score, insurance type and school type were significant in the adjusted model. Similar to HPV initiation, female adolescents had higher odds of receiving 3 HPV doses compared to male adolescents (aOR = 2.1, 95% CI [1.3, 3.4]). The odds of parents with higher scores on the HPV attitude and belief index reporting their child received 3 HPV doses increased by 1.2 compared with parents with lower scores (aOR = 1.2, CI [1.0, 1.5]). Adolescents in high school and with private insurance were at increased odds of completing 3 doses compared with middle school adolescents and adolescents with Medicaid coverage (aOR = 1.71, 95% CI [1.02, 2.86] and aOR = 1.71, 95% CI [1.00, 2.91], respectively) (Table 4).

Discussion

Parental knowledge, attitudes, and beliefs have been measured consistently in HPV studies. Several studies have shown that parental knowledge and awareness of HPV vaccine is a strong predictor of vaccine receipt or intention to vaccinate.^13,14 High attitude and belief scores were associated with HPV series completion as well. A cohort study by Fishman et al. that only measured HPV knowledge among parents and adolescents found knowledge was not a significant predictor of vaccination behavior and suggested that beliefs and attitudes are more predictive of vaccination behaviors.¹⁶ Our analysis of parental attitudes and beliefs found that higher scores predicted receiving an initial dose of HPV and series completion, which is consistent with these published studies.

HPV attitude and belief score was a consistent predictor for receipt of one and receipt of 3 HPV doses. Parents in 2 intervention arms (arm 2 and arm 3) received educational materials to improve their knowledge and attitude toward HPV vaccination, yet parents were less likely to report HPV vaccination in these intervention arms compared to the control group, which received no educational materials. These findings are similar to a study by Nyhan et al who evaluated measles-mumps-rubella (MMR) vaccine messages and found that pro-vaccine messages do not always work and may vary depending on parental attitudes toward vaccines.¹⁷ Vaccine message framing for parents is an important factor that should be considered in future studies using educational materials. Although we pilot tested the materials with a small group of parents and adolescent, we did not specifically measure vaccine messaging and framing in the parent and adolescent educational materials. Our efforts to include pro-vaccine messaging and address myths about adolescent vaccines may not have been received positively by the larger study population. Additionally, our findings suggest that there may be other influential sources of information from which parents receive information about HPV vaccination that we have not accounted for in this study. It is important to understand not only the range of potential sources of information but also the content and intent in order to develop appropriate message framing to improve HPV vaccination rates.

Additionally, parents may have concerns specifically about HPV vaccination. Both the tetanus, diphtheria and pertussis (Tdap) vaccine and quadrivalent meningococcal conjugate vaccine (MCV4) were introduced and recommend by ACIP for adolescents at similar times as the HPV vaccine; however, vaccination coverage rates for Tdap and MCV4 vaccines surpass coverage of HPV vaccination coverage.¹⁸ Despite numerous trials for both efficacy and safety, parental safety concerns persist about HPV vaccination. The prevalence of health concerns about vaccine side effects ranged from 6% to 12% among parents in previously published studies.^19,20 Additionally, the lack of discussion by physicians and information provided to parents about vaccine safety has been documented as a barrier to HPV vaccination.^20,21 Message framing and content are important in communicating vaccine safety information to parents.

Gender was a consistent predictor of HPV initiation and series completion. The results of this survey are consistent with epidemiological reports and other studies showing that HPV initiation and completion rates for adolescent females are higher than those for males. Receipt of one HPV dose among adolescent males in our sample was slightly lower than the Georgia rate among adolescent males (36% vs 40%, respectively). Parents in our sample reported higher series completion rates than the national and state averages among adolescent males who started the series (51% vs 48% vs 41%, respectively).¹⁸ Among female adolescents, initiation rates in our sample were only slightly higher than both national and state percentages (59% vs 57% vs 53%, respectively). We had similar proportions of adolescent females in our sample who reported completing the HPV series, 68% compared to 70% nationally and 65% in Georgia.¹⁸ Gender tailored messaging promoting vaccination for both male and female adolescents should be communicated to parents.

Our findings showed that adolescents in high school were more likely to complete the vaccine series. Previous studies have also found that older adolescents, 16–17 y old are more likely to have initiated and completed the HPV vaccine series compared to younger adolescents, who are in the optimal age range for vaccination.^18,22,23 Studies have also shown that physicians and parents are more comfortable vaccinating older adolescents.^24,25 Provider recommendation is one of the strongest predictors of HPV vaccine receipt but strategies to overcome barriers to provider recommendations are needed.^12,26 Previous research by this team also investigated the impact of provider recommendations on adolescent vaccine receipt and found that physician recommendation was important for HPV vaccination.⁶ This article highlights other correlates of HPV vaccine receipt in addition to provider recommendation, which has been demonstrated as important to HPV vaccination.

Finally, the results of this study showed that parents surveyed in the third year were more likely to report initiation of HPV vaccination. As national discussions and health marketing campaigns about the benefits of the vaccine continue, and potentially physicians provide stronger recommendations for the vaccine, parental attitudes may evolve to favor HPV vaccination as a social norm to protecting adolescents.

Limitations

This study has several limitations. First, these results are representative of one county in Georgia, and might not be generalizable to other areas in the state or larger populations. The rate of return of consent forms was low. Attempts to increase the sample size were made for the second year follow up due to the large amount of returned mail due to inaccurate addresses and low response rate during baseline. However, the low response rate is is comparable to similar school-based studies with that require “active parental consent,” in which the parent must return consent forms to the researchers indicating willingness to participate.^6,27-30 All of the data from the surveys were self-reported from parents and guardians of adolescents. A random sample of parents was drawn from each school annually, and there was the potential that respondents answered the survey in multiple years. Since we did not utilize a tracking method, we were unable to link responses to individual participants over the course of the study, and a repeated measures analysis could not be performed. Tracking students was infeasible due to restrictions from the Family Educational Rights and Privacy Act (FERPA) and the Institutional Review Board. Parents who participated in the survey may differ in important ways from the majority of parents who opted not to take the survey. Information was not collected on non-responders. The strength of physician recommendation was excluded from this analysis because results have previously been published.⁶ Finally, social desirability may have factored into parental responses when the survey was conducted by phone.

Conclusion

The HPV vaccine is a cancer prevention tool that can protect future generations. However, there is work to be done to improve coverage rates for both males and females. This study has highlighted the importance of several important predictors of HPV vaccination among adolescents, including parental attitudes and beliefs, as components of a strategy for increasing coverage rates. There is a paucity of research examining HPV vaccination among predominately African American parents. This study contributes to this limited literature in hopes that efficacious interventions will be developed to increase HPV vaccination rates. Future studies should utilize research designs to capture changes in parental attitudes and beliefs along with vaccination behaviors and utilize appropriate health behavior theories shown to modify parental attitudes and beliefs.

Methods

Study population

The study population consisted of a sample of parents of adolescents who were enrolled in one of the participating middle or high schools. Eleven schools (5 middle and 6 high schools) participated in a randomized control trial of a vaccination promotion intervention in one county in Georgia in 2011–2013. Each school was randomly assigned to one of 3 study arms: Arm 1) no intervention (control), Arm 2) an educational brochure about adolescent vaccines mailed home for parents (parent-only), and Arm 3) a curriculum implemented by science teachers in classrooms of adolescents, plus educational brochures used in arm 2 (parent and adolescent). A detailed description of the educational intervention has previously been published.³¹ Students received the intervention to which the school was assigned. In 2013, the county population was 54% African American, and almost 40% of children under 18 y of age lived below the poverty level.³² Eligibility criteria for the survey included: (1) residing in the target county, (2) having an adolescent(s) enrolled in a participating middle or high school, and (3) providing written or verbal consent to participate in the survey. Study protocols were reviewed and approved by the Emory Institutional Review Board (IRB) and the IRBs of collaborating institutions.

Data collection

Data for the present study were collected from telephone and online surveys administered to parents of students in our participating schools at 3 points: (1) baseline, prior to intervention implementation (November–December 2011), (2) first year follow-up, which was 3 to 5 months after the first year-intervention (April-June 2012), and (3) second year follow-up, which was 3 to 5 months after the second year intervention (April–July 2013). Parents were randomly sampled annually from participating schools. Due to the low response rate and to account for the large volume of returned consent packets from incorrect addresses during baseline year, the random sample size was expanded 50% for the second year follow-up. For each year's survey, packets were mailed to the sample of parents (or primary caregivers) of adolescent students.

The packets included an invitation letter briefly descibing the survey, a telephone number to call with questions or to take the survey, and a website address where the survey could be completed online. Verbal consent was obtained over the phone or online after reading the consent letter and checking that they understood and wanted to participate. The online survey was hosted through a secure website, and all stored data were password protected. Reminder calls were made to parents 2 weeks after the intial packets were mailed. A $20 giftcard was offered as compensation for their time. Parents were prohibited from completing the survey multiple times in one data collection period (baseline, intervention cycle 1 or 2). Telephone surveys were conducted by trained research assistants and were approximately 25 minutes in duration. A total of 6,606 survey invitation packets were mailed to parents/guardians of enrolled middle and high school students, and 686 parents completed the survey. At baseline, 117 surveys were completed by phone; 209 were completed at the 1 y follow-up, and 360 were completed at the second year follow up. The overall response rate was 10%.

Survey instrument

The Health Belief Model (HBM) was used to guide survey development, allowing assessment of the effects of psychosocial factors associated with vaccine acceptance among parents.^15,33 HBM guided questions were adapted from surveys with demonstrated reliability and validity among parents.³⁴ The survey was designed to assess 4 main HBM components: (1) perceived susceptibility to disease, (2) perceived severity of disease, (3) perceived benefits of vaccination. The Theory of Reasoned Action (TRA) was also used to to assess the social norm component of vaccination behaviors.³⁵ Table 2 contains the survey questions and how they map onto the constructs.

Measures

The main outcome of interest was parent reported receipt of HPV vaccine. Parents were asked, “Has your child ever received at least one dose of the HPV vaccine?.” If they answered yes, they were asked, “Have they received all 3 doses?.” Responses were dichotomous, yes (1) and no (0). HPV vaccine intentions were assessed by asking parents who had not initiated the series, “Do you ever plan to have your child receive the HPV vaccine?.”

The following demographic data were collected: child's gender (dichotomous), race (categorical), age (continuous), insurance coverage (categorical). Six attitude and belief questions about HPV vaccination were compiled into an attitude score ranging from 0–6, where each “true” response contributed one point to the total attitude score. Higher scores indicate more positive attitudes about HPV vaccination.

Data analysis

Descriptive analyses were performed. Independent T-test analyses were performed to test for mean differences in attitudes among groups. Unadjusted and adjusted logistic regression analyses were conducted to assess demographic and psychosocial correlates. All surveys collected online and over the phone were de-identified; as a result, we were unable to determine if participants completed the survey at more than one time point or track survey responses over the 3 survey time periods. We controlled for this in the analyses by including study year as a covariate. All analyses were conducted using SPSS v21. Missing data were not included.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

We would like to thank Dianne Miller at Emory University, Dr. Ketty M. Gonzalez, former District Health Director for the East Central Health District and Dr. Tara Vogt, our project officer at the Centers for Disease Control and Prevention. We would also like to thank the school district administrators, principals, science teachers, staff of the participating schools, and the students and parents who participated in the study.

Funding

This project is funded by the Centers for Disease Control and Prevention cooperative agreement 5UO11P000413. Dr. Sales was supported by grant K01 MH085506 from the National Institutes of Mental Health, NIH.