Human Papillomavirus Vaccine Uptake,Predictors of Vaccination,and Self-Reported Barriers to Vaccination

Introduction

Human papillomavirus (HPV) is the most common sexually transmitted infection (STI) in the United States, with lifetime prevalence among women estimated at 80%. ¹ The Food and Drug Administration (FDA) approved a quadrivalent vaccine against HPV in June 2006. Subsequently, the Advisory Committee of Immunization Practices (ACIP) recommended routine vaccination of girls 11–12 years of age, as well as catch-up vaccination of girls and women 13–26 years of age, regardless of prior sexual experience. ² The vaccine is highly effective in preventing persistent HPV infection and precancerous cervical lesions caused by HPV-16 and HPV-18 among women who are not infected with those types at the time of vaccination. ^3,4 Given that both high-risk HPV and cervical cancer prevalence rates are higher among poor, minority women in the United States, ^5,6 the vaccine has the potential to either narrow or widen disparities in cervical cancer, depending on vaccine access and uptake patterns among women.

Studies conducted before and just after licensing of the HPV vaccine suggested that adolescent and young adult women generally found vaccination to be acceptable and that gynecological history and beliefs about HPV vaccines were associated with intention to receive the vaccine. ^7,8 Little is known, however, about rates of HPV vaccination, predictors of vaccination, and self-reported barriers to vaccination during the first 2 years after vaccine licensing. It is especially important to explore these factors among low-income, minority women because they are at relatively high risk for HPV infection and for developing cervical cancer later in life. In addition, special attention should be paid to predictors of vaccination and barriers to vaccination in the age group eligible for catch-up vaccination, as they may experience specific barriers to vaccination, including lack of insurance coverage.

Thus, we designed a longitudinal study of predominantly low-income, minority, 13–26-year-old girls and women with the following aims: (1) to describe rates of HPV vaccination in this sample approximately 12–18 months after the quadrivalent HPV vaccine was licensed, (2) to identify which demographic, attitudinal, behavioral, and health systems-related factors measured at baseline predicted actual vaccine receipt at follow-up, and (3) to identify self-reported barriers to HPV vaccination among those who had not received the vaccine at follow-up.

Materials and Methods

The study sample comprised 13–26-year-old girls and women recruited from an urban, hospital-based teen health center who had previously participated in a multisite baseline study that assessed cervicovaginal HPV infection as well as attitudes toward HPV vaccination. ⁹ Inclusion criteria for the baseline study included a history of sexual contact with a man or a woman and ability to complete a survey in English or Spanish. The participants provided written consent to participate in the baseline study (a requirement for parental consent was waived) and were asked if they were willing to be contacted by phone at a later date to ask if they would be interested in participating in a follow-up study. When participants were recontacted for the follow-up study, verbal consent was requested, and those who consented completed a phone survey. The baseline cohort was enrolled between November 2006 and May 2007. The follow-up study was performed approximately 6 months after each participant was enrolled in the original study, between June and December 2007. At the beginning of the follow-up study, the HPV vaccine had been available in the clinic for approximately 6 months (Fig. 1). This study was approved by the Cincinnati Children's Hospital Medical Center Institutional Review Board.

OPEN IN VIEWER

FIG. 1.

Timeline: HPV vaccine licensing, vaccine availability, recruitment for baseline study, recruitment for follow-up study.

The survey completed by participants at baseline assessed demographic factors, gynecological history, risk behaviors, HPV knowledge, beliefs about HPV and HPV vaccines, HPV vaccination history, and both intention and self-confidence to receive the HPV vaccine. ⁹ Participants received written information about HPV infection and HPV vaccines. The phone survey conducted in this follow-up study assessed whether or not the participant had received any doses of HPV vaccine and, if so, dates of vaccination. All immunization dates were confirmed by reviewing the state immunization registry, and any discrepancies were resolved by chart review. Participants were considered late for the second or third dose of the vaccine if the interval between the first and second doses was >4 months or if the interval between the second and third doses was >6 months. Self-reported barriers to receiving the HPV vaccine were assessed only among participants who had yet to receive any doses of the HPV vaccine at the time of the follow-up interview. These participants were asked whether they had not returned to see the physician since enrollment, had returned but not been offered the vaccine, or had been offered the vaccine but did not receive it. Participants indicating that they did not receive the vaccine when offered were then asked why they did not. They were prompted from a list of 10 possible reasons, including an open-ended response category, and were allowed to select multiple reasons. Participants who had received at least one dose of vaccine but were late for their second or third vaccine doses (or both) were also asked to report one or more reasons why they were unable to get the vaccination on time; they were prompted from a list of 8 possible reasons, including an open-ended response category.

We used insurance information collected during the baseline study to create a new variable indicating whether each participant was likely to have coverage for the HPV vaccine through either private insurance, Medicaid, or the Vaccines for Children Program. Insurance plan information collected at baseline was highly correlated with insurance information collected at the time of the follow-up survey (r = 0.71, p < 0.0001).

The primary outcome variable was receipt of at least one HPV vaccination, and the secondary outcome variables were self-reported barriers to receiving at least one HPV vaccination. Exploratory outcomes included reasons why those who had been offered the vaccine between the baseline and follow-up study visits had not received it and reasons why participants had received a second or third dose late.

We first examined if demographic, attitudinal, and behavioral variables measured at baseline were associated with receipt of at least one HPV vaccine dose at follow-up, using chi-square analysis for dichotomous predictor variables and a Wilcoxon rank-sum test for ordinal variables. Those variables associated with vaccination at p < 0.05 were entered into a multivariable logistic regression model to identify baseline variables independently associated with receipt of the HPV vaccine; a stepwise procedure was used for variable selection. We found evidence for collinearity between age and the variable measuring coverage for the HPV vaccine and, therefore, did not include age in the multivariable model.

Results

Approximately 98% of those approached for the baseline study at all sites agreed to participate. ⁹ Of the 262 women who enrolled in the baseline study at the teen health center, 257 (98%) agreed to be recontacted for a follow-up study, and 189 (72%) were both recontacted and agreed to participate in the follow-up study. The baseline characteristics of the women who participated in the follow-up study did not differ significantly from those of women who did not (Table 1). Participants in this study were predominantly black, their mean age was 17.5 years, and the majority was insured by Medicaid at the time of enrollment. Insurance rates varied by age: 90% of those <18 years of age vs. 73% of those ≥18 years of age reported having some type of insurance (p < 0.01). Rates of gonorrhea and Chlamydia infection were high, as expected in a sample of urban, predominantly low-income young women, but no participant reported HIV infection. We calculated that 127 (70.6%) of the sample had coverage for the HPV vaccine through private insurance, Medicaid, or the Vaccines for Children program.

At follow-up, 68 participants (36%) had received at least one dose of the HPV vaccine, of whom 21 (31% of those vaccinated) had received only one vaccine dose, 38 (56%) had received two doses, and 9 (13%) had received three doses. Ten of the 68 patients who had started the vaccination series did so late enough in the follow-up period that they could not yet be considered late for the second vaccine dose at the time of follow-up. Of the remaining 58 patients, 26 (45%) were late for their second dose of vaccine. Among those who received a second vaccine, the mean interval between doses one and two was 89 days (standard deviation [SD] 45 days, range 28–209 days).

Bivariate associations among demographic, behavioral, and attitudinal predictors and receipt of at least one HPV vaccination are shown in Table 2. Factors associated with vaccination included younger age, coverage for HPV vaccination, use of a condom at last sexual intercourse, no history of an abnormal Pap test, no history of pregnancy, and normative beliefs (the perception that parents, partners, and one's clinician would approve of HPV vaccination). Intention to receive the vaccine was not predictive of vaccination. In the adjusted logistic regression model, coverage for vaccination and normative beliefs predicted vaccination, and a history of an abnormal Pap test was associated with reduced likelihood of receiving the vaccine (Table 3). Insurance coverage for vaccination was associated with more than five times the odds of having received the HPV vaccine, and a one-unit increase in the normative beliefs scale (e.g., from agree to strongly agree) was associated with more than twice the odds of having received the HPV vaccine.

Of the 121 participants who received no doses of the vaccine at follow-up, 54 (45%) reported they had not returned to the clinic since the baseline visit, 51 (42%) had returned to the clinic but had not been offered the vaccine, and 15 (12%) had been offered the vaccine but did not receive it. Refusal of vaccination because of concern about insurance coverage for the vaccine was the most frequently reported reason for not receiving the vaccine when offered (n = 13, 86.7% of those reporting reasons why they had refused vaccination). Of the 26 patients who were late to receive their second dose of vaccine, 16 (61.5%) reported they forgot to make an appointment or forgot to return for an appointment they made, and 10 (38.5%) reported they did not know they needed additional injections.

Discussion

In this study, we measured rates of HPV vaccination among a diverse sample of low-income women eligible for catch-up vaccination, identified attitudinal and behavioral predictors of HPV vaccination, and documented self-reported barriers to vaccination. Thirty-six percent of participants had received at least one HPV vaccination at follow-up, compared with 5% when the baseline study was conducted. Investigators from the Centers for Disease Control and Prevention (CDC) recently reported that surveillance systems in six states demonstrated that HPV vaccination rates among 11–18-year-old girls ranged from 6% to 15% in the third quarter of 2007, ¹⁰ and another U.S. survey demonstrated that 25% of 13–17-year-old girls received at least one HPV vaccine in 2007. ¹¹ A single-center study examining rates of vaccination of 11–17-year-old girls who were recruited for participation from a pediatric primary care clinic reported a vaccination rate of 26%. ¹² The comparatively high vaccination rate in our sample may be explained in part by participants' receipt of care at an adolescent clinic where clinicians were generally supportive of HPV vaccination. In addition, subjects who are willing to participate in a clinical study may tend to trust clinicians, which in turn could correlate with agreement to receive a vaccine recommended by their clinician. Delivery of HPV vaccines in school-based settings may lead to higher vaccination rates, as shown in one study from the United Kingdom in which 71% of girls received at least one HPV vaccine dose. ¹³

Coverage of the cost of vaccination, defined as either having insurance that covered the cost of vaccination or eligibility for vaccination through Medicaid or the Vaccines for Children Program, was the strongest predictor of HPV vaccination in this cohort. Younger age was associated with higher vaccination rates as well as higher rates of insurance coverage for HPV vaccination. This relationship between age and insurance coverage is consistent with U.S. population trends in insurance rates: lifetime insurance rates are highest among 13–14-year-olds but decrease to a nadir among young adults aged 23–24 years. ¹⁴ Low-income adolescents are less likely to be insured than higher-income adolescents across every age group. Therefore, lack of insurance coverage may be an important barrier to HPV vaccination in low-income young women eligible for catch-up vaccination.

Normative beliefs—a measurement of the participant's belief that her medical provider, her parents, and others would approve of her receiving the HPV vaccine—was the only attitudinal predictor of vaccination in this study. These findings are consistent with a number of studies demonstrating that social norms predict both intention to be vaccinated and vaccination. ^{15 –17} The importance of normative beliefs in our study population suggests that campaigns for catch-up HPV vaccination may be most successful if they include information about endorsement of HPV vaccination by physicians and trusted individuals and focus on vaccination as a social norm.

Failure to return to the physician within the study period was the most common self-reported barrier to vaccination. Vaccine recall and reminder systems—in which clinics contact patients by mail or phone to remind them to make or keep vaccine appointments—are known to be effective in increasing vaccination rates in both children and adults ¹⁸ and might be effective in increasing rates of HPV vaccination. Proactive outreach to girls in the target age group for catch-up immunization may be particularly important for HPV vaccination efforts because the vast majority of young women initiate sexual activity and are at high risk for acquiring HPV infection during the years when they are eligible for catch-up vaccination. ^19,20 In our sample, for example, 43% of girls reported having a new sexual partner in the past 3 months. As vaccination appears to have no therapeutic benefit in women already infected with vaccine-type HPV at the time of vaccination, ⁴ the individual and public health benefits of vaccination in the catch-up age group will depend on timely vaccination. Recall and reminder systems for vaccination were not in place at the state, local, or clinic level at the time of this study. Establishing such systems for young women in this age group may help maximize both vaccination rates and the health benefits of vaccination.

The second most commonly cited reason for not getting vaccinated was not being offered the vaccine despite returning to the physician's office during the study period. A clinician may not have recommended HPV vaccination during an office visit if a participant was moderately or severely ill, in which case vaccination is not recommended. In other cases, however, a clinician may have missed an opportunity to vaccinate a young woman seeking treatment for a mild illness or another reason. Given the importance of timely vaccination in this age group, office-based interventions may be helpful in maximizing vaccination rates, for example, office-based procedures that identify patients eligible for vaccination. ^21,22 Other reasons clinicians do not recommend HPV vaccination may have to do with their attitudes about the vaccine or specific barriers to recommending the vaccine. In prelicensing studies, providers reported largely positive attitudes about recommending HPV vaccines. ^{23 –26} However, little information is available concerning actual clinician recommendations for HPV vaccines. As clinician recommendation is one of the most important predictors of vaccination, ^17,27,28 further information about clinicians' attitudes, prescribing practices, and barriers to vaccination will be helpful in designing interventions to improve HPV vaccination recommendations.

Of the 121 subjects reporting barriers to vaccination, <5% reported concerns about safety and efficacy, echoing prelicensing studies that reported high acceptability of an HPV vaccine among young adults. ^29,30 Of the 10 subjects who reported concerns about insurance as a reason for refusing vaccine, 5 did in fact have insurance coverage for vaccination, consistent with prior research demonstrating that recipients of both private and public insurance plans may underestimate their benefits. ³¹ In addition to recall and reminder systems for patients and providers, healthcare personnel should help patients to understand their insurance benefits for vaccination if concerns arise.

Contrary to our expectations, intention to receive the HPV vaccine was not associated with having received the vaccine at the time of follow-up. The previously published baseline study in this same cohort examined the correlations among demographic, behavioral, and attitudinal factors and intention to vaccinate at baseline. ⁹ With the exception of history of an abnormal Pap test, all the predictors found to be significantly associated with intention to be vaccinated in the baseline study were also predictive of actual vaccination in this study. Given the short follow-up period and the fact that 90% of those not vaccinated had either not returned to the physician or had not been offered the vaccine, it is likely that many of those intending to receive the HPV vaccine simply did not have the opportunity to do so during the follow-up period. Studies with longer follow-up periods will likely be necessary to definitively assess whether intention to receive the vaccine predicts actual receipt of the vaccine.

Our study has several limitations. Participants were recruited from a hospital-based adolescent primary care clinic serving a racially and ethnically diverse, predominantly low-income population. Generalization from this population is limited both by the demographic composition of our sample and their connection to clinical care. However, initial recruitment took place during both ill visits and health maintenance visits, and the follow-up study did not require a repeat clinical visit; thus, there was at least some variation in terms of connectedness to care. Those who participated in the follow-up study may differ from those who did not participate; however, there were no baseline characteristics that differed in the two groups. The study sample consisted of those previously enrolled in a study of attitudes about HPV vaccines; thus, participants could have had inherently higher interest in HPV vaccination or be more likely to get the vaccine. In addition, the barriers to vaccination in this study are self-reported, and although they provide information about the young women's perceived barriers, they are subject to recall error. Finally, our study followed this population for only 6 months after initial recruitment and, therefore, reports relatively short-term predictors of vaccination. However, conducting such a study soon after the HPV vaccine was licensed should provide useful preliminary information for the timely design of interventions that address barriers and aim to improve vaccination rates.

Conclusions

Our data suggest that strategies to ensure that those who are eligible for catch-up immunization are vaccinated should include attention to young women's attitudes about vaccination and to the implementation of office-based procedures to ensure that clinicians do not miss opportunities for vaccination. Ensuring that all young women in this age group are able to afford vaccination will help to avoid an increase in existing racial and economic disparities in cervical cancer mortality.