Disparities in human papillomavirus (HPV) vaccine initiation and completion based on sexual orientation among women in the United States

ABSTRACT

Objectives

We compared HPV vaccine initiation and completion of heterosexual with lesbian and bisexual (LB) women.

Methods

We aggregated National Health and Nutrition Examination Survey data from 2009 to 2016 for 3,017 women aged 18 to 34 y in the United States. HPV vaccine initiation was defined as reported receipt of ≥1 dose of the vaccine and completion as receipt of the three recommended doses. Weighted percentages and multivariable logistic regression models were used to examine differences in HPV vaccine initiation and completion between heterosexual and LB women.

Results

Approximately 12% of respondents self-identified as LB women. Overall, a higher percentage of respondents (26%) had initiated the HPV vaccine than completed the three vaccine doses (17%). In the bivariate analysis, LB women had higher initiation ([35% of LB women versus 25% heterosexual]; p = .0012) and completion rates ([24% of LB women versus 17% heterosexual]; p = .0052) than heterosexual women. After adjusting for covariates, compared to heterosexual women, LB women were 60% (aOR = 1.60, 95% CI: 1.16–2.19) more likely to initiate and 63% (aOR = 1.63, 95% CI: 1.12–2.37) more likely to complete the HPV vaccine.

Conclusions

Although LB women had higher likelihood of HPV vaccine initiation and completion compared with heterosexual women, their HPV vaccine uptake was well below the Healthy People 2020 target (80%). Understanding differences in the drivers of vaccine uptake in the LB population may inform strategies that would further increase HPV vaccine uptake toward achieving the 80% completion target.

Introduction

Human papillomavirus (HPV) is spread via skin-to-skin contact and infects cutaneous and mucosal epithelial cells in many areas of the body and comprises over 150 subtypes.¹ An estimated 80–90% of sexually active people will acquire HPV in their lifetime.² Among women aged 15 to 59 y, approximately   14million new cases of sexually transmitted infections are attributable to HPV each year.³ For most people exposed, the virus is harmless and the body naturally clears the infection. However, for some, certain strains of the HPV infection are associated with cervical, anal, oropharyngeal, vaginal, vulvar, and penile cancers^4-6 and second primary cancers.^7–9 Specifically, HPV accounts for about 44,000 HPV-associated cancers per year: about 25,000 among women, and about 19,000 among men.¹⁰

The Advisory Committee on Immunization Practices recommends routine HPV vaccination for adolescents between 11 and 12 y of age.¹¹ Catch-up vaccination is also recommended for both males and females aged 13–26 y and shared clinical decision-making through age 45.¹¹ Despite the availability, safety, and efficacy of the HPV vaccines,¹² uptake has been suboptimal compared to other routine vaccinations in adolescents. In 2017, 65.5% of adolescents aged 13–17 y received at least one dose of HPV vaccine (68.6% for girls and 62.6% for boys) and 48.6% complete the series (53.1% for girls and 44.3% for boys).¹³ These rates fall short of the Healthy People 2020 goal of 80% completion of the three-dose vaccine among males and females 13–15 y olds.¹⁴ For young adults aged 18–26 y, this uptake is even lower, with national initiation rates of 40% for women and only 8% for men.¹⁵ Data from the 2013–2015 National Health Interview Survey data reported HPV vaccine initiation was 51.6% among bisexual females and 40.2% for heterosexual females.¹⁶

Sexual minority women (i.e., lesbians and bisexuals, LB) are an often-overlooked group despite their risk for HPV infection. Sexual minority women are at increased risk of HPV infection and cervical cancer from current or prior sexual partners, both female and male.^17–19 Previous studies report that about 30% of LBs have had a history of genital HPV infections compared with 23% of heterosexual women,²⁰ and about 12% have had a history of genital warts vs. 7% of heterosexual women.²¹ Moreover, sexual minority women may be at a higher risk for cervical cancer than heterosexual women because sexual minority women are less likely to receive Papanicolaou (Pap) smear screening,^22,23 and experience more sexual violence.^24,25 With a differential prevalence of HPV infection among sexual minority versus heterosexual women but lower prevalence of regular Pap test use, HPV vaccination among sexual minorities should be an important strategy for preventing HPV-related disease. Therefore, the aim of this study was to compare HPV vaccine initiation and completion between heterosexual and LB women using the National Health and Nutrition Examination Survey (NHANES) data from 2009 to 2016 which is a probability sampling design to provide national estimates of various health outcomes.

Methods

Data collection

We utilized publicly available data from the NHANES, pooled from 2009 to 2016 for 3,017 women aged 18 to 34 y. The censored upper age was selected because these women would have been eligible to have received the HPV vaccine at some point after it was licensed in the U.S. in 2006. The NHANES is a nationally representative cross-sectional survey of United States adults and children that assesses health and nutritional status by using in-home interviews and physical examinations. Starting in 1999, the NHANES surveys about 5000 new persons annually. Combining data across several years provides adequate sample size for hard-to-reach populations such as sexual minority women. Survey response rates for the NHANES-examined sample ranged from 71% to 80% for the 6-y period.²⁶ More details regarding the NHANES design and sampling strategies are described elsewhere.²⁷ This study did not require ethical approval from an institution review board since we used publicly available datasets.

Measures

HPV vaccination was assessed with the question “Human Papillomavirus (HPV) vaccine is given to prevent cervical cancer in girls and women. There are two HPV vaccines available called Cervarix and Gardasil. It is given in 3 separate doses over a 6 month period. How many doses have you received?” We defined HPV vaccine initiation as reported receipt of at least one dose of the vaccine. HPV vaccine completion was defined as reported receipt of three recommended doses. We used 3-dose as completion since the 2-dose vaccination series was not recommended until 2016 when the survey had already been conducted. Those who answered, “don’t know” (n = 101) were included in the “No” initiation/completion group. In addition, a sensitivity analysis was performed where the “don’t know” group were excluded.

The main independent variable for our study was sexual orientation. In NHANES, gender was assessed with the question “What is the gender of the respondent? Male or Female”. Sexual orientation for females aged 18–59 y was then assessed using the sexual orientation question: “Do you think of yourself as … heterosexual or straight (that is, sexually attracted only to men); homosexual or lesbian (that is, sexually attracted only to women); bisexual (that is, sexually attracted to men and women); something else; or you’re not sure?” For the present study, we defined LB as women who self-identified either as lesbian, bisexual, or something else/not sure. Reported sexual orientation was then dichotomized as heterosexual versus LB. We also conducted a sensitivity analysis which excluded women who answered something else/not sure. The following covariates based on a literature review were included in the model: participants’ age, race/ethnicity (non-Hispanic White, non-Hispanic Black, Hispanic, and other race – including multiracial), marital status (married/other), education (high school diploma or less, some college or 2-y degree, and 4-y college graduate or higher), and annual household income (≤$24 999, 25 000 USD to 54 999 USD, 55 000 USD to 99 999 USD, and 100 000+ USD).^28,29 HPV vaccine initiation and completion rates have been shown to vary significantly by rural-urban residence; however, we were unable to adjust for rural-urban status because NHANES do not have Metropolitan Statistical Area (MSA) status or Rural-Urban Continuum (RUC) codes.

Statistical analysis

Analysis was conducted between January and March 2019. We analyzed data by using SAS version 9.4 survey procedures (SAS Institute, Cary, NC), incorporating both the design information and weights as specified in the NHANES Analytic and Reporting Guidelines. The survey procedures, which incorporate survey-sampling weights to account for the complex sampling design of NHANES, provide representative estimates of the United States population. We compared demographic characteristics by sexual orientation using chi-square test for proportions and the student t-test for continuous variables. Weighted multivariable logistic regression models were used to examine differences in HPV vaccine initiation and completion. All variables were included in the multivariable models regardless of their association with the outcome variable.

To examine the effect of potential recall bias, we conducted three separate sensitivity analyses where we excluded women (n = 101) who answered don’t know/refuse to the HPV vaccination question (model 1); excluded women (n = 82) who answered something else/don’t know to the sexual orientation question (model 2); and treated the something else/don’t know as a category in the sexual orientation variable (Model 3). Weighted multivariable logistic regression models (a total of six) were used to examine differences in HPV vaccine initiation and completion for each model above. All tests were two sided, and an alpha of <0.05 or confidence intervals that exclude one were considered statistically significant.

Results

Of the 7129 women aged 18–34 y old, 3041 answered the sexual orientation question. Among the 3041 women, we excluded 24 who refused to answer the question, resulting in 3017 women being included in the study. Table 1 shows the characteristics of the study population overall and by sexual orientation. All percentages presented shown are weighted. The average age was 25.3 (SD = 5.0) y and approximately 377 (11.7%) women self-identified as LB. Overall, 25.9% of the women had initiated the HPV vaccine, and the initiation rate among LB women was significantly higher than heterosexual women (34.6% of LB women versus 24.8% of heterosexual women, p = .0012). Overall vaccine completion rate was 17.4%, and similarly the rate among LB women was significantly higher than heterosexual women (24.1% of LB women versus 16.5% of heterosexual, p = .0052). Heterosexual women were more likely to be married (p < .0001) and earn higher income (p < .0001) than their LB counterparts.

Table 1.

Characteristics of adult women aged 18–34 by sexual orientation, NHANES (2009–2016)

	Total cohort (N = 3017)	Heterosexual (N = 2640)	LB (N = 377)	P-value*
HPV vaccination
Initiation (Yes), n (%)	764 (25.9)	652 (24.8)	112 (34.6)	0.0012
Completion (Yes), n (%)	493 (17.4)	418 (16.5)	75 (24.1)	0.0052
Age, mean (sd)	25.3 (5.0)	25.4 (5.1)	24.4 (4.5)	<0.0001
Race/ethnicity, n (%)				0.0369
Non-Hispanic white	1060 (57.3)	930 (57.6)	130 (55.2)
Non-Hispanic black	677 (14.1)	572 (13.6)	105 (18.2)
Hispanic	852 (19.0)	769 (19.5)	83 (15.6)
Other race	428 (9.6)	369 (9.3)	59 (11.0)
Marital status, n (%)				<0.0001
Married	1232 (47.6)	1114 (49.2)	118 (35.3)
Other	1242 (42.0)	1055 (40.6)	187 (52.6)
Missing	543 (10.4)	471 (10.2)	72 (12.1)
Education level, n (%)				<0.0001
Missing	542 (10.4)	470 (10.2)	72 (12.1)
High school graduate or less	839 (26.4)	703 (25.3)	136 (35.1)
Some college or 2-y degree	978 (35.7)	864 (35.6)	114 (36.2)
4-y college graduate or higher	658 (27.5)	603 (28.9)	55 (16.6)
Household income, n (%)				<0.0001
Missing	116 (2.7)	102 (2.9)	14 (2.8)
<$24,999	1092 (30.7)	917 (28.0)	175 (43.8)
$25,000 to $54,999	871 (28.7)	764 (28.8)	107 (28.2)
$55,000 to $99,999	579 (21.7)	526 (22.7)	53 (14.5)
$100,000 or above	359 (16.0)	331 (16.7)	28 (10.6)

HPV, human papillomavirus; NHANES, National Health and Nutrition Examination Survey; LB, lesbians and bisexuals; n, unweighted frequencies; percentages shown in this table are weighted.

*P-values are based on t-test or Chi-square test. Comparison is between heterosexual and LB columns.

Table 2 presents the crude and adjusted odds of HPV vaccine initiation and completion based on sexual orientation. In the crude model, LB women were more likely to initiate (aOR = 1.61, 95% CI: 1.20–2.15) and complete (aOR = 1.61, 95% CI: 1.15–2.24) the HPV vaccine than heterosexual women. After adjusting for potential confounders, the results remained similar. Compared to heterosexual women, LB women were 60% more likely to initiate (aOR = 1.60, 95% CI: 1.16–2.19) and 63% more likely to complete (aOR = 1.63, 95% CI: 1.12–2.37) the HPV vaccine. In the adjusted model, a 1-y increase in age was associated with lower odds of initiating (aOR = 0.82, 95% CI: 0.80–0.85) and completing (aOR = 0.86, 95% CI: 0.83–0.89) the HPV vaccination. Women who were not married were more likely to initiate (aOR = 1.34, 95% CI: 1.02, 1.75) and complete (aOR = 1.61, 95% CI: 1.17, 2.22) the HPV vaccine compared to those that were married. The likelihood of initiating and completing the HPV vaccine declined steadily across education levels and was lowest for those with high school diploma or less compared to those with college degree or higher.

Table 2.

Weighted multivariable logistic regression models for HPV vaccine uptake among young adult women, NHANES (2009–2016) aged 18–34 y

	HPV vaccine initiation		HPV vaccine completion
	cOR (95% CI)	aOR (95% CI)	cOR (95% CI)	aOR (95% CI)
Sexual orientation
Heterosexual	1.0	1.0	1.0	1.0
LB	1.61 (1.20, 2.15)	1.60 (1.16, 2.19)	1.61 (1.15, 2.24)	1.63 (1.12, 2.37)
Age	0.85 (0.83, 0.87)	0.82 (0.80, 0.85)	0.87 (0.85, 0.90)	0.86 (0.83, 0.89)
Race
Non-Hispanic white	1.0	1.0	1.0	1.0
Non-Hispanic black	0.84 (0.66, 1.05)	0.79 (0.61, 1.02)	0.70 (0.53, 0.92)	0.67 (0.50, 0.90)
Hispanic	0.59 (0.47, 0.75)	0.64 (0.49, 0.83)	0.53 (0.40, 0.70)	0.61 (0.46, 0.83)
Other race	0.68 (0.50, 0.92)	0.58 (0.41, 0.82)	0.67 (0.46, 0.96)	0.59 (0.40, 0.88)
Marital status
Married	1.0	1.0	1.0	1.0
Other	2.04 (1.61, 2.58)	1.34 (1.02, 1.75)	2.16 (1.62, 2.87)	1.61 (1.17, 2.22)
Missing	3.99 (3.00, 5.30)	–	3.90 (2.80, 5.41)	–
Education level
Missing	2.52 (1.86, 3.42)	–	2.05 (1.46, 2.88)	–
High school grad or less	0.47 (0.34, 0.65)	0.30 (0.20, 0.44)	0.39 (0.26, 0.58)	0.30 (0.19, 0.47)
Some college or 2-y degree	1.24 (0.94, 1.62)	0.77 (0.56, 1.05)	0.97 (0.71, 1.32)	0.65 (0.46, 0.93)
4-y college graduate or higher	1.0	1.0	1.0	1.0
Household income
Missing	0.59 (0.32, 1.08)	0.66 (0.33, 1.35)	0.25 (0.11, 0.55)	0.28 (0.13, 0.63)
<$24,999	0.68 (0.49, 0.92)	0.64 (0.45, 0.92)	0.53 (0.37, 0.75)	0.54 (0.37, 0.80)
$25,000 to $54,999	0.73 (0.53, 1.01)	0.90 (0.63, 1.30)	0.59 (0.41, 0.85)	0.74 (0.50, 1.10)
$55,000 to $99,999	0.58 (0.41, 0.83)	0.77 (0.52, 1.14)	0.51 (0.35, 0.77)	0.68 (0.44, 1.05)
$100,000 or above	1.0	1.0	1.0	1.0

HPV, human papillomavirus; NHANES, National Health and Nutrition Examination Survey; cOR, crude odds ratio; aOR, adjusted odds ratio; CI, confidence interval; LB, lesbians and bisexuals.

Models are adjusted for age, race, marital status, education level, income level, and sexual orientation.

The sensitivity analyses are presented in Table 3. Overall, the findings from the sensitivity analyses were similar to that of the main analyses. In model 1 where we excluded women who answered don’t know/refuse to the HPV vaccination question, LB women were more likely to initiate (aOR = 1.59, 95% CI: 1.15–2.19) and complete (aOR = 1.63, 95% CI: 1.12–2.37) the HPV vaccine than heterosexual women. Similarly, in model 2 where we excluded women who answered something else/don’t know to the sexual orientation question, LB women were more likely to initiate (aOR = 1.75, 95% CI: 1.24–2.48) and complete (aOR = 1.76, 95% CI: 1.17–2.64) the HPV vaccine than heterosexual women. Finally, in model 3 where we treated the something else/don’t know as a category in the sexual orientation variable, compared to heterosexual women, LB women were more likely to initiate (aOR = 1.76, 95% CI: 1.24–2.48) and complete (aOR = 1.76, 95% CI: 1.17–2.64) the HPV vaccine. There were no differences between women who answered something else/don’t know and heterosexual women.

Table 3.

Weighted multivariable logistic regression models for HPV vaccine uptake among young adult women, NHANES (2009–2016) aged 18–34 y [Sensitivity Analyses]

	Weighted percentage	HPV vaccine initiation aOR (95% CI)	HPV vaccine completion aOR (95% CI)
Sexual orientationa
Heterosexual	88.3	1.0	1.0
LB	11.7	1.59 (1.15, 2.19)	1.63 (1.12, 2.37)
Sexual orientationb
Heterosexual	90.0	1.0	1.0
LB	10.0	1.75 (1.24, 2.48)	1.76 (1.17, 2.64)
Sexual orientationc
Heterosexual	88.7	1.0	1.0
LB	9.8	1.76 (1.24, 2.48)	1.76 (1.17, 2.64)
Something else/don’t know	1.5	0.61 (0.26, 1.39)	0.53 (0.18, 1.54)

HPV, human papillomavirus; NHANES, National Health and Nutrition Examination Survey; aOR, adjusted odds ratio; CI, confidence interval; LB, lesbians and bisexuals.

^aModel 1: Excludes women who answered don’t know/refuse to the HPV vaccination question.

^bModel 2: Excludes women who answered something else/don’t know/refuse to the sexual orientation question.

^cModel 3: Adds something else/don’t know/refuse as a category or level to the sexual orientation variable.

Each model is adjusted for age, race, marital status, education level, income level, and sexual orientation.

Discussion

Cancer risk and outcomes among LBs is an important public health concern, as formalized in multiple position statements (e.g. American Society of Clinical Oncology) and designation as a priority population by the National Institutes of Health.^30,31 In this study, we investigated the association between sexual orientation and HPV vaccine uptake among young adult women aged 18–34 y using a national sample. The vaccination rate overall was low with only 1-in-4 having initiated the vaccine and about 1-in-5 having completed the vaccine dose. After adjusting for covariates, our study showed that LB women were 60% and 63% more likely to initiate and complete the HPV vaccine, respectively, compared to heterosexual women.

This study found that HPV vaccination rate (both initiation and completion) was very low in both LB (35%, 25%) and heterosexual (24%, 17%) women even though it was higher in LB women. The initiation rate for LB was slightly higher than what Agénor et al.²⁹ reported (22.5%) but less than the 45% reported by McRee et al.³² In addition, the completion rate in our study was less than what was reported by McRee et al. (32%).³² It should be noted that McRee et al.³² used data from 543 LB and bisexual women ages 18–26 during Fall 2013 and Agenor et al.²⁹ used the 2006–2010 National Survey of Family Growth data for 3,253 women and girls aged 15–25 y. Our findings show that vaccination rates have been increasing compared to Agenor et al.²⁹ study, which is expected. It is hard to explain how the vaccination rates in McRee et al.³² were much higher than what we found in our study. These vaccination rates are far below the Healthy People 2020 goal of achieving 80% vaccination rates. After adjusting for covariates, we found that adult LB women were still more likely to initiate and complete the vaccine than their heterosexual peers. Interestingly, several studies have shown that LB women compared to heterosexual women are significantly less likely to receive a regular Pap smear screening, which is the most important screening tool for the detection of precancerous cervical abnormalities.^11,12 Therefore, HPV vaccination holds great promise and importance for the primary prevention of HPV-associated diseases in LB women.

There are several explanations for the higher HPV vaccination uptake observed among LB women. LB women are more likely to have a sexual encounter at an early age, have more male and female partners, and experience more sexual violence than their heterosexual peers.^16,17,33 Therefore, LB women might experience more sexual and reproductive health risks, which may encourage them to initiate and complete HPV vaccination as a preventive measure. In addition, the assertion that LB women have more sexual partners may influence health-care providers to recommend the HPV vaccine to this group.³³ Furthermore, previous studies and the current one have found that unmarried women are more likely to initiate and complete the HPV vaccination. Also, in our study, LB women were less likely to be married than heterosexual peers. This could, therefore, explain why they were more likely to initiate and complete the vaccine compared with heterosexual women.

Implications

With LB women at increased risk of HPV infection,^17–19 and having lower prevalence of Pap test use thereby having a higher risk of cervical cancer,^11,12,34 improving HPV vaccine uptake is paramount for preventing HPV-related disease. Effective policy interventions such as increasing coverage for HPV vaccination and implementing opt-out provisions in schools could help increase vaccine uptake in adolescents and young adults,³⁵ irrespective of sexual minority status. Future HPV prevention strategies should be enhanced due to the effect of vaccination on the cost-effectiveness of existing cervical screening programs. Regardless of sexual orientation identity, health-care providers should counsel women and offer HPV vaccination to every individual who meets age guideline since provider recommendation³⁶ is the main predictor of vaccine uptake. Furthermore, the 2010 Affordable Care Act (ACA)’s policy of increasing access to preventive services for HPV vaccine coverage without cost sharing for individuals aged 19–25 has affected the likelihood of initiating and completing the HPV vaccine among young adult women, increasing it by 7.7 and 5.8 percentage points, respectively.³⁷ A couple of studies have observed that 1 dose of the HPV vaccine may gain similar protection against HPV infection compared with those who received additional doses. There should be more clinical trials in the future to evaluate the efficacy of a single-dose regime compared with 2 or 3 doses.^38,39

Limitations

The study does have some limitations. Our study combined LB and bisexual women into one group, which precludes assessing differences between those sexual minority subgroups among women. We were unable to distinguish between these groups due to the small sample sizes. Second, the use of self-reported measures of HPV vaccine initiation and completion, as well as sexual minority status, could have introduced misclassification due to recall bias. However, previous studies have shown that recall and electronic medical records for HPV vaccine are comparable.⁴⁰ There was also a high non-response rate for the sexual orientation question which could have affected findings. But the survey weights generated by NHANES do account for non-response bias; therefore, we assume this effect will be minimized. Third, there could be a risk of social desirability bias which could affect our findings. Fourth, there is the potential of clustering of both individual-level and contextual factors that logistic regressions are unable to account for. This could lead to the underestimation of standard errors of regression coefficients, leading to an overestimation of observed effect sizes. However, this bias is minimized due to the fact that we accounted for the complex survey design used for NHANES which adjusts for oversampling, stratification, and clustering. Lastly, although we assessed and controlled for a limited number of factors that are reported in NHANES, the potential for residual confounding remains. For example, there was no information on provider recommendation, which has been shown to positively improve vaccination rates, ³⁶ nor rural-urban status of participants.

Conclusions

This study indicates that LB women are more likely to both initiate and complete the HPV vaccine than heterosexuals. Continued inclusion of sexual minority status on national surveys will allow for a more nuanced comparison of this population including sexual minority subpopulations compared to heterosexuals. Further research to understand the factors that increase uptake by LB women can aid in improving uptake in this sexual minority population and thus among all women irrespective of the sexual orientation.

Disclosure statement

All authors have no conflict of interest to declare.