Socioeconomic Disadvantage and Human Papillomavirus (HPV) Vaccination Uptake

Abstract

Importance:

Despite availability of safe and effective human papillomavirus (HPV) vaccines, vaccination uptake remains low in the U.S. Research examining the impact of neighborhood socioeconomic status on HPV vaccination may help target interventions.

Objective:

To examine the association between area deprivation and HPV vaccine initiation and completion.

Design, Setting, Participants:

Retrospective cohort study of individuals aged 11–18 years residing in the upper Midwest region. Receipt of HPV vaccination was examined over a three-year follow-up period (01/01/2016–12/31/2018).

Main Outcomes and Measures:

Outcomes of interest were initiation and completion of HPV vaccination. Demographic data were collected from the Rochester Epidemiology Project (REP). Area-level socioeconomic disadvantage was measured by calculating an Area Deprivation Index (ADI) score for each person, a measure of socioeconomic disadvantage derived from American Community Survey data. Multivariable mixed effect Cox proportional hazards models were used to examine the association of ADI quartiles (Q1-Q4) with HPV vaccine series initiation and completion given initiation.

Results:

Individuals residing in census block groups with higher deprivation had significantly lower likelihood of HPV vaccine initiation (Q2: HR=0.91, 0.84–0.99 Q3: HR= 0.83, 0.76–0.90; Q4: HR=0.84, 0.74–0.96) relative to those in the least-deprived block groups (Q1). Similarly, those living in block groups with higher deprivation had significantly lower likelihood of completion (Q2: HR=0.91, 0.86–0.97; Q3: HR= 0.87, 0.81–0.94; Q4: HR=0.82, 0.74–0.92) compared to individuals in the least-deprived block groups (Q1).

Conclusions and Relevance:

Lower probability of both HPV vaccine-series initiation and completion were observed in areas with greater deprivation. Our results can inform allocation of resources to increase HPV vaccination rates in our primary care practice and provide an example of leveraging public data to inform similar efforts across diverse health systems.

Introduction

Human papillomavirus (HPV) causes nearly all cervical cancers and some cancers of the vulva, vagina, anus, penis, and oropharynx ^1,2. In the U.S., an estimated average of 35,900 HPV-attributable cancers were diagnosed each year from 2013 to 2017 ³. Although most HPV-related cancers can be prevented by the HPV vaccine ^4,5, only 54.2% of adolescents 13 through 17 years of age in the U.S. have completed their HPV vaccination series ⁶. This rate remains well below the Healthy People 2020 goal of 80% ^6–9. Concerns about vaccine safety, limited understanding of the need for HPV infection prevention, and lack of health care provider recommendations have been identified as barriers to HPV vaccination ^10–12. In addition, sociodemographic characteristics have been associated with varying uptake, both initiation and completion, of HPV vaccination. Characteristics associated with lower uptake include male sex ^13,14, lower socioeconomic status ¹⁵, and rurality ¹⁶.

Over the last two decades, geographic disparities in rates of HPV cancers have persisted with rural areas experiencing greater incidence and mortality of cancers caused by HPV than urban areas ^17,18. The disproportionate burden of HPV-associated cancers among rural populations may be explained partly by lower knowledge of and adherence to evidence-based practices, including HPV vaccination for prevention of all forms of HPV-attributable cancers and early detection of cervical cancer through screening ^18–21. Lower HPV vaccine initiation and completion among rural populations may be attributed to lower access to care and vaccine availability and lack of health care provider recommendations ²¹. Addressing this inequitable burden requires focused action to eliminate cancer-related disparities through cancer control efforts informed by an understanding of the impact of social determinants of health on cancer-related behavior and outcomes ²².

Previous studies suggest that individual and area-level sociodemographic factors play a role in both HPV vaccine initiation and completion ¹⁵. One measure that effectively captures geospatial disparities and social determinants of health is the Area Deprivation Index (ADI) ^23,24. The ADI is a validated, area-based, composite indicator comprised of 17 U.S. Census variables spanning four domains: income, housing, employment, and education ^24,25. The ADI can be constructed at various levels of geography, the most granular census-based region being the block group ²⁴. Research conducted by our team suggests that more deprived and socioeconomically disadvantaged block groups in the Midwest have lower rates of cervical, breast, and colorectal cancer screenings ²⁴. While HPV-related cancer disparities have been described previously in rural populations ¹⁸, it is not known whether area-level deprivation ²⁶, independent of rurality, is associated with initiation and completion of HPV vaccination ²⁴.

The objective of this study was to examine HPV-vaccine related disparities as a function of area deprivation using patient-level data from persons residing in a largely rural, Upper Midwest region from January 1, 2016, to December 31, 2018. Understanding the social determinants of health that influence initiation and completion of the HPV vaccine will inform clinical practice and outreach interventions aimed at improving both initiation and completion of the vaccine series.

Materials and Methods

Study Population

This retrospective cohort study included individuals aged 11–18¹ on January 1, 2016, residing in the upper Midwest region captured by the Rochester Epidemiology Project (REP) ²⁷. The REP has been previously described. Briefly, the REP is a data-linkage research infrastructure that links medical record information from multiple healthcare providers to persons residing in a 27-county region of Southern Minnesota and West-Central Wisconsin. The REP captures data for approximately 60% of persons residing in this region and approximately 90% of persons residing in a smaller 7-county region. Demographic data (age, sex, race) and address information for the study population were obtained using the electronic REP indices. This work was reviewed and approved by the Mayo Clinic and Olmsted Medical Center Institutional Review Boards.

Cohort Acquisition

For the analysis of HPV vaccine completion, given HPV vaccine initiation, the cohort of individuals was ascertained on 01/01/2016, and only information occurring within the observation period of 01/01/2016 to 12/31/2018 was considered. eFigure illustrates fve fictitious individuals’ (A-F) timelines of HPV vaccination. Individuals A and B have both initiated and completed outside the window of observation and, as such, are not included in the analysis. Individual C initiated HPV vaccination during the observation period but completed outside of the study period and, as such, is censored on 12/31/2018 but included in the analysis. Individual D initiated and completed entirely within the observation period and is, therefore, counted. Individual F initiated prior to 01/01/2016 but completed their HPV vaccination during the observation window. As a result, their time prior to 01/01/2016 is removed, but the time after 01/01/2016 is included, as is the completion. eFigureB illustrates how the timelines in eFigureA are converted to days from initiation to completion for use in the mixed effects cox models.

HPV Vaccinations

Current procedure terminology (CPT) codes corresponding to HPV delivery (90649, 90650 and 90651) were extracted using the REP from January 1, 2016, through December 31, 2018. The “eligible-to-initiate” group had received no HPV vaccination prior to January 1, 2016, and the “eligible-to-complete” group had not met criteria for HPV vaccination series completion and had either: A) received one or two HPV vaccine doses before January 1, 2016, or B) received one HPV vaccine dose during the study period. Individuals in the “eligible-to-initiate” sample who initiated during the study period were counted in the “eligible-to-complete” sample. Appropriate spacing definitions for initiation and completion of the HPV vaccination were in accordance with current Advisory Committee on Immunization Practices standards for our time frame ^28,29.

Area Deprivation Index (ADI)

The ADI, a composite measure of socioeconomic disadvantage ^25,30,31 was calculated for all U.S. census block groups ²⁴. The modified 2016 ADI values include 17 variables from five-year American Community Survey (ACS) estimates ^24,32. A previous paper by Kurani et al. (2020) contains detailed information on the methodology for ADI derivation and the weights assigned to each variable (eTable) ²⁴. Block group level ADI quartiles were defined nationally from the least deprived 25% of block groups (Q1) to most deprived 25% of block groups (Q4). A Census Geocoder was used to obtain latitude and longitude points for each patient’s address ³³. These coordinates were spatially joined to a TIGER/Line census block group shapefile in ArcMap, version 10.7 (ESRI). Patient addresses that were not located in Minnesota or Wisconsin and those that could not be geocoded with a match score greater than 60, which represents how reliably the inputted address matches a candidate in the reference data, were excluded.

Rurality

Rural/urban status was defined using Rural Urban Commuting Area (RUCA) definitions at the zip-code level ³⁴. Zip-code level RUCA codes were used as census block group RUCA codes are not available through the U.S. Department of Agriculture website ^35,36. Rural/urban status was defined as rural (RUCA 4.0, 4.2, 5.0, 5.2, 6.0, 6.1,7.0, 7.2, 7.3, 7.4, 8.0, 8.2, 8.3, 8.4, 9.0, 9.1, 9.2, 10.0, 10.2, 10.3, 10.4, 10.5, 10.6) and urban (RUCA 1.0, 1.1, 2.0, 2.1, 3.0, 4.1, 5.1, 7.1, 8.1, 10.1) ³⁴. Individuals with missing zip codes were excluded.

Analysis

Demographic characteristics were summarized separately for those who were eligible to initiate the HPV vaccine series as of January 1, 2016, and those who were eligible to complete the vaccine series during the study period (January 1, 2016, through December 31, 2018). The number and percent of individuals who initiated or completed the vaccine series in the three-year follow-up period from January 1, 2016, through December 31, 2018, were similarly summarized in Table 1, along with three-year rates of vaccine initiation and completion. Initiation of HPV vaccination and completion of HPV vaccination, given initiation, was modeled by ADI quartiles and demographic characteristics individually, as well as jointly using mixed-effects Cox proportional hazards models implemented in the R package coxme v 2.2–16. For all models, clustering was accounted for using a random effect for census block level, and all other predictors were modeled as fixed effects. For vaccine initiation, time was started on January 1, 2016, as all adolescents included were in the risk set. For this cohort, age was modeled as age at ascertainment since all adolescents were eligible at the start of the follow-up interval. For vaccine completion, time was started at initiation; and for those individuals who initiated prior to ascertainment (January 1, 2016), their time prior to January 1, 2016, was removed from the risk set. In the analysis of HPV vaccine completion, age was modeled as a time-dependent covariate, as not all adolescents were in the risk set as of January 1, 2016. Modeling age as a time-dependent covariate allows for comparisons of age at any point after initiation during the follow-up period of January 1, 2016, to December 31, 2018. Results were presented as hazard ratios and 95% confidence intervals (CIs). All analyses were complete in SAS v9.4 or R v3.6.2.

Table 1.

Number and percent of eligible individuals who initiated and completed HPV vaccination series between January 1, 2016, and December 31, 2018

	Eligible to Initiate N=34,637	Eligible to Complete N=19,936
	Initiation N, %	Completion N, %1
		Initiated: 10,753, 31.0%	Completed: 10,121, 66.0%
	Total
ADI quartiles
First	2888, 35.3%	2942, 70.3%
Second	4627, 30.6%	4308, 65.7%
Third	2472, 28.6%	2230, 63.5%
Fourth	766, 28.5%	641, 59.1%
Age^
11–12	5662, 48.5%	4690, 75.6%
13–15	2941, 26.7%	3586, 67%
16–18	2150, 18.0%	1845, 48.1%
Sex
Boys	5672, 29.6%	5007, 63.8%
Girls	5081, 32.9%	5114, 68.4%
Race
White	9055, 31.7%	8687, 67.5%
Black	465, 26.7%	385, 54.2%
Asian	324, 37.4%	298, 65.2%
Other/Unknown	909, 26.4%	751, 57.5%
Ethnicity
Hispanic	894, 35.6%	9380, 66.5%
Not Hispanic	9859, 30.7%	741, 60.3%
Rurality
Rural	3336, 30.2%	2682, 62.7%
Urban	7417, 31.4%	7439, 67.1%

¹Total events, three-year rate

^{^}At cohort selection

ADI=Area Deprivation Index

Results

There were 34,637 individuals who were eligible to initiate (had no HPV vaccine doses prior to January 1, 2016) and 19,936 individuals who were eligible to complete the HPV vaccination series (had one or two doses prior to January 1, 2016, or initiated the series during the study period). Individuals eligible to initiate the vaccine series had an average age of 14 years (SD, 2.5), 45% were female, 93% were non-Hispanic, and 83% were White (Table 2). Individuals eligible to complete the series had similar demographic characteristics (Table 2). Overall, 31% of eligible individuals initiated, and 66% completed during the study period. Generally, initiation and completion during the observation period was more frequently observed among those in the first ADI quartile (least deprived) (35%, 70%, respectively), those ages 11–12 years (49%, 76%, respectively) and girls (33%, 68%, respectively) compared to relevant subgroups (Table 1). Initiation and completion were slightly more frequent among Hispanics (35%, 67%, respectively) compared to non-Hispanics (31%, 60%, respectively) (Table 1). Initiation and completion were less frequent among rural (30%, 63%, respectively) compared to urban populations (31%, 67%, respectively) (Table 1).

Table 2.

Baseline demographic characteristics of the study population on January 1, 2016, by HPV vaccination status

	Eligible to Initiate	Eligible to Complete
	N=34,637	N=19,936
	N, %	N, %
ADI quartile
First	8172, 23.6%	5334, 26.8%
Second	15131, 43.7%	8540, 42.8%
Third	8648, 25.0%	4616, 23.2%
Fourth	2686, 7.8%	1446, 7.3%
Age
11–12	11670, 33.7%	13707, 31.3%
13–15	10997, 31.8%	14837, 33.9%
16–18	11970, 34.6%	15280, 34.9%
Sex
Boys	19196, 55.4%	10248, 51.4%
Girls	15441, 44.6%	9688, 48.6%
Race
White	28585, 82.5%	16640, 83.5%
Black	1744, 5.0%	979, 4.9%
Asian	867, 2.5%	583, 2.9%
Other/Unknown	3441, 9.9%	1734, 8.7%
Ethnicity
Hispanic	2514, 7.3%	1636, 8.2%
Not Hispanic	32123, 92.7%	18300, 91.8%
Rurality
Rural	11051, 31.9%	5631, 28.2%
Urban	23586, 68.1%	14305, 71.8%

ADI=Area Deprivation Index

Association of Area-Level Deprivation with HPV Vaccination

Individuals living in more deprived block groups were significantly less likely to initiate and complete HPV vaccinations. The probability of vaccine series initiation was approximately 16% lower for those living in ADI quartile 4 (most deprived) compared to ADI quartile 1 (least deprived) (Q2: HR=0.91, 0.84–0.99 Q3: HR= 0.83, 0.76–0.90; Q4: HR=0.84, 0.74–0.96) (Table 3). Similarly, individuals living in block groups with higher deprivation had significantly lower probability of completion (Q2: HR=0.91, 0.86–0.97; Q3: HR= 0.87, 0.81–0.94; Q4: HR=0.82, 0.74–0.92) compared to individuals in the least deprived block groups (Q1) (Table 3).

Table 3.

Factors Associated with Initiation and Completion of the HPV Vaccination Series from January 1, 2016, through December 31, 2018

	Initiation of HPV vaccination		Completion of HPV vaccination given initiation
	Unadjusted	Multivariable*	Unadjusted	Multivariable*
	HR (95% CI)	HR (95% CI)	HR (95% CI)	HR (95% CI)
ADI quartiles
First	Referent	Referent	Referent	Referent
Second	0.89 (0.82, 0.97)	0.91 (0.84, 0.99)	0.91 (0.86, 0.97)	0.91 (0.86, 0.97)
Third	0.80 (0.73, 0.88)	0.83 (0.76, 0.90)	0.85 (0.79, 0.91)	0.87 (0.81, 0.94)
Fourth	0.78 (0.68, 0.88)	0.84 (0.74, 0.96)	0.77 (0.69, 0.86)	0.82 (0.74, 0.92)
Age^
11–12	Referent	Referent	Referent	Referent
13–15	0.46 (0.44, 0.49)	0.47 (0.45, 0.49)	0.82 (0.78, 0.87)	0.83 (0.79, 0.87)
16–18	0.30 (0.29, 0.32)	0.30 (0.29, 0.32)	0.52 (0.49, 0.55)	0.52 (0.49, 0.55)
19–21	-	-	0.26 (0.24, 0.29)	0.26 (0.24, 0.30)
Sex
Boys	Referent	Referent	Referent	Referent
Girls	1.16 (1.12, 1.21)	1.13 (1.09, 1.17)	1.16 (1.11, 1.2)	1.10 (1.05, 1.14)
Race
White	Referent	Referent	Referent	Referent
Black	0.80 (0.73, 0.89)	0.84 (0.76, 0.92)	0.69 (0.62, 0.77)	0.71 (0.63, 0.78)
Asian	1.09 (0.97, 1.22)	1.05 (0.94, 1.18)	0.90 (0.80, 1.01)	0.87 (0.77, 0.98)
Other/Unknown	0.79 (0.74, 0.85)	0.75 (0.69, 0.80)	0.76 (0.70, 0.82)	0.77 (0.71, 0.83)
Ethnicity
Not Hispanic	Referent	Referent	Referent	Referent
Hispanic	1.16 (1.08, 1.24)	1.23 (1.14, 1.33)	0.86 (0.80, 0.93)	0.92 (0.85, 1.00)
Rurality
Urban	Referent	Referent	Referent	Referent
Rural	0.91 (0.85, 0.97)	0.94 (0.88, 1.00)	0.96 (0.90, 1.01)	0.98 (0.93, 1.04)

^*Multivariable model includes all variables

^{^}Baseline for initiation, Time of observation for completion

HR=Hazard Ratio, 95% CI=95% confidence interval for Hazard Ratio, ADI=Area Deprivation Index

Association of Rurality with HPV Vaccination

Rurality was independently associated with vaccine initiation and completion for individuals living in rural areas. Individuals with rural residence had decreased probabilities of initiation compared to individuals living in urban areas, HR of 0.94 (95% CI: 0.88–1.00) (Table 3). Rurality was not significantly associated with vaccine completion.

Additional Patient Factors Associated with HPV Vaccination

Older individuals, males, and Black individuals were less likely to initiate and complete HPV vaccinations (Table 3). By contrast, persons of Hispanic ethnicity were more likely to initiate HPV vaccinations (Table 3). These associations remained statistically significant when adjusted for all other variables, including ADI quartile and rural residence.

Discussion

Widespread initiation and completion of the HPV vaccination could reduce HPV-related cancer morbidity and mortality ³⁷, improve health outcomes, and reduce medical costs ^38–40. Despite an available and effective vaccine, gaps in care continue to persist due to a lack of provider recommendations, hesitancy surrounding vaccines, and access to care ⁹. Our study builds on previous work that examined HPV vaccine uptake in rural populations ^18,38 and cervical cancer screening completion in the Midwest ²⁴. Consistent with our prior research using a novel housing-based, individual-level measure of socioeconomic status, the Housing-based Index of Socioeconomic Status (HOUSES index),^9,41 we identified a strong inverse association with area-level deprivation and HPV vaccine initiation and completion. Those living in the most deprived block groups were less likely to initiate and complete the HPV vaccination series, even after adjusting for rural status. Those living in rural areas were also less likely to initiate the HPV vaccine series; this finding is consistent with our observations using the HOUSES index⁹. These associations were adjusted for patient age, sex, race, and ethnicity. Coupled with our previous findings demonstrating decreased cancer screening among adults in deprived and rural areas, this study reinforces that suboptimal cancer prevention is independent of age in disadvantaged communities. The findings from this work suggest that widely- and publicly available area-level indicators, such as ADI, can be used to identify geographic disparities related to HPV vaccination and may help inform clinical guidelines and target community interventions. Understanding and addressing the barriers surrounding cancer prevention in underserved communities may help explain and decrease cancer morbidity and mortality reported in rural America ⁴². Findings from this work have informed the development of a planned intervention to increase HPV vaccination in clinical sites within our health system in areas of greatest deprivation and corresponding low levels of adolescent vaccination.

The results also highlight marked sex differences in vaccine uptake consistent with previous reports that females are more likely to initiate and complete HPV vaccination series than males ⁹. Initial testing, approval, marketing, and implementation of the HPV vaccine was targeted to girls, which likely contributed to the sex disparity in HPV vaccination rates ⁴³. Rates of HPV-attributable oropharyngeal cancer have increase dramatically over the past two decades, particularly among men ⁴⁴. As such, enhanced approaches to address parents of young boys regarding the HPV vaccination, perhaps through social marketing, should be prioritized ⁴⁵.

Lower rates of HPV vaccination initiation and completion were observed among those individuals aged 13 or older compared to those aged 11–12 years. Older individuals in our cohort who were either eligible to initiate or eligible to complete the HPV vaccine series obviously did not do so during when they were younger and likely represent a group less likely to accept the vaccine or perhaps less likely to have had a clinical encounter while aged 11–12 years.

Given the complex nature of social determinants of health and vaccine uptake, public health and policy efforts should take a multi-tiered approach to addressing known disparities. In addition to addressing geographic and socioeconomic disparities, interventions should also address issues surrounding misperceptions of the vaccine, lack of strong and consistent clinical recommendations ^46,47, and mistrust of the healthcare system. This can be accomplished through comprehensive initiatives promoting vaccination and system-wide efforts targeting both patients and healthcare providers and improving access to vaccines by partnering with community organizations and school systems to offer vaccines outside of the clinical encounter ⁴⁸.

Strengths and Limitations

To our knowledge, this was the first multi-site study examining the association of area deprivation and HPV vaccine initiation and completion across two Midwest states. This work was further strengthened by the granularity of the block group level data. Although ADI does not provide for an individual-level measure of socioeconomic status, our findings are consistent with those observed using an individual-level, housing-based measure of SES and ADI is a publicly available measure of SES. Our findings can be used to develop and target interventions aimed at increasing uptake of the HPV vaccination in populations and geographic regions with similar characteristics.

Due to the cross-sectional nature of this work and the use of survey data, we were unable to define causal relationships, and the ADI scores were susceptible to nonresponse and imputation bias. The population captured by the REP is similar in demographic characteristics to the population of the Upper Midwest ²⁷. However, no single population can ever represent all other populations, and we note that this population is less racially and ethnically diverse than the U.S. population. In addition, this population utilizes services from health care practices that specialize in primary to tertiary care. Study results may, therefore, differ for adolescents with different sociodemographic characteristics and access to care. As such, similar studies should be conducted in a wide range of study populations to identify social, cultural, and structural barriers to HPV vaccination in populations across the U.S.

Finally, we will have missed any HPV vaccinations that were delivered to our study population outside of the health care providers that partner to provide information to the REP. Therefore, initiation and completion rates may be underestimates of the true HPV vaccination rate in this population. However, the population in this region is stable, with >90% of the population in this age group having at least one follow-up visit within three years. Similarly, use of the REP allows capture of HPV vaccines delivered across multiple health care providers ⁴⁹, increasing the likelihood of capturing vaccines delivered by different providers or health systems. Finally, we conducted a sensitivity analysis of persons residing in a smaller, seven-county region with >90% coverage of the population. Results did not differ, suggesting that the results for the full study population are not substantially biased.

Conclusions

Area-level indicators, such as the ADI, can serve as powerful tools to identify populations with a lower likelihood of either initiating or completing HPV vaccinations and, as such, can be used to inform outreach and community engagement efforts. The findings from this work fill a critical gap in knowledge by explaining some of the increased cancer burden faced by socioeconomically disadvantaged and rural communities ²⁶.

Highlights.

• Human papillomavirus (HPV) vaccination uptake remains low in the U.S.

• Adolescents from disadvantaged areas are less likely to initiate and complete HPV vaccination

• This calls for focused attention on resource allocation and social determinants

Abbreviations:

HPV

human papillomavirus

REP

Rochester Epidemiology Project

CPT

current procedure terminology

ADI

area deprivation index

ACS

American Community Survey

RUCA

Rural Urban Commuting Area

CIs

confidence intervals

SD

standard deviation

HR

hazard ratio