Abstract
Introduction Latino community members have increased risk for human papillomavirus (HPV)-related cancers. HPV vaccination may be influenced by neighborhood-level social determinants of health and nativity. We examined differences in HPV vaccine coverage by social deprivation index (SDI) and nativity among Latino patients receiving care in community health centers (CHCs).
Methods We used electronic health record data for Latino patients aged 9-26 who received care in a multistate CHC network from 2012-2022 (n=212,611) to examine differences in HPV vaccine initiation and completion by SDI score tertile and nativity status, and by country of birth among those with country documented. Logistic regression was used to estimate covariate-adjusted odds ratios and prevalences.
Results While those living in higher social deprivation areas had higher prevalence of vaccine coverage, this was not consistent across all groups. Initiation was highest among foreign-born Latinos living in areas with medium deprivation (52.7%, 95% CI: 47.6-57.8%), but was similar for all US-born and foreign-born Latinos, and by SDI tertile. Completion was highest among US-born Latinos living in the areas with the highest deprivation (40.6%, 95% CI: 37.8-43.3%).
Conclusion In this multistate CHC sample of Latino patients, we observed more vaccination among those with a documented country of birth compared to those without that data, with low variation by neighborhood SDI and by country of birth. The lack of difference by SDI could be due to the CHC setting and suggests that interventions to increase HPV vaccination targeting Latinos overall, rather than by neighborhood social determinants, may be most effective.
- Cancer
- Community Health Centers
- Health Disparities
- Hispanic or Latino
- Human Papillomavirus Viruses
- Immunization
- Logistic Regression
- Papillomavirus Infections
- Papillomavirus Vaccines
- Social Determinants of Health
Introduction
Human papillomavirus (HPV) vaccination, which is recommended for all adolescents and young adults1 may be influenced by neighborhood-level social determinants of health, but existing research has not demonstrated consistent associations between neighborhood factors and HPV vaccination.2–5 Factors such as transportation, distance to clinics can be barriers to HPV vaccine access and uptake, as patients may have a difficult time getting to the clinics.2,6,7 Educational attainment and cost of vaccine or access to care due to economic reasons can also affect uptake, as patients may be unaware of or unable to afford vaccination.8,9 While some research has found neighborhood-level deprivation to be associated with lower vaccine series completion,2 other studies have found no difference by community or neighborhood once individual factors are taken into account,3 or no association at all between neighborhood factors and HPV vaccination.2 Specifically, more vaccination among those living in areas with lower income has been observed, potentially due to public health insurance programs, and the Center for Disease Control and Prevention’s Section 317 Grants Program which provide full coverage of recommended vaccines.4,5,10,11 Knowing whether or not there is an association between HPV vaccination and neighborhood-level deprivation will help direct resourcing and outreach efforts or reinforce existing policies that promote screening, and focusing on other individual factors or social risks that may be associated with HPV vaccination. This may help providers and healthcare systems better tailor care to the patients and communities for which they care.
Latinos are at higher risk for HPV-related cancers than non-Latino Whites12,13 thus HPV vaccination is a crucial cancer preventive service for this population. Individual, social, and cultural factors, as well as immigration itself can contribute to this risk, but lower access to screening and treatment is likely the primary driver for this disparity.12,14,15 Cervical and oral cancer are also prevalent in Latin America, affecting Latino immigrants to the US. Latinos may also experience more social risks (e.g., economic insecurity, housing needs) than other groups.16,17 Therefore, understanding how neighborhood social factors are associated with HPV vaccine coverage is imperative among Latinos. Research on HPV vaccination in general in Latinos is mixed: some report less HPV vaccination among Latinos,18 other findings have shown comparable or higher HPV vaccination.19 Language and nativity (i.e., patient-reported country of birth), are also relevant factors, since Latinos are not homogeneous with respect to country of birth and preferred language.20 Some research using survey data has found differences in awareness of the HPV vaccination by foreign-born status,21 and variable vaccine coverage has been observed among different Latino subgroups in the US, including differences by parental language preference.22
Community-based health centers (CHCs) are clinics where people can access healthcare regardless of ability to pay for it.23 CHCs serve a disproportionately high number of Latinos, and strive to provide equitable and comprehensive health care,24 often enhanced by offering services beyond clinical care such as wellness programs, social services, or legal services, and by employing people who live in the communities served.25,26 Given these factors, CHCs are key to studying Latino health disparities in HPV vaccination.
As aforementioned, Latinos face higher risk of social deprivation and HPV-related cancer. Therefore, we examined whether HPV vaccine coverage varied by neighborhood deprivation, nativity (as Latinos are not a homogeneous group, country of origin is known to affect the risk of certain cancers, and immigration can be a barrier to care), and language preference in a multistate US CHC network. This study expands beyond previous studies by using a large national sample of objective health record data, from a data network with a substantially greater absolute number of people with country of birth data than others of which we are aware. Findings may help clinics and clinicians to best communicate with patients and design local strategies for vaccine optimization.
Methods
Data and Inclusion Criteria
Data were sourced from the OCHIN Accelerating Data Value Across a National Community Health Center Network (ADVANCE) Epic electronic health record (EHR) system, specifically from 1,169 clinics located in 25 US states. ADVANCE partners serve Federally Qualified Health Centers, public health departments, and CHCs delivering care to people who are publicly insured, uninsured, or otherwise medically underserved. ADVANCE integrates outpatient EHR data and geocoded community-level data into a comprehensive common data model.27 The study sample consisted of Latino patients aged 9-26 years who received care during the study period (01/01/2012-09/14/2022). This age range was selected because these patients were eligible for the vaccine through the entire study period according to Advisory Committee on Immunization Practices (ACIP) guidelines.28 Clinics were selected based on their specialty, and included those specializing in primary care/family practice, pediatrics, public health, and dental care (dental clinics may provide HPV vaccination for oral cancer prevention29). School-based health centers were also included. This study was approved by the Oregon Health & Science University Institutional Review Board. Data in this study are collected during routine care and patient consent to use data for research is obtained from clinics during care initiation.
Outcome Variables
The outcomes were documented receipt of HPV vaccine coverage: 1) vaccine initiation: having at least one dose of the vaccine during or prior to the study period, beginning at age 9; and 2) vaccine completion: ≥2 doses during or prior to the study period for those who had the first vaccine under 15 years of age, and ≥3 doses for those who initiated vaccination at age 15 or older, according to ACIP guidelines.30
Independent Variables
The main independent variables included a set of indicators denoting nativity (i.e., foreign-born status) and if country of birth was not collected for patients, we disaggregated those by preferred language leading to the following mutually exclusive groups: (1) US-born, (2) foreign-born, (3) unknown country of birth, English-preferring and (4) unknown country of birth, Spanish-preferring. We disaggregated by preferred language among those without nativity data as a proxy for acculturation, as it is often used in research when nativity or other related data are unavailable.31–34 In secondary analyses, for a subset of patients who had country of birth documented in the EHR, we disaggregated the foreign-born group further by those born in the US, the Dominican Republic, El Salvador, Guatemala, Honduras, Mexico, or born in another Caribbean, Central, or South American country (including from Argentina, Belize, Bolivia, Brazil, Chile, Colombia, Costa Rica, Cuba, Ecuador, Guyana, Nicaragua, Panama, Peru, Suriname, Uruguay, and Venezuela), as these included sufficient sample size for regression analyses.
The other main independent variable was neighborhood-level social deprivation index (SDI) score. The SDI is a composite measure of multiple social determinants of health, (e.g., income, education level, transportation access) and is used to measure cumulative socioeconomic variation in geographic areas.35,36 SDI score was obtained from American Community Survey 2019 and 2020 estimates, and linked by patient address (of longest duration during the study period) at the census-tract level. We categorized the SDI score, which ranges from 1-100, into tertiles37 to allow comparison of vaccine coverage at low, medium, and high levels of neighborhood deprivation, first based on SDI scores from the sample of all patients (low deprivation neighborhoods: SDI score of 1-71, medium deprivation neighborhoods: 72-92, and high deprivation neighborhoods: 93-100), and again using SDI scores from only the sample of Latinos who had a country of birth in the EHR (low: 1-79, medium: 80-94, high: 95-100), to reflect the levels of deprivation most accurate to each sample. Each were used in their respective models.
Patient-level covariates included age category at first clinic visit within the study period (9-14 years, 15-17, and 18-26), sex (male/female), number of clinic visits per year (<1, 1-2, 3-4, ≥5), insurance over the study period (never insured, some public, some private, mixture of public and private), household income as a percent of the federal poverty level over the study period (FPL; always ≥138% FPL, always <138%, above and below 138%, or never documented to account for missing data), ever had contact with a community health worker (trusted source of HPV vaccine information9,38), and US region of the patients’ most frequented clinic (Midwest, Northeast, South, West; in the model with countries, the categories were Northeast and Other due to limited sample size in the other categories).
Statistical Analysis
Descriptive statistics were conducted to characterize patient demographics. For both the primary and secondary analyses, we conducted generalized estimating equations logistic regression models for each of the two binary outcomes (vaccine initiation and completion). These two logistic generalized estimating equation (GEE) models included the indicators for nativity groups, and included the covariates listed above and interaction effects between nativity groups and the categorical SDI to obtain covariate-adjusted odds ratios (OR) and corresponding 95% confidence intervals (CI). All GEE models used a robust sandwich variance estimator to account for patient clustering within the patients’ most frequented clinic and assumed exchangeable working correlation structure. Covariate-adjusted prevalences (predicted probabilities, derived from GEE parameter estimates) with 95% CIs are reported. Analyses were conducted in Stata version 15 and statistical significance was set at p-value<0.05.
Results
The sample includes 212,611 Latino patients aged 9-26 who received care in OCHIN clinics from 2012-2022 (Table 1, Appendix Table 1). About 11% of Latinos in the sample had a place of birth recorded (US-born Latinos: N=15,134; foreign-born Latinos N=8,133). Among those with no place of birth recorded, we observe a similar proportion of those that prefer English and Spanish (Spanish-preferring N=93,798; English-preferring N=95,546). The foreign-born groups included 2,649 patients born in Mexico, 1,768 born in Guatemala, 1,155 born in El Salvador, 917 born in the Dominican Republic, 902 born in Honduras, and 742 born in other Caribbean, Central, or South American countries.
Among the full sample, the unadjusted prevalence of those living in areas with the most deprivation (the highest tertile, with a score of 93-100) ranged from 28.1% among English-preferring Latinos with no country of birth documented, to 46.6% among foreign-born Latinos. Within the subsample of Latinos with a documented country of birth, the unadjusted prevalence of those in areas with the most deprivation (score of 95-100) ranged from 21.0% for those from other Caribbean, Central, or South American countries to 47.5% for those from the Dominican Republic.
The unadjusted prevalence of the outcomes of HPV vaccine initiation and completion were highest in Spanish-preferring Latinos without a country of birth documented in the EHR (initiation: 59.6%; completion: 43.0%), and among those from the Dominican Republic (initiation: 64.1%; completion: 39.8%). (Table 1, Appendix Table 1).
Figure 1a shows the adjusted prevalence of HPV vaccine initiation by SDI tertile among all Latino patients by language/nativity group. The prevalence was highest among foreign-born Latinos living in areas with medium deprivation (52.7%, 95% CI: 47.6-57.8%), but was similar for all US-born and foreign-born Latinos, and by SDI tertile. Latinos without a country of birth documented in the EHR had lower adjusted prevalence of initiation, with English-preferring patients living in the areas with the least deprivation having the lowest prevalence (44.0%, 95% CI: 42.4-45.6%) (Figure 1a, Table 2). Figure 1b shows the adjusted prevalence of HPV vaccine completion by SDI tertile among all Latinos. The prevalence of completion was highest among US-born Latinos living in the areas with the highest deprivation (40.6%, 95% CI: 37.8-43.3%), and lowest among Spanish-preferring Latinos living in the areas with the least deprivation (31.1%, 95% CI: 29.7-32.6%) (Figure 1b, Table 2).
Note: Generalized estimation equations logistic regression models adjusted for age, sex, clinic visits per year, insurance, income, ever having contact with a community health worker, US region. Abbreviations: No Country Documented (NCD), Social Deprivation Index (SDI)
Figure 2a shows the adjusted prevalence of HPV vaccine initiation by SDI tertile for Latinos with a country of birth documented in the EHR. The highest adjusted prevalence was among those born in El Salvador, living in the areas with the most deprivation (57.5%, 95% CI: 53.4-61.6%), and the lowest was in those born in the Dominican Republic (44.4%, 95% CI: 38.6-50.1%) (Figure 2a, Table 3). Figure 2b shows the adjusted prevalence of HPV vaccine completion by SDI tertile. The adjusted prevalence of completion was highest again among those born in El Salvador living in the areas with the most deprivation (39.1%, 95% CI: 35.0-43.2%), and lowest among those born in Guatemala living in neighborhoods with the least deprivation (24.3%, 95% CI: 19.4-29.2%) (Figure 2b, Table 3). Overall, while those living in areas with more deprivation had higher vaccine coverage, especially among the total sample, (Figures 1 & 2), this was not consistent across all groups.
Note: Generalized estimation equations logistic regression models adjusted for age, sex, clinic visits per year, insurance, income, ever having contact with a community health worker, US region.
Abbreviations: Dominican Republic (Dom. Rep.), Social Deprivation Index (SDI)
Discussion
This study used clinical data on nativity status, country of birth, and neighborhood-level social deprivation to examine the differences in HPV vaccine coverage among young Latinos receiving care in a multistate network of CHCs, which includes a large number of people with country of birth data. US- and foreign-born Latinos had a higher prevalence of vaccine initiation than those without a country of birth documented, with US-born Latinos also demonstrating a higher prevalence of vaccine completion. Those living in neighborhoods with more social deprivation had slightly higher prevalence of vaccination, however this was not seen in every group and was often only different by a couple percentage points. Among the subsample who had country of birth documented in the EHR, the prevalence of initiation and completion varied slightly, and SDI did not moderate the association between nativity and HPV outcomes.
The overall prevalence of HPV outcomes was suboptimal and below national prevalences,10 as well as guideline targets, which recommend HPV vaccination for all children and young adults.1,39 As the prevalence of initiation ranged from only 43.5% to nearly 60%, efforts are needed to improve HPV vaccine coverage for Latinos in CHCs in general. Given the persistent burden of HPV related cancers in Latinos,12 this vaccine is a pressing priority for cancer control.
Barriers to vaccination among Latino adolescents and their parents have previously been found to be related to social determinants of health, such as lack of access due to lack of transportation, educational attainment, and cost of vaccine or access to care due to economic reasons.40–43 Community health workers and healthcare providers working in US-Mexico border towns reported barriers to HPV knowledge among their patients were often financial or related to lack of education, or patriarchal attitudes (machismo) from husbands,9 and among immigrant farmworkers, vaccine barriers were related to insurance and access to vaccines.44 Many CHCs have programs in place to help with specific social risks, or other non-clinical factors.25 The comprehensive and equitable care models used in the clinics in this study may help attenuate the disparities seen in other settings, which could help explain the lack of differences by SDI score in our findings, which are contrary to some other research where social deprivation as associated with HPV vaccine coverage.2 The lack of differences by neighborhood suggest that interventions to broadly increase HPV vaccination for all Latinos, rather than by targeting neighborhood social determinants, may be most effective in tailoring care to patients and communities.
While differences by SDI category were not observed, differences between Latino groups (such as those with a country of birth documented, and those without) were seen. Those with country of birth documented at all had higher vaccine initiation compared to those with no country of birth documented. US-born Latinos had higher prevalence of completion compared to those without country of birth documented and the aggregated foreign-born group. Some research has found differences in awareness of the HPV vaccination by foreign-born status,21 and similar to our study, variable vaccine coverage has been observed among different Latino subgroups in the US, including differences by parental language preference.22 Previous research has also shown better health outcomes (e.g., fewer health insurance disparities, better mental health) among third generation vs first and second generation low-income Latino children.45,46 The higher prevalence of vaccination among those with country of birth documented could also be an artifact of the data collection (i.e., these patients had better historical data in general). Understanding the association of nativity data collection and nativity remains a pressing priority in health services research in Latinos.
Limitations
This paper has limitations, including that all data are from CHCs, so these results may not be generalizable to every health care setting, or even every area-level analysis in the US. As many of the patients in the clinics in this network are in low-income households, even patients in the category with the least amount of neighborhood deprivation still have a high amount of deprivation, as seen in the breakdown of SDI categories (with SDI scores in the 70s out of 100 in the lowest group). However, this setting is important, given the large amount of care Latinos receive at CHCs,24 and remains a vital setting in which to study, understand, and mitigate disparities. Future work can compare this to broader, more socioeconomically diverse care settings. Additionally, most of the sample did not have a country of birth documented, so we used a proxy (language), to try to disaggregate our data further, as we know not all Latinos have the same culture influences and experiences. Furthermore, immigration status is not included in the EHR. However, documentation of country of birth itself may be a useful care quality and access indicator. In 2019, HPV vaccination was also recommended up to age 45.28 However, the sample size of those over age 26 who were vaccinated in this time period was too small for analyses. Future work could examine HPV vaccination in the 27-45 age group. We sought to do state adjustments but given the data distribution, we were unable to include state as fixed effects. We included US region partially account for regional variability, as states in the same US regions tend to be more similar to each other with regard to health policies. We also did not have data on vaccine cost, but did include insurance in the modeling as coverage of vaccines can differ by insurance type. Furthermore, medical records do not contain information about parent perspectives, which for HPV vaccination specifically is a limitation, as there may be barriers related to beliefs about sexual activity, though these beliefs are often mitigated with more information and education about the vaccine.47
Conclusion
In this sample of Latino adolescents and young adults receiving care in CHCs, the lack of association between HPV vaccine prevalence and SDI may be a testament to the comprehensive and equitable care models of CHCs, as these community-based clinics may already have strategies in place to address social determinants of health. However, the overall vaccine prevalence is not as high as it could be, so there is still opportunity to address HPV vaccination in Latinos overall. This study suggests that tailoring interventions to increase the overall HPV vaccine rate broadly targeting the entire Latino population, rather than by neighborhood characteristics, may be most effective.
Conflicts of Interest
The authors have no conflicting or competing interests.
Corresponding Author
Jennifer A Lucas PhD, Department of Family Medicine, Oregon Health & Science University, lucasje{at}ohsu.edu
This article was externally peer reviewed.
Acknowledgements
The research reported in this work was powered by PCORnet®. PCORnet has been developed with funding from the Patient-Centered Outcomes Research Institute® (PCORI®) and conducted with the Accelerating Data Value Across a National Community Health Center Network (ADVANCE) Clinical Research Network (CRN). ADVANCE is a Clinical Research Network in PCORnet® led by OCHIN in partnership with Fenway Health, Health Choice Network, Oregon Health & Science University, and University of Washington. ADVANCE’s participation in PCORnet® is funded through the PCORI Award RI-OCHIN-01-MC.
Appendix
- Received for publication July 18, 2025.
- Accepted for publication October 13, 2025.








