Characterization of the Differential Adverse Event Rates by Race/Ethnicity Groups for HPV Vaccine by Integrating Data From Different Sources

Jing Huang, Jingcheng Du, Rui Duan, Xinyuan Zhang, Cui Tao, Yong ChenFrontiers in Pharmacology 2018 · DOI

Abstract: Monitoring vaccine safety through the Vaccine Adverse Event Reporting System (VAERS) is often hindered by a lack of individual-level demographic data due to privacy protections. This study introduces a novel statistical framework that integrates VAERS data with U.S. Census and CDC survey data to infer racial and ethnic disparities in adverse event (AE) reporting rates. By leveraging geographic variation across U.S. states, the researchers identified nine specific AEs for the HPV vaccine with significant reporting differences between non-Hispanic White females and other groups, offering a privacy-preserving method for enhanced pharmacovigilance.

Introduction

The Vaccine Adverse Event Reporting System (VAERS) is a cornerstone of post-market vaccine safety monitoring in the United States. However, to protect patient privacy, individual-level demographics such as race and ethnicity are often omitted or limited. This creates a significant gap in understanding how vaccine safety profiles might differ across diverse populations—a critical concern given that immune responses can vary based on genetic and ethnic backgrounds. This study sought to bridge this gap by integrating disparate data sources to detect population-level associations.

Methodology: Data Integration and Modeling

The researchers developed a generalized linear mixed effects model that uses U.S. states as the primary unit of analysis. By correlating state-level demographic distributions with state-level vaccine adverse event reporting rates, the model can infer associations that are missing at the individual level.

The study integrated four primary datasets from 2010 to 2016:

  1. VAERS: Reports of serious AEs following HPV vaccination.
  2. CDC NIS-Teen: Survey data providing state-level HPV vaccine coverage rates.
  3. U.S. Census Bureau: Annual population estimates for demographic distribution.
  4. MedDRA: Standardized coding for 2,864 unique adverse event "Preferred Terms."
Figure 1: State-level visualization of demographic distribution (A), vaccine coverage (B), and VAERS reporting rates (C) used to facilitate data integration.

Because serious adverse events are rare, the resulting dataset was highly sparse (99.2% zeros). To address this, the team employed a Zero-Inflated Poisson (ZIP) regression model, which specifically accounts for the high frequency of zero-report data points that would otherwise skew standard regression results.

Key Results: Identifying Disparities

After analyzing nearly 3,000 unique adverse event terms and applying a rigorous Bonferroni correction for multiple testing, the study identified nine adverse events with significantly different reporting rates between non-Hispanic White females and other race/ethnicity groups.

Figure 2: P-values for the differential reporting rates of nearly 3,000 Preferred Terms (PTs). Points above the red line indicate statistical significance.

Significant findings included:

  • Higher reporting rates in non-Hispanic White females: Vulvovaginal pain, vulvovaginal rash, lichen sclerosus (potentially related to autoimmune responses), ECG changes, and tension headaches.
  • Lower reporting rates in non-Hispanic White females (Higher in other groups): Fecal incontinence, absent pulse (potentially related to cardiovascular issues), bedridden status, and abnormal respiration.

Discussion and Limitations

This geographic proxy method demonstrates a powerful way to conduct sensitive demographic research without compromising individual privacy. However, the authors note several limitations. The "ecological fallacy" remains a risk; population-level associations do not necessarily dictate individual outcomes. Furthermore, the model assumes constant reporting rates across groups, which may not account for cultural or socioeconomic differences in how symptoms are reported.

Future work may involve applying this framework to other vaccines, such as influenza, and refining models to better account for potential reporting biases between different demographic groups.

Summary updated

Abstract

Data from the Vaccine Adverse Event Reporting System (VAERS) contain spontaneously reported adverse events (AEs) from the public. It has been a major data source for detecting AEs and monitoring vaccine safety. As one major limitation of spontaneous surveillance systems, the VAERS reports by themselves sometimes do not provide enough information to answer certain research questions. For example, patient level demographics are very limited in VAERS due to the protection of patient privacy, such that investigation of differential AE rates across race/ethnicity groups cannot be conducted using VAERS data only. For many vaccines, racial and ethnical difference in immune responses has been found in studies based on racially diverse cohorts. It is of great interest to characterize the differential AE rates by race and ethnicity groups for vaccines. In this study, we propose a novel statistical method to integrate VAERS data with data from other resources for vaccine pharmacovigilance research. Specifically, we integrate VAERS data with CDC survey data of vaccine coverage and U.S. census data of race/ethnicity distribution to quantify differential AE rates by race/ethnicity groups for HPV vaccine. We utilize the difference of race/ethnicity distributions across U.S. states to investigate the association between AE reporting rate and race/ethnicity groups at the population level. We identify 9 AEs with significantly different reporting rates between non-Hispanic White females and other race/ethnicity groups.
BibTeX
@inproceedings{huang2018characterization,
  author = {Jing Huang and Jingcheng Du and Rui Duan and Xinyuan Zhang and Cui Tao and Yong Chen},
  title = {{Characterization of the Differential Adverse Event Rates by Race/Ethnicity Groups for HPV Vaccine by Integrating Data From Different Sources}},
  booktitle = {Frontiers in Pharmacology},
  year = {2018},
  doi = {10.3389/fphar.2018.00539},
}
Figure