Quantifying and Adjusting for Confounding From Health-Seeking Behavior and Health Care Access in Observational Research

  1. Sophie Graham
  2. Jemma L Walker
  3. Nick Andrews
  4. William J Hulme
  5. Dorothea Nitsch
  6. Edward P K Parker
  7. Helen I McDonald
Read the full article See related articles

Discuss this preprint

Start a discussion What are Sciety discussions?

Listed in

This article is not in any list yet, why not save it to one of your lists.
Log in to save this article

Abstract

Background

Health-seeking behavior and health care access (HSB/HCA) are recognized confounders in many observational studies but are not directly measurable in electronic health records. We used proxy markers of HSB/HCA to quantify and adjust for confounding in observational studies of influenza and COVID-19 vaccine effectiveness (VE).

Methods

This cohort study used primary care data prelinked to secondary care and death data in England. We included individuals aged ≥66 years on 1 September 2019 and assessed influenza VE in the 2019–2020 season and early COVID-19 VE (December 2020–March 2021). VE was estimated with sequential adjustment for demographics, comorbidities, and 14 markers of HSB/HCA. Influenza vaccination in the 2019–2020 season was also considered a negative control exposure against COVID-19 before COVID-19 vaccine rollout.

Results

We included 1 991 284, 1 796 667, and 1 946 943 individuals in the influenza, COVID-19, and negative control exposure populations, respectively. Markers of HSB/HCA were positively correlated with influenza and COVID-19 vaccine uptake. For influenza, adjusting for HSB/HCA markers in addition to demographics and comorbidities increased VE against influenza-like illness from −1.5% (95% CI, −3.2% to .1%) to 7.1% (95% CI, 5.4%–8.7%) with a less apparent trend for more severe outcomes. For COVID-19, adjusting for HSB/HCA markers did not change VE estimates against infection or severe disease (eg, 2 doses of BNT162b2 against infection: 82.8% [95% CI, 78.4%–86.3%] to 83.1% [95% CI, 78.7%–86.5%]). Adjusting for HSB/HCA markers removed bias in the negative control exposure analysis (−7.5% [95% CI, −10.6% to −4.5%] vs −2.1% [95% CI, −6.0% to 1.7%] before vs after adjusting for HSB/HCA markers).

Conclusions

Markers of HSB/HCA can be used to quantify and account for confounding in observational vaccine studies.

Article activity feed

  1. Version published to 10.1093/ofid/ofae598
  2. Version published to 10.1101/2024.02.27.24303434 on medRxiv