Vaccine Effectiveness Among 5- to 17-year-old Individuals with Prior SARS-CoV-2 Infection: An EHR-Based Target Trial Emulation Study from the RECOVER Project

Jiajie Chen, Yuqing Lei, Qiong Wu, Ting Zhou, Bingyu Zhang, Michael J. Becich, Yuriy Bisyuk, Saul Blecker, Elizabeth A. Chrischilles, Dimitri Christakis, Lindsay G. Cowell, Mollie Cummins, Soledad Fernández, Daniel Fort, Sonia Pérez González, Sharon J. Herring, Benjamin D. Horne, Carol R. Horowitz, Mei Liu, Susan Kim, Parsa Mirhaji, Abu Saleh Mohammad Mosa, Jennifer A. Muszynski, Catharine I. Paules, Alice I. Sato, Hayden T. Schwenk, Soumitra SenGupta, Srinivasan Suresh, Bradley Taylor, David A. Williams, Yongqun He, Jeffrey S. Morris, Ravi Jhaveri, Christopher R. Forrest, Yong ChenmedRxiv 2025 · DOI

TL;DR

This study demonstrates that COVID-19 vaccination provides a significant "hybrid immunity" boost to children and adolescents who have already had a prior infection, but this protection is highly dependent on the viral variant. While vaccines were over 60% effective at preventing reinfection during the early Omicron (BA.1/2) era, that effectiveness plummeted to statistically insignificant levels as the XBB subvariants became dominant.

What Problem Does This Paper Address?

While the effectiveness of COVID-19 vaccines in "immunologically naive" populations (those never infected) is well-documented, the landscape has shifted. A vast majority of children and adolescents now have "natural immunity" from prior infections.

The researchers sought to answer a critical public health question: Does vaccinating a child who has already had COVID-19 provide meaningful additional protection against reinfection? Furthermore, how does this protection hold up against the rapid-fire succession of Omicron subvariants (BA.1/2, BA.4/5, and XBB)? Because randomized controlled trials for every new variant are impractical, the team turned to real-world data to find answers.

Methods

The researchers employed a Target Trial Emulation (TTE) design. This is a sophisticated statistical framework used to extract "trial-like" insights from observational data—in this case, Electronic Health Records (EHR).

Key technical features included:

  • Sequential 7-Day Trials: Instead of one long study, they "started" a new mini-trial every seven days. This allowed them to account for the exact timing of vaccination and changing variant prevalence.
  • Matching: To ensure the "vaccinated" and "unvaccinated" groups were comparable, they used a combination of exact matching (on gender, race, and prior vaccine doses) and propensity score matching (on age, obesity, and chronic conditions).
  • Cox Proportional Hazards Models: These were used to estimate the "hazard ratio" of reinfection, which was then converted into a Vaccine Effectiveness (VE) percentage.
  • Negative Control Experiments: To check for "hidden" bias (like health-seeking behavior), they ran the same analysis on 40 outcomes unrelated to COVID-19 (e.g., bone fractures) to ensure the vaccine didn't "appear" to prevent things it shouldn't.

Datasets

The study utilized data from the RECOVER (Researching COVID to Enhance Recovery) Initiative, a massive NIH-funded project.

  • Source: EHR data from 37 U.S. children’s hospitals and health institutions.
  • Standardization: Data was mapped to the OMOP Common Data Model (CDM), which allows researchers to analyze data from different hospital systems using a unified language.
  • Scale: The analysis included hundreds of thousands of participants across three variant-dominant periods.

Key Insights and Ideas

  • The "Hybrid" Advantage is Waning: The study confirms that "hybrid immunity" (infection + vaccine) was a powerhouse against early Omicron variants. However, the virus is evolving faster than the immunity can keep up, specifically with the XBB lineage.
  • The Age Gap: Younger children (5–11) generally showed higher and more durable protection than adolescents (12–17). The authors suggest this might be because adolescents were eligible for vaccines much earlier, meaning their "baseline" immunity may have waned more by the time they were reinfected.
  • The "Medically Detectable" Filter: Because the study relied on EHRs, it primarily captured infections that were serious enough to warrant a doctor's visit or a formal test. This means the VE reported is specifically for documented reinfection, not necessarily every sniffle caught at home.

Key Results

The study broke down Vaccine Effectiveness (VE) by age and variant period:

Variant Period Age Group Vaccine Effectiveness (95% CI)
BA.1/2 Children (5-11) 62.3% (38.3% – 77.0%)
Adolescents (12-17) 64.5% (32.4% – 81.3%)
BA.4/5 Children (5-11) 57.1% (24.6% – 75.6%)
Adolescents (12-17) 36.5% (Not Statistically Significant)
XBB & Later Children (5-11) 22.5% (Not Statistically Significant)
Adolescents (12-17) 34.5% (Not Statistically Significant)

The takeaway: Vaccination was highly effective against BA.1/2, moderately effective for younger kids against BA.4/5, and failed to show significant protection against reinfection in the XBB era for either group.

Limitations and Future Work

  • Underreporting: The study likely missed many mild or asymptomatic cases diagnosed via home rapid tests that were never entered into a medical record. This could lead to an overestimation of vaccine effectiveness if vaccinated families are less likely to seek formal testing.
  • Low Uptake: Vaccination rates among children with prior infections were notably low (under 17% in the BA.1/2 cohort and dropping to ~7% in later cohorts). This small sample size for the "vaccinated" group makes it harder to achieve statistical significance for newer variants.
  • Future Directions: The authors emphasize the need for ongoing surveillance as new variants emerge and suggest that future research should focus on the long-term effectiveness of updated (bivalent or variant-specific) boosters in pediatric populations.

Summary updated

Abstract

COVID-19 vaccination provides significant protection against reinfection for children and adolescents with prior infections during the early and mid-Omicron periods. This study also highlights the importance of addressing low vaccination rates in pediatric populations to enhance protection against emerging variants.
BibTeX
@misc{chen2025vaccine,
  author = {Jiajie Chen and Yuqing Lei and Qiong Wu and Ting Zhou and Bingyu Zhang and Michael J. Becich and Yuriy Bisyuk and Saul Blecker and Elizabeth A. Chrischilles and Dimitri Christakis and Lindsay G. Cowell and Mollie Cummins and Soledad Fernández and Daniel Fort and Sonia Pérez González and Sharon J. Herring and Benjamin D. Horne and Carol R. Horowitz and Mei Liu and Susan Kim and Parsa Mirhaji and Abu Saleh Mohammad Mosa and Jennifer A. Muszynski and Catharine I. Paules and Alice I. Sato and Hayden T. Schwenk and Soumitra SenGupta and Srinivasan Suresh and Bradley Taylor and David A. Williams and Yongqun He and Jeffrey S. Morris and Ravi Jhaveri and Christopher R. Forrest and Yong Chen},
  title = {{Vaccine Effectiveness Among 5- to 17-year-old Individuals with Prior SARS-CoV-2 Infection: An EHR-Based Target Trial Emulation Study from the RECOVER Project}},
  year = {2025},
  doi = {10.1101/2025.02.07.25321814},
  howpublished = {medRxiv},
}
Figure