Impact of COVID-19 vaccination on preventing long COVID: a population-based cohort study using linked NHS data

Date of Approval
Application Number
Technical Summary

While many people fully recover after COVID-19, a substantial proportion continues to suffer from long-term
complications. While vaccines have shown impressive efficacy to prevent severe COVID-19, their impact on long COVID is not investigated in ongoing trials.

1) To test if the proposed observational analyses adequately account for confounding.
2) To evaluate the impact of covid vaccination on the prevention of long COVID.
3) To evaluate the comparative effectiveness of covid vaccines to prevent long COVID.

Data: NHS records from CPRD GOLD and AURUM, linked to HES inpatient data, Index of Multiple Deprivation and publicly available regional infection/vaccination rates.
Participants: All adults registered in CPRD GOLD/AURUM for >180 days

a) To characterise long COVID in CPRD based on a positive test/clinical diagnosis with persistent symptoms
for >28 days. Relevant symptoms will be defined based on a systematic literature review.
b) To emulate the null effect of vaccines to prevent COVID-19 in the 10 days after first dose. To test this, 4 staggered
cohort studies will be conducted, with eligibility criteria based on vaccination priority groups. Cohort-specific
propensity score (PS) will be estimated, and PS-weighted Poisson models fitted to calculate rate ratios (RR) for
COVID-19 in the 10 days after first dose. Negative control outcomes will be used to calibrate for residual

To estimate the effect of vaccines on the development of long COVID (as defined in WP1a) by comparing vaccinated vs. unvaccinated persons using the study design “validated” in WP1b and cox regression. Sensitivity analyses considering e.g. different long COVID definitions will be conducted.

To study the comparative effectiveness of Oxford/AstraZeneca vs BioNTech/Pfizer vaccine to prevent long COVID using the methods in WP1b/WP2.

Understanding the impact vaccination has on preventing long COVID will inform future risk-benefit evaluations for COVID vaccines.

Health Outcomes to be Measured

WP1A: Diagnoses/symptoms characterising long COVID, which were recorded >28days after a positive test/clinical diagnosis of COVID-19 in CPRD;
WP1B: A first positive test/clinical diagnosis of COVID-19 recorded in the first 10 days after the first dose of a covid vaccine;
WP2 and WP3: Long COVID;


Daniel Prieto-Alhambra - Chief Investigator - University of Oxford
Annika Jodicke - Corresponding Applicant - University of Oxford
Albert Prats Uribe - Collaborator - University of Oxford
Antonella Delmestri - Collaborator - University of Oxford
Edward Burn - Collaborator - Oxford University Hospitals
Junqing Xie - Collaborator - University of Oxford
Kristin Kostka - Collaborator - University of Oxford
Martí Català Sabaté - Collaborator - University of Oxford
Trishna Rathod-Mistry - Collaborator - University of Oxford
Victoria Y Strauss - Collaborator - University of Oxford
Xintong Li - Collaborator - University of Oxford


HES Admitted Patient Care;Patient Level Index of Multiple Deprivation