OpenSAFELY study shows the effectiveness of Pfizer COVID booster doses

In a recent study posted to the medRxiv* pre-print server, researchers assessed the effectiveness of the Pfizer-BioNTech BNT162b2 coronavirus disease 2019 (COVID-19) vaccine in the seven million adults in the United Kingdom (UK) eligible for booster shots between September 16 and December 16 2021.

Study: Effectiveness of BNT162b2 booster doses in England : an observational study in OpenSAFELY-TPP.  Image Credit: Jo Panuwat D/ShutterstockStudy: Effectiveness of BNT162b2 booster doses in England : an observational study in OpenSAFELY-TPP. Image Credit: Jo Panuwat D/Shutterstock


As part of its national COVID-19 vaccination programme, the UK initially prioritized its high-risk population to receive booster shots in September 2021 before making it available to the general adult population. However, following the emergence of the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) variant of concern (VOC) Omicron, the availability of boosters was reduced to three months on December 8, 2021. Previously, boosters were available after six months from when an individual received their second COVID-19 vaccine dose.

About the study

In the current study, researchers estimated the effectiveness of booster shots of the BNT162b2 vaccine in adults who received two doses of vaccine between September 16 and December 16, 2021. The team used data from the OpenSAFELY-TPP database for the study analysis. The database contained information on 40% of UK GP practices linked to national data on coronavirus surveillance, hospitalization and death records. The study’s follow-up lasted 10 weeks or December 31, 2021, with the researchers comparing each booster recipient to a matched unboosted control subject based on booster priority and prior vaccination status.

The team assigned a matched control to each study participant based on the National Health Service (NHS) region, the start date of the study, the vaccine brand, and the date of the second vaccine dose. A Cox regression model adjusted for these factors to estimate hazard ratios (HRs), where 1-HR, expressed as a percentage, indicated the effectiveness of the BNT162b2 booster. The HR estimates compared the boosted to unboosted people, globally and individually, for days 1 to 28 and 29 to 70. In addition, the team estimated HRs for days one to seven, eight to 14, 15 to 28, 29 to 42, and 43 to 70. The study had several outcomes: a positive SARS-CoV-2 test, COVID-19-related death, COVID-19 hospitalization, and death unrelated to COVID-19.

Study findings

The study population consisted of 6,990,219 individuals who were matched as controls and then rematched in the booster group. As expected, the matching factors and the proportion of people with pre-existing morbidities were comparable between the study groups. Although the standardized mean difference between the two groups was consistently less than 0.1, subjects in the control group had more learning disabilities, a lower SARS-CoV-2 test frequency, and higher rates of serious mental illness and deprivation. The screening of 23,151,145 person weeks of follow-up yielded 123,378 SARS-CoV-2 positive tests, 3,672 COVID-19 hospitalizations, 588 COVID-19 deaths, and 6,990 non-COVID-19 deaths.

The boosted individuals showed an anomalous increase in the overall incidence of positive SARS-CoV-2 tests approximately seven days after receiving booster injections; however, they did not show a similar pattern for other severe COVID-19 outcomes. In addition, the differences in the overall incidence of positive SARS-CoV-2 tests between the study groups were apparent within the first few days.

Accordingly, the 10-week risk of a positive SARS-CoV-2 test in the boosted vs. unboosted group 47.3 and 84.0 per 1000 respectively, corresponding to a risk difference of 36.8 per 1000. The risk difference for hospitalization with COVID-19, Covid-19 death and non-COVID-19 death were 3.6, 1 and 8.1 respectively.

The (overall) effectiveness of the booster for days one to seven was 50.7%, 80.1%, 88.5% and 80.3% for SARS-CoV-2 positive test, COVID-19 hospitalization, COVID -19 mortality and non-COVID-19 mortality. In addition, booster effectiveness against all COVID-19 outcomes was generally lower during days one to 28 than between days 29 to 70. Within shorter time periods, booster effectiveness against positive SARS-CoV-2 tests appeared to decrease from about six weeks after booster vaccination, 68.2% vs. 45.3% between days 15 to 28 and days 43 to 70.


Overall, a booster injection of BNT162b2 reduced the number of positive SARS-CoV-2 tests by ~50% during the first 10 weeks after booster administration. In addition, the booster shots reduced rates of COVID-19 hospitalization, COVID-19 death, and non-COVID-19 death. More importantly, the booster’s effectiveness remained comparable against all severe COVID-19 outcomes, regardless of the brand of vaccine used for primary vaccination and prior infection.

In contrast, the estimated effectiveness of the booster was lower in subjects younger than 65 years of age and in clinically vulnerable populations. Since the current study showed evidence of decreasing effectiveness of boosters against positive SARS-CoV-2 tests, the authors emphasized the need to monitor the effectiveness of booster vaccines over time.

*Important announcement

medRxiv publishes preliminary scientific reports that have not been peer-reviewed and therefore should not be considered conclusive, that should guide clinical practice/health-related behavior or be treated as established information.


  • Effectiveness of BNT162b2 booster doses in England: an observational study in OpenSAFELY-TPP, William J Hulme, Elizabeth J Williamson, Elsie Horne, Amelia CA Green, Linda Nab, Ruth Keogh, Edward PK Parker, Venexia M Walker, Tom M Palmer, Helen J Curtis, Milan Wiedemann, Christine Cunningham, Alex J Walker, Louis Fisher, Brian MacKenna, Christopher T Rentsch, Anna Schultze, Krishnan Bhaskaran, John Tazare, Laurie A Tomlinson, Helen I McDonald, Caroline E Morton, Richard Croker, Colm D Andrews, Lisa EM Hopcroft, Robin Y Park, Jon Massey, Amir Mehrkar, Jessica Morley, Sebastian CJ Bacon, David Evans, Peter Inglesby, George Hickman, Simon Davy, Iaim Dillingham, Tom Ward, Viyaasan Mahalingasivam, Bang Zheng, Ian J Douglas, Stephen JW Evans, Christopher Bates, Jonathan AC Sterne, Miguel A Hernan, Ben Goldacre, medRxiv pre-print 2022, DOI:

Leave a Comment

Your email address will not be published.