select search filters
roundups & rapid reactions
before the headlines
Fiona fox's blog

expert reaction to preprint looking at the Pfizer-BioNTech COVID-19 vaccine efficacy in Israel

A preprint, an unpublished non-peer reviewed study, looks at Pfizer-BioNTech vaccine effectiveness against SARS-CoV-2 infection and COVID-19 in Israel.


Dr Peter English, Consultant in Communicable Disease Control, Former Editor of Vaccines in Practice Magazine, Immediate past Chair of the BMA Public Health Medicine Committee, said:

“This paper provides more excellent news about vaccination against Covid-19 disease.

“Israel has been quick off the mark with Covid vaccination, having launched a national vaccination campaign on 06 Dec 2020. This study analyses data collected up until 06 Feb 2021. The only vaccine used was the Pfizer-BioNTech (“BNT162b2”) vaccine.

“This study was concerned particularly with the efficacy of vaccination in fully vaccinated people – which they defined as ≥7 days after the second dose.

“It shows that the vaccine is highly effective at preventing Covid-19. Vaccine efficacy against all SARS-CoV-2 infections (including asymptomatic infection) was estimated to be 89.4%, rising to 93.3% for hospitalisations, 93.9% for severe illness, and – at least in the study period, 92.9% for deaths. (It is likely, however, that the efficacy for deaths is an underestimate, as it is too soon for all the Covid-19 deaths to have occurred and been recorded.)

“As with other countries, Israel started by vaccinating those most at risk (older people, healthcare workers, and long-term facilities residents), and they then worked their way down the age groups. As Covid-19 infection is thought to provide immunity for months at least, people known to have had the disease previously were considered to be at lower risk, so were not prioritised for vaccination in the initial phases of the vaccination campaign.

“Israel has an excellent health service, and a long track-record of excellent vaccination research. It has a good Identification Number (IDN) system, which allows “record-linkage” – so that illness episodes and Covid tests, can be linked to vaccination records, without unnecessary access to individuals’ personal details.

“The study benefitted from the fact that there was a high incidence of Covid-19 at the time of the study. It also coincided with a time when the more transmissible B.1.1.7 variant was predominant and increasing – so it demonstrated that the vaccine is effective against this variant. As the authors point out, they were unable to ascertain the vaccine’s efficacy against other strains, in particular the B.1.351 variant which has been causing some concern.

“The biggest shortcoming in the paper – for an international audience, at least – is that it is sketchy about Covid-19 testing. The paper states that ‘testing is free-of-charge and widely available in Israel’ – but it is not clear what triggers testing, who is tested and why, and so it is hard to make out what biases there might be on, in particular, collection of data on asymptomatic infections. (This is not necessarily a shortcoming in the study – it could be that this is well-described elsewhere, and editorial demands to keep the word count down precluded providing a more detailed description.)

“It seems likely that the vaccination campaign reduced infections overall – including asymptomatic infection; but we cannot infer too much on this from this particular study.

“Was ≥7 days after the second dose a reasonable period from which to expect full vaccine efficacy, given that we know that it takes about three weeks for the protection from a single dose to kick in? Yes, it was: the first dose is intended to “prime” the immune system; and this takes time. We know that it continues for well over 3 weeks – this is why you get a better response to the second (booster) dose after a longer interval. But the booster dose, given when the immune system has been primed, induces a much more rapid effect than the initial dose, so – although we don’t yet know this for Covid vaccines specifically – what we know about vaccination more generally (and we know a lot) tells us that we can expect to see near maximal effects by 7 days from the second dose. And, indeed, this study, as more data become available, can be expected to provide more information on this, as it included data on the timing of any infections related to vaccine doses.

“We know that vaccines tend to be most effective against the most severe forms of disease, and this study categorised infections into five categories: all confirmed infections (including asymptomatic infections); symptomatic cases; hospitalisations; severe and critical hospitalisations (including those who have died); and deaths.

“In addition to record-linkage, attempts were made to interview everybody with a confirmed infection, to ask about symptoms and illness severity, and it is a testimony to the country’s public health system that 94% of people with confirmed infection – an amazing 94,289 of 100,412 people with confirmed infections – were interviewed.

“By the end of the observation period, >5.4 million doses of vaccine had been were administered to >3.4 million persons; 27.4% of >15-year-olds, and 71.5% of >65-year-olds had been fully vaccinated. As more people were vaccinated, so the number of infections, hospital admissions, and so forth in the vaccinated age groups fell.

“The authors have written an excellent section on potential biases and other factors that might have distorted their results; but overall their findings are conservative, likely not to have overestimated the effects of vaccination. For example, one of the issues identified is that the date used for onset of disease was the date the laboratory confirmed the diagnosis. As this is likely to have been some time – maybe a few days – after the actual onset of infection, this might have misclassified a few people as having been fully vaccinated, when actually their infection will have started before the 7 days after the second dose had elapsed, before they will have benefited from the full effect of vaccination.”

Answers to questions:

Is this good quality research?  Are the conclusions backed up by solid data?

“Yes, this is very good quality research, backed up by very solid data.”

How does this compare to the recent findings in England and Scotland?

“The results are highly consistent with recent data from the UK – although the data from Scotland provide evidence for both the Pfizer-BioNTech and Astra-Zeneca vaccines, while this and the data published this week for England only provide data on the Pfizer-BioNTech vaccine.2 3 Note, also, that this study considers the impact of being fully vaccinated, whereas in the UK we do not yet have much information on this, only on the impact of a single dose.”

Have the authors accounted for confounders, such as impact of lockdown and selection bias?  Are there important limitations to be aware of?

“There is an excellent section on this in the paper. As an observational study, it is impossible to control for these factors entirely; but the authors have done a good job in discussing potential biases.”

What are the implications in the real world?  Is there any overspeculation?

“The conclusions seem to be quite conservative, and not to be over-optimistic. They genuinely provide excellent news: that the likelihood of being severely ill, hospitalised, or dying after being fully vaccinated is about 10% what it would have been if you hadn’t been vaccinated.”


  1. Haas EJ, Angulo FJ, McLaughlin JM, Anis E, Singer SR, Khan F, et al. BNT162b2 effectiveness against SARS-CoV-2 infections and COVID-19, Israel [pre-print]. Pre-print 2021.
  2. Public Health England. First real-world UK data show Pfizer-BioNTech vaccine provides high levels of protection against COVID-19 from the first dose. London: Public Health England, 2021 (22 Feb); 1-3.
  3. Public Health Scotland. Vaccine linked to reduction in risk of COVID-19 admissions to hospitals. Press release 2021; Updated 22 Feb 2021; Accessed: 2021 (22 Feb).



Preprint (not a paper): ‘BNT162b2 effectiveness against SARS-CoV-2 infections and COVID-19, Israel’ by Haas et al is expected to be published on a preprint server shortly. This work is not peer-reviewed.



All our previous output on this subject can be seen at this weblink:



Declared interests

Dr Peter English: “No interests to declare.”

in this section

filter RoundUps by year

search by tag