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

expert reaction to new stats from the ONS on deaths involving COVID-19 by vaccination status, England: deaths occurring between 2 January and 2 July 2021

The Office for National Statistics (ONS) have released new data for the number of deaths from COVID-19 in England in relation to vaccination status.


Prof Gino Martini, Chief Scientist for the Royal Pharmaceutical Society, said:

“It is extremely reassuring to see the latest statistics around the vaccination programme in England.  The data shows that people are much better protected from the virus after two doses, and hopefully this will inspire others to get the vaccination to protect themselves and others.”


Dr James Doidge, Senior Statistician, Intensive Care National Audit & Research Centre (ICNARC); and Honorary Associate Professor, London School of Hygiene and Tropical Medicine, said:

“While this report from the ONS does provide further evidence to support vaccination, it also has some important gaps: because vaccinated and unvaccinated individuals differ, particularly in age, the only useful part of the analysis are the age-standardised mortality rates (section 4).  However, while previous sections include non-COVID-19 deaths, this section focuses exclusively on deaths involving COVID-19.  Why the sudden exclusion of non-COVID-19 deaths?  Not only does this exclude potential adverse effects of vaccination and allow for possible bias from differential attribution of deaths to involving COVID-19 or not depending on vaccination status, but it could also mask important benefits of vaccination in terms of preventing known complications of COVID-19 that may not be recorded as a “death involving COVID-19” (strokes, heart attacks, etc).

“The other very important thing to note is that age-standardised deaths represent deaths for an entire population, which of course are dominated by deaths among the elderly.  If there were a different pattern for mortality rates in younger people, it would not be observable in the analysis provided.  The important question facing us at the moment is not “what is the impact of vaccination on mortality rates for deaths involving covid-19 in the elderly” but rather “what are impacts of vaccination on all-cause mortality rates for those at lower risk from COVID-19, i.e. the young and the previously infected”.  For this, we need to see analysis of all deaths, stratified by age-group and prior infection status.”


Prof Kevin McConway, Emeritus Professor of Applied Statistics, The Open University, said:

“This is a very useful data release from ONS.  It compares deaths involving Covid-19, and deaths not involving Covid-19, in people with different vaccination statuses – defined in terms of whether they have been vaccinated at all, and if so, how many doses they had had and how long ago.  And it does this using data linked from several different sources – death registrations, vaccination records, some other health records – in order to build up an informative picture.  On deaths, it mostly doesn’t just compare numbers of deaths, but takes into account that differing age and sex splits for people with different vaccination statuses so that we can compare like with like.  ONS have huge expertise in this sort of data analysis, and in my view the resulting figures can very much be trusted.

“The results emphasises several things that were already well known.  First, vaccines are effective against death from Covid-19 – being fully vaccinated reduces the chance of a death involving Covid-19 very considerably, compared to not being vaccinated at all.  Second, the vaccines aren’t perfect – some people do still die of Covid-19 even though they are fully vaccinated.  But no vaccine is 100% effective.  Third, it’s important to have both doses.  Death rates are consistently lower in people who have had both vaccine doses than in people who have had just one, or no vaccination at all, when the differing ages of people who have had different numbers of doses are taken into account by the process of calculation age-standardised mortality rates (ASMRs).  (Those calculations do take the sex balance into account as well.)  Fourth, of the people who, sadly, did die of or with Covid-19 after two vaccine doses, a greater percentage had a weakened immune system for some previously-known reason, than for people whose death involved Covid-19 but who weren’t fully vaccinated, or than for people who died of something else without any mention of Covid-19.  What this is reminding us is that there will always be people, perhaps because they have weakened immune systems, perhaps for some other reason, that aren’t protected as well by the vaccines than the rest of us.  That’s an important reason why levels of infection need to be kept low, so that other people don’t pass on the infection to those whom the vaccines don’t protect well – and that’s another reason why it’s important for as many people as possible to be vaccinated.  Being vaccinated does reduce the chance that you will be infected at all, and if you aren’t infected, you can’t pass the infection to someone else – whether that someone else has a weakened immune system or not.

Further information

“There are a couple of things that look slightly surprising in these figures, at first glance, though I’ll explain why in fact they aren’t (in my view) matters of concern.  First, since about May, the age-standardised death rate (ASMR) for deaths involving Covid-19 for people who have had only one vaccine dose is slightly higher, in some weeks, than the corresponding ASMR for unvaccinated people.  Does this mean the vaccines are harmful?  No, for at least three reasons.  First, the numbers of deaths involving Covid-19 in unvaccinated people and in people who have had only one dose are very low in May and later, because most people in those positions are relatively young.  This means that the ASMRs cannot be estimated very precisely.  So the best that can be said about these rates is what ONS say about them, that they are ‘similar’.  But, second, there are many differences between the people who have not been vaccinated at all and those who have had just one dose, apart from their vaccination status.  The calculation of the ASMRs allows for differences in age and sex, but all the other differences haven’t been allowed for in the calculation, and any difference in death rates could be caused by these other differences between the groups and not the vaccination status at all.  Third, if there were substantial numbers of deaths that are actually caused by the vaccination, that would show up in death registrations, and they have not showed up.  This ONS release refers to a different ONS data release1 on causes of death, in which the number of deaths involving adverse effects of Covid-19 vaccines throughout the pandemic up to July 2021 was only four.

“The other slightly surprising looking figure is that, in the very early stages of vaccination, so in January and the beginning of February 2021 in the data in this release, the age-standardised death rate (ASMR) in people vaccinated once but more than 21 days after their first dose, so that the immunity provided by the vaccine would have kicked in, was higher than that in people within 21 days of their first vaccination.  (The ASMR in unvaccinated people at that time was very much higher still.)  But the difference between the ASMRs before and after 21 days from the first dose is likely to be because particularly vulnerable people would have been vaccinated earlier, because they would have had priority for the vaccine.  Therefore people dying within 21 days from their first dose in, say, late January, would on average have been healthier than people dying more than 21 days from their first dose around then.  In any case, that difference reversed after early February, with higher death rates in people less than 21 days after one vaccine dose than in people more than 21 days after their first dose.”

1 Table 12 of




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


Declared interests

Prof Gino Martini: “No interests to declare.”

Dr James Doidge: “I have no conflicts of interest to declare.”

Prof Kevin McConway: “I am a Trustee of the SMC and a member of its Advisory Committee.  I am also a member of the Public Data Advisory Group, which provides expert advice to the Cabinet Office on aspects of public understanding of data during the pandemic.  My quote above is in my capacity as an independent professional statistician.”

in this section

filter RoundUps by year

search by tag