Queries regarding the COVID-19 vaccine effectiveness publication on 8 March 2023

FOI REF: FOI/2023/4988

You asked

Following the release of dataset 'COVID-19 vaccine effectiveness estimated using Census 2021 variables' on 8 March 2023, please state:

1. What was the total number of the population cohort that you describe as 'some people who were vaccinated but not included in the National Immunisation Management System (NIMS) data' excluding the 1,069 accounted for?

2. In the latest dataset, how were those people - not included in the NIMS data - counted within the dataset? Were the counted as unvaccinated, vaccinated or unknown?

3. Within Section 2 of the report you state: "unvaccinated (those with no vaccination or who were vaccinated with a first dose less than 21 days ago)".

This is an entirely disingenuous assumption. In the previous publication 'Death by Vaccination Status, England' the ONS confirmed that people were classed as being vaccinated from the moment the vaccine was administered. You have now gone backwards on that.

Can the ONS please provide all internal correspondence on how this decision was reached, to re-classify that cohort as 'Unvaccinated' even though they have received a vaccine. If anything, that population cohort should have been categorised under a separate category - it is certainly not appropriate to classify them as 'unvaccinated'.

4. ONS employees have been active on social media, using their private accounts to share ONS data. Please provide copies of internal ONS communication material, where employees discussed or requested authorisation to engage in social media on behalf of the ONS from their own personal accounts

We said

Thank you for your request.

Q1 and 2.

We receive a supplementary extract of NHS point of care data from NHS Digital, which includes people who died shortly after vaccination. These people would not appear in the National Immunisation Management System (NIMS) data we receive. We deduplicate this supplementary extract and link it to the NIMS dataset to identify and include the new entries that were not in the original NIMS data. If someone is in neither the NIMS dataset or the supplementary extract, they will be recorded as unvaccinated. We cannot know how many of these people were in fact vaccinated or why we do not have their record, but the vast majority of people vaccinated should be recorded routinely in NIMS.

Q3.

In response to your third question, you have referenced two different publications with different aims.

In the COVID-19 vaccine effectiveness release, we are looking at vaccine effectiveness. Therefore, we only want to look at the period of time after vaccination when the vaccine can be said to be effective, i.e. when it has had time to become more effective. People who were vaccinated with a first dose less than 21 days before were grouped with the unvaccinated to ensure we were reporting on cases where the vaccine had time to become effective.

Our Deaths by vaccination status release, reports on mortality rates rather than vaccine effectiveness. We provided the rates for unvaccinated and those vaccinated with a first dose less than 21 days ago separately, as there was interest in looking at mortality rates after vaccination particularly for those that had died where COVID-19 was not a contributing factor.

As these are different publications with different aims, no correspondence is held, as there was no such decision to reclassify.

Q4.

We do not hold internal communication where employees have discussed or requested specific authorisation for individual employees to post on social media. Our guidance is for all members of staff to abide by our Social Media Policy, available in the associated download.  A number of ONS staff are expressly authorised to post on social media in a professional capacity. In doing so they provide their expertise for public benefit in line with our statutory mission under the Statistics and Registration Service Act 2007.

All tweeting statisticians sign an agreement prior to their account being created, a copy of which is available as associated download copy of tweeting statistician - Agreement - redacted.pdf. Personal data from the documents has been redacted under Section 40 (2) of the Freedom of Information Act 2000.