The Cabinet Office

A response to Steerpike’s review of Michael Gove’s lockdown claims

A response to Steerpike's review of Michael Gove’s lockdown claims
(Getty images)
Text settings

Steerpike ran the rule over Michael Gove's article in the Times in which he argued ‘lockdown was the only way to stop the NHS being broken’. Here, the Cabinet Office responds to Steerpike:

1. Claim: Steerpike claims Michael Gove, the Chancellor of the Duchy of Lancaster, used his essay in the Times on 28 November to set out 'his new argument: the only strategy for stopping the NHS from being overwhelmed now is the government's new tougher three-tier system'.

Rebuttal: the argument that social distancing and non-pharmaceutical interventions are necessary for protecting the NHS is not a new argument. It has been at the heart of the UK Government’s approach since spring and is an approach shared by all the devolved administrations in the UK. Governments across Europe have taken similar measures to protect their health service.

2. Claim: Steerpike argues that Gove insists that 'cases were doubling just before lockdown — only there is plenty of data to suggest that cases were actually coming down'.

Rebuttal: R was estimated as above 1 in every region of the UK when, on 30 October, the Government took this decision to introduce new national restrictions in England. This means the epidemic was growing exponentially in every region of the UK. ONS figures showed that the infection was doubling on average every 15 days in the six weeks up to the announcement of the November measures (31 October). We also saw that doubling times in the previously lower prevalence areas like the East of England and Midlands were as low as 12-14 days over that period. The Government accepted at the time that the Local Covid Alert Levels, announced on 12 October, were having an effect on reducing transmission. The fact that R was estimated as above 1 in all regions of England meant that it was considered necessary to go further and prevent these areas of lower prevalence becoming areas of high prevalence.

3. Claim: Steerpike claims 'ministers were hit not by ‘data coming in from the frontline’ but by out-of-date ‘scenarios’ that had been drawn up weeks ago, then sprung on the cabinet without warning'.

Rebuttal: Steerpike presents only a small proportion of relevant data that was presented to Ministers, who consider a range of data, with daily monitoring of cases, hospitalisations, deaths and other figures. The data from the 'frontline' was stark: on 30 October the increase over the last two weeks was significant. The latest ONS survey available at that point, covering the week from 17 to 23 October, showed 1 in 100 people across England were infected, up from 1 in 160 in the survey released a fortnight prior; there were 11,248 Covid-19 patients in hospital in the UK, up from 6,054 on 16 October; and sadly the UK had recorded a further 3,324 deaths over this period, which meant in the last two weeks of October approximately the same number of people had died as in the preceding three-and-a-half months. 

There was new data, however, and this was the projections from the NHS, based on the latest SPI-M modelling, suggesting that were infections to continue growing at the current rate, Covid-19 demand would outstrip NHS capacity in England in early December. This analysis was presented by the UK Government’s Chief Scientific Adviser at the 31 October press conference, for which all of the underlying data has already been made available. This was also the first week in which the Office for National Statistics presented information suggesting that the percentage of the population testing positive for COVID-19 was above one per cent.

4. Claim: Steerpike claims that 'since the middle of November, actual deaths have been beneath the lowest bound of the forecast.' He argues that this was not due to lockdown as 'it takes three to four weeks for policy changes to show in COVID-19 fatality figures' and suggests the data was largely 'pre-lockdown infections. Things were nowhere near as bad as ministers - and the country – had been led to believe'.

Rebuttal: The short-term (six week ahead) SPI-M projections and NHS analysis showed that Covid-19 demand could exceed NHS capacity, even that which could be freed up from cancelling electives. Deaths continued to track above these projections in early November, showing that the position was indeed perilous. 

Steerpike conflates these short term projections with the illustrative winter scenarios, which projected four months into the future. The claim was not that the UK was exactly on one of the October scenarios at the time. Instead the argument was that there was the potential for very large numbers of deaths had the Government not acted to bring the growth of the virus under control. This has nothing to do with the NHS capacity projections which were based on short-term, up-to-date, high confidence modelling.

We were already exceeding the Reasonable Worst Case Scenario produced by SPI-M in the summer (which at the time had been widely dismissed). It is of course welcome that some of the most stringent measures in the October iteration of tiers had some effect on transmission, but that effect was not uniform, and at the time there was exponential growth in all regions of the country, with some of the lowest prevalence areas seeing the fastest doubling times. We had also seen some temporary flattening of rates in places like Spain and France, where cases then sadly began to rise again.

5. Claim: Steerpike argues that flattening hospital numbers show that the old tiers were working: 'if it takes three weeks for policy changes to affect hospital numbers, and they peaked at about 11 November (as government data currently suggests), that would suggest the old system of more restrained tiers was working'.

Rebuttal: The NHS estimates that it takes 7-10 days for policy changes to affect hospital numbers. This would imply that hospitalisations on 11 November were from infections caught 1 to 4 November – i.e. after the England-wide restrictions were announced on 30 October. 

In any event some changes in behaviour would be expected immediately following the announcement, even before the measures had legal force, due to people adapting their behaviour. The same dynamic was observed in March. It is also the case that in itself NHS admissions peaking does not mean the epidemic has peaked; variations in the age-profile of people who have the virus will vary the number of NHS admissions.

High rates of incidence should not be a source of complacency. In the latest ONS survey, which covered the week ending 23 October, and which was the latest available to Ministers at the time of the decision to move to England-wide national restrictions, it was estimated that 1 in 100 people were infected with Covid-19 (up from 1 in 470, on 24 September). This worsened further, with the ONS subsequently estimating in its release on 6 November that, in the week running up to 31 October, 1 in 90 people in England had Covid-19. In the most recent survey, the proportion stands at 1 in 85, having reached 1 in 80 in the week ending 14 November. It therefore remains above the level of late October. Restrictions therefore remain a regrettable necessity. 

The Government’s approach, informed by expert advice, has been to introduce new, tougher local restrictions which can control the spread of infection. Weaker restrictions, with such a high level of prevalence, would have left the NHS vulnerable to small increases in R to above 1.

6. Claim: Steepike claims in a discussion of a metaphor about NHS capacity and hospital admissions that 'by the time lockdown came, the bath was probably emptying'.

Rebuttal: SPI-M estimated R as above 1 in every region of England and nation of the UK when the UK Government took the decision to introduce a second, limited set of England-wide restrictions. When R is above 1 this means the epidemic is growing exponentially. The SAGE estimate of R has only just fallen with confidence to below 1 in the update published 27 November, and Covid-19 hospitalisations remain at around 16,000 versus the 11 September low of 740. 

In addition, it is also only recently that the ONS infection survey has shown the signs of a levelling in the infection rate, and still at a very high level. The bath is therefore only now definitively emptying albeit slowly. But – to continue the metaphor – it remains precariously full.

7. Claim: Steerpike claims that Ministers would have known that the epidemic was already slowing based on case rates.

Rebuttal: The Government has always said that the previous tier system slowed the growth of the epidemic. But they did not cause it to shrink. Without action, growth would have continued, and the NHS would have been placed under intolerable pressure. 

In the most up to date SPI-M/SAGE consensus statement published prior to the decision to implement the England-wide national restrictions (published on 27 October), R was estimated to be above 1 in every region of the country, i.e the epidemic was growing exponentially. As was set out at the time, the level of pressure on the NHS left little room for error particularly in certain regions of the country. Case rates are not a robust measure of the growth of the epidemic since they are affected by testing volumes. These volumes were in decline from 26 October to 2 November which would have affected cases recorded. For that reason the R rate is judged to be the more reliable estimate of epidemic growth. It is only in the wake of national England-wide measures, on the estimate published 28 November, that SAGE have judged R to be below 1.

8. Claim: Steerpike claims that 'the actual truths' show that 'the old tier system was effective enough'.

Rebuttal: This is not true. As the Government set out in its analysis document yesterday, and as detailed in the paper by the Task and Finish Group on the autumn interventions presented to SAGE last week, our best analysis of the Local Covid Alert Levels introduced on 12 October shows that:

  • many Lower Tier Local Authorities (LTLAs) in Tier 1 still had positive growth rates. Indeed, the regions with the lowest levels of infection were the places where the virus was growing most rapidly;
  • in Tier 2 the epidemic in most LTLAs was growing more slowly than before the interventions and was shrinking in many, but many local epidemics were still growing; and
  • and that in Tier 3, epidemics in all LTLAs had a lower growth rate than before tiers were introduced and most, but not all, were declining.

Despite the effectiveness of the England-wide restrictions, we needed further decisive action to suppress R to below 1 and will need to continue to drive down prevalence from its still-high level. With the old tiers not providing us with confidence that they would do the job everywhere that they were employed, the Government has moved to strengthen them.

9. Claim: Steerpike suggests that infection rates in Liverpool were falling for a month before mass testing.

Rebuttal: No one has claimed otherwise. Michael Gove was clear in the Times article that the strength of the tier helped reduce numbers in Liverpool: In Liverpool, the mayor Joe Anderson bravely adopted measures above and beyond the old basic Tier 3 and championed mass testing. The result: falling infections.

10. Claim: Steepike claimed that 'contrary to what Gove claims', Sweden has placed hardly any 'restrictions on its population'.

Rebuttal: Regrettably, Sweden is seeing a rise in new cases and deaths. As Steerpike acknowledges, Sweden has introduced a rule of eight for public gatherings. Although Sweden has been presented by some as exemplifying a more laissez-faire approach, anyone who organises an event that violates the new cap can face a fine or prison sentence of a maximum of six months. 

While many restrictions have been introduced in guidance, the Government has supplemented this with other stricter interventions. There is a ban on non-essential travel to Sweden from countries outside Europe. Alcohol cannot be served after 22:00. All food and drink has to be served at tables. 

While Sweden may not have a 'tiers' system analogous to those in place in many other countries, many regions have gone further than the advice proposed at a national level. In fact, all but one of the regions have now introduced tighter guidance following recommendations from their Public Health Agency, including advice against visiting indoor environments such as shopping centres and gyms, and to avoid physical contact with people outside their households.