Sage memo makes the case for lockdown

23 December 2021

4:55 AM

23 December 2021

4:55 AM

On Monday, Covid restrictions were rejected after the cabinet debated the issue robustly for the first time since the pandemic started. The Prime Minister said he’d revisit the decision, so the debate is very much still ongoing.

But it wasn’t just ministers meeting that day. Sage assembled its experts as well, with over 70 scientists and government officials in attendance. The minutes, seen by The Spectator, give an interesting summary of the official case for more lockdown restrictions.

Everyone is wrestling with two questions; if there are no more restrictions, how far will Omicron case numbers rise? And how will that translate into hospitalisations? If there is reason to believe that the NHS will be overwhelmed, this makes the case for more restrictions now. But if Omicron starts to level off and is so mild as to mean fewer hospitalisations than the earlier peak, as has been the case in South Africa, then restrictions could risk inflicting needless economic and social harm.

There is another twist to this. Sage is not the only outfit modelling Covid for the UK: there are now sophisticated independent forecasters as well. One is JP Morgan, whose extensive Covid research is being passed like samizdat papers between cabinet members who fear that Sage has a negative bias. ‘We anticipate the efficacy of boosters against hospitalisations holding up (with 90 per cent of over 65s already boosted) such that hospitalisations remain below 1,500 per day,’ JP Morgan says (the January peak was almost 4,000). In this scenario, British boosters save the day and Omicron is rebuffed by a wall of vaccinated people.

But the Sage memo paints a very different picture: of a vulnerable nation where more restrictions are needed for everyone quite quickly. Here are its main points:

1. Sage accepts that Omicron may have peaked in South Africa

In public, Sage officials talk about a ‘doubling time’ of Omicron to suggest that, without restrictions, its rise will be exponential. But it accepts that the virus is already levelling off in South Africa, the country with the most experience of Omicron which has not locked down. However, Sage says that assumptions about UK hospital admissions cannot be made from South Africa’s experience.

The number of infections and hospitalisation in Gauteng appears to be declining. The reasons for this are not clear and it cannot be assumed that this will be sustained. Nor can it be assumed that the wave in the UK will follow a similar pattern, given the different populations and epidemiological situations.

2. Sage says milder scenarios should ‘not be assumed for planning purposes’

The JP Morgan scenario had Omicron daily hospitalisations at 1,500. The PCCF model developed by Prof Philip Thomas (which predicted the third wave) suggests a peak of 2,800. Both envisage no more restrictions. But no such scenario appears in the Sage papers.

Indeed, ministers are explicitly told to forget about such scenarios ‘for planning purposes’. The latest Sage minutes suggest that a ‘range of values’ have been used to create scenarios for Omicron’s trajectory but that Omicron would need to be at least 90 per cent less severe than Delta to stop the NHS from being inundated.

In the absence of further interventions or significant behaviour change, intrinsic severity would need to be greatly reduced (by around 90 per cent) for hospitalisations to not reach the levels of previous peaks unless the wave peaks early for other reasons, which should not be assumed for planning purposes.

The 90 per cent figure appears to reference the University of Warwick’s modelling, which provides scenarios for 10, 20, 50, and 100 per cent severity, compared with Delta. Real-world data (the likes of which are not just coming in from South Africa) shows that Omicron is 80 per cent less likely to hospitalise; data from Denmark says 60 per cent. They are the only two countries to have worked this out, and it is not yet known what emphasis can be put on variant severity, vaccines at work and natural immunity. But neither findings are mentioned in the Sage minutes. Instead there is simply a reference to ‘uncertainty’.

The final sentence – the optimistic scenarios and other factors that could influence the extent of admissions ‘should not be assumed for planning purposes’ – unnerves ministers who want to know: why not? What are they being given: a range of plausible projections or just a selection of worse-case ‘scenarios’? Neither Sage documents nor the accompanying interpretation makes this clear.

3. Sage say acting now stops longer measures later

The Sage document calls for ‘earlier interventions’ as well as a ‘short intervention… if introduced early enough’ to flatten the Omicron curve.

It continues to be the case that the earlier interventions happen the greater the effect they will have. Even a short intervention could reduce both peak and total admissions, particularly if introduced early enough. The main benefit of a short intervention would be in flattening the peak of admissions, and to allow more people to receive boosters. If measures are implemented later, when hospital admissions have risen significantly, measures may need to be in place for longer and may be too late to avert a period with very high admissions.

4. Sage rejects the idea of more limited, ‘targeted measures’

Ministers are told that shielding and ‘targeted’ measures are not ‘effective’ compared to ‘population-wide’ measures. This suggests Sage leans more towards a circuit-breaker (a euphemism for lockdown) than less stringent interventions:

Protective measures are important for the individual, and this is particularly the case for those who are at higher risk. However, measures targeted at more at-risk groups only (e.g. ‘shielding’) are not an effective substitute for population-wide measures if the aim is to reduce overall hospitalisation rates (high confidence).

5. Sage cites ‘mental health’ as a reason to reject targeted measures

The document appears to look at one of the shielding options (that has been raised before by government): asking care home staff to shield, but not placing the same requirement on the rest of the population.

The group of people affected would be wider than just those the policy aimed to protect. For example, those coming into contact with residents (such as care home staff and visitors) may also need to reduce their contacts to prevent incursions of infections into the care home network. Reducing contacts may have significant negative impacts on the mental health and wellbeing of those asked to do so (high confidence). In addition, those who have previously been asked to shield may be reluctant to take similar steps again, reducing the effectiveness of such measures. If individuals are asked to reduce contacts more than the wider population, policies and messaging will need to consider potential reluctance and mental health impacts.

6. Omicron is spreading within the NHS

Hospitalisations with Omicron are increasing but the numbers remain uncertain. Pillar 1 testing data in Manchester indicate a doubling time of around 2–3 days for patients and staff testing positive for Omicron infection. This suggests that Omicron is getting into hospitals at a similar rate to its spread in the community.

This certainly reflects data from London hospitals.

7. What Sage does not say: length of hospital stay

As each day passes, new data on Omicron comes rolling in but it’s striking how little of the more encouraging news is referenced. Take, for example, the length of stay in hospital. Sage has this to say:

Hospital occupancy will be affected by length of stay as well as admissions. Changes to treatment plans e.g., treating more patients at home, or increased use of antivirals, may affect admissions and occupancy. Occupancy scales approximately linearly with length of stay.

In a ‘significant early finding’ South Africa found hospital stays with Omicron were closer to 2.8 days than 8.5 days, so this is a major variable which could have a game-changing effect on whether the NHS can cope. Again, ministers are given no hint of this in the Sage briefing note. Length of hospital stay was one of the variables that allowed JP Morgan to critique the LSHTM modelling and come up with a less alarming scenario where no more restrictions would be needed.

The document is phrased as if Sage is collectively making the case for lockdown: in reality Sage is a huge number of experts most of whom wish to provide advice but not lobby government for anything. But over the pandemic, Sage advice has tended to be edited and summarised by those inside government lobbying for more restrictions.

Nor are Sage attendees the decision-makers who ultimately decide whether or not we lock down. This morning, when talking about lockdown, Prof Graham Medley, chair of the Sage forecasters, said: ‘I am not making the decisions.’

It’s an important point: Prof Medley is an academic asked to model various scenarios by government officials, mindful of how risk lies on both sides. He has done more than most to open a window on how the whole process works. The academics on Sage are not responsible for how their advice is presented to ministers.

It’s cabinet that ultimately makes the call. The question they must ask is if they are being given the full picture by Sage, how to balance the risks involved and how many still-important questions need an answer before any decision to lock down again can be taken.

Got something to add? Join the discussion and comment below.

Show comments