Jodie Simpson: Highlights of the SAGE minutes

This is an article by Jodie Simpson (not her real name), who has gone through the minutes of the meetings held by the UK’s Scientific Advisory Group of Emergencies (SAGE).

I’ve read through the SAGE minutes, from March to the present date.  Below are selected quotes, with very little input from me.  This is just my take on it – someone else might find other elements more noteworthy, so a selection bias should be taken into account.  My italics throughout.

My personal impression from reading the minutes is that:

1. There is a noticeable shift in approach after the Imperial team release their modelling.

2. New data are mentioned – but the advice seems to continue to be based on the modelling rather than on these data, even when they appear to conflict.

3. The claim that the government is ‘following the science’ was initially true.  However, post-Lockdown, the minutes seem to show that SAGE is now fully on board with the government’s extremely cautious approach; tunnel vision is the order of the day.


3rd March

“Social distancing” for over 65s is mentioned, but not as a distance rule – instead it means staying out of circulation as much as possible.

“SAGE noted the importance of assessing the wider health implications of these interventions, e.g. the effect of self-isolation on mental health.”

“There is currently no evidence that cancelling large events would be effective.”


5th March

“There are currently no scientific grounds to move away from containment efforts in the UK.”

“Individual home isolation” of symptomatic individuals is recommended – SAGE’s modelling assumes 50% compliance.

“No evidence that banning very large gatherings would reduce transmission”; “preventing all social interaction in public spaces would have an effect… but would be very difficult to implement”.


10th March

Summary mentions social distancing, again as a general practice regarding vulnerable groups and no one else. SAGE notes that “a balance needs to be struck between interventions that theoretically have significant impacts and interventions that the public can feasibly and safely adopt in sufficient numbers over long periods”.

The three measures agreed upon for dealing with the epidemic at this point are:

1. Home isolation of symptomatic cases;

2. Whole household isolation; [HD: ie. if a household member develops symptoms]

3. Social distancing for over 70s and vulnerable groups.


13th March

“SAGE is considering further social distancing interventions… to reduce demand below NHS capacity to respond.”

“The behavioural science suggests openly explaining to the public where the greatest risks lie and what individuals can do to reduce their own risk and risk to others… Greater transparency will help people understand personal risk and enable personal agency, send useful signals about risk in general and build public trust. Citizens should be treated as rational actors, capable of taking decisions for themselves and managing personal risk.

“There is no strong evidence for public compliance rates changing during a major emergency.  There is however, a link between public anxiety and protective behavioural change.”

‘There are no strong scientific grounds to hasten or delay implementation of either household isolation or social distancing of the elderly or the vulnerable in order to manage the epidemiological curve compared to previous advice.  However, there will be some minor gains from going early and potentially useful reinforcement of the importance of taking personal action if symptomatic. Household isolation is modelled to have the biggest effect of the three interventions currently planned, but with some risks.”

“ACTION: DHSC Moral and Ethical Advisory Group (MEAG) to be invited to consider the ethical ramifications of household quarantine, given the increased risk to other residents where one resident is symptomatic.”

SAGE was unanimous that measures seeking to completely suppress spread of Covid 19 will cause a second peak.


16th March

“The risk of one person within a household passing the infection to others within the household is estimated to increase during household isolation, from 50% to 70%.”

“The objective is to avoid critical cases exceeding NHS intensive care and other respiratory support bed capacity… It is vital to understand numbers of cases regionally relative to NHS capacity, to know where local more stringent interventions might need to be introduced. The science suggests additional social distancing measures should be introduced as soon as possible.”

“Compliance with the measures by the public is key.  It is expected to take two to three weeks before the impacts of measures are observed (this needs to be monitored carefully and the appropriate metrics need to be in place).”

“While SAGE’s view remains that school closures constitutes one of the less effective single measures to reduce the epidemic peak, it may nevertheless become necessary to introduce school closures in order to push demand for critical care below NHS capacity. However, school closures could increase the risks of transmission at smaller gatherings and for more vulnerable groups as well as impacting on key workers including NHS staff. As such it was agreed that further analysis and modelling of potential school closures was required (demand/supply, and effects on spread). SAGE agreed that its advice on interventions should be based on what the NHS needs and what modelling of those interventions suggests, not on the (limited) evidence on whether the public will comply with the interventions in sufficient numbers and over time.”


17th March – Neil Fergusons’s Imperial team release their modelling, which predicts a potential 500,000 deaths in the UK.


18th March

“SAGE advises that available evidence now supports implementing school closures on a national level as soon as practicable to prevent NHS intensive care capacity being exceeded.”

“Modelling suggests that, without mitigation, London could reach Covid 19 intensive care capacity by early April.”

“SAGE considered the modelling now supports school closures on a national level and that the effect would be greatest if instituted early… SAGE discussed behavioural science considerations on school closures. With limited evidence, SAGE considered the importance of clear public messaging and of drawing on the views of teachers on keeping schools open for key workers or vulnerable groups. There is a risk that even if schools remain open for the above groups, children may not attend.”

On London: “Measures with the strongest support, in terms of effect, were closure of a) schools, b) places of leisure (restaurants, bars, entertainment and indoor public spaces) and c) indoor workplaces. Modelling is unlikely to be able to analyse the impact of these interventions with great precision. Transport measures such as restricting public transport, taxis and private hire facilities would have minimal impact on reducing transmission.”


There are no minuted meetings for four days.


23rd March – LOCKDOWN


23rd March

“Estimated Covid 19 fatalities are anticipated to overlap with those who are likely to be within the final year of their lives.  It is important to get an accurate excess deaths estimate, including potential deaths due to the measures taken…  Given the clear links between poverty and long-term ill health, health impacts associated with the economic consequences of interventions also need to be investigated.”

“SAGE noted that social distancing behaviours have been adopted by many but there is uncertainty whether they are being observed at the level required to bring the epidemic within NHS capacity. Key areas for further improvement include reducing contact with friends and family outside the household, and contact in shops and other areas… Compliance levels vary throughout the country; higher levels of compliance are being observed in London.”

Number of cases arriving from other countries are estimated to be insignificant in comparison with domestic cases, comprising approximately 0.5%.


26th March

“SAGE will consider how to minimise potential harms from the interventions, including from those arising from postponement of normal services, mental ill health and reduced ability to exercise.  It needs to consider in particular health impacts on poorer people.”

“Spare bed capacity is at roughly 20%, including in London.  Surge capacity planning for London is underway.”

“SAGE advises that there are currently conflicting data concerning potential treatments, such as chloroquine. No drug is completely safe, and it is vital not to make hasty decisions regarding treatments based on poor data.  All cases should be used in some form of clinical trial.”

“SPI-M are reviewing 2 scenarios today using a consensus model from the Imperial group: the reasonable worst case and a more optimistic scenario… SAGE advises that, of these two scenarios, the reasonable worst case is the less likely.”


29th March

“Vast majority of admissions to ICU and high dependency are aged between mid-40s and 70.  There are fewer admissions among the over 70s.  ICU care may not reflect the full burden of the disease, as now many patients are being cared for in other settings.  NHS reported that critical care bed occupancy is not yet reaching saturation levels, London included.”

“Reasonable worst case and optimistic scenario: SAGE endorsed the document under review… further work is required to understand how best to release measures and the scale of the second epidemic peak.”


31st March

SAGE noted that the trends in ICU admissions and deaths appeared consistent with a straight line increase rather than an exponential increase.

“NHS reported that critical care bed occupancy has not yet reached saturation levels, with around 1,000 beds in London, but that surge capacity was being used…”

R is estimated to be around 0.6 with an upper bound of 0.9.


2nd April

“SAGE agreed that it is unlikely before week 13th April it can start to advise whether the interventions in place are having enough of an effect.  SAGE does not currently recommend that changes be made at that point.  There is a danger that lifting measures too early could cause a second wave of exponential epidemic growth – requiring measures to be re-imposed.

“CO-CIN data is signalling nosocomial infection more strongly than previously.”

“SAGE agreed that the reasonable worst case scenario remains valid.”


7th April

There is no current evidence that transmission is accelerating; it may be slowing. ICU admission doubling times are lengthening, particularly in London, and are now at 8.8. days in London and 6.5 days elsewhere. There remains NHS capacity in London and elsewhere, with the Nightingale hospitals available in addition to this.”


9th April

ICU numbers appear to be flattening and new admissions to hospitals stabilising.  Doubling times in hospitals continue to lengthen. Calls to NHS 111 and 999 appear to have peaked and be on the decline. The epidemic may be reaching its peak, but could remain at a plateau for some time.”

“WHO has concluded that there is currently no conclusive evidence that facemasks are beneficial for community use.”


14th April

The number of deaths is plateauing, with transmission in the community highly likely to be declining. Nosocomial transmission accounts for an increasing proportion of cases.

“Relatively small changes to social distancing measures could push R back above 1 in the community. It is therefore too early to recommend releasing any measures.”

Risk of outdoor transmission is significantly lower than indoors.

“Data indicate that hospital death numbers are still high… with a possible decline in London.”

“There is a decline in hospital admissions newly confirmed with Covid 1”

Transmission in the community has slowed and it is highly likely that R in the community is less than 1… There is significant transmission in hospitals. This may have been masking the decline in cases in the community.

“Overall, the evidence that masks could prevent spread is weak, but probably marginally in favour of a small effect. If there are benefits, these are only likely in specific circumstances.”


16th April – Lockdown extended for ‘at least’ another three weeks.


16th April

“There is some regional variation in compliance with distancing measures – with London having the highest compliance and the South West of England and Wales the lowest. There appears to be a relationship between compliance levels and epidemic growth. It was noted that the epidemic entered the South West of England last.”


21st April

“NHS remains well within bed capacity…”

There is no indication that R is greater than 1 across any region, but there could still be more localised outbreaks.”


23rd April

“Public Order: SAGE noted the importance of continually considering and testing the legitimacy and equity of lockdown measures, as well as thinking about approaches to addressing this. SAGE also noted the importance of developing evaluation strategies before measures are lifted.” [SAGE underlining.]


28th April

“The ‘2 metre rule’ remains appropriate, though closer contacts of a short duration are likely to pose a very low risk.”


30th April

“Hospital admissions are declining consistently across the country.”

“SAGE noted that evidence concerning the role of children in transmission and their susceptibility to infection remains inconclusive.”

“SAGE advised that, in addition to the importance of developing a vaccine for Covid 19, a clear UK plan is required for the seasonal flu vaccine for winter 2020-2021, including consideration of whether to vaccinate the entire UK population.


1st May

“The consequences of changes in behaviour or contacts outside of schools as a result of schools reopening (such as changes to adherence to measures and to working patterns) are likely to have a larger effect on R than the effect of the schools themselves. These consequences are complex and highly uncertain. Even a short period of reopening may result in some of these occurring, which may persist even after schools close again for holidays.”

SAGE discussed the test and trace system in development. It agreed that at least 80% of contacts of an index case would need to be contacted for a system to be effective.

“There is currently insufficient evidence to determine whether the testing of index case contacts would significantly impact the epidemic compared with isolation alone (nor is it clear when to test to avoid false negatives).”

It is considered essential that this testing capability is reached before the autumn/winter flu season when a large number of those reporting symptoms may not have Covid 19.

There is also a lack of information on modes of transmission in the UK… both in hospitals in the community. A case control study is urgently needed.


5th May

R is in the range 0.5-0.9. If health and social care settings are excluded, it is likely to be at the lower end of this range… SAGE advises that, based on current data, focus should be maintained on reducing transmission in health and care settings.”

“The overall epidemic can be considered as three separate, but interacting, epidemics…”

“Better data are needed from care homes… Data suggest that urgent action should be taken in settings where it is not already underway… such as avoiding movements of patients or staff between establishments, separating people as far as is practical, and testing extensively.”

“Preliminary swabbing results indicate that a significant proportion of infections are associated with healthcare workers… in both COVID and non-COVID areas.”

“The idea of ‘bubbles’ has many merits and should be explored further. There are both positive and negative behavioural aspects to be considered.”


7th May

SAGE noted the important contribution made by Neil Ferguson over the course of the response and agreed the importance of continuing to draw upon the work of the Imperial College London team.

“A safe approach to bubbles would need to include isolation of all members of a bubble in the case of one member showing symptoms. This would lead to increased frequency of isolation for people, particularly in the winter months.

Concerns are expressed over the resilience of SAGE participants as they “continue to work under intense pressure for many more months… The need for pastoral support to be available to participants was noted.

Update: Submissions still wanted. This is especially the time to submit if you are a sciencey type who would never dream of submitting an article to a generalist magazine. Live a little, hey? But also please submit if you have a totally non-Covid piece, eg. monetary symbolism in the poetry of Ewan McTeagle. Also: Please donate! Keep me from having to eat my children (I need expensive medicine to stop the cravings).

Share this article on social media:

20 thoughts on “Jodie Simpson: Highlights of the SAGE minutes

  1. “New data are mentioned – but the advice seems to continue to be based on the modelling rather than on these data, even when they appear to conflict.”

    Agree, that’s how it appears. Before the modelling data, everything was quite sound.

    After the modelling. It really appears the model was the real world to them, and everything else was noise.

    The modelling seems to set the frame in the minds of everyone there, and I suspect still does. They are not looking at what this virus really is, with the actual data there is.

  2. I agree; the modelling data appears to have spooked them into more severe restrictions or ‘lockdown’, but probably in combination with two other things: first; reports from Italy and Spain of many more deaths, and second, mounting hysteria from some media, a plethora of stories about ‘covidiots’, Piers Morgan’s spittle-flecked rants, and that journalist from the Daily Mail asking ‘when are you going to get the police involved?’

    Actually, I might add a third: an impatience, probably from the government rather than SAGE, that existing measures weren’t showing results quickly enough.

  3. It’s interesting that there are no caveats around the modelling, and so they seem to be committing the reification fallacy.

  4. “The behavioural science suggests openly explaining to the public where the greatest risks lie and what individuals can do to release their own risk and risk to others…”

    That sounds like simple common sense.

    Here is a common sense suggestion:

    1. When “behavioural science” agrees with common sense, accept it.

    2. When “behavioural science” disagrees with common sense, go with common sense.

    They can go on chattering about “behavioural science” if they like; it does no harm and might keep some of the dimwits out of the way of the people who have work to do.

  5. Dene, I have observed such behaviour throughout my working life. (I am now retired).

    Managers and other blockheads find reality much too complicated to cope with. So they create immensely oversimplified models of reality, which leave out 99% of the important facts.

    Then they kneel down and worship the models, and treat them as if they were reality.

    We need to find some way of confining such people to appropriate institutions, instead of putting them in charge as we do at present.

  6. “ACTION: DHSC Moral and Ethical Advisory Group (MEAG) to be invited to consider the ethical ramifications of household quarantine, given the increased risk to other residents where one resident is symptomatic.”

    They might also have considered the “ethical ramifications” of destroying the economy and millions of people’s jobs and lives.

    But it seems they didn’t. Probably not “within our remit”.

  7. Seems very odd no minutes in the 4 days prior to lockdown on 23 March but were almost every other day in the week until 18 March. David Starkey interview mentioned that reports of ICU in Harrow overwhelmed also a factor in weekend before Monday 23 March.

  8. Any mention of why the UK leads the world in deaths per case, currently standing at around seven? Any examination of the NHS’s part in the treatment of cases especially the way they do nothing until hospital admission? Is it OK, or could they do better?

  9. “There is currently no evidence that …” is an entirely pernicious argument for inaction whenever, in the nature of things with a novel virus, there can be no evidence.

  10. “SAGE noted the important contribution made by Neil Ferguson over the course of the response and agreed the importance of continuing to draw upon the work of the Imperial College London team.”

    We’ve been here before. Ferguson got an OBE for his work on Foot & Mouth. Some of that work has since been called into question. I know they can only work with the data available at the time but the problem is the repeated reliance on modelling which is subsequently shown to be dubious. Humans have an unerring ability to repeat the same mistakes and reward the same people for making them. It would be funny if not so serious.

  11. Tom: “Dene, I have observed such behaviour throughout my working life. (I am now retired).”

    I too am retired, thankfully. My work wasn’t in modelling, but I did experience irrationality when things like office politics took precedence over doing what made sense. One situation in particular left me a lot more cynical about organisations and people.

  12. Not one piece of evidence that supports the latest face mask fascism .
    Agree with previous comments about computer modelling, I sometimes wonder if those sci-fi stories I read as a lad about computers being banned may yet be prophetic.

  13. Jodie, Thank you

    My highlights

    13th March
    SAGE was unanimous that measures seeking to completely suppress spread of Covid 19 will cause a second peak.

    23rd March – LOCKDOWN

    31st March
    R is estimated to be around 0.6 with an upper bound of 0.9.

    2nd April
    CO-CIN data is signalling nosocomial infection more strongly than previously

    14th April
    Transmission in the community has slowed and it is highly likely that R in the community is less than 1… There is significant transmission in hospitals. This may have been masking the decline in cases in the community

    21st April
    There is no indication that R is greater than 1 across any region

    13th March: That seemed glaringly obvious to me when lockdown being demanded; unless locked down until cure/vaccine found

    R was never above 1 in community, only where nosocomial infection rampant*. Thus Gov’t & media have been lying

    Politicians, Ferguson, ICL, SAGE, media etc need punished. Mass sackings, and 50% pay cuts and 20% tax surcharge sounds appropriate

    *However, no/little nosocomial infection rampant on military bases. One near me still open, audible firing range activity & troops in supermarket normal

    @rhoda klapp
    SImples: if Doctor thinks CV19 might be present, death due to it
    NI goes further: if CV19 in last 28 days death due to CV19 even if shot, run over by bus,

  14. Pcar, I know that, but it is not enough to think it answers the question. It might do so, but how can we know. Further, in whose interest was it to puff up the figures?

    What’s the number on nosocomial infections? Anyone know? Or how we are still getting so many new cases after so long a lockdown?

    SAGE: Makes me think of this Churchill quote:

    “You may take the most gallant sailor, the most intrepid airman, or the most audacious soldier, put them at a table together—what do you get? The sum of their fears.” —16 November 1943

  15. Have missed out points 38 to 40 of SAGE minutes of 23rd March. Prof John Anston, the HO Chief Scientific Adviser, is instructed to “lead actuarial work on establishing excess death” to “estimate numbers of deaths caused indirectly by Covid19, including those caused by the social interventions.” Some people have looked for this but it seems no one can find it. Everyone should be asking what has been the result of Prof Anston’ work and why has it not been published 11 weeks later? SAGE minutes recognised the importance of the work. Does it suggest an inconvenient truth?

  16. Pcar,
    “Figures inflated to induce fear and compliance with lockdown rules and justify lockdown, hospitals empty etc,”
    This is absolutely right. The Coronavirus Act 2020 specifically altered (eased) the method by which medics could alter the way deaths were reported.
    One only has to ask, “Why was this done?” or, “What reason can there be for doing so, other than to be able to inflate the figures in order to frighten the populace and, latterly, justify the government’s balls-up?”
    I do think this was done by the government with little or no forethought (surprise, surprise…) but the numbers then grew and got completely out of control, ending up with the 40,000 we have now.
    My other question (albeit one I assume is unanswerable) to this useless “administration” would be, “How many of that 40,000 ACTUALLY DIED OF CV19?” Not “with”, but “OF”.
    Inflating the numbers in this manner would explain why the UK death numbers appear so large. Doing it this way, by Act, also got medical staff off the hook, insofar as they could not be held responsible for misdiagnosis.
    If anyone can offer alternative reasons for government officially altering/adjusting the manner in which deaths are now reported, I’d be happy to hear them.

Comments are closed.