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

expert reaction to SAGE documents published last night summarising the effectiveness and harms of different non-pharmaceutical interventions

The Scientific Advisory Group for Emergencies (SAGE) have released the latest batch of minute sand documents. 


Prof Paul Hunter, Professor in Medicine, UEA, said:

“I have read the ‘Summary of the effectiveness and harms of different non-pharmaceutical interventions, 16 September 2020’: and its associated table.

“These documents seek to summarise the evidence for and against the different non-pharmaceutical interventions for their likely impact on controlling COVID-19 and any potential harm.

“From the document it is very difficult to get an understanding of what the strength of evidence is behind each of the possible interventions being presented.  For several of these interventions the conclusions drawn are not that surprising and are likely to have already achieved consensus in the wider scientific and medical communities.  Indeed all such interventions are likely to have some effect but no single one alone is likely to be sufficient.  One issue that does need more consideration and thought is the short duration lockdown “circuit-breaker”.  Whilst there is no doubt that should one be implemented it will have an impact, at least in the short-term, its longer-term value is much less clear.  Part of the problem is that back in March/April case numbers increased much more rapidly as we approached the peak then they declined after the peak.  So, a short one or two week circuit breaker may delay any subsequent peak but would it be sufficient to suppress the eventual size of the peak?  From the information being shared in the SAGE document it is far from clear how strong the evidence is for this intervention.  What were the assumptions behind any modelling in favour of this option?

“Whilst much of the comments and conclusions presented in this document are not that surprising or controversial it is not generally clear what was the scientific evidence base underpinning many of these recommendations.  In medicine we have come to realise that all that we do has to be based on evidence of its effectiveness and safety, this is known as evidence-based medicine.  Evidence based medicine relies not only on what the scientific evidence says but whether the quality of that evidence is strong or weak.  In general, we work to an agreed hierarchy of evidence that puts systematic reviews of good quality randomised controlled trials at the top and expert opinion at the bottom.  In this framework modelling studies do not generally feature as they are themselves based on assumptions for which there may be strong or poor evidence.  After all experts are only human themselves and different experts can interpret and represent scientific evidence in different ways – this doesn’t mean they are biased people, but scientific bias is a recognised part of assessing evidence and that is why multiple people look at evidence.  Even when conclusions are based on consensus amongst groups of scientists we don’t always know whether there was unanimous agreement among the group or whether different individuals thought differently or carried more weight.

“In 2007 in The Lancet, Oxman and colleagues wrote a review of the way that the World Health Organization formulated its recommendations (Oxman AD, Lavis JN, Fretheim A. Use of evidence in WHO recommendations. The Lancet. 2007 Jun 2;369(9576):1883-9.)  The authors were rightly very critical of WHO because they found that much of its policy development at the time was based on consensus within expert committees and that systematic reviews and concise summaries of findings were rarely used for developing recommendations.  WHO dramatically reformed its policy making and guideline development processes in the light of this article and now most new guidelines have a much stronger scientific underpinning.  These will usually include a much clearer statement of the strength of evidence behind each statement.

“Many of these issues are even more problematic during an epidemic like COVID-19 where there are still large areas of uncertainty of the science.  In such a case, it is even more important to be able to distinguish between advice that is the opinion of a small group of experts from that which is based on consensus of the scientific and medical community as a whole from that which is based on high quality scientific evidence adequately brought together in a well conducted systematic review.  That does not mean that we can always wait for high quality scientific evidence to become available.  When faced with such as threat as COVID-19 we need to act on best available evidence even if that evidence is of poor quality.  But we should always make it clear the strength of evidence otherwise we run the risk of uncritically accepting poor quality solutions when other better options may be available.”


Prof Robert Dingwall, Professor of Sociology, Nottingham Trent University, said:

“Predictably, many readers have seized on particular passages in the two SAGE documents to justify their particular standpoints.  Looking at them as a whole, however, what is striking is the degree of caution about the quality of evidence and likely benefit of almost any intervention.  Harms are acknowledged but it is accepted that evidence for them has not been systematically collected, which makes it difficult to balance them against benefits.  Face coverings are a good example, where the benefits are assessed as minimal, especially outdoors, relative to the harms, so that the main justification appears to be for maintaining population levels of commitment, which some of us might translate as population levels of fear and anxiety.  The ethics of this are not questioned.  From the outside we can’t know whether SAGE were asked to come up with recommendations despite the degree of confidence in the strength of evidence, and likely effectiveness, for some of the interventions being low.

“Perhaps the most telling section, though, is the long list of areas where research is needed and what it says about the lack of investment in evaluating NPIs relative to evaluating therapies or vaccines.  With a novel virus, social and behavioural interventions are the first line of defence in preventing transmission and buying time for biomedical science to catch up.  This was a lesson that should have been learned from the experience with HIV or even with the more recent Ebola outbreak in West Africa, where effective NPIs brought transmission down well before any effective medical intervention was available.  Indeed, this is the case with most epidemic diseases, as Thomas McKeown pointed out in the 1960s.  It was acceptable for the initial response in February/March to be based to some extent on inference from pre-existing theories and evidence from other infections.  Six months on, we should have a much more solid evidence base, with contributions from a range of social sciences on the social, economic and psychological impacts of the interventions that have been tried and a reasonable sense of what works, why it works and how it might inform ideas about other possible measures to try out in the future.”


Prof James Naismith FRS FRSE FMedSci, Director of the Rosalind Franklin Institute, and University of Oxford, said:

“We are not stuck in a loop but the winter will be difficult.

“Science and medicine are moving at record speed, there will be a vaccine, there will be highly effective medicines and already discoveries have lowered the death rate from infection.

“Scientists advise and in a democracy it is politicians, as our elected representatives, who decide.

“What is being decided now is how best to reduce harm between now and spring of 2021.

“SAGE is doing its job analysing the data and giving advice.

“SAGE has analysed the test track and isolate system, stating it has had at best a ‘marginal’ impact.  Since the minutes were taken, the performance of track trace isolate has not improved.

“It is re-assuring that SAGE are clear eyed about this and not misleading themselves about the effectiveness of the current system as a means to control viral spread.  It is for politicians, not SAGE, to decide how much money should continue to be invested in the system and what if anything needs done about it.

“SAGE’s advice was that a nationwide lockdown was more likely to work in terms of reducing viral spread.

“The government has opted to wait and see if less stringent measures can avoid a severe second wave.

“Of course, we all hope that the current measures will be enough.  As SAGE clearly states, lockdowns and social restrictions cause serious harm.

“There is a risk that we will end up having to lock down again (perhaps with a different name but in effect the same thing).  If we do so the duration of lock down will be likely be longer as a result of delay.

“Despite wishing otherwise, there are only difficult choices and competing risks.”


Dr James Gill, Honorary Clinical Lecturer, Warwick Medical School, said:

“Looking over the SAGE documents, it is important to recognise that data analysis and expert opinions are the guiding factors in their recommendations.

“The team outlines that making some choices, such as keeping schools open are possible, but necessitate harder choices in response – such as activating the two week “circuit breaker” lock down to curb case rises.

“In many ways it’s simple maths:  If Choice A allows for increased in Covid 19 transmission, then it should be paired with Action 1, to attempt to mitigate that risk.

“It does make sense – schools are a potential mixing pot for both people and virus.  So stopping households mixing in response would appear to have a direct effect of mitigating any school induced spread.

“Unfortunately the politics of these decisions is difficult, and thus action slow, almost in direct contrast to the SAGE documents discussing need for rapid intervention.  The team acknowledges harms will occur with any of their advised actions, but stress their recommendations balance the risk of COVID19 fighting measures against their consequential health and social harms.

“Discussions about Covid and the expected winter impact have been on going for most of the year.  The latest SAGE documents double down their previous advice for caution and strengthening of COVID interventions, which only appears to have been given lip service in the latest government measures.  I hope that the SAGE recommendations are rapidly given the appropriate weighting in further government decision making, as currently, their scientific base suggestions appear regrettably diminished by comparison to the politics.”



‘Summary of the effectiveness and harms of different non-pharmaceutical interventions, 16 September 2020’:

‘NPIs table, 17 September 2020’:

‘Fifty-eighth SAGE meeting on COVID-19, 21 September 2020’:


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


Declared interests

None received.

in this section

filter RoundUps by year

search by tag