The future of education equity policy: ‘neoliberal’ versus ‘social justice’ approaches

This post summarises Cairney and Kippin’s qualitative systematic review of peer-reviewed research on education equity policy. See also: The future of equity policy in education and health: will intersectoral action be the solution? and posts on ‘Heath in All Policies’ and health inequalities.

Governments, international organisations, and researchers all express a high and enduring commitment to ‘education equity’. Yet, this is where the agreement ends.

The definition of the problem of inequity and the feasibility of solutions is highly contested, to the extent that it is common to identify two competing approaches:

1. A ‘neoliberal’ approach, focusing on education’s role in the economy, market-based reforms, and ‘new public management’ reforms to schools.

2. A ‘social justice’ approach, focusing on education’s role in student wellbeing and life opportunities, and state-led action to address the wider social determinants of education outcomes.

Almost all of the research included in our review suggests that the neoliberal approach dominates international and domestic policy agendas at the expense of the wider focus on social justice.

We describe education equity researchers as the narrators of cautionary tales of education inequity. Most employ critical policy analysis to challenge what they call the dominant stories of education that hinder meaningful equity policies.

First, many describe common settings, including a clear sense that unfair inequalities endure despite global and domestic equity rhetoric.

They also describe the multi-level nature of the governance of education, but with less certainty about relationships across levels. A small number of international organisations and countries are key influencers of a global neoliberal agenda and there is discretion to influence policy at local and school levels. In that context, some studies relate the lack of progress to the malign influence of one or more levels, such as global and central government agendas undermining local change, or local actors disrupting central initiatives.

Second, studies describe similar plots. Many describe stymied progress on equity caused by the negative impacts of neoliberalism: undermining equity by (1) equating it with narrow definitions of equal access to well-performing schools and test-based attainment outcomes, and (2) taking attention from social justice to focus on economic competitiveness.

Many describe policymakers using a generic focus on equity as a facade, to ignore and reproduce inequalities in relation to minoritized populations. Or, equity is a ‘wicked’ issue that defies simple solutions. Many plots involve a contrast between agency-focused narratives that emphasise hopefulness (e.g. among ‘change agents’) and systemic or structural narratives that emphasise helplessness.

Third, they present common ideas about characters. In global narratives, researchers challenge the story by international organisations that they are the heroes providing funding backed by crucial instructions to make educations systems and economies competitive. Most education articles portray neoliberal international organisations and central governments as the villains: narrowing equity to simplistic measures of performance at the expense of more meaningful outcomes.

At a national and local level, they criticise the dominant stories of equity within key countries, such as the US, that continue to reproduce highly unequal outcomes while projecting a sense of progress. The most vividly told story is of white parents, who portray their ‘gifted’ children as most deserving of advantage in the school system, and therefore the victims of attempts to widen access or redistribute scarce resources (high quality classes and teachers). Rather, these parents are the villains standing – sometimes unintentionally, but mostly intentionally – in the way of progress.

The only uncertainty regards the role of local and school leaders. In some cases, they are the initially-heroic figures, able to find ways to disrupt a damaging national agenda and become the ‘change agents’ that shift well-established rules and norms before being thwarted by community and parental opposition. In others, they are perhaps-unintentional villains who reproduce racialised, gendered, or class-based norms regarding which students are ‘gifted’ and worthy of investment versus which students need remedial classes or disrupt other learners.

Fourth, the moral of the story is mostly clear. Almost all studies criticise the damaging impact of neoliberal definitions of equity and the performance management and quasi-market techniques that support it. They are sold as equity measures but actually exacerbate inequalities. As such, the moral is to focus our efforts elsewhere: on social justice, the social and economic determinants of education, and the need to address head-on the association between inequalities and minoritized populations (to challenge ‘equity for all’ messages). However, it is difficult to pinpoint the source of much-needed change. In some cases, strong direction from central governments is necessary to overcome obstacles to change. In others, only bottom-up action by local and school leaders will induce change.

Perhaps the starkest difference in approaches relates to expectations for the future. For ‘neoliberal’ advocates, solutions such as market incentives or education system reforms will save schools and the next generation of students. In contrast, ‘social justice’ advocates expect these reforms to fail and cause irreparable damage to the prospect of education equity.

The future of public health policymaking after COVID-19: lessons from Health in All Policies

Behold Paul’s attempt to make our faces the same size

Paul Cairney, Emily St Denny, Heather Mitchell 

This post summarises new research on the health equity strategy Health in All Policies. As our previous post suggests, it is common to hope that a major event will create a ‘window of opportunity’ for such strategies to flourish, but the current COVID-19 experience suggests otherwise. If so, what do HIAP studies tell us about how to respond, and do they offer any hope for future strategies? The full report is on Open Research Europe, accompanied by a brief interview on its contribution to the Horizon 2020 project – IMAJINE – on spatial justice.

COVID-19 should have prompted governments to treat health improvement as fundamental to public policy

Many had made strong rhetorical commitments to public health strategies focused on preventing a pandemic of non-communicable diseases (NCDs). To do so, they would address the ‘social determinants’ of health and health inequalities, defined by the WHO as ‘the unfair and avoidable differences in health status’ that are ‘shaped by the distribution of money, power and resources’ and ‘the conditions in which people are born, grow, live, work and age’.

COVID-19 reinforces the impact of the social determinants of health. Health inequalities result from factors such as income and social and environmental conditions, which influence people’s ability to protect and improve their health. COVID-19 had a visibly disproportionate impact on people with (a) underlying health conditions associated with NCDs, and (b) less ability to live and work safely.

Yet, the opposite happened. The COVID-19 response side-lined health improvement

Health departments postponed health improvement strategies and moved resources to health protection.

This experience shows that the evidence does not speak for itself

The evidence on social determinants is clear to public health specialists, but the idea of social determinants is less well known or convincing to policymakers.

It also challenges the idea that the logic of health improvement is irresistible

Health in All Policies (HIAP) is the main vehicle for health improvement policymaking, underpinned by: a commitment to health equity by addressing the social determinants of health; the recognition that the most useful health policies are not controlled by health departments; the need for collaboration across (and outside) government; and, the search for high level political commitment to health improvement.

Its logic is undeniable to HIAP advocates, but not policymakers. A government’s public commitment to HIAP does not lead inevitably to the roll-out of a fully-formed HIAP model. There is a major gap between the idea of HIAP and its implementation. It is difficult to generate HIAP momentum, and it can be lost at any time.

Instead, we need to generate more realistic lessons from health improvement and promotion policy

However, most HIAP research does not provide these lessons. Most HIAP research combines:

1. functional logic (here is what we need)
2. programme logic (here is what we think we need to do to achieve it), and
3. hope.

Policy theory-informed empirical studies of policymaking could help produce a more realistic agenda, but very few HIAP studies seem to exploit their insights.

To that end, this review identifies lessons from studies of HIAP and policymaking

It summarises a systematic qualitative review of HIAP research. It includes 113 articles (2011-2020) that refer to policymaking theories or concepts while discussing HIAP.

We produced these conclusions from pre-COVID-19 studies of HIAP and policymaking, but our new policymaking context – and its ironic impact on HIAP – is impossible to ignore.

It suggests that HIAP advocates produced a 7-point playbook for the wrong game

The seven most common pieces of advice add up to a plausible but incomplete strategy:

1. adopt a HIAP model and toolkit
2. raise HIAP awareness and support in government
3. seek win-win solutions with partners
4. avoid the perception of ‘health imperialism’ when fostering intersectoral action
5. find HIAP policy champions and entrepreneurs
6. use HIAP to support the use of health impact assessments (HIAs)
7. challenge the traditional cost-benefit analysis approach to valuing HIAP.

Yet, two emerging pieces of advice highlight the limits to the current playbook and the search for its replacement:

1. treat HIAP as a continuous commitment to collaboration and health equity, not a uniform model; and,
2. address the contradictions between HIAP aims.

As a result, most country studies report a major, unexpected, and disappointing gap between HIAP commitment and actual outcomes

These general findings are apparent in almost all relevant studies. They stand out in the ‘best case’ examples where: (a) there is high political commitment and strategic action (such as South Australia), or (b) political and economic conditions are conducive to HIAP (such as Nordic countries).

These studies show that the HIAP playbook has unanticipated results, such as when the win-win strategy leads to  HIAP advocates giving ground but receiving little in return.

HIAP strategies to challenge the status quo are also overshadowed by more important factors, including (a) a far higher commitment to existing healthcare policies and the core business of government, and (b) state retrenchment. Additional studies of decentralised HIAP models find major gaps between (a) national strategic commitment (backed by national legislation) and (b) municipal government progress.

Some studies acknowledge the need to use policymaking research to produce new ways to encourage and evaluate HIAP success

Studies of South Australia situate HIAP in a complex policymaking system in which the link between policy activity and outcomes is not linear.  

Studies of Nordic HIAP show that a commitment to municipal responsibility and stakeholder collaboration rules out the adoption of a national uniform HIAP model.

However, most studies do not use policymaking research effectively or appropriately

Almost all HIAP studies only scratch the surface of policymaking research (while some try to synthesise its insights, but at the cost of clarity).

Most HIAP studies use policy theories to:

1. produce practical advice (such as to learn from ‘policy entrepreneurs’), or
2. supplement their programme logic (to describe what they think causes policy change and better health outcomes).

Most policy theories were not designed for this purpose.

Policymaking research helps primarily to explain the HIAP ‘implementation gap’

Its main lesson is that policy outcomes are beyond the control of policymakers and HIAP advocates. This explanation does not show how to close implementation gaps.

Its practical lessons come from critical reflection on dilemmas and politics, not the reinvention of a playbook

It prompts advocates to:

• Treat HIAP as a political project, not a technical exercise or puzzle to be solved.
• Re-examine the likely impact of a focus on intersectoral action and collaboration, to recognise the impact of imbalances of power and the logic of policy specialisation.
• Revisit the meaning-in-practice of the vague aims that they take for granted without explaining, such as co-production, policy learning, and organisational learning.
• Engage with key trade-offs, such as between a desire for uniform outcomes (to produce health equity) but acceptance of major variations in HIAP policy and policymaking.
• Avoid reinventing phrases or strategies when facing obstacles to health improvement.

We describe these points in more detail here:

Our Open Research Europe article (peer reviewed) The future of public health policymaking… (europa.eu)

Paul summarises the key points as part of a HIAP panel: Health in All Policies in times of COVID-19

ORE blog on the wider context of this work: forthcoming

What have we learned so far from the UK government’s COVID-19 policy?

This post first appeared on LSE British Politics and Policy (27.11.20) and is based on this article in British Politics.

Paul Cairney assesses government policy in the first half of 2020. He identifies the intense criticism of its response so far, encouraging more systematic assessments grounded in policy research.

In March 2020, COVID-19 prompted policy change in the UK at a speed and scale only seen during wartime. According to the UK government, policy was informed heavily by science advice. Prime Minister Boris Johnson argued that, ‘At all stages, we have been guided by the science, and we will do the right thing at the right time’. Further, key scientific advisers such as Sir Patrick Vallance emphasised the need to gather evidence continuously to model the epidemic and identify key points at which to intervene, to reduce the size of the peak of population illness initially, then manage the spread of the virus over the longer term.

Both ministers and advisors emphasised the need for individual behavioural change, supplemented by government action, in a liberal democracy in which direct imposition is unusual and unsustainable. However, for its critics, the government experience has quickly become an exemplar of policy failure.

Initial criticisms include that ministers did not take COVID-19 seriously enough in relation to existing evidence, when its devastating effect was apparent in China in January and Italy from February; act as quickly as other countries to test for infection to limit its spread; or introduce swift-enough measures to close schools, businesses, and major social events. Subsequent criticisms highlight problems in securing personal protective equipment (PPE), testing capacity, and an effective test-trace-and-isolate system. Some suggest that the UK government was responding to the ‘wrong pandemic’, assuming that COVID-19 could be treated like influenza. Others blame ministers for not pursuing an elimination strategy to minimise its spread until a vaccine could be developed. Some criticise their over-reliance on models which underestimated the R (rate of transmission) and ‘doubling time’ of cases and contributed to a 2-week delay of lockdown. Many describe these problems and delays as the contributors to the UK’s internationally high number of excess deaths.

How can we hold ministers to account in a meaningful way?

I argue that these debates are often fruitless and too narrow because they do not involve systematic policy analysis, take into account what policymakers can actually do, or widen debate to consider whose lives matter to policymakers. Drawing on three policy analysis perspectives, I explore the questions that we should ask to hold ministers to account in a way that encourages meaningful learning from early experience.

These questions include:

Was the government’s definition of the problem appropriate?
Much analysis of UK government competence relates to specific deficiencies in preparation (such as shortages in PPE), immediate action (such as to discharge people from hospitals to care homes without testing them for COVID-19), and implementation (such as an imperfect test-trace-and-isolate system). The broader issue relates to its focus on intervening in late March to protect healthcare capacity during a peak of infection, rather than taking a quicker and more precautionary approach. This judgment relates largely to its definition of the policy problem which underpins every subsequent policy intervention.

Did the government select the right policy mix at the right time? Who benefits most from its choices?

Most debates focus on the ‘lock down or not?’ question without exploring fully the unequal impact of any action. The government initially relied on exhortation, based on voluntarism and an appeal to social responsibility. Initial policy inaction had unequal consequences on social groups, including people with underlying health conditions, black and ethnic minority populations more susceptible to mortality at work or discrimination by public services, care home residents, disabled people unable to receive services, non-UK citizens obliged to pay more to live and work while less able to access public funds, and populations (such as prisoners and drug users) that receive minimal public sympathy. Then, in March, its ‘stay at home’ requirement initiated a major new policy and different unequal impacts in relation to the income, employment, and wellbeing of different groups. These inequalities are list in more general discussions of impacts on the whole population.

Did the UK government make the right choices on the trade-offs between values, and what impacts could the government have reasonably predicted?

Initially, the most high-profile value judgment related to freedom from state coercion to reduce infection versus freedom from the harm of infection caused by others. Then, values underpinned choices on the equitable distribution of measures to mitigate the economic and wellbeing consequences of lockdown. A tendency for the UK government to project centralised and ‘guided by the science’ policymaking has undermined public deliberation on these trade-offs between policies. The latter will be crucial to ongoing debates on the trade-offs associated with national and regional lockdowns.

Did the UK government combine good policy with good policymaking?

A problem like COVID-19 requires trial-and-error policymaking on a scale that seems incomparable to previous experiences. It requires further reflection on how to foster transparent and adaptive policymaking and widespread public ownership for unprecedented policy measures, in a political system characterised by (a) accountability focused incorrectly on strong central government control and (b) adversarial politics that is not conducive to consensus seeking and cooperation.

These additional perspectives and questions show that too-narrow questions – such as was the UK government ‘following the science’ – do not help us understand the longer term development and wider consequences of UK COVID-19 policy. Indeed, such a narrow focus on science marginalises wider discussions of values and the populations that are most disadvantaged by government policy.

_____________________

The UK government’s lack of control of public policy

This post first appeared as Who controls public policy? on the UK in a Changing Europe website. There is also a 1-minute video, but you would need to be a completist to want to watch it.

Our #AcademicintheSpotlight series highlights social scientists doing innovative & dynamic research that you really ought to be aware of.

This week, it’s the turn of @CairneyPaul. Paul works on understanding the mechanics of public policy and policymaking. Interested? Read on… pic.twitter.com/bMs8RoQi2a

— The UK in a Changing Europe (@UKandEU) September 30, 2020

Most coverage of British politics focuses on the powers of a small group of people at the heart of government. In contrast, my research on public policy highlights two major limits to those powers, related to the enormous number of problems that policymakers face, and to the sheer size of the government machine.

First, elected policymakers simply do not have the ability to properly understand, let alone solve, the many complex policy problems they face. They deal with this limitation by paying unusually high attention to a small number of problems and effectively ignoring the rest.

Second, policymakers rely on a huge government machine and network of organisations (containing over 5 million public employees) essential to policy delivery, and oversee a statute book which they could not possibly understand.

In other words, they have limited knowledge and even less control of the state, and have to make choices without knowing how they relate to existing policies (or even what happens next).

These limits to ministerial powers should prompt us to think differently about how to hold them to account. If they only have the ability to influence a small proportion of government business, should we blame them for everything that happens in their name?

My approach is to apply these general insights to specific problems in British politics. Three examples help to illustrate their ability to inform British politics in new ways.

First, policymaking can never be ‘evidence based’. Some scientists cling to the idea that the ‘best’ evidence should always catch the attention of policymakers, and assume that ‘speaking truth to power’ helps evidence win the day.

As such, researchers in fields like public health and climate change wonder why policymakers seem to ignore their evidence.

The truth is that policymakers only have the capacity to consider a tiny proportion of all available information. Therefore, they must find efficient ways to ignore almost all evidence to make timely choices.

They do so by setting goals and identifying trusted sources of evidence, but also using their gut instinct and beliefs to rule out most evidence as irrelevant to their aims.

Second, the UK government cannot ‘take back control’ of policy following Brexit simply because it was not in control of policy before the UK joined. The idea of control is built on the false image of a powerful centre of government led by a small number of elected policymakers.

This way of thinking assumes that sharing power is simply a choice. However, sharing power and responsibility is borne of necessity because the British state is too large to be manageable.

Governments manage this complexity by breaking down their responsibilities into many government departments. Still, ministers can only pay attention to a tiny proportion of issues managed by each department. They delegate most of their responsibilities to civil servants, agencies, and other parts of the public sector.

In turn, those organisations rely on interest groups and experts to provide information and advice.

As a result, most public policy is conducted through small and specialist ‘policy communities’ that operate out of the public spotlight and with minimal elected policymaker involvement.

The logical conclusion is that senior elected politicians are less important than people think. While we like to think of ministers sitting in Whitehall and taking crucial decisions, most of these decisions are taken in their name but without their intervention.

Third, the current pandemic underlines all too clearly the limits of government power. Of course people are pondering the degree to which we can blame UK government ministers for poor choices in relation to Covid-19, or learn from their mistakes to inform better policy.

Many focus on the extent to which ministers were ‘guided by the science’. However, at the onset of a new crisis, government scientists face the same uncertainty about the nature of the policy problem, and ministers are not really able to tell if a Covid-19 policy would work as intended or receive enough public support.

Some examples from the UK experience expose the limited extent to which policymakers can understand, far less control, an emerging crisis.

Prior to the lockdown, neither scientists nor ministers knew how many people were infected, nor when levels of infection would peak.

They had limited capacity to test. They did not know how often (and how well) people wash their hands. They did not expect people to accept and follow strict lockdown rules so readily, and did not know which combination of measures would have the biggest impact.

When supporting businesses and workers during ‘furlough’, they did not know who would be affected and therefore how much the scheme would cost.

In short, while Covid-19 has prompted policy change and state intervention on a scale not witnessed outside of wartime, the government has never really known what impact its measures would have.

Overall, the take-home message is that the UK narrative of strong central government control is damaging to political debate and undermines policy learning. It suggests that every poor outcome is simply the consequence of bad choices by powerful leaders. If so, we are unable to distinguish between the limited competence of some leaders and the limited powers of them all.

The UK Government’s COVID-19 policy: assessing evidence-informed policy analysis in real time

On the 23rd March 2020, the UK Government’s Prime Minister Boris Johnson declared: ‘From this evening I must give the British people a very simple instruction – you must stay at home’. He announced measures to help limit the impact of COVID-19 , including new regulations on behaviour, police powers to support public health, budgetary measures to support businesses and workers during their economic inactivity, the almost-complete closure of schools, and the major expansion of healthcare capacity via investment in technology, discharge to care homes, and a consolidation of national, private, and new health service capacity (note that many of these measures relate only to England, with devolved governments responsible for public health in Northern Ireland, Scotland, and Wales). Overall, the coronavirus prompted almost-unprecedented policy change, towards state intervention, at a speed and magnitude that seemed unimaginable before 2020.

Yet, many have criticised the UK government’s response as slow and insufficient. Criticisms include that UK ministers and their advisors did not:

• take the coronavirus seriously enough in relation to existing evidence (when its devastating effect was increasingly apparent in China in January and Italy from February)
• act as quickly as some countries to test for infection to limit its spread, and/ or introduce swift measures to close schools, businesses, and major social events, and regulate social behaviour (such as in Taiwan, South Korea, or New Zealand)
• introduce strict-enough measures to stop people coming into contact with each other at events and in public transport.

They blame UK ministers for pursuing a ‘mitigation’ strategy, allegedly based on reducing the rate of infection and impact of COVID-19 until the population developed ‘herd immunity’, rather than an elimination strategy to minimise its spread until a vaccine or antiviral could be developed. Or, they criticise the over-reliance on specific models, which underestimated the R (rate of transmission) and ‘doubling time’ of cases and contributed to a 2-week delay of lockdown.

Many cite this delay, compounded by insufficient personal protective equipment (PPE) in hospitals and fatal errors in the treatment of care homes, as the biggest contributor to the UK’s unusually high number of excess deaths (Campbell et al, 2020; Burn-Murdoch and Giles, 2020; Scally et al, 2020; Mason, 2020; Ball, 2020; compare with Freedman, 2020a; 2020b and Snowden, 2020).

In contrast, scientific advisers to UK ministers have emphasised the need to gather evidence continuously to model the epidemic and identify key points at which to intervene, to reduce the size of the peak of population illness initially, then manage the spread of the virus over the longer term (e.g. Vallance). Throughout, they emphasised the need for individual behavioural change (hand washing and social distancing), supplemented by government action, in a liberal democracy in which direct imposition is unusual and, according to UK ministers, unsustainable in the long term.

We can relate these debates to the general limits to policymaking identified in policy studies (summarised in Cairney, 2016; 2020a; Cairney et al, 2019) and underpinning the ‘governance thesis’ that dominates the study of British policymaking (Kerr and Kettell, 2006: 11; Jordan and Cairney, 2013: 234).

First, policymakers must ignore almost all evidence. Individuals combine cognition and emotion to help them make choices efficiently, and governments have equivalent rules to prioritise only some information.

Second, policymakers have a limited understanding, and even less control, of their policymaking environments. No single centre of government has the power to control policy outcomes. Rather, there are many policymakers and influencers spread across a political system, and most choices in government are made in subsystems, with their own rules and networks, over which ministers have limited knowledge and influence. Further, the social and economic context, and events such as a pandemic, often appear to be largely out of their control.

Third, even though they lack full knowledge and control, governments must still make choices. Therefore, their choices are necessarily flawed.

Fourth, their choices produce unequal impacts on different social groups.

Overall, the idea that policy is controlled by a small number of UK government ministers, with the power to solve major policy problems, is still popular in media and public debate, but dismissed in policy research .

Hold the UK government to account via systematic analysis, not trials by social media

To make more sense of current developments in the UK, we need to understand how UK policymakers address these limitations in practice, and widen the scope of debate to consider the impact of policy on inequalities.

A policy theory-informed and real-time account helps us avoid after-the-fact wisdom and bad-faith trials by social media.

UK government action has been deficient in important ways, but we need careful and systematic analysis to help us separate (a) well-informed criticism to foster policy learning and hold ministers to account, from (a) a naïve and partisan rush to judgement that undermines learning and helps let ministers off the hook.

To that end, I combine insights from policy analysis guides, policy theories, and critical policy analysis to analyse the UK government’s initial coronavirus policy. I use the lens of 5-step policy analysis models to identify what analysts and policymakers need to do, the limits to their ability to do it, and the distributional consequences of their choices.

I focus on sources in the public record, including oral evidence to the House of Commons Health and Social Care committee, and the minutes and meeting papers of the UK Government’s Scientific Advisory Group for Emergencies (SAGE) (and NERVTAG), transcripts of TV press conferences and radio interviews, and reports by professional bodies and think tanks.

The short version is here. The long version – containing a huge list of sources and ongoing debates – is here. Both are on the COVID-19 page.

COVID-19 policy in the UK: Table 2: Summary of SAGE minutes, January-June 2020

This post is part 8 of COVID-19 policy in the UK: Did the UK Government ‘follow the science’? Reflections on SAGE meetings

The table is too big to reproduce here, so you have the following options:

Table 2 in PDF

Table 2 as a word document

Or, if you prefer not to read the posts individually:

The whole thing in PDF

The whole thing as a Word document

The full list of SAGE posts:

COVID-19 policy in the UK: yes, the UK Government did ‘follow the science’

Did the UK Government ‘follow the science’? Reflections on SAGE meetings

The role of SAGE and science advice to government

The overall narrative underpinning SAGE advice and UK government policy

SAGE meetings from January-June 2020

SAGE Theme 1. The language of intervention

SAGE Theme 2. Limited capacity for testing, forecasting, and challenging assumptions

SAGE Theme 3. Communicating to the public

COVID-19 policy in the UK: Table 2: Summary of SAGE minutes, January-June 2020

COVID-19 policy in the UK: SAGE Theme 3. Communicating to the public

This post is part 7 of COVID-19 policy in the UK: Did the UK Government ‘follow the science’? Reflections on SAGE meetings

SAGE’s emphasis on uncertainty and limited knowledge extended to the evidence on how to influence behaviour via communication:

‘there is limited evidence on the best phrasing of messages, the barriers and stressors that people will encounter when trying to follow guidance, the attitudes of the public to the interventions, or the best strategies to promote adherence in the long-term’ (SPI-B Meeting paper 3.3.20: 2)

Early on, SAGE minutes described continuously the potential problems of communicating risk and encouraging behavioural change through communication (in other words, based on low expectations for the types of quarantine measures associated with China and South Korea).

• It sought ‘behavioural science input on public communication’ and ‘agreed on the importance of behavioural science informing policy – and on the importance of public trust in HMG’s approach’ (28.1.20: 2).
• It worried about how the public might interpret ‘case fatality rate’, given the different ways to describe and interpret frequencies and risks (4.2.20: 3).
• It stated that ‘Epidemiological terms need to be made clearer in the planning documents to avoid ambiguity’ (11.2.20: 3).
• Its extensive discussion of behavioural science (13.2.20: 2-3) includes: there will be public scepticism and inaction until first deaths are confirmed; the main aim is to motivate people by relating behavioural change to their lives; messaging should stress ‘personal responsibility and responsibility to others’ and be clear on which measures are effective’, and ‘National messaging should be clear and definitive: if such messaging is presented as both precautionary and sufficient, it will reduce the likelihood of the public adopting further unnecessary or contradictory behaviours’ (13.2.20: 2-3)
• Banning large public events could signal the need to change behaviour more generally, but evidence for its likely impact is unavailable (SPI-M-O, 11.2.20: 1).

Generally speaking, the assumption underpinning communication is that behavioural change will come largely from communication (encouragement and exhortation) rather than imposition. Hence, for example, the SPI-B (25.2.20: 2) recommendation on limiting the ‘risk of public disorder’:

• ‘Provide clear and transparent reasons for different strategies: The public need to understand the purpose of the Government’s policy, why the UK approach differs to other countries and how resources are being allocated. SPI-B agreed that government should prioritise messaging that explains clearly why certain actions are being taken, ahead of messaging designed solely for reassuring the public.
• This should also set clear expectations on how the response will develop, g. ensuring the public understands what they can expect as the outbreak evolves and what will happen when large numbers of people present at hospitals. The use of early messaging will help, as a) individuals are likely to be more receptive to messages before an issue becomes controversial and b) it will promote a sense the Government is following a plan.
• Promote a sense of collectivism: All messaging should reinforce a sense of community, that “we are all in this together.” This will avoid increasing tensions between different groups (including between responding agencies and the public); promote social norms around behaviours; and lead to self-policing within communities around important behaviours’.

The underpinning assumption is that the government should treat people as ‘rational actors’: explain risk and how to reduce it, support existing measures by the public to socially distance, be transparent, explain if UK is doing things differently to other countries, and recognise that these measures are easier for some more than others (13.3.20: 3).

In that context, SPI-B Meeting paper 22.3.20 describes how to enable social distancing with reference to the ‘behaviour change wheel’ (Michie et al, 2011): ‘There are nine broad ways of achieving behaviour change: Education, Persuasion, Incentivisation, Coercion, Enablement, Training, Restriction, Environmental restructuring, and Modelling’ and many could reinforce each other (22.3.20: 1). The paper comments on current policy in relation to 5 elements:

1. Education – clarify guidance (generally, and for shielding), e.g. through interactive website, tailored to many audiences
2. Persuasion – increase perceived threat among ‘those who are complacent, using hard-hitting emotional messaging’ while providing clarity and positive messaging (tailored to your audience’s motivation) on what action to take (22.3.20: 1-2).
3. Incentivisation – emphasise social approval as a reward for behaviour change
4. Coercion – ‘Consideration should be given to enacting legislation, with community involvement, to compel key social distancing measures’ (combined with encouraging ‘social disapproval but with a strong caveat around unwanted negative consequences’ (22.3.20: 2)
5. Enablement – make sure that people have alternative access to social contact, food, and other resources for people feeling the unequal impact of lockdown (particularly for vulnerable people shielding, aided by community support).

Apparently, section 3 of SPI-B’s meeting paper (1.4.20b: 2) had been redacted because it was critical of a UK Government ‘Framework; with 4 new proposals for greater compliance: ‘17) increasing the financial penalties imposed; 18) introducing self-validation for movements; 19) reducing exercise and/or shopping; 20) reducing non-home working’. On 17, it suggests that the evidence base for (e.g.) fining someone exercising more than 1km from their home could contribute to lower support for policy overall. On 17-19, it suggests that most people are already complying, so there is no evidence to support more targeted measures. It is more positive about 20, since it could reduce non-home working (especially if financially supported). Generally, it suggests that ministers should ‘also consider the role of rewards and facilitations in improving adherence’ and use organisational changes, such as staggered work hours and new use of space, rather than simply focusing on individuals.

Communication after the lockdown

SAGE suggests that communication problems are more complicated during the release of lockdown measures (in other words, without the ability to present the relatively-low-ambiguity message ‘stay at home’). Examples (mostly from SPI-B and its contributors) include:

• Address potential confusion, causing false concern or reassurance, regarding antigen and antibody tests (meeting papers 1.4.20c: 3; 13.4.20b: 1-4; 22.4.20b: 1-5; 29.4.20a: 1-4)
• When notifying people about the need to self-isolate, address the trade-offs between symptom versus positive test based notifications (meeting paper 29.4.20a: 1-4; 5.5.20: 1-8)
• If you are worried about public ‘disorder’, focus on clear, effective, tailored communication, using local influencers, appealing to sympathetic groups (like NHS staff), and co-producing messages between the police and public (in other words, police via consent, and do not exacerbate grievances) (meeting papers 19.4.20: 1-4; 21.4.20: 1-3; 4.5.20: 1-11)
• Be wary of lockdowns specific to very small areas, which undermine the ‘all in it together’ message (REDACTED and Clifford Stott, no date: 1). If you must to it, clarify precisely who is affected and what they should do, support the people most vulnerable and impacted (e.g. financially), and redesign physical spaces (meeting paper SPI-B 22.4.20a)
• When reopening schools (fully or partly), communication is key to the inevitably complex and unpredictable behavioural consequences (so, for example, work with parents, teachers, and other stakeholders to co-produce clear guidance) (29.4.20d: 1-10)
• On the introduction of Alert Levels, as part of the Joint Biosecurity Centre work on local outbreaks (described in meeting paper 20.5.20a: 1-9): build public trust and understanding regarding JBC alert levels, and relate them very clearly to expected behaviour (SAGE 28.5.20). Each Alert Level should relate clearly to a required response in that area, and ‘public communications on Alert Levels needs many trusted messengers giving the same advice, many times’ (meeting paper 27.5.20b: 3).
• On transmission between social networks, ‘Communicate two key principles: 1. People whose work involves large numbers of contacts with different people should avoid close, prolonged, indoor contact with anyone as far as possible … 2. People with different workplace networks should avoid meeting or sharing the same spaces’ (meeting paper 27.5.20b: 1).
• On outbreaks in ‘forgotten institutional settings’ (including Prisons, Homeless Hostels, Migrant dormitories, and Long stay mental health): address the unusually low levels of trust in (or awareness of) government messaging among so-called ‘hard to reach groups’ (meeting paper 28.5.20a: 1).

See also:

SPI-M (Meeting paper 17.3.20b: 4) list of how to describe probabilities. This is more important than it looks, since there is a potentially major gap between the public and advisory group understanding of words like ‘probably’ (compare with the CIA’s Words of Estimative Probability).

The full list of SAGE posts:

COVID-19 policy in the UK: yes, the UK Government did ‘follow the science’

Did the UK Government ‘follow the science’? Reflections on SAGE meetings

The role of SAGE and science advice to government

The overall narrative underpinning SAGE advice and UK government policy

SAGE meetings from January-June 2020

SAGE Theme 1. The language of intervention

SAGE Theme 2. Limited capacity for testing, forecasting, and challenging assumptions

SAGE Theme 3. Communicating to the public

COVID-19 policy in the UK: Table 2: Summary of SAGE minutes, January-June 2020

COVID-19 policy in the UK: SAGE Theme 2. Limited capacity for testing, forecasting, and challenging assumptions

This post is part 6 of COVID-19 policy in the UK: Did the UK Government ‘follow the science’? Reflections on SAGE meetings

Limited testing

Oral evidence to the Health and Social Care committee highlights the now-well-documented limits to UK testing capacity and PPE stocks (see also NERVTAG on PPE). SAGE does not discuss testing capacity much in the beginning, although on 10.3.20 it lists as an action point: ‘Plans for how PHE can move from 1,000 serology tests to 10,000 tests per week’ and by 16.3.20 it describes the urgent need to scale up testing – perhaps with commercial involvement and to test at home (if can ensure accuracy) – and to secure sufficient data to track the epidemic well enough to inform operational decisions. From April, it highlights the need for a ‘national testing strategy’ to cover NHS patients, staff, an epidemiological survey, and the community (2.4.20), and the need for far more testing is a feature of almost every meeting from then.

Limited contact tracing

Initially, SAGE describes a quite-low contact tracing capacity: ‘Currently, PHE can cope with five new cases a week (requiring isolation of 800 contacts). Modelling suggests this capacity could be increased to 50 new cases a week (8,000 contact isolations)’ (18.2.20: 1).

Previously, it had noted that the point would come when transmission was too high to make contact tracing worthwhile, particularly since many (e.g. asymptomatic) cases may already have been missed (20.2.20: 2) and the necessary testing capacity was not in place (16.4.20): ‘PHE to work with SPI-M to develop criteria for when contact tracing is no longer worthwhile. This should include consideration of any limiting factors on testing and alternative methods of identifying epidemic evolution and characteristics’ (11.2.20: 3; see also Testing and contact tracing).

It returned to the feasibility question after the lockdown, with:

• SPI-M (meeting paper 4.20d: 1-3) estimating that effective contact tracing (80% of non-household cases, in 2 days) could reduce the R by 30-60% if you could quarantine many people, multiple times; and,
• SPI-B (meeting paper 4.20a: 1-3) advising on the need to clarify to people how it would work and what they should do, redesign physical spaces, and conduct new qualitative research and stakeholder engagement to ‘help us to understand more clearly the specific drivers, enablers and barriers for new behavioural recommendations’ to address an unprecedented problem in the UK (22.4.20a: 2). SPI-B also describes the trade-offs between app-informed systems (notification based on symptoms would suit people seeking to be precautionary, but could reduce compliance among people who believe the risk to be low) (see meeting papers 29.4.20: 3 and 5.5.20: 1-8)
• SAGE noting ongoing work on clusters and super-spreading events, which necessitate cluster-based contact tracing (11.6.20: 3)
• A more general message that contact tracing will be overwhelmed if lockdown measures are released too soon, raising R well above 1 and causing incidence to rise too quickly (e.g. 14.5.20)

Low capacity to achieve high levels of information necessary for forecasting

This type of discussion exemplifies a general and continuous focus on the lack of data to inform advice:

‘24. Real-time forecasting models rely on deriving information on the epidemic from surveillance. If transmission is established in the UK there will necessarily be a delay before sufficiently accurate forecasts in the UK are available. 25. Decisions being made on whether to modify or lift non-pharmaceutical interventions require accurate understanding of the state of the epidemic. Large-scale serological data would be ideal, especially combined with direct monitoring of contact behaviour. 26. Preliminary forecasts and accurate estimates of epidemiological parameters will likely be available in the order of weeks and not days following widespread outbreaks in the UK (or a similar country). While some estimates may be available before this time their accuracy will be much more limited. 27. The UK hospitalisation rate and CFR will be very important for operational planning and will be estimated over a similar timeframe. They may take longer depending on the availability of data’ (Meeting paper 2.3.20: 3-4).

A limited capacity to reach a relatively cautious consensus?

These limitations to information contributed to the difference between SAGE’s estimate on UK transmission (such as in comparison with Italy) and the UK’s much faster rate of transmission:

‘the UK likely has thousands of cases – as many as 5,000 to 10,000 – which are geographically spread nationally … The UK is considered to be 4-5 weeks behind Italy but on a similar curve (6-8 weeks behind if interventions are applied)’ (10.3.20: 1)

‘Based on limited available evidence, SAGE considers that the UK is 2 to 4 weeks behind Italy in terms of the epidemic curve’ (18.3.20: 1)

Rather, the UK was under 2 weeks behind Italy on the 10^th March, suggesting that its lockdown measures were put in place too late.

At the heart of this estimate was the under-estimated doubling time of infection (‘the time it takes for the number of cases to double in size’, Meeting paper 3.2.20a):

• although described as 3-4 days (28.1.20: 1) then 4-6 days (Meeting paper 2.3.20) based on Wuhan, and 3-5 days based on Hubei (Meeting paper 3.2.20a),
• SAGE estimates ‘every 5-6 days’ (16.3.20: 1) and states that ‘Assuming a doubling time of around 5-7 days continues to be reasonable’ (18.3.20: 1).
• Only by meeting 18 does SAGE estimate the doubling time (ICU patients) at 3-4 days (23.3.20). By meeting 19, it describes the doubling time in hospitals as 3.3 days (26.3.20: 1).

Kit Yates suggests that (a) the UK exhibited a 3-day doubling time during this period (Huffington Post), and (b) although many members of SAGE and SPI-M would have preferred to model on the assumption of 3-days:

Having spoken to some of the modellers on SPI-M, not all of them were missing this. Many of the groups had fitted models to data and come up with shorter and more realistic doubling times, maybe around the 3-day mark, but their estimates never found consensus within the group, so some members of SPI-M have communicated their concerns to me that some of the modelling groups had more influence over the consensus decision than others, which meant that some opinions or estimates which might have been valid, didn’t get heard, and consequently weren’t passed on up the line to SAGE, and then further towards the government, so an over-reliance on certain models or modelling groups might have been costly in this situation (interview, Kit Yates, More or Less, 10.6.20: 4m47s-5m27s)

Yates then suggests that the most listened-to model – led by Neil Ferguson, published 16.3.20 –  estimates a doubling time of 5-days, based on early data from Wuhan, using estimate of R2.4 (and generation time of 6.5 days), ‘which we now know to be way too low’ when we look at the UK data:

‘If they had just plotted the early trajectory of the epidemics against the current UK data at that point, they would have seen [by 14.3.20] that their model was starting to underestimate the number of cases and then the number of deaths which were occurring in the UK’ (interview, Kit Yates, More or Less, 10.6.20: 7m2s-7m15s)

Yates’ account highlights not only

1. the effect of uncertainty and limited capacity to generate more information, but also
2. the wider effect of path dependence, in which the (a) written and unwritten rules and norms of organisations, and (b) enduring ways of thinking (in individuals and groups, and political systems) place limits on new action. These limits are often necessary and beneficial, and often unnecessary and harmful.

Compare with Vallance’s oral evidence to the Health and Social Care committee (17.3.20: q96):

‘If you thought SAGE and the way SAGE works was a cosy consensus of agreeing scientists, you would be very mistaken. It is a lively, robust discussion, with multiple inputs. We do not try to get everybody saying exactly the same thing’.

The full list of SAGE posts:

COVID-19 policy in the UK: yes, the UK Government did ‘follow the science’

Did the UK Government ‘follow the science’? Reflections on SAGE meetings

The role of SAGE and science advice to government

The overall narrative underpinning SAGE advice and UK government policy

SAGE meetings from January-June 2020

SAGE Theme 1. The language of intervention

SAGE Theme 2. Limited capacity for testing, forecasting, and challenging assumptions

SAGE Theme 3. Communicating to the public

COVID-19 policy in the UK: Table 2: Summary of SAGE minutes, January-June 2020

COVID-19 policy in the UK: SAGE Theme 1. The language of intervention

This post is part 5 of COVID-19 policy in the UK: Did the UK Government ‘follow the science’? Reflections on SAGE meetings

There is often a clear distinction between a strategy designed to (a) eliminate a virus/ the spread of disease quickly, and (b) manage the spread of infection over the long term (see The overall narrative).

However, generally, the language of virus management is confusing. We need to be careful with interpreting the language used in these minutes, and other sources such as oral evidence to House of Commons committees, particularly when comparing the language at the beginning (when people were also unsure what to call SARS-CoV-2 and COVID-19) to present day debates.

For example, in January, it is tempting to contrast ‘slow down the spread of the outbreak domestically’ (28.1.20: 2) with a strategy towards ‘extinction’, but the proposed actions may be the same even if the expectations of impact are different. Some people interpret these differences as indicative of a profoundly different approach (delay versus eradicate); some describe the semantic differences as semantics.

By February, SAGE’s expectation is of an inevitable epidemic and inability to contain COVID-19, prompting it to describe the inevitable series of stages:

‘Priorities will shift during a potential outbreak from containment and isolation on to delay and, finally, to case management … When there is sustained transmission in the UK, contact tracing will no longer be useful’ (18.2.20: 1; its discussion on 20.2.20: 2 also concludes that ‘individual cases could already have been missed – including individuals advised that they are not infectious’).

Mitigation versus suppression

On the face of it, it looks like there is a major difference in the ways on which (a) the Imperial College COVID-19 Response Team and (b) SAGE describe possible policy responses. The Imperial paper makes a distinction between mitigation and suppression:

1. Its ‘mitigation strategy scenarios’ highlight the relative effects of partly-voluntary measures on mortality and demand for ‘critical care beds’ in hospitals: (voluntary) ‘case isolation in the home’ (people with symptoms stay at home for 7 days), ‘voluntary home quarantine’ (all members of the household stay at home for 14 days if one member has symptoms), (government enforced) ‘social distancing of those over 70’ or ‘social distancing of entire population’ (while still going to work, school or University), and closure of most schools and universities. It omits ‘stopping mass gatherings’ because ‘the contact-time at such events is relatively small compared to the time spent at home, in schools or workplaces and in other community locations such as bars and restaurants’ (2020a: 8). Assuming 70-75% compliance, it describes the combination of ‘case isolation, home quarantine and social distancing of those aged over 70’ as the most impactful, but predicts that ‘mitigation is unlikely to be a viable option without overwhelming healthcare systems’ (2020a: 8-10). These measures would only ‘reduce peak critical care demand by two-thirds and halve the number of deaths’ (to approximately 250,000).
2. Its ‘suppression strategy scenarios’ describe what it would take to reduce the rate of infection (R) from the estimated 2.0-2.6 to 1 or below (in other words, the game-changing point at which one person would infect no more than one other person) and reduce ‘critical care requirements’ to manageable levels. It predicts that a combination of four options – ‘case isolation’, ‘social distancing of the entire population’ (the measure with the largest impact), ‘household quarantine’ and ‘school and university closure’ – would reduce critical care demand from its peak ‘approximately 3 weeks after the interventions are introduced’, and contribute to a range of 5,600-48,000 deaths over two years (depending on the current R and the ‘trigger’ for action in relation to the number of occupied critical care beds) (2020a: 13-14).

In comparison, the SAGE meeting paper (26.2.20b: 1-3), produced 2-3 weeks earlier, pretty much assumes away the possible distinction between mitigation versus suppression measures (which Vallance has described as semantic rather than substantive – scroll down to The distinction between mitigation and suppression measures). In other words, it assumes ‘high levels of compliance over long periods of time’ (26.2.20b: 1). As such, we can interpret SAGE’s discussion as (a) requiring high levels of compliance for these measures to work (the equivalent of Imperial’s description of suppression), while (b) not describing how to use (more or less voluntary versus impositional) government policy to secure compliance. In comparison, Imperial equates suppression with the relatively-short-term measures associated with China and South Korea (while noting uncertainty about how to maintain such measures until a vaccine is produced).

It is taking forever, but I think I am now managing to get into the rhythm of the February SAGE minutes (in relation to media interviews and oral evidence in March), at least enough to identify the need to interpret the text in a relatively thoughtful/ sympathetic way ….

— Professor Paul Cairney (@CairneyPaul) June 29, 2020

One reason for SAGE to assume compliance in its scenario building is to focus on the contribution of each measure, generally taking place over 13 weeks, to delaying the peak of infection (while stating that ‘It will likely not be feasible to provide estimates of the effectiveness of individual control measures, just the overall effectiveness of them all’, 26.2.20b: 1), while taking into account their behavioural implications (26.2.20b: 2-3).

• School closures could contribute to a 3-week delay, especially if combined with FE/ HE closures (but with an unequal impact on ‘Those in lower socio-economic groups … more reliant on free school meals or unable to rearrange work to provide childcare’).
• Home isolation (65% of symptomatic cases stay at home for 7 days) could contribute to a 2-3 week delay (and is the ‘Easiest measure to explain and justify to the public’).
• ‘Voluntary household quarantine’ (all member of the household isolate for 14 days) would have a similar effect – assuming 50% compliance – but with far more implications for behavioural public policy:

‘Resistance & non-compliance will be greater if impacts of this policy are inequitable. For those on low incomes, loss of income means inability to pay for food, heating, lighting, internet. This can be addressed by guaranteeing supplies during quarantine periods.

Variable compliance, due to variable capacity to comply, may lead to dissatisfaction.

Ensuring supplies flow to households is essential. A desire to help among the wider community (e.g. taking on chores, delivering supplies) could be encouraged and scaffolded to support quarantined households.

There is a risk of stigma, so ‘voluntary quarantine’ should be portrayed as an act of altruistic civic duty’.

• ‘Social distancing’ (‘enacted early’), in which people restrict themselves to essential activity (work and school) could produce a 3-5 week delay (and likely to be supported in relation to mass leisure events, albeit less so when work activities involve a lot of contact.

[Note that it is not until May that it addresses this issue of feasibility directly (and, even then, it does not distinguish between technical and political feasibility: ‘It was noted that a useful addition to control measures SAGE considers (in addition to scientific uncertainty) would be the feasibility of monitoring/ enforcement’ (7.5.20: 3)]

As theme 2 suggests, there is a growing recognition that these measures should have been introduced by early March (such as via the Coronavirus Act 2020 not passed until 25.3.20), and likely would if the UK government and SAGE had more information (or interpreted its information in a different way). However, by mid-March, SAGE expresses a mixture of (a) growing urgency, but also (b) the need to stick to the plan, to reduce the peak and avoid a second peak of infection). On 13^th March, it states:

‘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. SAGE therefore thinks there is scientific evidence to support household isolation being implemented as soon as practically possible’ (13.3.20: 1)

‘SAGE further agreed that one purpose of behavioural and social interventions is to enable the NHS to meet demand and therefore reduce indirect mortality and morbidity. There is a risk that current proposed measures (individual and household isolation and social distancing) will not reduce demand enough: they may need to be coupled with more intensive actions to enable the NHS to cope, whether regionally or nationally’ (13.3.20: 2)

On 16^th March, it states:

‘On the basis of accumulating data, including on NHS critical care capacity, the advice from SAGE has changed regarding the speed of implementation of additional interventions. SAGE advises that there is clear evidence to support additional social distancing measures be introduced as soon as possible’ (16.3.20: 1)

Overall, we can conclude two things about the language of intervention:

1. There is now a clear difference between the ways in which SAGE and its critics describe policy: to manage an inevitably long-term epidemic, versus to try to eliminate it within national borders.
2. There is a less clear difference between terms such as suppress and mitigate, largely because SAGE focused primarily on a comparison of different measures (and their combination) rather than the question of compliance.

See also: There is no ‘herd immunity strategy’, which argues that this focus on each intervention was lost in radio and TV interviews with Vallance.

The full list of SAGE posts:

COVID-19 policy in the UK: yes, the UK Government did ‘follow the science’

Did the UK Government ‘follow the science’? Reflections on SAGE meetings

The role of SAGE and science advice to government

The overall narrative underpinning SAGE advice and UK government policy

SAGE meetings from January-June 2020

SAGE Theme 1. The language of intervention

SAGE Theme 2. Limited capacity for testing, forecasting, and challenging assumptions

SAGE Theme 3. Communicating to the public

COVID-19 policy in the UK: Table 2: Summary of SAGE minutes, January-June 2020

COVID-19 policy in the UK: SAGE meetings from January-June 2020

This post is part 4 of COVID-19 policy in the UK: Did the UK Government ‘follow the science’? Reflections on SAGE meetings

SAGE began a series of extraordinary meetings from 22nd January 2020. The first was described as ‘precautionary’ (22.1.20: 1) and includes updates from NERVTAG which met from 13^th January. Its minutes state that ‘SAGE is unable to say at this stage whether it might be required to reconvene’ (22.1.20: 2). The second meeting notes that SAGE will meet regularly (e.g. 2-3 times per week in February) and coordinate all relevant science advice to inform domestic policy, including from NERVTAG and SPI-M (Scientific Pandemic Influenza Group on Modelling) which became a ‘formal sub-group of SAGE for the duration of this outbreak’ (SPI-M-O) (28.1.20: 1). It also convened an additional Scientific Pandemic Influenza subgroup (SPI-B) in February. I summarise these developments by month, but you can see that, by March, it is worth summarising each meeting. The main theme is uncertainty.

January 2020

The first meeting highlights immense uncertainty. Its description of WN-CoV (Wuhan Coronavirus), and statements such as ‘There is evidence of person-to-person transmission. It is unknown whether transmission is sustainable’, sum up the profound lack of information on what is to come (22.1.20: 1-2). It notes high uncertainty on how to identify cases, rates of infection, infectiousness in the absence of symptoms, and which previous experience (such as MERS) offers the most useful guidance. Only 6 days later, it estimates an R between 2-3, doubling rate of 3-4 days, incubation period of around 5 days, 14-day window of infectivity, varied symptoms such as coughing and fever, and a respiratory transmission route (different from SARS and MERS) (28.1.20: 1). These estimates are fairly constant from then, albeit qualified with reference to uncertainty (e.g. about asymptomatic transmission), some key outliers (e.g. the duration of illness in one case was 41 days – 4.2.20: 1), and some new estimates (e.g. of a 6-day ‘serial interval’, or ‘time between successive cases in a chain of transmission’, 11.2.20: 1). By now, it is preparing a response: modelling a ‘reasonable worst case scenario’ (RWC) based on the assumption of an R of 2.5 and no known treatment or vaccine, considering how to slow the spread, and considering how behavioural insights can be used to encourage self-isolation.

February 2020

SAGE began to focus on what measures might delay or reduce the impact of the epidemic. It described travel restrictions from China as low value, since a 95% reduction would have to be draconian to achieve and only secure a one month delay, which might be better achieved with other measures (3.2.20: 1-2). It, and supporting papers, suggested that the evidence was so limited that they could draw ‘no meaningful conclusions … as to whether it is possible to achieve a delay of a month’ by using one or a combination of these measures: international travel restrictions, domestic travel restrictions, quarantine people coming from infected areas, close schools, close FE/ HE, cancel large public events, contact tracing, voluntary home isolation, facemasks, hand washing. Further, some could undermine each other (e.g. school closures impact on older people or people in self-isolation) and have major societal or opportunity costs (SPI-M-O, 3.2.20b: 1-4). For example, the ‘SPI-M-O: Consensus view on public gatherings’ (11.2.20: 1) notes the aim to reduce duration and closeness of (particularly indoor) contact. Large outdoor gatherings are not worse than small, and stopping large events could prompt people to go to pubs (worse).

Throughout February, the minutes emphasize high uncertainty:

• if there will be an epidemic outside of China (4.2.20: 2)
• if it spreads through ‘air conditioning systems’ (4.2.20: 3)
• the spread from, and impact on, children and therefore the impact of closing schools (4.2.20: 3; discussed in a separate paper by SPI-M-O, 10.2.20c: 1-2)
• ‘SAGE heard that NERVTAG advises that there is limited to no evidence of the benefits of the general public wearing facemasks as a preventative measure’ (while ‘symptomatic people should be encouraged to wear a surgical face mask, providing that it can be tolerated’ (4.2.20: 3)

At the same time, its meeting papers emphasized a delay in accurate figures during an initial outbreak: ‘Preliminary forecasts and accurate estimates of epidemiological parameters will likely be available in the order of weeks and not days following widespread outbreaks in the UK’ (SPI-M-O, 3.2.20a: 3).

This problem proved to be crucial to the timing of government intervention. A key learning point will be the disconnect between the following statement and the subsequent realisation (3-4 weeks later) that the lockdown measures from mid-to-late March came too late to prevent an unanticipated number of excess deaths:

‘SAGE advises that surveillance measures, which commenced this week, will provide

actionable data to inform HMG efforts to contain and mitigate spread of Covid-19’ … PHE’s surveillance approach provides sufficient sensitivity to detect an outbreak in its early stages. This should provide evidence of an epidemic around 9- 11 weeks before its peak … increasing surveillance coverage beyond the current approach would not significantly improve our understanding of incidence’ (25.2.20: 1)

It also seems clear from the minutes and papers that SAGE highlighted a reasonable worst case scenario on 26.2.20. It was as worrying as the Imperial College COVID-19 Response Team report dated 16.3.20 that allegedly changed the UK Government’s mind on the 16^th March. Meeting paper 26.2.20a described the assumption of an 80% infection attack rate and 50% clinical attack rate (i.e. 50% of the UK population would experience symptoms), which underpins the assumption of 3.6 million requiring hospital care of at least 8 days (11% of symptomatic), and 541,200 requiring ventilation (1.65% of symptomatic) for 16 days. While it lists excess deaths as unknown, its 1% infection mortality rate suggests 524,800 deaths. This RWC replaces a previous projection (in Meeting paper 10.2.20a: 1-3, based on pandemic flu assumptions) of 820,000 excess deaths (27.2.20: 1).

As such, the more important difference could come from SAGE’s discussion of ‘non-pharmaceutical interventions (NPIs)’ if it recommends ‘mitigation’ while the Imperial team recommends ‘suppression’. However, the language to describe each approach is too unclear to tell (see Theme 1. The language of intervention; also note that NPIs were often described from March as ‘behavioural and social interventions’ following an SPI-B recommendation, Meeting paper 3.2.20: 1, but the language of NPI seems to have stuck).

March 2020

In March, SAGE focused initially (Meetings 12-14) on preparing for the peak of infection on the assumption that it had time to transition towards a series of isolation and social distancing measures that would be sustainable (and therefore unlikely to contribute to a second peak if lifted too soon). Early meetings and meeting papers express caution about the limited evidence for intervention and the potential for their unintended consequences. This approach began to change somewhat from mid-March (Meeting 15), and accelerate from Meetings 16-18, when it became clear that incidence and virus transmission were much larger than expected, before a new phase began from Meeting 19 (after the UK lockdown was announced on the 23^rd).

Meeting 12 (3.3.18) describes preparations to gather and consolidate information on the epidemic and the likely relative effect of each intervention, while its meeting papers emphasise:

• ‘It is highly likely that there is sustained transmission of COVID-19 in the UK at present’, and a peak of infection ‘might be expected approximately 3-5 months after the establishment of widespread sustained transmission’ (SPI-M Meeting paper 2.3.20: 1)
• the need the prepare the public while giving ‘clear and transparent reasons for different strategies’ and reducing ambiguity whenever giving guidance (SPI-B Meeting paper 3.2.20: 1-2)
• The need to combine different measures (e.g. school closure, self-isolation, household isolation, isolating over-65s) at the right time; ‘implementing a subset of measures would be ideal. Whilst this would have a more moderate impact it would be much less likely to result in a second wave’ (Meeting paper 4.3.20a: 3).

Meeting 13 (5.3.20) describes staying in the ‘containment’ phase (which, I think, means isolating people with positive tests at home or in hospital) , and introducing: a 12-week period of individual and household isolation measures in 1-2 weeks, on the assumption of 50% compliance; and a longer period of shielding over-65s 2 weeks later. It describes ‘no evidence to suggest that banning very large gatherings would reduce transmission’, while closing bars and restaurants ‘would have an effect, but would be very difficult to implement’, and ‘school closures would have smaller effects on the epidemic curve than other options’ (5.3.20: 1). Its SPI-B Meeting paper (4.3.20b) expresses caution about limited evidence and reliance on expert opinion, while identifying:

• potential displacement problems (e.g. school closures prompt people to congregate elsewhere, or be looked after by vulnerable older people, while parents to lose the chance to work)
• the visibility of groups not complying
• the unequal impact on poorer and single parent families of school closure and loss of school meals, lost income, lower internet access, and isolation
• how to reduce discontent about only isolating at-risk groups (the view that ‘explaining that members of the community are building some immunity will make this acceptable’ is not unanimous) (4.3.20b: 2).

Meeting 14 (10.3.20) states that the UK may have 5-10000 cases and ‘10-14 weeks from the epidemic peak if no mitigations are introduced’ (10.3.20: 2). It restates the focus on isolation first, followed by additional measures in April, and emphasizes the need to transition to measures that are acceptable and sustainable for the long term:

‘SAGE agreed that a balance needs to be struck between interventions that theoretically have significant impacts and interventions which the public can feasibly and safely adopt in sufficient numbers over long periods’ …’the public will face considerable challenges in seeking to comply with these measures, (e.g. poorer households, those relying on grandparents for childcare)’ (10.3.20: 2)

Meeting 15 (13.3.20: 1) describes an update to its data, suggesting ‘more cases in the UK than SAGE previously expected at this point, and we may therefore be further ahead on the epidemic curve, but the UK remains on broadly the same epidemic trajectory and time to peak’. It states that ‘household isolation and social distancing of the elderly and vulnerable should be implemented soon, provided they can be done well and equitably’, noting that there are ‘no strong scientific grounds’ to accelerate key measures but ‘there will be some minor gains from going early and potentially useful reinforcement of the importance of taking personal action if symptomatic’ (13.3.20: 1) and ‘more intensive actions’ will be required to maintain NHS capacity (13.3.20: 2).

*******

On the 16^th March, the UK Prime Minister Boris Johnson describes an ‘emergency’ (one week before declaring a ‘national emergency’ and UK-wide lockdown)

*******

Meeting 16 (16.3.20) describes the possibility that there are 5-10000 new cases in the UK (there is great uncertainty on the estimate’), doubling every 5-6 days. Therefore, to stay within NHS capacity, ‘the advice from SAGE has changed regarding the speed of implementation of additional interventions. SAGE advises that there is clear evidence to support additional social distancing measures be introduced as soon as possible’ (16.3.20: 1). SPI-M Meeting paper (16.3.20: 1) describes:

‘a combination of case isolation, household isolation and social distancing of vulnerable groups is very unlikely to prevent critical care facilities being overwhelmed … it is unclear whether or not the addition of general social distancing measures to case isolation, household isolation and social distancing of vulnerable groups would curtail the epidemic by reducing the reproduction number to less than 1 … the addition of both general social distancing and school closures to case isolation, household isolation and social distancing of vulnerable groups would be likely to control the epidemic when kept in place for a long period. SPI-M-O agreed that this strategy should be followed as soon as practical’

Meeting 17 (18.3.20) marks a major acceleration of plans, and a de-emphasis of the low-certainty/ beware-the-unintended-consequences approach of previous meetings (on the assumption that it was now 2-4 weeks behind Italy). It recommends school closures as soon as possible (and it, and SPIM Meeting paper 17.3.20b, now downplays the likely displacement effect). It focuses particularly on London, as the place with the largest initial numbers:

‘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. … Transport measures such as restricting public transport, taxis and private hire facilities would have minimal impact on reducing transmission’ (18.3.20: 2)

Meeting 18 (23.3.20) states that the R is higher than expected (2.6-2.8), requiring ‘high rates of compliance for social distancing’ to get it below 1 and stay under NHS capacity (23.3.20: 1). There is an urgent need for more community testing/ surveillance (and to address the global shortage of test supplies). In the meantime, it needs a ‘clear rationale for prioritising testing for patients and health workers’ (the latter ‘should take priority’) (23.3.20: 3) Closing UK borders ‘would have a negligible effect on spread’ (23.3.20: 2).

*******

The lockdown. On the 23^rd March 2020, the UK Prime Minister Boris Johnson declared: ‘From this evening I must give the British people a very simple instruction – you must stay at home’. He announced measures to help limit the impact of coronavirus, including police powers to support public health, such as to disperse gatherings of more than two people (unless they live together), close events and shops, and limit outdoor exercise to once per day (at a distance of two metres from others).

*******

Meeting 19 (26.3.20) follows the lockdown. SAGE describes its priorities if the R goes below 1 and NHS capacity remains under 100%: ‘monitoring, maintenance and release’ (based on higher testing); public messaging on mass testing and varying interventions; understanding nosocomial transmission and immunology; clinical trials (avoiding hasty decisions’ on new drug treatment in absence of good data) and ‘how to minimise potential harms from the interventions, including 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’ (26.3.20: 1-2). The optimistic scenario is 10,000 deaths from the first wave (SPIM-O Meeting paper 25.3.20: 4).

Meeting 20 Confirms RWC and optimistic scenarios (Meeting paper 25.3.20), but it needs a ‘clearer narrative, clarifying areas subject to uncertainty and sensitivities’ and to clarify that scenarios (with different assumptions on, for example, the R, which should be explained more) are not predictions (29.3.20).

Meeting 21 seeks to establish SAGE ‘scientific priorities’ (e.g. long term health impacts of COVID-19, including socioeconomic impact on health (including mental health), community testing, international work (‘comorbidities such as malaria and malnutrition) (31.3.20: 1-2). NHS to set up an interdisciplinary group (including science and engineering) to ‘understand and tackle nosocomial transmission’ in the context of its growth and urgent need to define/ track it (31.3.20: 1-2). SAGE to focus on testing requirements, not operational issues. It notes the need to identify a single source of information on deaths.

April 2020

The meetings in April highlight four recurring themes.

First, it stresses that it will not know the impact of lockdown measures for some time, that it is too soon to understand the impact of releasing them, and there is high risk of failure: ‘There is a danger that lifting measures too early could cause a second wave of exponential epidemic growth – requiring measures to be re-imposed’ (2.4.20: 1; see also 14.4.20: 1-2). This problem remains even if a reliable testing and contact tracing system is in place, and if there are environmental improvements to reduce transmission (by keeping people apart).

Second, it notes signals from multiple sources (including CO-CIN and the RCGP) on the higher risk of major illness and death among black people, the ongoing investigation of higher risk to ‘BAME’ health workers (16.4.20), and further (high priority) work on ‘ethnicity, deprivation, and mortality’ (21.4.20: 1) (see also: Race, ethnicity, and the social determinants of health).

Third, it highlights the need for a ‘national testing strategy’ to cover NHS patients, staff, an epidemiological survey, and the community (2.4.20). The need for far more testing is a feature of almost every meeting (see also The need to ramp up testing).

Fourth, SAGE describes the need for more short and long-term research, identifying nosocomial infection as a short term priority, and long term priorities in areas such as the long term health impacts of COVID-19 (including socioeconomic impacts on physical and mental health), community testing, and international work (31.3.20: 1-2).

Finally, it reflects shifting advice on the precautionary use of face masks. Previously, advisory bodies emphasized limited evidence of a clear benefit to the wearer, and worried that public mask use would reduce the supply to healthcare professionals and generate a false sense of security (compare with this Greenhalgh et al article on the precautionary principle, the subsequent debate, and work by the Royal Society). Even by April: ‘NERVTAG concluded that the increased use of masks would have minimal effect’ on general population infection (7.4.20: 1), while the WHO described limited evidence that facemasks are beneficial for community use (9.4.20). Still, general face mask use but could have small positive effect, particularly in ‘enclosed environments with poor ventilation, and around vulnerable people’ (14.4.20: 2) and ‘on balance, there is enough evidence to support recommendation of community use of cloth face masks, for short periods in enclosed spaces where social distancing is not possible’ (partly because people can be infectious with no symptoms), as long as people know that it is no substitute for social distancing and handwashing (21.4.20)

May 2020

In May, SAGE continues to discuss high uncertainty on relaxing lockdown measures, the details of testing systems, and the need for research.

Generally, it advises that relaxations should not happen before there is more understanding of transmission in hospitals and care homes, and ‘until effective outbreak surveillance and test and trace systems are up and running’ (14.5.20). It advises specifically ‘against reopening personal care services, as they typically rely on highly connected workers who may accelerate transmission’ (5.5.20: 3) and warns against the too-quick introduction of social bubbles. Relaxation runs the risk of diminishing public adherence to social distancing, and to overwhelm any contact tracing system put in place:

‘SAGE participants reaffirmed their recent advice that numbers of Covid-19 cases remain high (around 10,000 cases per day with wide confidence intervals); that R is 0.7-0.9 and could be very close to 1 in places across the UK; and that there is very little room for manoeuvre especially before a test, trace and isolate system is up and running effectively. It is not yet possible to assess the effect of the first set of changes which were made on easing restrictions to lockdown’ (28.5.20: 3).

It recommends extensive testing in hospitals and care homes (12.5.20: 3) and ‘remains of the view that a monitoring and test, trace & isolate system needs to be put in place’ (12.5.20: 1)

June 2020

In June, SAGE identifies the importance of clusters of infection (super-spreading events) and the importance of a contact tracing system that focuses on clusters (rather than simply individuals) (11.6.20: 3). It reaffirms the value of a 2-metre distance rule. It also notes that the research on immunology remains unclear, which makes immunity passports a bad idea (4.6.20).

It describes the result of multiple meeting papers on the unequal impact of COVID-19:

‘There is an increased risk from Covid-19 to BAME groups, which should be urgently investigated through social science research and biomedical research, and mitigated by policy makers’ … ‘SAGE also noted the importance of involving BAME groups in framing research questions, participating in research projects, sharing findings and implementing recommendations’ (4.6.20: 1-3)

See also: Race, ethnicity, and the social determinants of health

The full list of SAGE posts:

COVID-19 policy in the UK: yes, the UK Government did ‘follow the science’

Did the UK Government ‘follow the science’? Reflections on SAGE meetings

The role of SAGE and science advice to government

The overall narrative underpinning SAGE advice and UK government policy

SAGE meetings from January-June 2020

SAGE Theme 1. The language of intervention

SAGE Theme 2. Limited capacity for testing, forecasting, and challenging assumptions

SAGE Theme 3. Communicating to the public

COVID-19 policy in the UK: Table 2: Summary of SAGE minutes, January-June 2020

COVID-19 policy in the UK: The overall narrative underpinning SAGE advice and UK government policy

This post is part 3 of COVID-19 policy in the UK: Did the UK Government ‘follow the science’? Reflections on SAGE meetings (update: see the notes on Dominic Cummings’ tweets at the end)

I discuss the UK government’s definition of the COVID-19 policy problem in some other posts (1. in a now-dated post on early developments, and 2. in relation to oral evidence to the Health and Social Care committee). It includes the following elements:

• We need to use a suppression strategy to reduce infection enough to avoid overwhelming health service capacity, and shield the people most vulnerable to major illness or death caused by COVID-19, to minimize deaths during at least one peak of infection.
• We need to maintain suppression for a period of time that is difficult to predict, subject to compliance levels that are difficult to predict and monitor.
• We need to avoid panicking the public in the lead up to suppression, avoid too-draconian enforcement, and maintain wide public trust in the government.
• We need to avoid (a) excessive and (b) insufficient suppression measures, either of which could contribute to a second wave of the epidemic of the same magnitude as the first.
• We need to transition safely from suppression measures to foster economic activity, find safe ways for people to return to work and education, and reinstate the full use of NHS capacity for non-COVID-19 illness.
• In the absence of a vaccine, this strategy will likely involve social distancing and (voluntary) track-and-trace measures to isolate people with COVID-19.

This understanding in the UK, informed strongly by SAGE, also informs the ways in which SAGE (a) deals with uncertainty, and (b) describes the likely impact of each stage of action.

Manage suppression during the first peak to avoid a second peak

Most importantly, it stresses continuously the need to avoid excessive suppressive measures on the first peak that would contribute to a second peak [my emphasis added]:

• ‘Any combination of [non-pharmaceutical] measures would slow but not halt an epidemic’, 25.2.20: 1).
• ‘Mitigations can be expected to change the shape of the epidemic curve or the timing of a first or second peak, but are not likely to reduce the overall number of total infections’. Therefore, identify whose priorities matter (such as NHS England) on the assumption that, ‘The optimal shape of the epidemic curve will differ according to sectoral or organisational priorities’ (27.2.20: 2).
• ‘A combination of these measures [school closures, household isolation, social distancing] is expected to have a greater impact: implementing a subset of measures would be ideal. Whilst this would have a more moderate impact it would be much less likely to result in a second wave. In comparison combining stringent social distancing measures, school closures and quarantining cases, as a long-term policy, may have a similar impact to that seen in Hong Kong or Singapore, but this could result in a large second epidemic wave once the measures were lifted’ (Meeting paper 4.3.20a: 3).
• ‘SAGE was unanimous that measures seeking to completely suppress spread of Covid-19 will cause a second peak. SAGE advises that it is a near certainty that countries such as China, where heavy suppression is underway, will experience a second peak once measures are relaxed’ (also: ‘It was noted that Singapore had had an effective “contain phase” but that now new cases had appeared) (13.3.20: 2)
• Its visual of each possible peak of infection emphasises the risk of a second peak (Meeting paper 4.3.20: 2).

• ‘The objective is to avoid critical cases exceeding NHS intensive care and other respiratory support bed capacity’ … SAGE ‘advice on interventions should be based on what the NHS needs’ (16.3.20: 1)
• The fewer cases that happen as a result of the policies enacted, the larger subsequent waves are expected to be when policies are lifted (SPI-M-O Meeting paper 25.3.20: 1)
• ‘There is a danger that lifting measures too early could cause a second wave of exponential epidemic growth – requiring measures to be re-imposed’ (2.4.20: 1)

Avoid the unintended consequences of epidemic suppression

This understanding intersects with (c) an emphasis of the loss of benefits caused by certain interventions (such as schools closures).

• SPI-B (Meeting paper 4.3.20b: 1-4) expresses reluctance to close schools, partly to avoid the unintended consequences, including: displacement problems (e.g. school closures prompt children to be looked after by vulnerable older people, or parents to lose the chance to work); and, the unequal impact on poorer and single parent families (loss of school meals, lost income, lower internet access, exacerbating isolation and mental ill health). It then states that: ‘The importance of schools during a crisis should not be overlooked. This includes: Acting as a source of emotional support for children; Providing education (e.g. on hand hygiene) which is conveyed back to families; Provision of social service (e.g. free school meals, monitoring wellbeing); Acting as a point of leadership and communication within communities’ (4.3.20b: 4).
• ‘Long periods of social isolation may have significant risks for vulnerable people … SAGE agreed that a balance needs to be struck between interventions that theoretically have significant impacts and interventions which the public can feasibly and safely adopt in sufficient numbers over long periods. Input from behavioural scientists is essential to policy development of cocooning measures, to increase public practicability and likelihood of compliance … the public will face considerable challenges in seeking to comply with these measures, (e.g. poorer households, those relying on grandparents for childcare)’ (10.3.20: 2).
• After the lockdown (23.3.20), SAGE describes a priority regarding: ‘how to minimise potential harms from the interventions, including 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’ (26.3.20: 1-2).

Exhort and encourage, rather than impose

It also intersects with (d) a primary focus on exhortation and encouragement rather than the imposition of behavioural change (Table 1), largely based on the belief that the UK government would be unwilling or unable to enforce behavioural change in ways associated with China. In that context, the government’s willingness and ability to enforce social distancing and business closure from the 23^rd March is striking.

Examples include:

• when recommending ‘individual home isolation (symptomatic individuals to stay at home for 14 days) and whole family isolation (fellow household members of symptomatic individuals to stay at home for 14 days after last family member becomes unwell)’, it assumes a 50% compliance rate, and notes that ‘closing bars and restaurants ‘would have an effect, but would be very difficult to implement’ (5.3.20: 1).

See also: oral evidence to the Health and Social Care committee, which suggests that the UK government and SAGE’s problem definition contrasts with approaches in countries such as South Korea (described by Kim et al, and Kim).

It also contrasts with the approach described by several of the UK’s (expert) critics, including Professor Devi Sridhar (Professor of Global Public Health), who is critical of SAGE specifically, and more generally of the UK government’s rejection of an ‘elimination’ strategy:

Eradication -> getting rid of every case of COVID on the planet. Only possible in long-term w/vaccine.

Elimination -> moving towards 0 infections within national borders. Yes, imported cases will occur, but if identified quickly, can mean return to largely ‘normal’ daily life.

— Prof. Devi Sridhar (@devisridhar) June 22, 2020

We need a concerted push across the 4 nations for elimination. Here is how New Zealand did it & why it's feasible over the summer with political will, leadership & public willingness to comply. Only way to get back to 'normal' life. https://t.co/7BXf5gZJjz https://t.co/Qk7PM13XrO

— Prof. Devi Sridhar (@devisridhar) June 14, 2020

Too much short-sightedness & focus on 2 weeks ahead instead of 6-8 months ahead. Same issue back in Jan 2020.

— Prof. Devi Sridhar (@devisridhar) June 26, 2020

Table 1 sets out one way to describe the distinction between these approaches:

• The UK government is addressing a chronic problem, being cautious about policy change without supportive evidence, identifying trigger points to new approaches (based on incidence), and assuming initially that the approach is based largely on exhortation.
• One alternative is to pursue elimination aggressively, adopting a precautionary principle before there is supportive evidence of a major problem and the effectiveness of solutions, backed by measures such as contact tracing and quarantine, and assuming that the imposition of behaviour should be a continuous expectation.

One approach highlights the lack of evidence to support major policy change, and therefore gives primacy to the status quo. The other is more preventive, giving primacy to the precautionary principle until there is more clarity or certainty on the available evidence.

In that context, note (in Table 2) how frequently the SAGE minutes state that there is limited evidence to support policy change, and that an epidemic is inevitable (in other words, elimination without a vaccine is near-impossible). Both statements tend to support a UK government policy that was, until mid-March, based on reluctance to enforce a profound lockdown to impose social distancing.

As the next post describes, the chronology of Table 2 is instructive, since it demonstrates a degree of path dependence based on initial uncertainty and hesitancy. This approach was understandable at first (particularly when connected to an argument about reducing the peak of infection then avoiding a second wave), before being so heavily criticised only two months later.

The full list of SAGE posts:

COVID-19 policy in the UK: yes, the UK Government did ‘follow the science’

Did the UK Government ‘follow the science’? Reflections on SAGE meetings

The role of SAGE and science advice to government

The overall narrative underpinning SAGE advice and UK government policy

SAGE meetings from January-June 2020

SAGE Theme 1. The language of intervention

SAGE Theme 2. Limited capacity for testing, forecasting, and challenging assumptions

SAGE Theme 3. Communicating to the public

COVID-19 policy in the UK: Table 2: Summary of SAGE minutes, January-June 2020

Update 24.5.21

Dominic Cummings’ tweets 38-55 (22-24 May 2021) describe much of the initial UK Government approach (described above) as a ‘herd immunity’ strategy:

38/ Media generally abysmal on covid but even I’ve been surprised by 1 thing: how many hacks have parroted Hancock’s line that ‘herd immunity wasn’t the plan’ when 'herd immunity by Sep' was *literally the official plan in all docs/graphs/meetings* until it was ditched

— Dominic Cummings (@Dominic2306) May 22, 2021

51/ It was in week of 9/3 that we started to figure out Plan B to dodge herd immunity until vaccines. Even AFTER we shifted to PlanB, COBR documents had the ‘OPTIMAL single peak strategy’ graphs showing 260k dead cos the system was so confused in the chaos, see below pic.twitter.com/kXsgfkQdbF

— Dominic Cummings (@Dominic2306) May 23, 2021

I discuss here why I think ‘herd immunity’ has become a damagingly ambiguous term, used too loosely and misleadingly by too many people to help us understand what happened:

3. Defining the policy problem: ‘herd immunity’, long term management, and the containability of COVID-19 | Paul Cairney: Politics & Public Policy (wordpress.com)

However, clearly these tweets are crucial to our understanding of the influence of initial advice and strategies, based on the idea of acting to mitigate a first peak while avoiding a second.

COVID-19 policy in the UK: The role of SAGE and science advice to government

This post is part 2 of COVID-19 policy in the UK: Did the UK Government ‘follow the science’? Reflections on SAGE meetings

The issue of science advice to government, and the role of SAGE in particular, became unusually high profile in the UK, particularly in relation to four factors:

1. Ministers described ‘following the science’ to project a certain form of authority and control.
2. The SAGE minutes and papers – including a record of SAGE members and attendees – were initially unpublished, in line with the previous convention of government to publish after, rather than during, a crisis.

‘SAGE is keen to make the modelling and other inputs underpinning its advice available to the public and fellow scientists’ (13.3.20: 1)

When it agrees to publish SAGE papers/ documents, it stresses: ‘It is important to demonstrate the uncertainties scientists have faced, how understanding of Covid-19 has developed over time, and the science behind the advice at each stage’ (16.3.20: 2)

‘SAGE discussed plans to release the academic models underpinning SAGE and SPI-M discussions and judgements. Modellers agreed that code would become public but emphasised that the effort to do this immediately would distract from other analyses. It was agreed that code should become public as soon as practical, and SPI-M would return to SAGE with a proposal on how this would be achieved. ACTION: SPI-M to advise on how to make public the source code for academic models, working with relevant partners’ (18.3.20: 2).

SAGE welcomes releasing names of SAGE participants (if willing) and notes role of Ian Boyd as ‘independent challenge function’ (28.4.20: 1)

SAGE also describes the need for a better system to allow SAGE participants to function effectively and with proper support (given the immense pressure/ strain on their time and mental health) (7.5.20: 1)

3. There were growing concerns that ministers would blame their advisers for poor choices (compare Freedman and Snowdon) or at least use science advice as ‘an insurance policy’, and
4. There was some debate about the appropriateness of Dominic Cummings (Prime Minister Boris Johnson’s special adviser) attending some meetings.

Therefore, its official description reflects its initial role plus a degree of clarification on the role of science advice mechanisms during the COVID-19 pandemic. The SAGE webpage on the gov.uk sites describes its role as:

‘provides scientific and technical advice to support government decision makers during emergencies … SAGE is responsible for ensuring that timely and coordinated scientific advice is made available to decision makers to support UK cross-government decisions in the Cabinet Office Briefing Room (COBR). The advice provided by SAGE does not represent official government policy’.

Its more detailed explainer describes:

‘SAGE’s role is to provide unified scientific advice on all the key issues, based on the body of scientific evidence presented by its expert participants. This includes everything from latest knowledge of the virus to modelling the disease course, understanding the clinical picture, and effects of and compliance with interventions. This advice together with a descriptor of uncertainties is then passed onto government ministers. The advice is used by Ministers to allow them to make decisions and inform the government’s response to the COVID-19 outbreak …

The government, naturally, also considers a range of other evidence including economic, social, and broader environmental factors when making its decisions…

SAGE is comprised of leading lights in their representative fields from across the worlds of academia and practice. They do not operate under government instruction and expert participation changes for each meeting, based on the expertise needed to address the crisis the country is faced with …

SAGE is also attended by official representatives from relevant parts of government. There are roughly 20 such officials involved in each meeting and they do not frequently contribute to discussions, but can play an important role in highlighting considerations such as key questions or concerns for policymakers that science needs to help answer or understanding Civil Service structures. They may also ask for clarification on a scientific point’ (emphasis added by yours truly).

Note that the number of participants can be around 60 people, which is more like an assembly with presentations and a modest amount of discussion, than a decision-making function (the Zoom meeting on 4.6.20 lists 76 participants). Even a Cabinet meeting is about 20 and that is too much for coherent discussion/ action (hence separate, smaller, committees).

Further, each set of now-published minutes contains an ‘addendum’ to clarify its operation. For example, its first minutes in 2020 seek to clarify the role of participants. Note that the participants change somewhat at each meeting (see the full list of members/ attendees), and some names are redacted. Dominic Cummings’ name only appears (I think) on 5.3.20, 14.4.20, and two meetings on 1.5.20 (although, as Freedman notes, ‘his colleague Ben Warner was a more regular presence’).

More importantly, the minutes from late February begin to distinguish between three types of potential science advice:

1. to describe the size of the problem (e.g. surveillance of cases and trends, estimating a reasonable worst case scenario)
2. to estimate the relative impact of many possible interventions (e.g. restrictions on travel, school closures, self-isolation, household quarantine, and social distancing measures)
3. to recommend the level and timing of state action to achieve compliance in relation to those interventions.

SAGE focused primarily on roles 1 and 2, arguing against role 3 on the basis that state intervention is a political choice to be taken by ministers. Ministers are responsible for weighing up the potential public health benefits of each measure in relation to their social and economic costs (see also: The relationship between science, science advice, and policy).

Example 1: setting boundaries between advice and strategy

• ‘It is a political decision to consider whether it is preferable to enact stricter measures at first, lifting them gradually as required, or to start with fewer measures and add further measures if required. Surveillance data streams will allow real-time monitoring of epidemic growth rates and thus allow approximate evaluation of the impact of whatever package of interventions is implemented’ (Meeting paper 26.2.20b: 1)

This example highlights a limitation in performing role 2 to inform 3: SAGE would not be able to compare the relative impact of measures without knowing their level of imposition and its impact on compliance. Further, the way in which it addressed this problem is crucial to our interpretation and evaluation of the timing and substance of the UK government’s response.

In short, it simultaneously assumed away and maintained attention to this problem by stating:

• ‘The measures outlined below assume high levels of compliance over long periods of time. This may be unachievable in the UK population’ (26.2.20b: 1).
• ‘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’ (16.3.20: 1)

The assumption of high compliance reduces the need for SAGE to make distinctions between terms such as mitigation versus suppression (see also: Confusion about the language of intervention and stages of intervention). However, it contributes to confusion within wider debates on UK action (see Theme 1. The language of intervention).

Example 2: setting boundaries between advice and value judgements

• ‘SAGE has not provided a recommendation of which interventions, or package of interventions, that Government may choose to apply. Any decision must consider the impacts these interventions may have on society, on individuals, the workforce and businesses, and the operation of Government and public services’ (Meeting paper 4.3.20a: 1).

To all intents and purposes, SAGE is noting that governments need to make value-based choices to:

1. Weigh up the costs and benefits of any action (as described by Layard et al, with reference to wellbeing measures and the assumed price of a life), and
2. Decide whose wellbeing, and lives, matter the most (because any action or inaction will have unequal consequences across a population).

In other words, policy analysis is one part evidence and one part value judgement. Both elements are contested in different ways, and different questions inform political choices (e.g. whose knowledge counts versus whose wellbeing counts?).

[see also:

• ‘Determining a tolerable level of risk from imported cases requires consideration of a number of non-science factors and is a policy question’ (28.4.20: 3)
• ‘SAGE reemphasises that its own focus should always be on providing clear scientific advice to government and the principles behind that advice’ (7.5.20: 1)]

Future reflections

Any future inquiry will be heavily contested, since policy learning and evaluation are political acts (and the best way to gather and use evidence during a pandemic is highly contested).  Still, hopefully, it will promote reflection on how, in practice, governments and advisory bodies negotiate the blurry boundary between scientific advice and political choice when they are so interdependent and rely so heavily on judgement in the face of ambiguity and uncertainty (or ‘radical uncertainty’). I discuss this issue in the next post, which highlights the ways in which UK ministers relied on SAGE (and advisers) to define the policy problem.

The full list of SAGE posts:

COVID-19 policy in the UK: yes, the UK Government did ‘follow the science’

Did the UK Government ‘follow the science’? Reflections on SAGE meetings

The role of SAGE and science advice to government

The overall narrative underpinning SAGE advice and UK government policy

SAGE meetings from January-June 2020

SAGE Theme 1. The language of intervention

SAGE Theme 2. Limited capacity for testing, forecasting, and challenging assumptions

SAGE Theme 3. Communicating to the public

COVID-19 policy in the UK: Table 2: Summary of SAGE minutes, January-June 2020

COVID-19 policy in the UK: Did the UK Government ‘follow the science’? Reflections on SAGE meetings

SAGE is the Scientific Advisory Group for Emergencies. The text up there comes from the UK Government description. SAGE is the main venue to coordinate science advice to the UK government on COVID-19, including from NERVTAG (the New and Emerging Respiratory Virus Threats Advisory Group, reporting to PHE), and the SPI-M (Scientific Pandemic Influenza Group on Modelling) sub-groups on modelling (SPI-M) and behavioural public policy (SPI-B) which supply meeting papers to SAGE.

I have summarized SAGE’s minutes (41 meetings, from 22 January to 11 June) and meeting/ background papers (125 papers, estimated range 1-51 pages, median 4, not-peer-reviewed, often produced a day after a request) in a ridiculously long table. This thing is huge (40 pages and 20000 words). It is the sequoia table. It is the humongous fungus. Even Joey Chestnut could not eat this table in one go. To make your SAGE meal more palatable, here is a series of blog posts that situate these minutes and papers in their wider context. This initial post is unusually long, so I’ve put in a photo to break it up a bit.

Did the UK government ‘follow the science’?

I use the overarching question Did the UK Government ‘follow the science’? initially for the clickbait. I reckon that, like a previous favourite (people have ‘had enough of experts’), ‘following the science’ is a phrase used by commentators more frequently than the original users of the phrase. It is easy to google and find some valuable commentaries with that hook (Devlin & Boseley, Siddique, Ahuja, Stevens, Flinders, Walker, Lancet Infectious Diseases, FT; see also Vallance) but also find ministers using a wider range of messages with more subtle verbs and metaphors:

• ‘We will take the right steps at the right time, guided by the science’ (Prime Minister Boris Johnson, 3.20)
• ‘We will be guided by the science’ (Health Secretary Matt Hancock, 2.20)
• ‘At all stages, we have been guided by the science, and we will do the right thing at the right time’ (Johnson, 3.20)
• ‘The plan is driven by the science and guided by the expert recommendations of the 4 UK Chief Medical Officers and the Scientific Advisory Group for Emergencies’ (Hancock, 3.20)
• ‘The plan does not set out what the government will do, it sets out the steps we could take at the right time along the basis of the scientific advice’ (Johnson, 3.20).

Still, clearly they are saying ‘the science’ as a rhetorical device, and it raises many questions or objections, including:

1. There is no such thing as ‘the science’.

Rather, there are many studies described as scientific (generally with reference to a narrow range of accepted methods), and many people described as scientists (with reference to their qualifications and expertise). The same can be said for the rhetorical phrase ‘the evidence’ and the political slogan ‘evidence based policymaking’ (which often comes with its notionally opposite political slogan ‘policy based evidence’). In both cases, a reference to ‘the science’ or ‘the evidence’ often signals one or both of:

• a particular, restrictive, way to describe evidence that lives up to a professional quality standard created by some disciplines (e.g. based on a hierarchy of evidence, in which the systematic review of randomized control trials is often at the top)
• an attempt by policymakers to project their own governing competence, relative certainty, control, and authority, with reference to another source of authority

2. Ministers often mean ‘following our scientists’

When Johnson (12.3.20) describes being ‘guided by the science’, he is accompanied by Professor Patrick Vallance (Government Chief Scientific Adviser) and Professor Chris Whitty (the UK government’s Chief Medical Adviser). Hancock (3.3.20) describes being ‘guided by the expert recommendations of the 4 UK Chief Medical Officers and the Scientific Advisory Group for Emergencies’ (Hancock, 3.3.20).

In other words, following ‘the science’ means ‘following the advice of our scientific advisors’, via mechanisms such as SAGE.

As the SAGE minutes and meeting papers show, government scientists and SAGE participants necessarily tell a partial story about the relevant evidence from a particular perspective (note: this is not a criticism of SAGE; it is a truism). Other interpreters of evidence, and sources of advice, are available.

Therefore, the phrase ‘guided by the science’ is, in practice, a way to:

• narrow the search for information (and pay selective attention to it)
• close down, or set the terms of, debate
• associate policy with particular advisors or advisory bodies, often to give ministerial choices more authority, and often as ‘an insurance policy’ to take the heat off ministers.

3. What exactly is ‘the science’ guiding?

Let’s make a simple distinction between two types of science-guided action. Scientists provide evidence and advice on:

1. the scale and urgency of a potential policy problem, such as describing and estimating the incidence and transmission of coronavirus
2. the likely impact of a range of policy interventions, such as contact tracing, self-isolation, and regulations to oblige social distancing

In both cases, let’s also distinguish between science advice to reduce uncertainty and ambiguity:

• Uncertainty describes a lack of knowledge or a worrying lack of confidence in one’s knowledge.
• Ambiguity describes the ability to entertain more than one interpretation of a policy problem.

Put both together to produce a wide range of possibilities for policy ‘guided by the science’, from (a) simply providing facts to help reduce uncertainty on the incidence of coronavirus (minimal), to (b) providing information and advice on how to define and try to solve the policy problem (maximal).

If so, note that being guided by science does not signal more or less policy change. Ministers can use scientific uncertainty to defend limited action, or use evidence selectively to propose rapid change. In either case, it can argue – sincerely – that it is guided by science. Therefore, analyzing critically the phraseology of ministers is only a useful first step. Next, we need to identify the extent to which scientific advisors and advisory bodies, such as SAGE, guided ministers.

The role of SAGE: advice on evidence versus advice on strategy and values

In that context, the next post examines the role of SAGE.

It shows that, although science advice to government is necessarily political, the coronavirus has heightened attention to science and advice, and you can see the (subtle and not subtle) ways in SAGE members and its secretariat are dealing with its unusually high level of politicization. SAGE has responded by clarifying its role, and trying to set boundaries between:

• Advice versus strategy
• Advice versus value judgements

These aims are understandable, but difficult to do in theory (the fact/value distinction is impossible) and practice (plus, policymakers may not go along with the distinction anyway). I argue that it also had some unintended consequences, which should prompt some further reflection on facts-versus-values science advice during crises.

The ways in which UK ministers followed SAGE advice

With these caveats in mind, my reading of this material is that UK government policy was largely consistent with SAGE evidence and advice in the following ways:

1. Defining the policy problem

This post (and a post on oral evidence to the Health and Social Care Committee) identifies the consistency of the overall narrative underpinning SAGE advice and UK government policy. It can be summed up as follows (although the post provides a more expansive discussion):

1. coronavirus represents a long term problem with no immediate solution (such as a vaccine) and minimal prospect of extinction/ eradication
2. use policy measures – on isolation and social distancing – to flatten the first peak of infection and avoid overwhelming health service capacity
3. don’t impose or relax measures too quickly (which will cause a second peak of infection)
4. reflect on the balance between (a) the positive impact of lockdown (on the incidence and rate of transmission), (b) the negative impact of lockdown (on freedom, physical and mental health, and the immediate economic consequences).

While SAGE minutes suggest a general reluctance to comment too much on the point 4, government discussions were underpinned by 1-3. For me, this context is the most important. It provides a lens through which to understand all of SAGE advice: how it shapes, and is shaped by, UK government policy.

2. The timing and substance of interventions before lockdown, maintenance of lockdown for several months, and gradual release of lockdown measures

This post presents a long chronological story of SAGE minutes and papers, divided by month (and, in March, by each meeting). Note the unusually high levels of uncertainty from the beginning. The lack of solid evidence, available to SAGE at each stage, can only be appreciated fully if you read the minutes from 1 to 41. Or, you know, take my word for it.

In January, SAGE discusses uncertainty about human-to-human transmission and associates coronavirus strongly with Wuhan in China (albeit while developing initially-good estimates of R, doubling rate, incubation period, window of infectivity, and symptoms). In February, it had more data on transmission but described high uncertainty on what measures might delay or reduce the impact of the epidemic. In March, it focused on preparing for the peak of infection on the assumption that it had time to transition gradually towards a series of isolation and social distancing measures. This approach began to change from mid-March when it became clear that the number of people infected, and the rate of transmission, was much larger and faster than expected.

In other words, the Prime Minister’s declarations – of emergency on 16.3.20 and of lockdown on 23.3.20 – did not lag behind SAGE advice (and it would not be outrageous to argue that it went ahead of it).

It is more difficult to describe the consistency between UK government policy & SAGE advice in relation to the relaxation of lockdown measures.

SAGE’s minutes and meeting papers describe very low certainty about what will happen after the release of lockdown. Their models do not hide this unusually high level of uncertainty, and they use models (built on assumptions) to generate scenarios rather than estimate what will happen. In this sense, ‘following the science’ could relate to (a) a level of buy-in for this kind of approach, and (b) making choices when scientific groups cannot offer much (if any) advice on what to do or what will happen. The example of reopening schools is a key example, since SPI-M and SPI-B focused intensely on the issue, but their conclusions could not underpin a specific UK government choice.

There are two ways to interpret what happened next.

First, there will always be a mild gap between hesitant SAGE advice and ministerial action. SAGE advice tends to be based on the amount and quality of evidence to support a change, which meant it was hesitant to recommend (a) a full lockdown and (b) a release from lockdown. Just as UK government policy seemed to go ahead of the evidence to enter lockdown on the 23^rd March, so too does it seem to go ahead of the cautious approach to relaxing it.

Second, UK ministers are currently going too far ahead of the evidence. SPI-M papers state repeatedly that the too-quick release of measures will cause the R to go above 1 (in some papers, it describes reaching 1.7; in some graphs it models up to 3).

3. The use of behavioural insights to inform and communicate policy

In March, you can find a lot of external debate about the appropriate role for ‘behavioural science’ and ‘behavioural public policy’ (BPP) (in other words, using insights from psychology to inform policy). Part of the initial problem related to the lack of transparency of the UK government, which prompted concerns that ministers were basing choices on limited evidence (see Hahn et al, Devlin, Mills). Oliver also describes initial confusion about the role of BPP when David Halpern became mildly famous for describing the concept of ‘herd immunity’ rather than sticking to psychology.

External concern focused primarily on the argument that the UK government (and many other governments) used the idea of ‘behavioural fatigue’ to justify delayed or gradual lockdown measures. In other words, if you do it too quickly and for too long, people will tire of it and break the rules.

Yet, this argument about fatigue is not a feature of the SAGE minutes and SPI-B papers (indeed, Oliver wonders if the phrase came from Whitty, based on his experience of people tiring of taking medication).

Rather, the papers tend to emphasise:

• There is high uncertainty about behavioural change in key scenarios, and this reference to uncertainty should inform any choice on what to do next.
• The need for effective and continuous communication with citizens, emphasizing transparency, honesty, clarity, and respect, to maintain high trust in government and promote a sense of community action (‘we are all in this together’).

John and Stoker argue that ‘much of behavioural science lends itself to’ a ‘top-down approach because its underlying thinking is that people tend to be limited in cognitive terms, and that a paternalistic expert-led government needs to save them from themselves’. Yet, my overall impression of the SPI-B (and related) work is that (a) although SPI-B is often asked to play that role, to address how to maximize adherence to interventions (such as social distancing), (b) its participants try to encourage the more deliberative or collaborative mechanisms favoured by John and Stoker (particularly when describing how to reopen schools and redesign work spaces). If so, my hunch is that they would not be as confident that UK ministers were taking their advice consistently (for example, throughout table 2, have a look at the need to provide a consistent narrative on two different propositions: we are all in this together, but the impact of each action/inaction will be profoundly unequal).

Expanded themes in SAGE minutes

Throughout this period, I think that one – often implicit – theme is that members of SAGE focused quite heavily on what seemed politically feasible to suggest to ministers, and for ministers to suggest to the public (while also describing technical feasibility – i.e. will it work as intended if implemented?). Generally, it seemed to anticipate policymaker concern about, and any unintended public reactions, to a shift towards more social regulation. For example:

‘Interventions should seek to contain, delay and reduce the peak incidence of cases, in that order. Consideration of what is publicly perceived to work is essential in any decisions’ (25.2.20: 1)

Put differently, it seemed to operate within the general confines of what might work in a UK-style liberal democracy characterised by relatively low social regulation. This approach is already a feature of The overall narrative underpinning SAGE advice and UK government policy, and the remaining posts highlight key themes that arise in that context.

They include how to:

• describe different types of intervention, such as contain, delay, mitigate and suppress
• deal with limitations, such as to (a) test to inform epidemiological forecasts and policy, and (b) challenge assumptions and conclusions within SAGE
• communicate risk and policy change to the public.

Delaying the inevitable

All of these shorter posts delay your reading of a ridiculously long table summarizing each meeting’s discussion and advice/ action points (Table 2, which also includes a way to chase up the referencing in the blog posts: dates alone refer to SAGE minutes; multiple meeting papers are listed as a, b, c if they have the same date stamp rather than same authors).

The full list of SAGE posts:

COVID-19 policy in the UK: yes, the UK Government did ‘follow the science’

Did the UK Government ‘follow the science’? Reflections on SAGE meetings

The role of SAGE and science advice to government

The overall narrative underpinning SAGE advice and UK government policy

SAGE meetings from January-June 2020

SAGE Theme 1. The language of intervention

SAGE Theme 2. Limited capacity for testing, forecasting, and challenging assumptions

SAGE Theme 3. Communicating to the public

COVID-19 policy in the UK: Table 2: Summary of SAGE minutes, January-June 2020

Further reading

It is part of a wider project, in which you can also read about:

• The early minutes from NERVTAG (the New and Emerging Respiratory Virus Threats Advisory Group)
• Oral evidence to House of Commons committees, beginning with Health and Social Care

I hope to get through all of this material (and equivalent material in the devolved governments) somehow, but also to find time to live, love, eat, and watch TV, so please bear with me if you want to know what happened but don’t want to do all of the reading to find out.

If you would rather just read all of this discussion in one document:

The whole thing in PDF

Table 2 in PDF

The whole thing as a Word document

Table 2 as a word document

If you would like some other analyses, compare with:

• Freedman (7.6.20) ‘Where the science went wrong. Sage minutes show that scientific caution, rather than a strategy of “herd immunity”, drove the UK’s slow response to the Covid-19 pandemic’. Concludes that ‘as the epidemic took hold the government was largely following Sage’s advice’, and that the government should have challenged key parts of that advice (to ensure an earlier lockdown).
• More or Less (1.7.20) ‘Why Did the UK Have Such a Bad Covid-19 Epidemic?’. Relates the delays in ministerial action to inaccurate scientific estimates of the doubling time of infection (discussed further in Theme 2).
• Both Freedman and More or Less focus on the mishandling of care home safety, exacerbated by transfers from hospital without proper testing.
• Snowden (28.5.20) ‘The lockdown’s founding myth. We’ve forgotten that the Imperial model didn’t even call for a full lockdown’. Challenges the argument that ministers dragged their feet while scientists were advising quick and extensive interventions (an argument he associates with Calvert et al (23.5.20) ‘22 days of dither and delay on coronavirus that cost thousands of British lives’). Rather, ministers were following SAGE advice, and the lockdown in Italy had a far bigger impact on ministers (since it changed what seemed politically feasible).
• Greg Clark MP (chair of the House of Commons Science and Technology Committee) Between science and policy – Scrutinising the role of SAGE in providing scientific advice to government

As part of our #Covid19 series, @draccoops and @cptyler (@UCLSTEaPP) look at UK’s science advisory system and at the lessons we should learn from the current crisis, to improve how science and #evidence inform public policy making #scipol #sciadvice https://t.co/XpZ1iXQKKF

— CfSocialScience (@CfSocialScience) June 1, 2020

COVID-19 policy in the UK: yes, the UK Government did ‘follow the science’

In this post, ‘following the science’ describes UK ministers taking the advice of their scientific advisers and SAGE (the Scientific Advisory Group for Emergencies).

If so, were UK ministers ‘guided by the expert recommendations of the 4 UK Chief Medical Officers and the Scientific Advisory Group for Emergencies’?

The short answer is yes.

They followed advice in two profoundly important ways:

1. Defining coronavirus as a policy problem.

My reading of the SAGE minutes and meeting papers identifies the consistency of the overall narrative underpinning SAGE advice and UK government policy. It can be summed up as follows:

1. coronavirus represents a long term problem with no immediate solution (such as a vaccine) and minimal prospect of extinction/ eradication
2. use policy measures – on isolation and social distancing – to flatten the first peak of infection and avoid overwhelming health service capacity
3. don’t impose or relax measures too quickly (which will cause a second peak of infection)
4. reflect on the balance between (a) the positive impact of lockdown (on the incidence and rate of transmission), (b) the negative impact of lockdown (on freedom, physical and mental health, and the immediate economic consequences).

If you examine UK ministerial speeches and SAGE minutes, you will find very similar messages: a coronavirus epidemic is inevitable, we need to ease gradually into suppression measures to avoid a second peak of infection as big as the first, and our focus is exhortation and encouragement over imposition.

2. The timing and substance of interventions before lockdown

I describe a long chronological story of SAGE minutes and papers. Its main theme is unusually high levels of uncertainty from the beginning. The lack of solid evidence, available to SAGE at each stage, should not be dismissed.

In January, SAGE discusses uncertainty about human-to-human transmission and associates coronavirus strongly with Wuhan in China. In February, it had more data on transmission but described high uncertainty on what measures might delay or reduce the impact of the epidemic. In March, it focused on preparing for the peak of infection on the assumption that it had time to transition gradually towards a series of isolation and social distancing measures. This approach began to change from mid-March when it became clear that the number of people infected, and the rate of transmission, was much larger and faster than expected.

Therefore, the Prime Minister’s declarations – of emergency on 16.3.20 and of lockdown on 23.3.20 – did not lag behind SAGE advice. It would not be outrageous to argue that it went ahead of that advice, at least as recorded in SAGE minutes and meeting papers (compare with Freedman, Snowden, More or Less).

The long answer

If you would like the long answer, I can offer you 35280 words, including a 22380-word table summarizing the SAGE minutes and meeting papers (meetings 1-41, 22.1.20-11.6.20).

It includes:

• The obligatory discussion of following the science: there is no such thing as the science, these phrases are rhetorical, they mean being guided by scientific advisers, they use this language to project governing competence, and they often do not take the advice of scientists.
• The ways in which SAGE tries to address its unusually politicized role, such as to try to distinguish between advice versus policy (strategy and values)
• A more nuanced (yes, I mean very long) account of policy and scientific advice, including the ways in which UK ministers are going ahead of the evidence to relax lockdown measures.
• Wider themes that arise from the SAGE minutes and other sources: the confusing language of intervention, the limitations to scientific advice (such as limited data and testing at key moments), and the role of ‘behavioural science’ in public communication.
• A detailed breakdown of each meeting and meeting paper (125 papers, estimated range 1-51 pages, median 4 pages, not-peer-reviewed, often produced a day after a request). At the end is a note on the abbreviations and all the words I had to google (including nosocomial, paucisymptomatic, and battle rhythm).

The full list of SAGE posts:

COVID-19 policy in the UK: yes, the UK Government did ‘follow the science’

Did the UK Government ‘follow the science’? Reflections on SAGE meetings

The role of SAGE and science advice to government

The overall narrative underpinning SAGE advice and UK government policy

SAGE meetings from January-June 2020

SAGE Theme 1. The language of intervention

SAGE Theme 2. Limited capacity for testing, forecasting, and challenging assumptions

SAGE Theme 3. Communicating to the public

COVID-19 policy in the UK: Table 2: Summary of SAGE minutes, January-June 2020

Further reading

So far, the wider project includes:

• The early minutes from NERVTAG (the New and Emerging Respiratory Virus Threats Advisory Group)
• Oral evidence to House of Commons committees, beginning with Health and Social Care

I am also writing a paper based on this post, but don’t hold your breath.

8. Race, ethnicity, and the social determinants of health

The beginning of this section comes from: Coronavirus and the ‘social determinants’ of health inequalities: lessons from ‘Health in All Policies’ initiatives

A ‘social determinants’ focus shows that the most profound impacts on population health can come from (a) environments largely outside of an individual’s control (e.g. in relation to threats from others, such as pollution or violence), (b) access to high quality education and employment, and (c) economic inequality, influencing access to warm and safe housing, high quality water and nutrition, choices on transport, and access to safe and healthy environments.

In that context, the coronavirus also provides stark examples of major inequalities in relation to self-isolation and social distancing: some people have access to food, private spaces to self-isolate, and open places to exercise away from others; many people have insufficient access to food, no private space, and few places to go outside.

Corburn et al’s (2014) study of Richmond, California’s, focusing on ‘coproducing health equity in all policies’ highlights the strong connection between health and income and wealth, which differs markedly according to race and immigration status. It reports on a series of community discussions to identify key obstacles to health:

“emerging from the workshops and health equity discussions was that one of the underlying causes of the multiple stressors experienced in Richmond was structural racism. By structural racism we meant that seemingly neutral policies and practices can function in racist ways by disempowering communities of color and perpetuating unequal historic conditions” (2014: 627-8).

In the UK, there has been some political attention devoted on the impact of coronavirus according to race and ethnicity, albeit generally described with the problematic catch-all term BAME (Black, Asian, and minority ethnic) to refer to all non-white populations.

Most notably, the PHE report Disparities in the risk and outcomes of COVID-19 highlights the unequal impact of coronavirus, with an action plan delayed, but expected to follow.

This inequality is discussed somewhat in committee proceedings, including in relation to:

• Walton (1.5.20: q3) on concerns for BAME pregnant women and NHS staff
• Owen (5.5.20: q424) on the social determinants of health inequalities
• Owen (14.5.20: q95, q100) on the poor fit of PPE for women and BAME women
• Owatemi (14.5.20: q99):

‘In a survey of over 2,000 BAME NHS staff, 50% stated that there was a culture of discrimination within the NHS. They felt that they were unable to speak up due to the lack of BAME representation in leadership roles. Currently, only 6% of NHS leadership positions are BAME staff’

"Beyond the data" The government has now published the 70 page literature review and community engagement exercise that was part of the Public Health England review of ethnic disparities, a fortnight after publishing the first part of the reporthttps://t.co/mHeaDqKxcs

— Sunder Katwala (@sundersays) June 16, 2020

COVID-19 policy in the UK: oral evidence to the Health and Social Care Committee (5th March- 3rd June 2020)

1. The need to ramp up testing (for many purposes)
2. The inadequate supply of personal protective equipment (PPE)
3. Defining the policy problem: ‘herd immunity’, long term management, and the containability of COVID-19
4. Uncertainty and hesitancy during initial UK coronavirus responses
5. Confusion about the language of intervention and stages of intervention
6. The relationship between science, science advice, and policy
7. Lower profile changes to policy and practice
8. Race, ethnicity, and the social determinants of health

7. Lower profile changes to policy and practice

A future series of posts will describe the many ways in which policy will (or should) change in practice, as public sector and other organizations change the way they do things in response to crisis. Current examples include the relaxation or postponement of high-political-stakes issues such as NHS targets and testing in schools, and as-important but lower stakes issues such as mental health self-management and the local authority obligation to provide social care.

In relation to oral evidence, examples include:

• Stevens (17.3.20: q150-3) discusses relaxations on rules for GP prescribing, paying GPs upfront in relation to contracts, reducing the regularity of Care Quality Commission checks, and taking a more flexible approach to A&E and other targets to avoid their inevitable unintended consequences. The changes take place in the context of a reduced-capacity NHS and growing waiting list for services not met during the lockdown (see Hopson, Dixon (Chief Executive The Health Foundation), and Edwards (Chief Executive The Nuffield Trust) 14.5.20: q75-84).
• Taiwo Owatemi MP (17.4.20: q371) on medicinal cannabis for children.
• Hancock (17.4.20: q373) on the availability and operation of cancer services (compare with Rosie Cooper MP, 17.4.20: q380; Palmer, National Cancer Director NHS England, 1.5.20: q21-35; Murray, Chief Executive The King’s Fund, 14.5.20: q73)
• Dean Russell MP (17.4.20: q386):

‘Last November, the Committee [Joint Committee on Human Rights] identified that human rights were being abused for people with learning disabilities and/or autism in mental health hospitals. As part of that, one of the concerns is that with coronavirus, family visits are currently being restricted and routine inspections have been suspended, which in turn potentially increases the young people’s isolation and also makes them more vulnerable to abuse of their rights.’

• Walton (Chief Executive, Royal College of Midwives) (1.5.20: q11) on ‘domestic abuse increases during pregnancy’ and ‘it appears that during lockdown domestic abuse and control issues have increased’.
• Murdoch (National Mental Health Director, NHS England) (1.5.20: q47-51) on the temporary reduction in referrals to child and adult mental health services, followed by a general consensus (MPs and witnesses, 14.5.20: q87-90) that adult and child mental health services were poorly funded anyway, with too few staff, dealing mostly with emergencies, so the post-pandemic provision is a major worry since there will be the latent demand plus new causes of mental health problems.

COVID-19 policy in the UK: oral evidence to the Health and Social Care Committee (5th March- 3rd June 2020)

1. The need to ramp up testing (for many purposes)
2. The inadequate supply of personal protective equipment (PPE)
3. Defining the policy problem: ‘herd immunity’, long term management, and the containability of COVID-19
4. Uncertainty and hesitancy during initial UK coronavirus responses
5. Confusion about the language of intervention and stages of intervention
6. The relationship between science, science advice, and policy
7. Lower profile changes to policy and practice
8. Race, ethnicity, and the social determinants of health

6. The relationship between science, science advice, and policy

There is a lot written in general about the extent to which UK policy is evidence-based (go to EBPM and this article and I’ll see you in a few days).

This issue was initially a big feature of the UK government rhetoric in March, in which the idea of ministers ‘following the science’ (or the advice of advisers and bodies such as Royal Colleges – Hancock, 17.4.20: q312) can be used to project a certain form of authority and control (see Weible et al).

It prompted regular debate on the extent to which scientific advisory bodies were subject to group-think and drawn from too-narrow pools of expertise (see for example Dingwall, Today programme 10.6.20, from 838am), and Vallance’s (17.3.20: q96) response:

‘If you thought SAGE and the way SAGE works was a cosy consensus of agreeing scientists, you would be very mistaken. It is a lively, robust discussion, with multiple inputs. We do not try to get everybody saying exactly the same thing. The idea is to look at the evidence and come up with the answers as best we can. There are sub-groups that work and feed into SAGE. The membership of SAGE changes, depending on what we are discussing. It is not as though it is the same group of people who always discuss all the topics; there are members who come for specific items’.

Then, when things began to go very wrong, commentators speculated about the extent to which ministers would blame their advisers for their policy and its timing. The latter problem became a regular feature in oral evidence. For example:

• Vallance (5.5.20: 392-6) states that (a) ‘SAGE does not make decisions. SAGE gives advice; it is an advisory body and Ministers of course have to make decisions’, and (b) they need some confidentiality to make sure that ministers get the information to make choices first (‘to be allowed time to make those decisions’).
• Vallance (5.5.20: q406) is heavy on the line that scientists only give advice, not make policy: ‘we give science advice and then Ministers have to make their decisions. All I can say is that the advice that we have given has been heard and has been taken by the Government. Clearly, what we do not give advice on is absolutely precise policy decisions or absolute timings on things. Those are decisions that Ministers must take on the basis of the science. The correct way of saying it is that the decisions are informed by science. They are not led by science, as you said in opening the question’.
• Vallance (5.5.20: Q407) describes how that advice may be presented when the scientists do not agree: ‘… our output is very much in the form of options, in the form of uncertainty and in the form of what could be done and what the potential consequences might be, not, “Here is the answer. Get on and do this.” That is not how it works.’

The UK’s nascent blame-game problem makes Costello’s (17.4.20: q298) suggestion of ‘a no-blame audit’ (‘where were the system errors that led us to have probably the highest death rates in Europe?’), to inform planning for the second wave, seem unrealistic. Open debates may be common in some scientific conferences (albeit not the ones I attend), but such learning is competitive and contested in adversarial political systems (see Dunlop, and Dunlop & Radaelli). I think this limitation helps explain Vallance’s (5.5.20: q390) reluctance to reflect openly on what he would do differently if he had better data on the doubling time of the virus in March (see also Harries, 5.5.20: q414-7 on excess deaths).

COVID-19 policy in the UK: oral evidence to the Health and Social Care Committee (5th March- 3rd June 2020)

1. The need to ramp up testing (for many purposes)
2. The inadequate supply of personal protective equipment (PPE)
3. Defining the policy problem: ‘herd immunity’, long term management, and the containability of COVID-19
4. Uncertainty and hesitancy during initial UK coronavirus responses
5. Confusion about the language of intervention and stages of intervention
6. The relationship between science, science advice, and policy
7. Lower profile changes to policy and practice
8. Race, ethnicity, and the social determinants of health

5. Confusion about the language of intervention and stages of intervention

The Health and Social Care Committee sessions highlight two types of confusion about how best to describe stages of government policy. In particular, the difference between mitigate/suppress represents either a profound change in UK government policy or trivial semantics.

The distinction between contain and delay measures.

The Committee probes the idea of a shift in early March from contain to delay, which taps into wider debates in which critics suggest that the UK gave up on containment too soon (discussed in post 1 in relation to testing). In two sessions, Whitty (5.3.20: q2-4 and q55; 16.3.20) emphasizes that the measures to contain and delay are very similar, so the distinction is misleading:

‘At this point in the early stage of delay, the actions are primarily ones of case finding and isolating cases that come from high-risk areas, which is to try to reduce the possibility of seeding into the community and therefore slow down the initiation of an epidemic, then try to pick up cases early and isolate them in hospital environments so we minimise the chance of transmission within hospitals’ (5.3.20: q55).

Vallance (5.5.20: q41) states that the initial plan was to isolate and track to contain the virus. Then, the UK had a ‘massive influx of cases, not from China but from all sorts of other places, partly because of the huge connectivity of the UK … once it went beyond that to being a pandemic you didn’t know where it was going to come from, and we got a very large number of cases coming in right the way across the country from multiple European sources somewhere around late February and detected in early March’.

The distinction between mitigation and suppression measures

It is common for UK media and social media discussions to highlight a U-turn from mitigation to suppression measures in mid-March, summed up to some extent by the Prime Minister’s exhortation to stay home (16.3.20) and then obligation to stay home (23.3.20).

During this time, the Imperial College COVID-19 Response Team report dated 16.3.20 described (a) ‘mitigation’ measures as likely to reduce UK deaths only from 500000 to 250000, and therefore not viable, prompting (b) the need for ‘suppression’ measures to reduce deaths to 5,600-48,000 over two years.

In contrast, Vallance (17.3.20: q67) describes the distinction as mere semantics:

‘It is a semantic difference, whether you call it suppression, delay or mitigation. The aim is exactly the same, which is, how do you keep this thing down, how do you keep it below the level at which you want to keep it, and how do you keep it down for long enough to ensure that you have managed to achieve suppression?’

Vallance suggests that the UK took suppression measures from 16.3.20 without using the word suppress (17.3.20: q84; see also 5.5.20: q432 ‘we are definitely in “mitigate” now’, which suggests that he says ‘mitigation’ when most might say ‘suppression’). Similarly, Hunt (17.3.20: q69) describes ‘the very dramatic social distancing measures that have now been announced’. Whitty is also keen in the 5.3.20 sessions to downplay the idea that there are distinct categories of action associated with different terms.

For me, these discussions highlight two main issues.

First, initial UK government policy was often confusing because its communication was poor. In this case, it seems that the meaning of each term was not agreed from the outset, contributing to some confusion among adviser, advisee, and commentator.

Second, and more importantly, they betray a lack of appreciation of the difference between measures in relation to their likely levels of implementation. Most notably, several discussions (17.3.20: q84-5) equate UK policy from 16.3.20 as the kinds of suppression measures associated with China or South Korea, despite a huge gulf in their level of enforcement (see also Harries, 5.5.20: q440: ‘People seem to think there was a lockdown moment, whereas in fact a series of interventions based on science were recommended’). In contrast, Costello (17.4.20: q298) favours a more compulsory form of isolation associated with ‘a lot of the Asian states’: ‘Just asking people to self-isolate will not achieve the quarantine that you want’.

COVID-19 policy in the UK: oral evidence to the Health and Social Care Committee (5th March- 3rd June 2020)

1. The need to ramp up testing (for many purposes)
2. The inadequate supply of personal protective equipment (PPE)
3. Defining the policy problem: ‘herd immunity’, long term management, and the containability of COVID-19
4. Uncertainty and hesitancy during initial UK coronavirus responses
5. Confusion about the language of intervention and stages of intervention
6. The relationship between science, science advice, and policy
7. Lower profile changes to policy and practice
8. Race, ethnicity, and the social determinants of health

4. Uncertainty and hesitancy during initial UK coronavirus responses

Vallance (17.3.20: q114) ‘I do not think any of us have seen anything like this. It is a first in not just a generation but potentially the first for 100 years. None of us has seen this. … This is a daily changing and unique situation where we are learning as we go along’.

Early UK discussions are characterized by the expression of uncertainty about what was happening (based on limited data and the questionable accuracy of the most-used model), and hesitancy about how quickly and substantively to respond. This combination of uncertainty and hesitancy informs continuous discussions about why the UK appeared to pursue a lockdown too late, contributing to an unusually high number of excess deaths.

However, it is worth keeping them separate – analytically – to compare uncertainty about (a) what is happening, and (b) what ministers and the public are willing to do about it (as described in the previous post, in relation to problem definition), which inform hesitancy in different ways. Either way, the wider context is that the UK government eventually introduced measures on social regulation that would have seemed unthinkable in the UK before 2020.

The NERVTAG notes show how much uncertainty there was in January 2020, with initial assessments of low risk before the virus spread to other countries and then the UK. Even by the early stages, and still in March, there was some hesitancy about recommending quarantine-style measures, and a tendency to focus on low impact or low social compliance as a way to reject new measures.

• Compare with Freedman 7.6.20 (‘Where the science went wrong. Sage minutes show that scientific caution, rather than a strategy of “herd immunity”, drove the UK’s slow response to the Covid-19 pandemic).

TROUBLE IN THE TROUGH: A story of certainty, science advice & the UK's lockdown policy.

1⃣ Before I start, decisions are hard, hindsight is easy.

This thread is to show how the process could be improved, and that science advice could do better in the future #scipol pic.twitter.com/7xXc81QefF

— Warren Pearce (@WarrenPearce) June 11, 2020

The oral evidence to the Health and Social Care committee

In the first oral evidence session in March, Whitty (5.3.20: q1) was still describing the virus in relation to China and only providing an initial mild warning that the chances of containment in China (followed by minimal global spread) are ‘slim to zero’ since it is ‘highly likely that there is some level of community transmission of this virus in the UK now’.

Similarly, Willett (Director for Acute Care, NHS England) (17.3.20: q175) described the sense that there was no perceived emergency (in WHO and UK statements) by the end of January, followed by the sense that information, advice, and policy was changing ‘literally every few days’.

The initial oral evidence shows that the science advice was primarily about how to inform and persuade people to change their behavior, focusing heavily on regular handwashing, followed by exhortations to self-isolate at home if feeling symptoms.

Whitty (5.3.20: 2-4) describes delaying the peak of the epidemic via ‘changes to society’ to (a) avoid it coinciding with ‘winter pressures on the NHS’ and boost the capacity to respond, (b) understand the virus better, and (c) hope that it ‘if you move into spring and summer, the natural rate of transmission may go down’ (as with respiratory viruses ‘like flu, colds and coughs’, in which people are less often in small enclosed spaces).

These early discussions emphasise the need for parliamentary and public discussion on more impositional measures, but with no strong push for anything like a lockdown (and, for example, some concern about the measures in South Korea not being acceptable in the UK – Whitty, 5.3.20: q5).

Even on 17.3.20, Vallance (q72) was describing waiting 2-3 weeks to find out the effect of the Prime Minister’s 16.3.20 message, hoping that it could keep the number of ‘excess deaths’ down to 20000 (and Vallance and Whitty had been describing pre-lockdown measures as quite extreme). The same day, Stevens (and Powis, National Medical Director, NHS England) described the PM’s hope that people would act according to the ‘good judgment and altruistic instincts of the British people’ without the need to impose social distancing (17.3.20: q176).

COVID-19 policy in the UK: oral evidence to the Health and Social Care Committee (5th March- 3rd June 2020)

1. The need to ramp up testing (for many purposes)
2. The inadequate supply of personal protective equipment (PPE)
3. Defining the policy problem: ‘herd immunity’, long term management, and the containability of COVID-19
4. Uncertainty and hesitancy during initial UK coronavirus responses
5. Confusion about the language of intervention and stages of intervention
6. The relationship between science, science advice, and policy
7. Lower profile changes to policy and practice
8. Race, ethnicity, and the social determinants of health

3. Defining the policy problem: ‘herd immunity’, long term management, and the containability of COVID-19

Note: it is worth reading this post together with COVID-19 policy in the UK: The overall narrative underpinning SAGE advice and UK government policy since both examine how the UK government defined the policy problem.

Frankly, the widespread and intense focus on the misleading phrase ‘herd immunity’ was a needless distraction, sparked initially by government advisors but then nitrous-turbo-boosted, gold-plated, and covered in neon lights during a series of ridiculous media and social media representations of ill-worded statements.

This initial focus took attention away from a much more profound discussion of what the UK government thinks is feasible, which informs a very stark choice: to define the COVID-19 problem as (a) a short term pandemic to be eradicated (as in countries like South Korea) or (b) a long term pandemic to be expected and managed every year (the definition in countries like the UK).

There is no ‘herd immunity strategy’

The key thing to note is that ministers and their advisors:

• Did talk in general terms about the idea of ‘herd immunity’ in March (best summed up as: herd immunity is only possible if there is a vaccine or enough people are infected and recover)
• Did not recommend an extreme non-intervention policy in which most of the population would be infected quickly to achieve herd immunity (in February, advisers described this outcome as the Reasonable Worst Case Scenario; see also Whitty, 5.3.20: q15-17)

Rather, describing the idea of herd immunity as an inevitability (not determined by choice) is key to understanding the UK approach. It helps us question the idea that there was a big policy U-turn in mid-March. Policy did change in the short term, but a sole focus on the short term distracts from the profound implications of its long-term strategy (in the absence of a vaccine) associated with phrases such as ‘flatten the curve’ (rather than ‘eradicate the virus’).

Examples of UK government representatives talking about herd immunity

1.      Wilful misrepresentation, often put to music

Full Fact’s challenge to the wilful misrepresentation of Prime Minister Boris Johnson’s appearance on the ITV programme This Morning (10.3.20): Here is the transcript of what Boris Johnson said on This Morning about the new coronavirus

These video stinkers, in which people (a) cut quotes so that you don’t hear the context, and provide a misleading headline, or (b) put a bunch of cut interviews in sequence and combine them with a tune that sounds like a knock-off version of the end credits to the TV Series The Hulk (in other words, people design these messages to get an emotional reaction).

2.      Headlines stoking the idea of herd immunity during a time when everyone should have been careful about how explain and interpret early discussions

British government wants UK to acquire coronavirus ‘herd immunity’, writes Robert Peston (12.3.20)

3.      The accentuation of a message not being emphasised by government spokespeople, at the expense of a message that requires more attention.

This interview is described by Sky News (13.3.20) as: ‘The government’s chief scientific adviser Sir Patrick Vallance has told Sky News that about 60% of people will need to become infected with coronavirus in order for the UK to enjoy “herd immunity”’. You might be forgiven for thinking that he was on Sky extolling the virtues of a strategy to that end. This was certainly the write-up in respected papers like the FT (UK’s chief scientific adviser defends ‘herd immunity’ strategy for coronavirus). Yet, he was saying nothing of the sort. Rather, when prompted, he discussed herd immunity in relation to the belief that COVID-19 will endure long enough to become as common as seasonal flu.

See Vallance’s interview on the same day (13.3.20) during Radio 4’s Today programme (transcribed by the Spectator and headlined as “How ‘herd immunity’ can help fight coronavirus” as if it is his main message). The Today Programme also tweeted only 30 seconds to single out that brief exchange. Yet, clearly his overall message – in this and other interviews – was that some interventions (e.g. staying at home; self-isolating with symptoms) would have bigger effects than others (e.g. school closures; prohibiting mass gatherings) during the ‘flattening of the peak’ strategy (‘What we don’t want is everybody to end up getting it in a short period of time so that we swamp and overwhelm NHS services’). Rather than describing ‘herd immunity’ as a strategy, he is really describing how to deal with its inevitability.

[PAC: Note that these examples are increasingly difficult to track, because people take the herd immunity argument for granted or cite reference to it misleadingly. For example, Scalley et al state “To widespread criticism, he floated an approach to “build up some degree of herd immunity” founded on an erroneous view that the vast majority of cases would be mild, like influenza”. Their citation takes you here, in which there is no reference to herd immunity or the quotation]

In oral evidence, Vallance (17.3.20: q70) compares these measures as follows (while describing the ‘confidence intervals’ as ‘quite wide’, q77):

‘The interventions we have made have all been modelled out – it is just modelling; we need to be aware of that – to say what effect they would have on the peak. In the first one we introduced, case isolation, you would expect to bring the peak down by about 20%. In the second one, whole household quarantine, you would expect to bring it down by about 25%. The social shielding of the elderly has less of an effect on the peak but a much bigger effect on the mortality, where you might expect it to be between 20% and 30%. General social distancing measures—as you said, quite extreme ones have now been introduced—would be expected to reduce the peak by about 50%. They are not necessarily all completely additive, but it tells you that together we should expect those to have a very significant effect on the peak, and we should start to see the rates come down in two or three weeks’ time. The ambition in any outbreak is to try to get the R0 value down below one. That is the value, on average, of what one person would do in infecting others. At the moment, the R0 value is somewhere between two and three, and the aim is to get it below one, at which time things start to decrease’

Further discussion of herd immunity in oral evidence

Throughout the evidence sessions, some MPs raise the herd immunity idea, but not energetically, and perhaps largely to allow government advisers to clarify their initial statements (e.g. 17.3.20: q112; 17.4.20: q303).

In May, Vallance (5.5.20: q389 and q404) makes it extra-clear that he was not advocating the ‘herd immunity’ idea associated with no intervention:

‘Q389 Chair: It sounds like there is still a degree of uncertainty as to whether an antibody leads to an immune response. Back on 13 March, you said the aim was to build up some kind of herd immunity where lots of people in the country had had the virus so that they could not catch it again. When you said that, which was nearly two months ago, what was your evidence then for the existence of that kind of immunity?

Sir Patrick Vallance: I should be clear about what I was trying to say, and, if I did not say it clearly enough, I apologise. What I was trying to say was that, in the absence of a therapeutic, the way in which you can stop a community becoming susceptible to this is through immunity. Immunity can be obtained either by vaccination or by people who have had the infection. We don’t know, as I said, exactly what degree of protection you get from natural infection, and we don’t yet have a vaccine. The second thing is that the higher the proportion of people in the community with immunity, the easier it is to control the disease and, ultimately, the easier it is to release measures. So vaccination becomes an important part of how you end up with protection, assuming you can get a decent immune response with vaccines, which we also do not know yet, but you would expect there to be some degree of immunity. The expectation is that antibody responses will correlate with immunity to some degree—maybe very high or maybe not so high. As to the degree of protection, whether it is to reduce the severity of the disease or to reduce the overall effect of the disease and the ability to catch the disease, we still have some work to do to find out about that.

Q404 ‘To reiterate, as I said at the beginning to Jeremy Hunt, my points about immunity were not actually about getting immunity through that route. My point has been clear from the outset that we need to suppress the peak, and keep the peak down flat below the level at which the NHS can cope, to protect the NHS and to make sure that we reduce deaths. That has been the strategy.’

See also:

“It’s not possible to stop everybody getting it and it’s also not desirable because you want some immunity in the population”

Sir Patrick Vallance, the UK's chief scientific adviser, says main aim is to protect vulnerable and elderly from coronavirushttps://t.co/uwJOIBtUWY pic.twitter.com/jQNw2g1HYN

— BBC News (UK) (@BBCNews) March 12, 2020

Government emails on herd immunity

[Update 15.10.20] The BBC used a Freedom of Information request to secure ‘every email sent by Sir Patrick [Vallance] and chief medical officer for England, Professor Chris Whitty, from the start of February to the start of June, containing the words “herd immunity”‘.

The BBC narrative is that some people think that the government was in favour of pursuing ‘herd immunity’ via high infection rates in the population (60%) that would contribute to hundreds of thousands of deaths.

For me, this is a misleading story.

The emails largely show that these advisors regret ever using the term ‘herd immunity’ because it allows people to jump to wild conclusions based on patching together minimal evidence (and ignoring more convincing evidence that advisors pushed strongly for suppression measures).

Dominic Cummings: herd immunity was the plan

(update 24.5.21) On the 22nd May 2021, Dominic Cummings (Prime Minister Boris Johnson’s former Special Advisor) tweeted to confirm that ‘herd immunity’ was government policy before a major policy shift in early March 2020:

As such, I think we are now at the stage of insiders arguing with insiders about (a) what happened, and (b) what ‘herd immunity’ means. It has prompted people to argue that their initial suspicions or reporting has been vindicated (e.g. Robert Peston). It has also prompted rejections of Cummings’ narrative by insiders such as Dr Jenny Harries. At the heart of this discussion is a combination of intentional and unintentional confusion about what ‘herd immunity’ means in relation to policy. For the most energetic government critics, it means an extreme hands-off measure to allow the epidemic to sweep through a population as quickly as possible. For me, it was a phrase used far too loosely to describe aspects of the mitigation strategy in place before lockdown (e.g. cocooning the most vulnerable during a wave of infection). For the government itself, it forms part of the emerging definition described below.

[See also: Who can you trust during the coronavirus crisis? ]

[Update 16.6.21 Cummings also released a lot of information on his substack (7000 words) in relation to his oral evidence (7 hours) to the House of Commons Science and Technology/ Health and Social Care committees]

Defining the COVID-19 problem in the UK

Greater clarity on key terms is essential. It allows us to think more about the implications of the UK government’s problem definition. In a much larger paper (that’s right – its completion is on the to-do list), I suggest that these elements inform the UK government’s definition of the policy problem by mid-March 2020:

1. We are responding to an epidemic that cannot be eradicated. Herd immunity is only possible if there is a vaccine or enough people are infected and recover.
2. We need to use a suppression strategy to reduce infection enough to avoid overwhelming health service capacity, and shield the people most vulnerable to major illness or death caused by COVID-19, to minimize deaths during at least one peak of infection.
3. We need to maintain suppression for a period of time that is difficult to predict, subject to compliance levels that are difficult to predict and monitor.
4. We need to avoid panicking the public in the lead up to suppression, avoid too-draconian enforcement, and maintain wide public trust in the government.
5. We need to avoid (a) excessive and (b) insufficient suppression measures, either of which could contribute to a second wave of the epidemic of the same magnitude as the first.
6. We need to transition safely from suppression measures to foster economic activity, find safe ways for people to return to work and education, and reinstate the full use of NHS capacity for non-COVID-19 illness. In the absence of a vaccine, this strategy will likely involve social distancing and (voluntary) track-and-trace measures to isolate people with COVID-19.
7. Any action or inaction has a profoundly unequal impact on social groups.

In other words, UK government policy is about reducing or moving the initial peak of infection, followed by longer term management to ensure that the NHS always has capacity to treat. The short-term focus emphasized the need to get the timing right in relation to the balance between public health benefits and social and economic cost (rather than to adopt a precautionary principle):

‘There is also timing. There will be quite a long period between knowing that we have an epidemic running at a reasonable rate and the actual peak. We are keen not to intervene until the point when we absolutely have to, so as to minimise the economic and social disruption on people, and then to stop it again as soon as we can afterwards. It is both the combination of what we need to do—in later questions we might want to go into some details about the things we can do—and the timing. The timing is critical. It is important that we minimise the social disruption while doing what we can to make sure we maximise the public health impact’ (Whitty, 5.3.20: q18)

[See also q39 on the unresolved difficulties of isolating vulnerable people physically without producing too high costs socially, and

q59 on ‘no need at this stage to be stocking up on anything. … this is going to be a marathon not a sprint. This is going to be a long period. There is going to be a lead time before the serious take-off of this comes, which we will be able to indicate … There is nothing in the current environment that would rationally lead someone to want to go out and stock up on stuff’.]

It is difficult to tell exactly what ministers and advisors expect to happen long-term in the absence of a vaccine (although Vallance 17.3.20: q102 is clear that the initial suppression measures will take an indeterminate number of months, not a few weeks). For example, are they managing infections and expecting regular deaths (assuming a mortality rate at approximately 1%) or expecting a high NHS capacity to reduce that mortality rate? Most discussions in public refer generally to the peak and NHS capacity but not the specifics:

‘Overall, the goal is clearly to bend the curve down and to make sure that the NHS capacity is there, and at the same time to do the work to try to improve our ability as a society to cope with this disease, with the goal of lifting some of the incredibly restrictive methods that we have had to place on the population in order to get the disease under control’ (Hancock, 17.4.20: q306)

The long-term implications of ‘flatten the curve’

In that context, Costello (17.4.20: q303) sums up my niggling concerns about the ‘flatten the curve’ message. This phrase suggests that we

• keep transmission low enough to make sure that the number of relevant cases does not overwhelm the NHS (an approach with high support), and
• accept that transmission will remain at a lower but significant rate until a vaccine is found (an approach that is not discussed as much, but it implies the continuation of deaths at a lower but regular rate):

‘The recent estimates, even from the chief scientific officer, are that after this wave, where we could see 40,000 deaths by the time it is over, we could have maybe only 10% to 15% of the population infected or covered, so the idea of herd immunity would mean maybe another five or six more waves to get to 60%. I do not think we should be using phrases such as “flatten the curve” because it implies continuing. We have to suppress this right down.’ (Costello, 17.4.20: q303)

The other side of this coin is that government advisers were initially working on the assumption that they could keep the initial number of deaths to 20000, which suggests a population infection rate well below 10% of the population (2/60m people, assuming the 1% mortality rate described by Whitty, 5.3.20: q11) and no expectation of herd immunity in the short term.

Comparing the UK definition with approaches in South Korea and China

The oral evidence sessions, probing the UK government’s longer term vision, help make key aspects of this definition somewhat clearer in two main ways.

First, they help confirm that UK policy is built on the assumption that COVID-19 will be a regular or seasonal problem (in the absence of a vaccine and culture change). For example, Hunt (17.3.20: q105) suggests that some country leaders think there will be (a) a peak of infection, then (b) containment, followed by (c) fizzling out:

‘China has officially announced that it thinks it is past the peak. The South Korean Foreign Minister was on TV at the weekend saying that she thinks South Korea has passed the peak, and it seems to think it has passed the peak with less than 1% of the population being infected. It is very realistic that there could be second or third waves, but it does not seem to be expecting it as much as you do … The Chinese and Korean view seems to be that it could be something like SARS, for example, which just burns itself out when the reproduction rate gets below one. Why is it that you are, unfortunately, so certain that it will come back?’

In contrast, Vallance (e.g. 17.3.20: q104-5) suggests strongly that COVID-19’s properties indicate high transmissibility and continuous recurrence (we may have to plan for a ‘spike every year’). This definition of the problem underpins the UK government’s expectation of long term management and, I think, is one of several reasons that ministers and advisers describe evaluation as premature.

Second, they suggest that this approach is built on a further assumption of what it feasible in the UK in relation to social behaviour.

A key element of international comparison relates to very different assumptions about social behaviour in each country. For example, the committee heard from respondents about experiences in South Korea, Hong Kong, Taiwan, and Singapore, in which previous pandemics – such as SARS – had a profound effect on government preparation and public behaviour (e.g. Comas-Herrera, 19.5.20: q446; Lum (Professor of Social Work and Social Administration, Hong Kong University) 19.5.20: q450, 456, 463; Chen (Former Vice-President of the Republic of China (Taiwan), 3.6.20: 492-504).

As a result, in many countries, you can expect widespread mask use and routine temperature checks, relatively invasive test and tracing measures, and obligatory isolation, to form part of a government’s response (such as to act quickly on regional ‘hotspots’ to prevent nationwide spread; compare with Hancock, 17.4.20: q318 on the connection of a UK lockdown to national unity, and Vallance, 5.5.20: q410-11 lukewarm on regional approaches, but also Harries, 5.5.20; q416 on the UK addressing hotspots in the earliest phase).

All of these measures and behaviours can contain the transmission of coronavirus in a way that seems to be far less feasible in the UK. For example, Doyle (26.3.20: q199-202) suggests that the South Korean system involves a degree of personal invasion not expected in UK, including giving bank details to government and being tested in public places like restaurants.

Further, even if these measures are possible, there is scepticism about their long-term impact: ‘the Chinese state and people are still doing some pretty extraordinary things’, with the potential that ‘when they take their foot off the brake the epidemic will surge back again’ (Whitty, 5.3.20: q13-4).

Instead, the emphasis from UK government respondents is initially (from March) about recommending the measures with the highest positive public health impact and lowest negative social and economic impact (handwashing). For example, Whitty (5.3.20: q18, see also q25, and Vallance 17.3.20: q92) compares measures:

‘ranging from those with almost no economic impact and high efficacy – top of the range being washing your hands and second being covering your mouth with a tissue when you cough – all the way down to those that have major societal impact, such as closing schools, which obviously affects children but also parents, potentially employment and particular sectors of the economy. It is very easy to choose a package of measures that is quite dramatic but has relatively little impact on the epidemic. We are very keen to avoid that, so we are modelling out all the combinations that we can because people’s livelihoods depend on it’.

Throughout, there is an emphasis on what might work in a UK-style liberal democracy characterised by relatively low social regulation, reinforced with reference to behavioural public policy:

‘All the behavioural science would suggest that we have to get the transparency right. We have to get the communication right. We have to trust that people want to know things, they want to know about this and they want to be able to be empowered to make their own decisions’ (Vallance, 17.3.20: 98).

COVID-19 policy in the UK: oral evidence to the Health and Social Care Committee (5th March- 3rd June 2020)

1. The need to ramp up testing (for many purposes)
2. The inadequate supply of personal protective equipment (PPE)
3. Defining the policy problem: ‘herd immunity’, long term management, and the containability of COVID-19
4. Uncertainty and hesitancy during initial UK coronavirus responses
5. Confusion about the language of intervention and stages of intervention
6. The relationship between science, science advice, and policy
7. Lower profile changes to policy and practice
8. Race, ethnicity, and the social determinants of health