Nafeez Ahmed reveals how minutes from the NERVTAG estimated up to of 1.3 million fatalities from the Coronavirus as far back as 21 February.
Figures from the Government’s New and Emerging Respiratory Virus Threats Advisory Group (NERVTAG) reveal that, in late February, Government figures suggested that up to 1.3 million people could die in a “reasonable worst-case” scenario from the COVID-19 pandemic – yet no lockdown was declared until three weeks later.
NERVTAG is an expert committee of the Department of Health and Social Care (DHSC), which advises the Chief Medical Officer (CMO) and, through the CMO, ministers, DHSC and other Government departments. It is also part of the Government’s Scientific Advisory Group on Emergencies (SAGE).
On 16 March, ministers received advice from Imperial College modellers led by NERVTAG member Professor Neil Ferguson, who found that the Government’s stated approach at the time of simply ‘slowing’ the virus could result in as many as 250,000 deaths. The Imperial College paper had further warned that, in the absence of any control measures, an unmitigated epidemic in the UK would lead to 510,000 deaths in Britain.
It is widely believed that the formulation of these figures was the first time the Government had visibility of the catastrophic loss of life that could result from its own previous strategy. However, NERVTAG minutes dated 21 February reveal that, according to working assumptions at the time about the rate and deadliness of transmission provided to Government by its scientific advisors, anywhere between 833,313 and 1,333,330 Britons would potentially die.
Despite this, the Government did not pursue major social distancing actions for several weeks, allowing the Coronavirus to become endemic in the UK in that period. The NERVTAG minutes indicate that no advice had been sought from the scientific group on social distancing in that period up to the end of February.
I cross-checked these figures with Dr Deepti Gurdasani, a clinical epidemiologist and statistical geneticist at Barts and The London School of Medicine and Dentistry’s William Harvey Research Institute, Queen Mary University, and formerly a senior staff-scientist at the Wellcome Sanger Institute.
Dr Gurdasani, co-author of a new paper in The Lancet Global Health on how contact tracing could have suppressed the virus, told me that those large figures are “correct, based on their [NERVTAG’s] assumptions” at the time.
The NERVTAG minutes indicate that, in the Government’s reasonable worst-case scenario, some 80% of Britons would be infected with the virus. Of these, 50% would display illness, leaving 30% of the population asymptomatic – which accords to 37% of those infected.
At this time, NERVTAG’s estimation of the case fatality rate was 2.5-4% of people who become infected with COVID-19 and display symptoms. Applying that rate to NERVTAG’s figures indicates that, on 21 February, the Government’s scientific advisors’ own working assumptions suggested that between 833,313 and 1,333,330 could die in a reasonable worst-case scenario in the absence of action. These were the figures that NERVTAG had endorsed “for modelling purposes”.
‘Almost Certain there will be Sustained Transmission’
Despite this, there is no evidence from the NERVTAG minutes available up to the end of February that the group was ever asked about social distancing.
In fact, the only reference to social distancing was made on 3 February, when the meeting minutes indicate that NERVTAG is not “expected to comment” on the matter until possibly a later time: “JR [James Rubin] asked if there is a point about not attending school or work when ill? It was clarified that NERVTAG is not currently expected to comment on the question of social distancing at this time but could be asked this by DHSC at a later date”.
Yet the working assumptions about the virus discussed on 21 February did not prompt DHSC to ask the group about social distancing.
By 2 March, the Scientific Pandemic Influenza Group on Modelling (SPI-M) produced a consensus document describing estimates of transmission and fatality rates for the virus. The document shows that, at this time, the Government was told that the ‘Infection Fatality Rate’ – the number of people infected by COVID-19 who die – “is in the range of 0.5% to 1%, ranging from 0.01% in the under 20s to 8% in the over 80s”.
Applying the overall infection fatality rate to the expected 80% of people anticipated by NERVTAG to become infected, implied between 266,600 and 533,200 deaths. Those estimates are startlingly similar to the upper and lower bounds of the Imperial College model produced by Neil Ferguson and his team on 16 March – more than two weeks after this SPI-M analysis.
The SPI-M document also suggested a slightly revised case fatality rate (which would apply only to the number of symptomatic cases) of between 0.25 and 4%. As NERVTAG estimated some 50% of the population to become symptomatically infected, that put the lowest possible number of deaths at around 83,312, and the highest at 1.3 million, with considerable uncertainty – the final outcome likely depending on a range of other variables including healthcare facilities, intensive care unit (ICU) availability and so on.
The document also put forward a mortality rate of some 12% of those who are hospitalised. Once again, as NERVTAG was assuming that well over 4% of symptomatic cases (50% of the population) would result in hospitalisation, this indicated potentially more than 1.3 million people requiring hospital treatment. By SPI-M’s estimated 12% hospitalisation mortality rate, out of these, some 159,960 people would die.
The NERVTAG and SPI-M documents reveal that, based on their own working assumptions at the time, the Government and its scientific advisors were well aware of the risk of hundreds of thousands of deaths in the absence of a meaningful social distancing strategy, long before Imperial College publicly released its own modelling to this effect. This evidence was available to the Government as early as 21 February, some three weeks before a lockdown was finally declared, and became more refined by 2 March.
On that date, SPI-M’s consensus document warned: “It is highly likely that there is sustained transmission of COVID-19 in the UK at present. It is almost certain that there will be sustained transmission in the UK in the coming weeks.”
The question, then, is why did the Government refuse to introduce stronger social distancing measures until weeks later?
‘Feasibility of Effectiveness’
One answer might be provided by an earlier SPI-M document discussing behavioural interventions, dated 26 February – just five days after NERVTAG’s reasonable worst-case scenario indicated potentially 1.3 million deaths in the UK.
The document said: “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. It will likely not be feasible to provide estimates of the effectiveness of individual control measures, just the overall effectiveness of them all.
“An additional strategy would be to apply more intense measures on those age or risk groups at most risk of experiencing severe disease (e.g. household isolation of those over 65, special measures around care homes). The majority of the population would then develop immunity, hopefully preventing any second wave, while reducing pressure on the NHS. However, SPI-M-O has not looked at the likely feasibility or effectiveness of such methods.”
The document laid out three options – start with stricter measures then weaken them over time, start with weaker measures then strengthen them over time, or focus only on at risk groups allowing “the majority” of people to develop “immunity”.
But the document also revealed that the Government had no scientific basis to consider the “feasibility of effectiveness” of this approach – not least because there is not yet sufficient scientific evidence available to determine the quality of immunity acquired after recovering from COVID-19 infection, including how long it lasts and how effective it is. Without that evidence, there is no firm basis to conclude that ‘herd immunity’ is actually possible.
‘A Pragmatic Balance’
NERVTAG also already ruled out any kind of port of entry screening at UK borders on 13 January for travellers arriving from Wuhan, on the basis that: “If there is already exit screening in place in Wuhan, additional entrance screening in the UK is likely to have a low yield”. But, the minutes made clear that NERVTAG was unable to confirm details about exit screening in China and whether it was really effective.
By 28 January, although advisors expected that the process would be “robust”, NERVTAG still had no real idea what was going in Wuhan in terms of exit screening: “NERVTAG felt there was a lack of clarity on the exit screening process in Wuhan… as noted, there were no data on the implementation of this programme”.
Despite that, NERVTAG did not advise border screening, noting that “the benefit is very unlikely to outweigh the substantial effort, cost and disruption”.
The “cost” of preventive action may also have played a role in the Government’s decision not to ramp up contact tracing.
The NERVTAG minutes indicate that, despite recognising the importance of contact tracing to prevent and contain the epidemic, the Government eventually lost interest in doing so even as it began to expect that a large majority of Britons would get infected.
On 30 January, according to the minutes, Deputy Chief Medical Officer Dr Jonathan Van Tam “explained that the aim of case detection at this stage of the epidemic was to do everything possible strategically and operationally to prevent the establishment of community transmission in the UK – detection and isolation being the critical things”.
By 7 February, this goal was significantly watered down. The original language around doing “everything possible strategically and operationally” to prevent establishment of the disease in the UK was removed. Instead, the strategic objective outlined by the DHSC at the meeting was changed to preventing “early establishment of community transmission in the UK without overwhelming the health system.”
The goal had thus shifted from the wholesale prevention of the establishment of the disease in the UK to instead simply delaying it in a way to ensure that the health system would not be overwhelmed.
Background ‘cost’ calculations appear to have led the Government to adopt a limited case definition approach due to concerns about lack of testing capability.
On 28 January, Neil Ferguson asked DHSC officials at a NERVTAG meeting whether the goal is “to attempt to control the spread and limit it, in which case it would be logical to extend the definition to the whole of China regardless of testing capability”. The ensuing discussion converged on the decision that the case definition would be restricted to various provinces in China due to “capacity” issues. The minutes record the Chief Medical Officer and Deputy Chief Medical Officer insisting “that there would need to be a pragmatic balance between the sensitivity of the case definition and the capacity of the health system under the winter pressures”.
The goal was described by NERVTAG chair, Peter Horby, as being solely “to identify all symptomatic cases if possible and to interrupt transmission” – not to identify all cases including asymptomatic ones.
But, according to a peer-reviewed study by Dr Deepti Gurdasani that has just been published in The Lancet Global Health, the Government’s stance on testing seems to have been based on a deeply flawed mathematical model produced by some of its scientific advisors.
The new paper, co-authored with Dr Hisham Ziauddeen of the Wellcome Trust–Medical Research Council Institute of Metabolic Science, University of Cambridge, critiqued that earlier model led by Dr Joel Hellewell from the London School of Hygiene and Tropical Medicine.
The Hellewell paper is listed on the SAGE website as informing Government policy, and SAGE and NERVTAG member Dr John Edmunds, is a co-author on the model. “It is possible that this model may have informed the Government’s decision to cease testing and contact tracing on the 12 March,” said Dr Gurdasani.
On 13 March, Government advisor and model co-author Dr Edmunds appeared on Channel 4 News justifying the Government’s lack of action by explicitly referring to the idea of “herd immunity”. During the interview, he dismissed the goal of containing the virus through contact tracing, on the grounds that tracking down and isolating every infected person would be “incredibly difficult” to do for a “mild disease” such as COVID-19.
“When the genie is out of the bottle, the virus is all around the world and spreading, the only other way that the epidemic is going to come to a stop is achieving herd immunity,” Edmunds said. While acknowledging that many people could die after being infected, he claimed that “there’s no way out of that”. As a result, he said, the Government’s approach is to about “trying to stage it as best we can”.
The Hellewell paper of which Edmunds was a co-author claimed that “in most plausible scenarios”, “near perfect contact tracing” would be required for effective pandemic control using a case detection, contact tracing and isolation approach. The paper concluded that “contact tracing and isolation might not contain outbreaks of COVID-19 unless very high levels of contact tracing are achieved”.
But, according to Dr Gurdasani’s new paper, she told me that Hellewell and his team “didn’t consider all plausible scenarios, as they considered at least a 3.8 day delay between case detection and isolation, which is quite long. We’ve argued that even at that point a shorter turnaround was possible, not just due to development of rapid testing, but the fact that case detection and isolation can be done based on symptoms alone and doesn’t even require testing”.
If they had considered this, she said, “their own model would have suggested that pandemic outbreaks can be controlled effectively, even without near perfect contact tracing – their own model suggests 30-60% contacts traced would have been sufficient to achieve control”.
‘Interventions Beholden by Capacity’
Gurdasani and Ziaudeen’s work further suggests that the role of active testing, contact tracing, isolation and quarantine measures is responsible for “vast differences” in the number of deaths between, for instance, Italy and South Korea.
“South East Asian countries implemented testing, contact tracing, isolation and quarantine measures early on, with lockdown and school closures coming into place much later, if at all,” Gurdasani said. “The epidemic curve was flattened well before stringent lockdown measures were put in place, suggesting the measures taken prior to this, rather than lockdown, were effective in controlling spread. Testing, contact tracing and isolation has formed a key part of their pandemic response.”
The new paper has also elicited a response commentary from the team advising Government, but according to Gurdasani it misses the point: “We essentially contest their claim about this not being very effective in pandemic control without ‘near perfect contact tracing’. I agree that any intervention is more likely to be effective in combination with another intervention. That they mentioned this doesn’t take away from the fact that they likely underestimated the effectiveness of this intervention by not considering a scenario where people could be isolated quickly.”
But the response commentary provides a clue as to why the Government’s science advisors took this approach: “Governments and public health agencies will have to consider what scale of contact-tracing effort is logistically possible. In the USA, for example, an estimated 100,000 new contact-tracing workers would be needed to manage future COVID-19 epidemics.”
In short, Hellewell and his co-authors claim that Gurdasani and Ziaudeein’s argument does not factor in feasibility in terms of the number of staff needed to achieve adequate contact tracing – in other words, to quote from phrases that also come up in the NERVTAG minutes, “capacity” and “cost”.
“We rebut this point – as we believe that this approach was feasible, given other countries who didn’t have capacity scaled this up in response to what was needed,” Dr Gurdasani told me. As the cases in south-east Asia prove, “capacity” and “cost” were ramped up by the state to save lives and did not need to hamstring the pandemic response.
“We don’t think that scientists advising on utility of interventions should be beholden by capacity at a given time, when there is a pathway to scaling up, and others have clearly done so,” said Gurdasani.
The question, then, is: who in Government told its scientists to keep their models about pandemic response options “beholden” to concerns about costs to the state?