Judging from some of today’s front pages (‘Time to fire up the engines, says PM’), you’d think Boris had announced the end of the lockdown yesterday. In fact, he announced… bugger all. For the time being, he said, the status quo would remain in place.

The message was the same as the one being peddled by Dominic Raab and other cabinet ministers at the daily Downing Street press briefings: the lockdown will remain in place until the government’s five tests have been met and in the meantime it would be irresponsible to set out an exit timetable. According to Boris, the most important of those tests is eliminating the risk of a “second spike”. If the easing of social distancing measures led to a second wave of infections that would not only be a humanitarian disaster, rendering all our sacrifices for nothing, it would also necessitate re-imposing restrictions and that, in turn, would be “an economic disaster”. So in order to protect the NHS, save lives and rebuild the economy, we have to remain under virtual house arrest for the foreseeable future. (You can read the full text of Boris’s speech here.)

This was deeply unimpressive from the prime minister. Where’s the evidence that phasing out the lockdown, but keeping more modest social distancing measures in place, would lead to a second wave? As Lyman Stone pointed out in an excellent essay in Public Discourse last week, the burden of proof isn’t on lockdown sceptics to show that they don’t work. Rather, the burden of proof is on the lockdown advocates to prove that they do.

“If you’re going to essentially cancel the civil liberties of the entire population for a few weeks, you should probably have evidence that the strategy will work,” he writes. “And there, lockdown advocates fail miserably, because they simply don’t have evidence.”

Stone is an economist specialising in population and demography and he looked at the effect of the lockdowns in Spain, Italy and France and compared them with the mitigation strategies pursued in Sweden and Holland. His conclusion? Lockdowns have made no difference when it comes to reducing Covid-19 deaths. To check this, he crunched the data from the US, comparing Covid mortality rates in those states that have locked down with those that haven’t and found that for every two weeks a “stay-at-home” order was in place the death rate actually increased by one person per 100,000. That is to say, lockdowns actually have a negative effect on Covid mortality. “We don’t need to have a national debate about whether the economic costs of lockdowns outweigh their public health benefits, because lockdowns do not provide public health benefits,” he concludes.

Stone’s analysis complements that of Carl Heneghan, Professor of Evidence-Based Medicine at Oxford, who argues that infections peaked in the UK before the lockdown was imposed. According to Professor Heneghan, the data shows infection rates halved when the government launched a public information campaign on 16 March urging people to wash their hands and keep two metres apart. That’s clear from the fact that deaths peaked on 8 April, almost exactly three weeks later, which is the lag time between infection and death. He believes that the government “lost sight” of the evidence and rushed into a nationwide quarantine on 23 March after being poorly advised by scientific experts who have been “consistently wrong” during the crisis. Professor Heneghan praises Sweden for “holding its nerve” and avoiding a “doomsday scenario”.

the rise and fall of infections and deaths follows the same pattern wherever the virus has struck, regardless of whether the country or region in question has imposed a lockdown or how severe that lockdown is.

In addition to this, there are the analyses showing that the rise and fall of infections and deaths follows the same pattern wherever the virus has struck, regardless of whether the country or region in question has imposed a lockdown or how severe that lockdown is. I’m thinking of this piece of work by Professor Isaac Ben-Israel, the head of the security studies programme at Tel Aviv university and the chairman of Israel’s National Council for Research and Development, showing that the virus largely vanishes 70 days after it strikes, no matter what strategies are put in place to contain it. There’s also this analysis by Phil Nuttridge, an ex-scientist with a masters degree in statistics, showing that infections peak between 31 and 33 days after cases first start appearing. He looked at six different European countries and detected the same pattern in all six irrespective of their varying population densities, testing rates, case levels and mortality rates, and in spite of the severity of the lockdowns they’ve imposed and when they imposed them.

Of course, there are plenty of epidemiologists and virologists who still believe lockdowns have been effective, such as Imperial College’s Professor Neil Ferguson, who seems to be emerging as the UK’s answer to Dr Strangelove. In an interview in UnHerd on Saturday, Professor Ferguson refused to row back on his previous claim that absent a lockdown 250,000 Britons would die of Covid-19. If we ease off now, he said, more than 100,000 would die before the end of the year, even if we continue to quarantine the elderly and the vulnerable. He reiterated that the only viable exit strategy is a vaccine.

In the interview, Professor Ferguson said a majority of epidemiologists are on his side in this debate – and that is probably true. But how reliable are the predictive computer models they’re using? The more scrutiny the Imperial College model is subjected to, the less well it holds up.

Followers of this debate will already be familiar with the scepticism of John Ioannidis, Professor of Medicine at Stanford, who says some of the assumptions and estimates built into the Imperial model are substantially inflated. But some may not be familiar with a new working paper published by the National Bureau of Economic Research. Jointly authored by a team of health economists from Harvard University and MIT, it’s highly critical of the predictive models used by epidemiologists in general – and particularly Professor Ferguson’s (you can read their paper here.). It’s quite technical, but thankfully Phillip Magness has summarised the main criticism in an article for the American Institute for Economic Research.

The gist is that the Imperial model predicts how many people are likely to die from Covid-19 in various different scenarios without allowing for the way in which people adapt to the risk of infection over time. For instance, the model claimed 510,000 people in the UK would die if the government did nothing to mitigate the impact of the virus – and 2.2 million in the US – ignoring the fact that as people learned more about the disease they would take more and more precautions, whether told to by the government or not. The authors of the NCER paper say of the ICL paper and four others based on epidemiological models, “in sum, the language of these papers suggests a degree of certainty that is simply not justified.”

But let’s suppose Professor Ferguson is right and an easing of the lockdown would see a rise in infections. Is that a good reason not to do it, as Boris Johnson said yesterday? The six million dollar question is whether the NHS would be overwhelmed if we returned to a mitigation strategy and the 16 March Imperial model is flawed in that respect because it underestimated the NHS’s emergency surge capacity (another instance of the model being insufficiently dynamic). Critical care capacity has now been significantly increased – the number of ICU beds in the UK has doubled, not counting the ones in the new Nightingale hospitals. According to the Imperial model, that means you could ease back on some of the extreme social distancing measures without overwhelming the NHS.

Take a look at the graph below in Professor Ferguson’s 16 March paper. The red line represents the NHS’s critical care capacity as it was then – eight per 100,000 members of the population. Even if we confine ourselves to the critical care capacity in existing hospitals and ignore the Nightingales, we can still move the red line up to 16 per 100,000 people. That pushes it above the orange line, which represents the estimated demand for critical care in a less extreme lockdown scenario, with schools and universities remaining open. Based on this, Boris Johnson could have announced yesterday that schools would reopen immediately without any risk of the NHS being overwhelmed.

Incidentally, schools are reopening in numerous countries that have imposed lockdowns as it becomes clear that children are at virtually no risk of dying of Covid-19. The argument for keeping schools closed is that asymptomatic schoolchildren might infect teachers, who are more vulnerable to the disease, and they might go on to infect others, thereby helping to trigger a second spike. But the latest research suggests otherwise. According to a paper by the National Centre for Immunisation Research and Surveillance (NCIRS) in Australia tracking the spread of Covid-19 in schools in New South Wales, the risk of schoolchildren infecting other, more vulnerable groups is negligible. “Our investigation found no evidence of children infecting teachers,” says Professor Kristine Macartney, Director of the NCIRS and the lead author of the paper. You can read more about that study here.

Okay, let’s suppose that, contrary to the evidence, reopening schools and switching off some of the other extreme measures did, in fact, overwhelm the NHS. Why not simply switch them back on again? What would be so economically disastrous about that? As far as I can tell, this argument crept in to the Number 10 press briefings last week without any accompanying explanation of how the government or its economic advisors reached this conclusion. They can’t fall back on the excuse that they’re just following the scientific advice, either. Professor Ferguson and his team recommended a strategy of switching measures on and off to manage infection flow after we’ve flattened the first wave in their 16 March paper. How could that possibly pose a greater risk to the economy than leaving the lockdown in place, given how catastrophic we know that will be? In his speech yesterday, Boris said, “I want to share all our working and our thinking, my thinking, with you, the British people.” Well, you could start by explaining that, prime minister.

Professor Ferguson claimed two-thirds of the people who’ve already succumbed to the virus would have died by the end of the year anyway, given their age and underlying health conditions.

In his UnHerd interview, Professor Ferguson claimed that more than 100,000 people would die of Covid-19 by the end of the year if we end the lockdown now and merely isolate the elderly and the vulnerable. But as he told the House of Commons Science and Technology Committee, two-thirds of the people who’ve already succumbed to the virus would have died by the end of the year anyway, given their age and underlying health conditions. Do we know the same doesn’t apply to those that would die of coronavirus if the lockdown was lifted? If it does, then Professor Ferguson’s argument is not that more than 100,000 people will die by the end of the year if we ease the lockdown, but about a third of that number. And for that to be a persuasive argument for maintaining the status quo, he’d have to show that keeping the lockdown in place will result in a lower loss of life, which is implausible given the impact on public health of the looming economic catastrophe.

Even if we ignore the toll on human life that the economic contraction will cause – not just here, but in the developing world, where a global economic recession will likely see tens of millions of people starve to death – there’s plenty of evidence that the lockdown has already had a negative effect on public health.

There was a good article in yesterday’s FT by Sebastian Payne documenting the costs the lockdown has incurred in terms of mental health, domestic violence and people failing to seek medical help for strokes, heart attacks and cancer symptoms. As he points out, one internal government estimate put the number of lives likely to be lost by the neglect of non-COVID-19 patients at 150,000. Among the costs he documents are:

• Refuge, a charity supporting victims of domestic violence, has seen a 49% rise to about 400 calls daily to its helpline since the lockdown began on 23 March.
• The Internet Watch Foundation, a charity based in Cambridge that monitors images of child abuse circulating online, has found a threefold increase in people seeking such content in the UK since March 23rd — from 100,000 to 300,000.
• The National Society for the Prevention of Cruelty to Children (NSPCC) has reported a dramatic spike in demand for its services. In the first week of April alone, the NSPCC took calls from 363 children suffering physical, sexual or emotional abuse or neglect.
• According to Mind, a charity that specialises in mental health, a third of the 4,000 people surveyed in a recent poll described their mental health as “poor or very poor,” while two-thirds said their state of mind had worsened over the past two weeks.
• St Thomas’, the hospital which treated the prime minister for coronavirus, has seen an 80% reduction in its minor injury work.
• Cancer Research UK says there has been a 75% drop in urgent referrals from GPs for those suspected of having the disease.
• According to Karol Sikora, dean of medicine at the university of Buckingham and an oncology consultant for 40 years, only 5,000 patients will be diagnosed with cancer in April 2020, down from 30,000 in a normal month.

The only straw to clutch at in Boris’s speech – the one being clutched by some of this morning’s papers – was the promise that preparations for a phased exit strategy were already under way. Let’s hope he recovers his nerve and announces that strategy later this week. The lockdown needs to end before it wreaks any more damage.