June 9, 2020

I want to start this piece by saying that I’ve donated my fee for it to the Malaria Consortium. I’ll get into exactly why at the end.  

Over the last week or so, we’ve seen pictures from all over the world of people gathering in large crowds to protest the murder of George Floyd in Minneapolis. In that time, the United States has seen its 100,000th death from Covid-19, and the UK has seen around a hundred new deaths announced each day.

The obvious question – to me, at least – is whether the risks of joining a protest are greater than the likely benefits. I thought that it might be useful for people to go through the risks as best I can. The key questions, it seems to me, are: how likely is it to spread in outdoor situations; and how many people are likely to have the virus in a large crowd.

First, it should be noted that the SARS-Cov2 virus seems to spread much less easily outside. Studies looking at superspreader events find far fewer such incidents outdoors than indoors — this one, in China in early April, looked at 318 different events where one person infected three or more, and found none happened outdoors. That said, Dr Babak Javid, an infectious disease specialist at Cambridge and Tshinghua universities, suggests that these studies were necessarily “severely biased, because the whole country had gone into some form of lockdown”, so there just weren’t many outdoor gatherings going on. It doesn’t mean it’s wrong, just that it needs to be treated with caution. Javid does say it seems very likely that being outdoors is of much lower risk.

Early in the UK outbreak, there were two big sporting events that went ahead despite widespread concerns — horseracing’s Cheltenham Festival, and the football match between Liverpool FC and Atletico Madrid, both in March. It struck me that they would be relatively useful parallels — large crowds, relatively tightly packed but outdoors — for the protests.

Unfortunately, there’s not a great deal that we know about them, despite media reports of extra deaths. As far as I can tell, there haven’t been any actual studies looking at them, apart from something by a consultancy firm called Edge Health. They claim a total of almost 13,000 infections and 78 deaths from the two events, but as far as I can tell they just compared the hospital nearest to the ground with another similar hospital in the region and saw whose deaths went up most. It seems extremely vulnerable to noise.

Dr Bharat Pankhania, a consultant in communicable diseases control at Exeter University, says that it’s an inherently difficult thing to study — people come to Cheltenham, and to Anfield, from all over the country, so any cases that developed from them will be recorded in their home NHS trust. “We didn’t have the mechanisms to collect data from the cases and ask them where they’d been,” he says. “So we haven’t been able to identify a common link to Cheltenham, which is a pity.” Good tracing of infected people’s movements wasn’t in place then, if it is in place now.

That said, he says it seems very likely that there was some infection and probably some deaths. Being outdoors is definitely a protection: “In an outdoor situation, there is a considerable dilution in virus particles.” You are just less exposed to it if the airborne water droplets are less concentrated.

But a large part of the benefit of being outdoors is the space. “Say you are out in the open at point A and I am sitting at point B three metres away, then it is highly unlikely that you are going to infect me,” says Dr Pankhania. “On the other hand if I am infectious at point A and I have somebody sitting by me within say a foot, cheek to jowl almost, then of course the situation is different.” 

Professor Tim Spector, an epidemiologist at King’s College London who heads their Covid symptom tracker app team, says that their data did show unusual spikes around Gloucestershire and Merseyside after those events. But there are differences between then and now which may make it less dangerous – partly behavioural, such as masks, and partly weather (viruses don’t like warm weather), but notably that there is simply less of the virus around. “Things that were spreading at the beginning of April would probably spread faster than they would now, when levels are low.” You’re simply less likely to come into contact with someone who has the virus now than you were at the peak of the infection.

That said, the levels aren’t that low. Dr Pankhania estimates that there are about 50,000 new cases — so roughly one person in every 1,000 in the UK — every week. And Prof Spector says his best guess for the prevalence of the disease in general is about one person in every 400, with wide variation — so perhaps one in every 200 in Manchester, and one in every 800 in the south-west.

Prof Spector used those numbers for some very quick thought experiments. Say you’re at a protest in an average city in the UK, and you come into prolonged close contact with 10 people during that time; there’d be about a one in 40 chance one of them would have the disease. We don’t know how likely you would be to catch it, but he suggested as a ballpark figure a 10 to 20% chance – let’s say 10%. That would suggest that your chance of getting the disease is about one in 400; higher in Manchester, lower in Bristol. Of course people vary in how likely they are to catch the disease, as well as how dangerous it is to them, but that’s the average.

I thought I’d take the thought experiment a little further. First, we’re interested in how likely you’d be to die, if you catch it; I think it’s probably pretty unlikely, since most people at the protests are probably young and healthy.

But of more concern is how likely you are to pass it on. Since the R value is hovering around 1 at the moment, that would imply that on average, you’ll pass it to one person. Of course it’s more complicated than that, and R in the community is different from R in care homes and hospitals, but let’s use it. We know that actually an R of 1 conceals wide variation — some people spread it to lots of people, most spread it to none — but again, let’s simplify.

That would imply that by attending a protest in an average UK city, you’d have about a one in 400 chance of getting the disease and passing it to another person. The infection fatality rate of the disease is hugely uncertain, but let’s take what I’d call a lower-end estimate of 0.5%. That would give us about a one in 80,000 chance of killing someone by attending a protest.

There are lots of uncertainties here; I’ve tried to err on the side of caution, but I could easily be off by an order of magnitude in either direction. And there are things people can do to reduce risk — wearing masks, obviously; definitely don’t go to any protests if you have any symptoms or have been in contact with a Covid-19 case. Dr Pankhania says “My advice to people who’ve been out protesting is that they sequester themselves, take themselves out of circulation, for 14 days.” 

It’s also worth noting that Dr Javid said that the impacts of the protests on a potential second wave would be real, but “less than what people will fear. The shouting and chanting may be bigger risks than the crowding”, and Prof Spector said that if a second wave does come, the protests will be just one small factor among several, such as the relaxation of lockdown. Dr Pankhania also emphasised that we shouldn’t “miss the woods for the trees” – he thinks the lifting of lockdown is premature, so “I wouldn’t want anyone to say the peak is only because of the protests”.

Still, it is absolutely the case that there is a risk. The question that everyone needs to answer for themselves is whether that risk is worth the potential benefit.

I can’t answer that for you. I can give you a number to work with – US police officers kill about 200 black people a year and some unknown number killed by other means. Last year, at least nine of these were unarmed. Hopefully the protests will prevent some percentage of them; you have to estimate what a realistic percentage is, and how likely it is that your attendance at a protest will make the difference. And, of course, hopefully it will have wider impacts on society, not just in the US, beyond just the deaths it prevents — although, of course, a new outbreak of coronavirus will have wider impacts as well, perhaps requiring a longer lockdown, greater economic hardship for the poor, and so on.

None of this is to say it’s not worth it. But I think it is vital to have a clear-eyed look at what the risks and benefits are likely to be. Having done that, it is entirely reasonable to conclude that the benefit is worth the risk, but you need to do that assessment first.

Now I’ll come back to what I said in the first paragraph. We all want to do good in the world. Right now, it seems almost impossible to know what the best way to do that is. Protesting racism is a good thing, but it could kill people; staying indoors to control the virus is a good thing, but you want to show solidarity with murdered black men.

I don’t know the right thing to do. When I plug my own assumptions in about my likely impact, I think the maths comes out pretty clearly that I’d be more likely to kill someone than save them, but your assumptions may be different.

But there are other ways to do good. Some of them are direct — Dr Pankhania suggests that we “bombard your MPs, your elected officials, your council representatives with emails”. But I suggest that, as well as that, if you can afford it, you donate what money you can to the charity that you think will do the most good.

There are plenty of excellent charities which work to combat racism in the UK – the Stephen Lawrence Charitable Trust and Runnymede are two which leap to mind. There are many equivalent ones in the US too.

Those are excellent choices, and I applaud them. But even though it may seem disconnected I’m going for something different. Charitable donation goes further in the developing world, so I’ve donated my fee for this column to the Malaria Consortium; it’s one of the charity evaluator GiveWell’s top charities, and it provides malaria prophylaxis and treatment to poor countries, mainly in sub-Saharan Africa. If you want to save black lives, and you don’t mind that it’s not in the UK or USA, then that strikes me as the best way to go.

GiveWell estimates that the charity saves one life for every $2,300 (£1,800) donated. Think about that, in the context of the numbers we were talking about earlier: your attendance at a rally has perhaps a one in 80,000 chance of killing someone, and, I would think, a rather lower chance of saving someone. But if you donate £100 to the Malaria Consortium, that’s equivalent to a one in 18 chance of saving a child’s life. That strikes me as an incredibly powerful statement in the light of the horror of George Floyd’s death. You can’t solve the world, but you can help make it better.

I don’t know if this will help others, but it helped me; it is a horrible sensation, feeling conflicted and powerless, and doing this made me feel a bit less of both. I realise this is virtue signalling, but I hope that the signal will encourage others to do something similar. 


*This article now says that Dr Pankhania estimates 50,000 new infections a week. An earlier version misstated this figure.