Professor Adam Kucharski, who is an Associate Professor and the Sir Henry Dale Fellow at the London School of Hygiene and Tropical Medicine, specialises in the mathematical analysis of infectious disease outbreaks. This puts him at the very heart of the debate on the global coronavirus pandemic , especially its spread and containment and mitigation questions.
He spoke to Narayan Lakshman about what we can learn from modelling this and prior infectious outbreaks, including parallels to the spread of fake news. Edited transcript:
In your book, The Rules of Contagion: Why Things Spread – and Why They Stop , you talk about Ronald Ross, a British doctor who in the late 19th century discovered that mosquitoes spread malaria and was awarded the 1902 Nobel Prize in physiology or medicine. He was credited with applying mathematics, biology and an understanding of societal forces to predict the progression pattern of an epidemic. How can we apply that to the coronavirus pandemic today, specifically for predicting an endgame scenario?
One of the key insights that Ross made was regarding the control of infection. He won the Nobel Prize for identifying that mosquitoes spread malaria. But then he was really interested in using mosquito control to reduce malaria and ideally cause it to come to zero in a particular area. At the time, people had this idea that you couldn’t control malaria until you could remove every single last mosquito. Ross used these very simple, conceptual models though, thinking that if you have so many mosquitoes you have a certain probability of them biting someone, and that person might get bitten by another mosquito and then it spreads. He thought about these processes and realised that actually you don’t need to remove every last mosquito. Once you get the density of mosquitoes low enough, the chances are that someone who is infected would recover before they pass it on to others.
Listen : The mathematics of an outbreak | The Hindu In Focus Podcast
We can think of a similar idea with a lot of the physical distancing measures that are coming in. Initially, there may still be some infectious people out there, but because these measures are in place, they will have the infection and recover before they pass it on to someone else. There will still be some interactions occurring between the population, because you have key industries, healthcare workers and other such groups that do need to interact with other people. But because the level of interactions is particularly low, that can control the infection. That’s what we’ve seen in places like Wuhan.
One idea that you’ve explored in your work is the concept of “dependent” and independent” happenings, particularly in the context of identifying what an epidemic is, and its potential to spread fast through a population. Please talk to us a bit about whether we have too many “dependencies” today – for example, relating to hyper-urbanisation, transportation networks, communications via the Internet, and interlocking economies and finance – does that make the contagion effects associated with an epidemic far worse?
Those kinds of connections can certainly help amplify the effects we see of an outbreak. In terms of dependent and independent happenings, we might think of independent happenings as random accidents that occur, and they are not linked, and one thing doesn’t feed into another. It was again Ronald Ross who came up with the idea of what we call “dependent happenings.” What happens to you depends on what happens to other people. Obviously, biological contagion is a good example of this, but we do see these knock-on effects happening in other ways, for example the knock-on effect on economies when you have a financial crisis, and even in a biological outbreak like COVID-19. When you have that connectivity, what happens in a local area of China can quickly influence what happens in other countries. That is really a transition that we’ve had in the last decade or two. Even if you look at flight volumes out of China since SARS back in 2003, it’s about a three-fold increase. Potentially, what would have been a very small outbreak a few decades ago that wouldn’t have affected other places, now [affects] what happens elsewhere, such as the UK, the U.S., or India.
How, in terms of the mathematics of epidemiology, does COVID-19 differ from other major infectious outbreaks such as the Spanish flu, SARS or Ebola?
There are a number of broad principles that we can use to understand these infections. These can apply across a number of pathogens, particularly in understanding the magnitude of spread. So, each person who gets infected, on average, how many people are they giving the virus to? For COVID-19, it is about two or three. We can also think about the time scales: if you have one case, how long on average is it until the person infected then shows symptoms. For COVID-19, that’s about five days or so. So, you get this of time scale, which is probably doubling every five days, in terms of your outbreak.
The nuances of what is driving those numbers will vary between pathogens in terms of when the transmission is actually happening and what interactions are causing them. One of the big challenges or differences of COVID-19, as opposed to SARS or Ebola, is that a lot of transmission seems to be happening very early on in the infection, when people don’t have symptoms or have very mild symptoms. One of the reasons why infections like Ebola and SARS have been easier to control is that a lot of people who are highly infectious have very distinctive symptoms. That means you can identify them, look at whom they’ve come in contact with and make sure those people have been quarantined. However, for COVID-19, a lot of the transmission happens among people who might feel perfectly well or might have a slight cough, for example. That makes it very hard to pick up all the infections, and is why we’ve seen in many countries, even if they’ve been trying to detect cases coming in, many cases have managed to go undetected.
Yes, and this is what makes me worry about a country like India that has a relatively low number of confirmed cases and correspondingly fewer recorded deaths linked to COVID-19. Given the size and density of our population, does this seem a little suspicious to you, perhaps owing to the limited extent of testing? Or do you think maybe India acted fast in closing its international borders or something structural within its population has limited its spread?
I think in the early stages it is quite hard to untangle those two things, when you have a very small number of cases – is it just by chance or population structure or other features that mean that transmission hasn’t quite taken off yet? In Europe, we saw a lot of early transmission in Italy and that probably increased the chance that that country got hit before others. Once we start to get more severe cases, especially deaths, you can then start to get a clearer picture of how much infection you might be seeing. Deaths are obviously generally well-identified in the data. We estimate that about 1% of people with clear symptoms will have a fatal outcome for this. If you imagine that we have 100 deaths, we would really be expecting something in the range of about 10,000 cases. If you’re not seeing that in your data, then that suggests that not all symptomatic cases are showing up. I want to emphasise that it’s not just one or two countries that have this situation. We recently did some rough estimates and we think that many countries that have clear outbreaks now are probably reporting less than 20% of their symptomatic cases. That is partly due to testing and other logistical things. That situation is very common. But that does mean that it is hard to get a handle on the outbreak as it is now, because if you’re looking at deaths to understand what’s going on in your outbreak, those people who have died may have well been infected several weeks earlier. So, you’re really are looking into the past.
Listen : In conversation with top epidemiologist Jayaprakash Muliyil | The Hindu In Focus Podcast
Can you tell us a bit about the idea of herd immunity, the notion that over time a certain critical mass of a population will acquire immunity to a pathogen and that negates the need for universal countermeasures such as immunisation? Is this approach, which seemed to gain some currency in the U.K., applicable worldwide to the current pandemic? Or does it imply a death toll that would be politically unacceptable?
The concept of herd immunity goes back to Ronald Ross’ work on mosquitoes, and it is the idea that you don’t need to get your entire population infected to stop your outbreak. Once you have enough immune people, then when you have someone infectious, it is more likely that they will meet someone who is immune than someone who is susceptible and that will slow transmission without needing all of your population to be immune. That’s how vaccination works. We vaccinate a majority and that stops the outbreak, even though there are newborns and some [other such] groups that cannot be vaccinated. But as you say, for COVID-19, this can have a really severe impact on health systems, it can lead to a lot of fatalities. Having herd immunity presented as an aim really isn’t helpful and doesn’t reflect the situation we face.
In the U.K., there was this period of time where I was getting asked a lot of questions such as “Is the U.K.’s strategy just to sit back and let everybody get infected in the country?” That certainly wasn’t the strategy, as I saw it, and it wasn’t the modelling work that we’d been doing to provide some evidence about what the scenarios might be. It was very clear, if you look at the transmission of the disease, at its severity, that it was going to very quickly overwhelm health systems. I think we might be in a situation, especially in countries where it takes longer to slow down transmission, where infections will keep occurring. We estimate that probably in the hundreds of thousands of infections have now occurred in the U.K., even though we’ve only got a few thousand confirmed cases. This is a lot of infections, but it’s not enough to protect the population through this herd immunity mechanism. We could see some immunity-building in populations just because these outbreaks are happening, but I don’t think we should go into it with the aim of getting large numbers of people infected. If that happens in a short period of time, you’re very quickly going to overrun your hospitals.
Interestingly, you have also explored a pandemic that we have all learned to live with in the digital ecosystems of the 21st century – the surge of fake news. In India, like other countries, we have seen so much fake news in recent days relating to the coronavirus, including regarding unscientific treatment methods. Do you see any parallels between how fake news transmission happens and deadly pandemics such as COVID-19?
There certainly are parallels. One of the actual similarities is this: if you look at the individual level of sharing viral content online, it seems quite similar to the spread of COVID-19, in terms of individual steps. For COVID-19, on average, in the early stage, each case would give the infection to a couple of others. In a study of viral Facebook content, it was found that each person, on average, who shares a viral post, will lead to about two more people sharing it. But the big difference is the time scale. With COVID-19, it takes a few days for that transmission and that new infection to occur, whereas online we might be talking about 30 seconds. Very quickly, you get this much faster growth.
The other similarity between many of these infections is the variation at the individual level. As we have seen, for COVID-19, there have been some outbreaks where a single event sparks a large number of secondary infections. A lot of content that becomes popular online tends to have that feature too. We see messages spreading on WhatsApp, but often there will be some high-profile person online or some media outlet that will have sparked the outbreak and shared to a large number of people. Then, you see these smaller but secondary clusters of sharing happening. It is worth considering those elements as well. It is not just this idea that everyone is sharing with a couple of friends on a message group on WhatsApp, but it is often those more prominent organisations and people that are feeding into that.
