COVID deaths: three times the official toll

Host: Benjamin Thompson

Welcome back to the Nature Podcast. This week: estimating the mortality associated with the COVID-19 pandemic.

Host: Nick Petrić Howe

And the lack of diversity in UK academia. I’m Nick Petrić Howe.

Host: Benjamin Thompson

And I’m Benjamin Thompson.

[Jingle]

Interviewer: Benjamin Thompson

This week, a team of researchers working with the World Health Organization have published a paper in Nature that has used statistical modelling to estimate the number of excess deaths associated with the COVID-19 pandemic. Someone who has written extensively about this topic is Nature’s Richard Van Noorden, who joins me to talk about it. Richard, hi. Thank you so much for joining me today.

Interviewee: Richard Van Noorden

Hi, Ben.

Interviewer: Benjamin Thompson

Richard, let's get up to speed quickly before we unpack the actual paper then. What do we mean when we talk about excess deaths in this context?

Interviewee: Richard Van Noorden

Quite simply, we mean how many deaths would you expect to see in the world in a normal year, and how many deaths have you actually seen on top of that? And the extra death is the ‘excess’ death. So, what these WHO researchers have done is trying to tot up all the excess deaths that there were in 2020 and 2021, above what you'd expect to see. And they've said, ‘Well, this is probably because of the pandemic.’ And this could be deaths directly because of an infection of COVID-19, but it could be any deaths associated with the disruption of the pandemic. For example, if someone couldn't get cancer screening because health services were disrupted and they subsequently died earlier than you'd expect. That would count as one of these excess deaths.

Interviewer: Benjamin Thompson

And before we talk about the actual hard numbers from this paper, Richard, why is this a useful metric for researchers?

Interviewee: Richard Van Noorden

So, in general terms, excess deaths tell you if there's been an unexpected spike of mortality, and that's what we're finding has happened. The WHO researchers say there's been an extra spike of about 15 million deaths over those two years. Now, if you look at the number of reported official COVID-19 deaths over those years, it was 5.5 million.

Interviewer: Benjamin Thompson

And there's quite a discrepancy then between the two. What might be the root cause of that?

Interviewee: Richard Van Noorden

Well, first of all, even among rich countries, they think they're recording everything, but they don't always record deaths by COVID-19 in the same way. And in those sorts of frantic first few months of the pandemic, in many countries, many deaths occurred that may not have been correctly attributed to COVID-19. And in many countries around the world, deaths are not recorded at all. And so, we will have missed a huge number of deaths and also a number of COVID-19 deaths. They just won't be there in the official statistics.

Interviewer: Benjamin Thompson

And in this paper then, the team have made estimates of these excess deaths for 2020 and 2021 at around 4.5 million and 10.4 million respectively. What are some of the other key headlines from this research?

Interviewee: Richard Van Noorden

Well, one of the big headlines is that this suggests that COVID was the biggest killer around the world in 2021 – bigger than coronary heart disease, which is ordinarily the largest cause of death, so that's pretty striking. But one of the main things to come out of this modelling estimation exercise is that if you look at the official numbers, the high-income countries seem to have been the most affected by COVID. But if you look at this modelling and these estimates, it suggests that middle- and lower-middle-income countries were harder hit than the highest or at least equally as affected. That's a really big takeaway because some people have suggested there's a bit of a mystery and that we weren't seeing these deaths being reported from lower-middle-income countries. Well, probably these countries were hit just as hard – we just didn't hear about all the deaths that were going on.

Interviewer: Benjamin Thompson

And, obviously, in this work, they've made a lot of estimates. And I guess this has to be one of the flaws to this metric, right? It's only as good as the data that can be fed in.

Interviewee: Richard Van Noorden

Right, so most countries around the world just don't report deaths full stop. So, by some estimates, around 40% of the deaths around the world just go completely unrecorded. In fact, the whole business of estimating mortality is an estimate. And it's normally done sort of 3 or 4 years later. Maybe surveys and censuses – you just ask people, ‘Has there been a death in your family in the last few years’ – that can help fill in what is missed by the official records. Here we're asking for much more sort of real-time estimates. So, what the researchers had to do in many, many countries that don't record deaths, they just had to model it. They looked at the characteristics of those countries, for instance, the reported COVID-19 death rate that we know about, how many cases we know about, what restrictions those countries put in, what the prevalence is of other diseases like diabetes or cardiovascular disease in those countries. All kinds of factors that it seems from the data we do know correlate quite well with recorded COVID deaths and excess deaths. And then they simply said, ‘Well, assuming that these other countries of which we have no data behaved the same way, this is the burden of mortality that we'd expect to see.’ So, it is only a modelling exercise, and it is a very crude and rough one. I talked about 15 million deaths, but the uncertainty intervals around this is very wide – anything between 13 and 16.5 million. The other problem is that even for the rich countries, we have all the data, every month they record their deaths. You're estimating this weird idea of how many deaths would you expect, and that means you've got to guess at, if past trends from 2019 and 2018 continued, what deaths would you have expected to see. And this is a tricky modelling exercise in itself, and that just adds further uncertainty to the estimates here.

Interviewer: Benjamin Thompson

So, if this is an estimation exercise, a modelling exercise with a variety of moving parts, as you've described there, what's the value of it then? I mean, what have researchers told you about that?

Interviewee: Richard Van Noorden

So, the value of it is, first of all, just observing that we've seen way more deaths than the official recorded numbers. So, this pandemic just has had a much bigger impact than you would know, if you just looked at the official figures. And secondly, the value is that these countries that we didn't hear much about, they were hit hard as well, probably, and you would miss that if you insisted that because we can never know the real number, let's just not embark on this exercise. So, it's kind of, if we don't try this at all, we won't even have an idea of the burden of COVID-19.

Interviewer: Benjamin Thompson

Now, Richard, it is worth pointing out that this isn't the only attempt to measure excess deaths, but it seems like this latest attempt has offered some new insights. But of course, the pandemic is still ongoing. And sadly, there is the real possibility that another one will come along. This isn't the last pandemic that we're going to see. What lessons have people said can be learned from exercises like this to prepare for the next whatever it may be?

Interviewee: Richard Van Noorden

So, one of the lessons is how difficult it is to do this, which shows that we need to improve systems for tracking mortality and disease in general. And demographers knew that many deaths around the world are never captured or recorded, but this pandemic has really underlined how little we know. There's a pandemic treaty that the WHO is preparing to sort of strengthen our resilience to future pandemics, and people told us that this really boring but vital point of creating better mortality recording systems has got to be part of this. The other thing that people said is, of course, that the impact of the pandemic is not just on mortality but also its burdens on the health of survivors, and that is an even harder exercise. And it's just sort of showing how unprepared, again, countries were to deal with the burden of this pandemic.

Interviewer: Benjamin Thompson

Nature’s Richard Van Noorden there. For more on this paper, look out for links in the show notes.

Host: Nick Petrić Howe

Coming up, we’ll be hearing how the leaky pipeline in academia affects some communities more than others. Right now, though, it's time for the Research Highlights, read this week by Dan Fox.

[Jingle]

Dan Fox

Exceptionally well-preserved fossils of the armoured dinosaur Zuul crurivastator have revealed that their bony tail clubs likely evolved to battle rivals from the same species instead of predators. Ankylosaurid dinosaurs like Zuul had stiff tails ending in massive bony knobs, which were long thought to be defensive weapons for warding off tyrannosaurs and other predators. But today's most impressive animal weapons – like antlers in the deer family – are used mainly for battles with members of the same species over mates. Now, researchers are arguing that the same is true for tail clubs, as juveniles – theoretically the most vulnerable to predators – don't have clubs. The team have also presented an example of preserved Zuul crurivastator skin pitted with battle scars that they say could only be caused by a rival’s tail. Confirmation that these ancient creatures engaged in ritualised combat for social dominance would imply a rich and complex social world. If you only caught the tail end of that research, don't worry, you can read it in full in Biology Letters.

[Jingle]

Dan Fox

A pair of 11,000-year-old carved panels uncovered in Turkey appear to be the earliest known portrayal of a narrative scene. The carvings were found on the side of a limestone bench at an excavation at the Sayburç archaeological site. The right panel features a male facing forwards – its shape protruding from the flat surface. The individual is flanked on each side by a leopard gazing towards it. In the left panel, another male figure holds a snake or rattle while approaching a bull. Because the panels sit side by side and portray similar themes – people encountering dangerous animals – they probably represent a progressing scene from a story. The researcher behind this discovery says that these works are the first known examples of an extended narrative. If that story has you gripped, read the research in full in Antiquity.

[Jingle]

Interviewer: Nick Petrić Howe

Next up on the show: the lack of ethnic diversity in academia. You may be familiar with the concept of the ‘leaky pipeline’ in academia. This describes the gradual loss of people at every stage of academia, due the environment being less than hospitable. Few people become postdocs, and fewer still become professors, despite excellent performance. Well, if the pipeline is ‘leaky’ for some people, it appears to be pouring for others, as a new series of Nature features is diving into, by gathering data on how ethnically diverse, or not, academia really is. The first of the features is focused on the UK and shows that people from Black, Bangladeshi and Pakistani communities are vastly underrepresented in science compared to the general population in the UK. And it’s not for lack of attainment. Whilst 8% of science undergraduates in the UK are Black, only 0.6% are professors. To get a sense of the problem – especially for early career researchers – I reached out to Mahrukh Shameem, a PhD student in immunology and an advocate for equity, diversity and inclusion at the University of Sheffield. I started by asking her about her own experience with the lack of diversity in academia, as a Kashmiri woman and an early career researcher.

Interviewee: Mahrukh Shameem

Throughout my undergrad, I wouldn't see many female lecturers, never mind South Asian lecturers. And when it came to certain case studies or just understanding different topics, it always came from a very skewed perspective. And so, we would talk about certain things and I would think, ‘Well, how does that affect some of my genetic background, or people from certain countries that aren't as represented here?’ And so, you don't feel represented in general, with the people that are teaching you or that you're learning from, but even from the things that you're being taught, and it kind of makes you feel excluded ever so slightly. And I realised that more within my postgraduate degree when I would talk about things, or even say, ‘Oh, I don't need to really take Easter off, I have my own religious holidays,’ and having to explain what that meant. It ends up being a burden without people realising, and it's a burden that you have to take on in educating people, rather than people taking that initiative to understand maybe why you're not eating for this month of Ramadan because you have to take time to explain that to people. So, it has become a burden, ever so slightly, unfortunately.

Interviewer: Nick Petrić Howe

And so, the feature this week in Nature is showing that people, especially from certain backgrounds, are leaving or prevented from progressing in academia. Do you have any thoughts as to why this might be?

Interviewee: Mahrukh Shameem

It's a very interesting question, and I think one of the reasons why we don't understand it is that the data doesn't really exist for it first of all, so I don't think I can definitively answer. But from talking to my friends who come from these backgrounds, a lot of the people don't see how they can find themselves fitting into research or fitting into higher levels of academia. They don't see it, so they don't think that it's attainable. A lot of people don't actually understand what research is. I didn't know that a PhD could be funded until someone told me. This is information that is very broadly available to a lot of people. However, if you're not really exposed to it at a very young age or throughout your undergrad even, you don't even know those options exists for you. Lower-income schools, for example, just don't have this information disseminated out to you. You don't have as many career fairs. You don't have as many people coming into your schools and telling you, ‘Hey, did you know that there are more careers than just being a doctor or lawyer,’ which is quite sad. And I think a lot of people don't seem to go for these careers because they genuinely don't know they exist or don't know how they can fit into them with their lifestyles.

Interviewer: Nick Petrić Howe

And so, one thing that's touched on in the feature is that certain ethnic groups tend to have lower household incomes, and so this may sway students to not do academia because it's not the most financially stable and secure job. Do you think this is a factor?

Interviewee: Mahrukh Shameem

Yes, you talk to any academic and there's so much stress with constantly applying for grants and the instability of jobs. And especially for me, and a lot of people I know, is when your family are immigrants, the idea is you will support them in some way. And so, having a career where it might take 20 or 30 years for you to then have some job stability or be able to then provide financially for them isn't something that is doable for a lot of first-generation immigrants.

Interviewer: Nick Petrić Howe

And so, for you then personally, knowing some of these challenges that are within academia, what made you decide to pursue it anyway?

Interviewee: Mahrukh Shameem

It’s going to sound incredibly cheesy, but I chose it because I loved it. And I don't think that just because there aren't many people who look like me, I should be afraid to do it. I think that's the complete opposite. I'd love to be part of that representation. And it's been very difficult for me to have to explain certain things that I don't think I have to or should be part of my job. But if it's an uphill battle for me, maybe it'll be a tiny bit easier for other people. I've had so many other people who are in much earlier stages of their career or even their degrees, who've come up to me and asked me how I got to the position I'm in so far, but it's great to be able to kind of feed that knowledge and information back to them. I'm not going to let the fact that science might be slightly behind not let me do what I love.

Interviewer: Nick Petrić Howe

But that sort of representation that you are doing, and could have been there for you, would make things potentially easier for people to be a part of academia.

Interviewee: Mahrukh Shameem

I 100% agree, yes. I've had amazing, amazing mentors, and they've been women, men, through many different stages of their career, who have either been part of the minority or haven't at all, but it's still so invaluable to be able to talk to them and understand why they're advocates, and how they are making changes in their small labs or part of their larger organisations to enforce change. And so, that's quite hopeful, and I think the fact that I haven't had, or haven't had the ability to see, as much representation hasn't meant that I haven't had amazing mentors and amazing advocates who really do want me to do well.

Interviewer: Nick Petrić Howe

Well, speaking of these sorts of changes, what, to your mind, are the things that we can do to improve things?

Interviewee: Mahrukh Shameem

I think, first of all, there needs to be data collected on understanding what background people come from, what difficulties they've had. I think gathering that data and understanding where these researchers end up really is invaluable to know where we may be lacking. And then I think having open forums is just needed more in general, whether that be in academia and industry where you have early career researchers who are able to voice out areas that they think want or need improving, and that should be backed up by senior voices. I also think having programmes which genuinely promote active change is needed, whether that be mentorship schemes, or scholarship schemes, or making applications blind. All these little things that really do add up to ensure that you're making the process as fair as possible and also supporting students who just might not have privileges that other people have had. And they don't necessarily have to be huge, ground-breaking things that topple society in any way, but small, very actionable changes matter, and you can see they’re making incredible change.

Interviewer: Nick Petrić Howe

That was Mahrukh Shameem from the University of Sheffield in the UK. To find out more about this topic and how it affects every stage of academia, check out the feature in the show notes, and be sure to check back for more in the future.

Host: Benjamin Thompson

Finally on the show, it's time for the Briefing Chat, where we talk about a couple of articles that have appeared in the Nature Briefing. Nick, why don't you go first this week? What have you got?

Host: Nick Petrić Howe

Well, Ben, this week I've been looking into a story about AI. I was reading this in Nature. So, you may be familiar with GPT-3, which is this sort of language-based programme. Well, there's a new version of it called ChatGPT, which is available freely, and some people worry it could spell the end of the essay.

Host: Benjamin Thompson

Yeah, ChatGPT then, Nick. This has been all over the internet over the past few weeks and, what, you give it a short prompt and it'll give you a long and seemingly quite eloquent response to it. And I've seen things like gym schedules, I've seen things like computer programmes, all sorts of things. And there are a lot of people who are quite worried about this. And it seems like one of the things that people are concerned about then maybe is essays, is that right?

Host: Nick Petrić Howe

Yeah, that's exactly right. So, if you cast your mind way back to undergraduate, you may remember being tested by being given an essay to write, and it would have like a question at the start, like, tell me about this topic and use this or something like that. And those are the exact sort of things you can put into ChatGPT, and it will spit out a response. So, the concern is that students can just use this tool, which is currently free, and just get it to write essays for them. So, one person who was interviewed for this article, Lilian Edwards, who studies law, innovation and society at Newcastle University, said that, ‘At the moment, it's looking a lot like the end of essays as an assignment for education.’

Host: Benjamin Thompson

I mean, there’ll be some students who are sort of punching the air right now. But I guess that does change things significantly. And so, this system hasn't been out in the wild that long but I guess, potentially, there already could be essays that are being created by it without anyone knowing. How would you go about sort of checking to see if something's been written by an AI, or whether it's been written by a human?

Host: Nick Petrić Howe

Well, that's the thing: as these tools get more and more sophisticated, it’s much harder to determine who is writing it. And so, GPT-3 – what ChatGPT is based on – tended to write answers which were a bit more cold and computer-like, whereas ChatGPT, as the name suggests, is a bit more chatty. So, it sounds a little bit more casual and as if a human is writing it. So, in order to check for it, it is going to be really difficult. There are a few flaws in what ChatGPT creates. So, it may say things that are factually incorrect. It is just using a lot of text that it’s found from the web to come up with associations, which may not necessarily actually be true. And, for instance, I asked it, when I was playing around with it before, to write a news story with me with some quotes, and it had just manufactured some quotes. The people were real people, who were real experts in the field I asked it to write about, but the quotes were completely made up. So, there are factual inaccuracies that it can maybe push out, which is a way that you could potentially detect it. But also, not everyone is convinced that this really is a game changer.

Host: Benjamin Thompson

Really? Why is that then?

Host: Nick Petrić Howe

Well, the reason for this is that you could have done a version of this for a very long time. So, essay mills have existed for years, and this is where you basically pay someone some money, and they pump out an essay for you. So, if students really wanted to get an essay that was a passing grade or whatever grade they wanted, there were ways in which they could do this. And the other thing as well is maybe we need to shift what we're asking students to do. So, some people who were interviewed for the article said that instead of asking students to just simply write about a topic, we need to prioritise getting them to think about it, like ask critical thinking or reasoning, which is something that ChatGPT can't really do that well, and may actually be better for the student to sort of engage those critical faculties.

Host: Benjamin Thompson

So, it may be then that the essay as we know it isn’t quite done for just yet.

Host: Nick Petrić Howe

Perhaps not. And the other thing is that this tool may not be free forever. So, there was a tweet from the CEO of OpenAI, who are the people who created this ChatGPT, saying that it's using a heck of a lot of computing resources, so they're going to have to figure out a way to monetise it at some point. And students may well balk at paying quite a lot of money to use a tool like this and it may be cheaper to just actually go learn some stuff.

Host: Benjamin Thompson

Well, an interesting story, Nick, and I'm sure one that's going to run and run for a while. For my story this week, well, it's something that I've been following myself, actually, for several weeks, and it's a story that I read about in Nature and it's a space story. And it's all about NASA's Orion spacecraft – the capsule part of which arrived safely back on Earth after going to the Moon and back. And, in fact, it did so 50 years to the day that humans last walked on the Moon as part of the Apollo 17 mission. So, really nice kind of alignment of the stars there, I suppose.

Host: Nick Petrić Howe

Well, I do like that. And this is part of the Artemis I mission, right? So, what's it been up to since the launch?

Host: Benjamin Thompson

Yeah, absolutely right, Nick. So, the Orion spacecraft was sitting on top of a giant rocket, which launched back in November. And you may remember from the podcast a few weeks ago, I got up super early to watch along with the launch with our colleague, Alex Witze. And shortly after we recorded our chat, the Orion spacecraft did some manoeuvres and orientated itself and flew off to the Moon. And what's happened since then is this uncrewed mission then has been orbiting the Moon. And, in fact, it set a distance record of over 430,000 kilometres away from Earth, so that's the furthest that any potentially crewed spacecraft has been from Earth. But it made its way back after sort of 3 weeks, a 2.2-million-kilometre trip, so a fair distance, I would say. But all that would have been for naught if the very last part of the mission – which is re-enter Earth’s atmosphere and land kind of serenely in the ocean – hadn't gone to plan, right. And, thankfully, it did, and this is a very, very difficult thing to do indeed.

Host: Nick Petrić Howe

And so, the idea of doing this is to one day take human beings back to the Moon, so I guess this is a big success.

Host: Benjamin Thompson

Yeah, it does seem like a big success, Nick, and I think everyone was delighted that this has happened. And actually, in a strange way, it's now, right, let's get back to work. And so, what's happened in this instance as well is they were doing a lot of tests here, obviously, as this is a new spacecraft. And so, when it returned to Earth, the capsule portion of this spacecraft kind of separated and the spacecraft bit kind of burnt up in the atmosphere. And the capsule had to slow down from 40,000 kilometres an hour down to 32 kilometres an hour, which is a heck of a thing, as you might imagine, and then slowly glide down on some parachutes to splash down into the ocean. And the brunt of this kind of re-entry was done by the Orion’s heat shield, right, which endured temperatures of around 2,800 °C, right, so incredibly hot. And it's made of the same stuff that was used during the Apollo programme, but it's been applied in a different way. And the capsule also came back in a different way to the Apollo missions. In this case, what the Orion capsule did was it did this kind of skipping manoeuvre, almost like a stone skipping on a pond through the atmosphere, and this hasn't been tried before on a crewed or potentially crewed space mission, right. And what it allows NASA to do is really closely pinpoint where it's going to land on Earth. And it does a bunch of other things as well, like, hopefully, it will subject astronauts to less G forces, things like that. But it really was a bit of a test, and it seemed that all went well, Nick. It splashed down off the coast of Mexico, less than 10 kilometres away from a waiting US Navy ship, right, so that's gone pretty well. And now, the next steps to Artemis II begin.

Host: Nick Petrić Howe

And so, what are those next steps, and how soon are people going to be on the Moon again?

Host: Benjamin Thompson

Yeah, well, that's a huge question, Nick. I mean, the stage is now set for the Artemis II mission in 2024. But what needs to happen is obviously they need to do a bunch of tests of what's come back. So, some time will be needed, obviously, to test parts of the capsule that will be included in Artemis II to make sure they're working okay. Particularly things like the flight computers, which are actually being re-used from this mission to the next mission. That was a decision made a long time ago to save money. And there are a bunch of things in the capsule as well like radiation sensors and so forth, so they need to test those to see what radiation levels were like in the capsule. So, they want to make things as safe as they possibly can for the next mission, which, as I say, 2024 is the plan, and that is going to be with people. It's going to be a crewed mission. The same sort of setup, so the rocket will take off, the Orion spacecraft will go around the Moon, and then it will arrive back. And then if that goes well, I mean, they're talking 2025 for maybe landing on the Moon, but obviously that's a long old way away, Nick.

Host: Nick Petrić Howe

Well, it certainly seems to be getting closer every day, so I'll be keeping an eye out on that. Thanks, Ben. And listeners, for more details on the stories we discussed and where to sign up for the Nature Briefing to get more like them delivered straight to your inbox, check out the show notes for some links.

Host: Benjamin Thompson

And that's all for this week. But keep your ears out in seven days’ time for next week's instalment of the Nature Podcast. It'll be our last regular show of the year and, as always, we'll be going out on a festive note, so there'll be high jinks, there'll be seasonal songs, there'll be all the stuff you expect from us at this time of year, so look out for that.

Host: Nick Petrić Howe

In the meantime, you can keep in touch with us on Twitter. We’re @NaturePodcast. Or you can send an email to podcast@nature.com. I’m Nick Petrić Howe.

Host: Benjamin Thompson

And I’m Benjamin Thompson. Thanks for listening.