By Charvy Narain
What do a mathematician, an epidemiologist, a vaccine developer, a protein crystallographer and a whole bevy of immunologists and infectious disease specialists have in common? Answer: they’re just some of the Oxford University researchers coming together to fight the novel Coronavirus outbreak which has (to date) killed more than 1,350 people across the globe, with over 60,000 people infected.
The outbreak, which the World Health Organisation has now declared a global health emergency, is caused by a new type of an old foe: coronaviruses are common enough to be one of the causes of the common cold. But they can cause a range of respiratory symptoms from mild to serious – it was a coronavirus that was responsible for the 2002-2004 outbreak of the severe acute respiratory syndrome (SARS), though the novel coronavirus (until recently known as 2019-nCoV, though now dubbed SARS-CoV-2) has already outstripped the SARS death toll in the three months since it has been active.
Like the SARS virus, which was traced back to civet cats, this previously undescribed SARS-CoV-2 is also likely to have been transmitted from an animal to humans – most people in the first cluster of cases worked at or were frequent visitors to one single seafood market in Wuhan in China. But it is now clear that SARS-CoV-2 can be also transferred from an infected person to another person, and these human-to-human transmissions are how the outbreak is currently spreading.
Mapping the disease
Dr Moritz Kraemer is the Branco Weiss Fellow at the Oxford University Department of Zoology, and part of the Oxford Martin Programme on Pandemic Genomics. Like many of the Oxford University researchers currently working on the SARS-CoV-2 outbreak, he is a veteran when it comes infectious disease: he has previously crowdsourced data to track the spread of Ebola and Zika in real time, and his modelling of yellow fever in Angola showed how ecological and demographic factors contributed to that outbreak.
Dr Kraemer is a spatial epidemiologist interested in how the spread of infectious diseases interacts with geography. Together with researchers at Harvard Medical School, Northeastern University in the US, the Boston Children’s Hospital and Tsinghua University, he has produced a real-time map of all confirmed cases of COVID-19 (the disease caused by SARS-CoV-2), with all of the data publically available: you can watch how the virus spread from Wuhan in China to the 28 countries across the world that have so far reported COVID-19 cases.
What sets this map apart is that instead of being based on total counts of how many cases of COVID-19 are found in each country, it is based on a ‘line list’ – detailed information about the demographics of each confirmed case of COVID-19. This includes whatever information is available about their age, sex, the date their symptoms started, where they live, and where they might have travelled to.
What I hope to do is to build a baseline for evidence-based decisions.
This kind of information isn’t always easily available in the midst of an outbreak, but analysing it can yield all sorts of insights. Dr Kraemer says: “We can, for example, analyse this data to find what the early signals of local transmission might be, such as a change in age distribution shifting from people in their early 40s up to people in their late 40s.”
Based on this kind of detailed line list data, researchers have been able to estimate the incubation of the SARS-CoV-2 virus and the age distribution of people affected, as well as track how time elapsing from symptoms appearing to hospitalisation and testing is changing as the outbreak evolves.
All of this is not just an academic exercise and having this information can help governments and policymakers make the most effective decisions. For example, by combining the data on the number of cases in each Chinese province plus its population size with air-traffic patterns, Dr Moritz and his colleagues were able to work out the risk for the SARS-CoV-2 being introduced to countries in Africa.
There are no reported cases of COVID-19 in Africa yet, but the limited health infrastructure in many countries means that an outbreak here may have particularly devastating consequences. By combining information about risk of transmission versus the country’s preparedness, the researchers were able to identify that Ethiopia and Nigeria may be particularly vulnerable.
In the lab
One of the scientists helping make sure there is accurate, detailed data about SARS-CoV-2 is Dr Peter Horby, Professor of Emerging Infectious Disease Global Health at the Nuffield Department of Medicine. Dr Horby went to Vietnam, which hosts a large scale University health research unit, for a short WHO secondment as part of the response to the SARS outbreak – and ended up staying for nine years. He came back just in time for the Ebola outbreak, and set up clinical trials for a candidate Ebola treatment in the middle of an active epidemic.
He now leads the Epidemic Research Group at Oxford, which aims to reduce the impact of epidemic infections through ways of doing research that work even in epidemics, and which is currently working with the Chinese government and researchers. One of the things that this group is doing is developing and distributing an electronic case record form, which will help get the detailed and accurate data that maps and models are so dependent on.
Dr Horby and others are also a part of a band of researchers working away in labs to understand the virus, and hopefully, develop a treatment. Dr Horby is part of a clinical trial currently testing two potential drugs for treating COVID-19, using the same technology they used to generate a vaccine (currently in testing) for the 2012 Middle East Respiratory Syndrome outbreak, which was also caused by a coronavirus.
These are different approaches, but as the SARS-CoV-2 outbreak unfolds, scientists are keen to have many potential weapons in their arsenal to fight it.
And while Professor Sarah Gilbert and her team (also at the Nuffield Department of Medicine) are currently working on a potential vaccine (they’re using the same technology they’ve already used to generate an in-testing vaccine for the 2012 Middle East Respiratory Syndrome outbreak, which was also caused by a coronavirus), other researchers at the Nuffield Department of Medicine are going right back to basics: Professor Dave Stuart and Yvonne Jones have been collaborating with researchers in China to successfully decode key structures related to SARS-CoV-2 at the atomic level. They are now starting work to understand the structure of the SARS-CoV-2 ‘spike’ protein, which will help map antibodies to the virus.
Understanding these spike protein antibodies is also the main focus of immunologist Professor Alain Townsend. Professor Townsend, based at the Radcliffe and Nuffield Departments of Medicine, is also working on a vaccine based on this same spike protein.
The maths of disease
How the outbreak might unfold is one of the questions that Dr Robin Thompson, Junior Research Fellow at Christ Church, is interested in. Dr Thompson is based at the Mathematics Institute and like Dr Kraemer, Dr Thompson is an epidemiologist.
But while Dr Kraemer is interested in using the tools of epidemiology to capture trends in an outbreak as it happens, Dr Thompson is interested in using maths to develop models of what might happen in a disease outbreak. To do this, researchers take real-life data about an outbreak, and then build a mathematical ‘model’ that is consistent with the data and captures its key parameters. But in a sort of thought experiment, you can also run these models forward in time, to get a forecast of what might happen in the future.
We can use our mathematical simulations to show that if you can quickly isolate infected people after they develop symptoms, you are likely to prevent sustained outbreaks in new countries. This is true even if some people might be infectious before they have any symptoms.
Since the model is a mini-simulation of the world, researchers can try out all sorts of potential interventions and find out what effect these would have.
This is not to say that mathematical epidemiology can provide a Minority Report-style accurate future prediction. Dr Thompson says: “One challenge we have is that in the real world, there is only one realisation of what might happen, while a model gives us several possible scenarios. So it’s hard to make specific predictions early in the outbreak about precisely how many cases of COVID-19 there will be, or exactly when an outbreak will be peak.”
However, what these models do yield is a range of predictions, and what they are useful for is estimating probabilities after specific events – such as how likely a sustained outbreak might be if SARS-CoV-2 travels to a new country.
The current UK government advice is for anybody showing even mild symptoms to self-isolate for 14 days, and Dr Thompson thinks that this is good advice.
He is more equivocal about the current advice for people to self-isolate if they’re coming from Wuhan or Hubei province, even if they have no symptoms. Dr Thompson says: “It’s quite a drastic measure. While containment is easiest when case numbers are low, this needs to be balanced against the effect on people who are quarantined when they are almost certainly not infected.”
This answer from mathematical modelling, like a lot of research in this area, depends crucially on open sharing of accurate data. For example, in the unlikely scenario that COVID-19 transmission from non-symptomatic people becomes common (there are now doubts about even the one confirmed case of non-symptomatic transmission), the answers from the modelling will change substantially, and Oxford University researchers are prepared – what happens in the case of non-symptomatic SARS-CoV-2 transmission is being looked at by one of Dr Thompson’s students.
But whatever the eventual scenarios, open data sharing will continue to be crucial, and is of benefit to researchers across the globe. “Open data sharing from very early on was one of the key features of our map,” says Dr Kraemer. “We made the data behind it immediately available.” This data has now been used by researchers from another group, for example, to show that once a place has three cases of COVID-19, there is a 50% chance of the infection becoming established in that population.
Professor Peter Horby and his colleagues are aiding this data sharing effort by developing and freely distributing a free toolkit of SARS-CoV-2 clinical research resources to anyone studying the outbreak. This set of flexible research protocols aims to help the research community generate more precise and robust conclusions faster.
At some point, the SARS-CoV-2 outbreak will end, and researchers will need to be ready for the next big one. Sharing information is likely to be crucial to doing this, and an editorial from Nature had a simple message for researchers: “Work hard to understand and combat this infectious disease; make that work of the highest standard; and make results quickly available to the world.”
Please note, the information in this blog is correct at time of posting. The University will communicate significant research developments as they emerge.