Research in the time of a pandemic: How a 2018 BBC documentary helped advance COVID-19 research
In our latest blog, Dr. Josh Firth, researcher at the Department of Zoology at the University of Oxford and currently working on the COVID-19 pandemic through the Royal Society RAMP Scheme, discusses the use of data from a 2018 BBC documentary to demonstrate COVID-19 spread patterns and assess the effectiveness of different interventions in curtailing the spread. Moreover, we talk about different types of technology like smart phone apps, tracking and detection systems and how they are being utilised in COVID-19 research, as well as the increasing importance of pre-prints to facilitate faster sharing. In addition we take a look at the UK’s research efforts and how different disciplines come together to address the pandemic. And finally, we assess how the current social events have made imperative the promotion of diversity and equality in STEM.
Can you give us a bit of a background
about the Contagion! The BBC Four Pandemic citizen science experiment that took place in 2018? How have you used the data from that experiment in your latest research to simulate outbreaks of COVID-19? What have been the results and learnings so far?
The 2018 documentary ‘Contagion! The BBC Four Pandemic’ saw Dr Hannah Fry lead a large-scale data collection effort, in collaboration with Cambridge University and London School of Hygiene and Tropical Medicine (LSHTM), aimed at predicting the impact of the next pandemic in the UK. Part of this included recruiting residents of a UK town (Haslemere, Surrey) to agree to be tracked via their mobile phones over a period of three full days. This provided one of the most complete fine-scale social interaction datasets available for humans. Coincidentally, Haslemere also ended up being where the first case of COVID-19 transmission within the UK was recorded too.
In our recent research, we used this data to build social networks, detailing the number and strength of close contacts between people. Importantly, we then applied epidemic models based on what we already know about the intricacies of COVID-19 (i.e. incubation periods, asymptomatic transmission etc) to assess how COVID-19 would spread in this real-world network, and how effective different interventions (contacting tracing, testing, physical distancing) would work.
So far, the results have not only demonstrated the value of using real social interactions to predict the spread of this disease, but also clearly illustrated that it is important to consider the interactions between potential interventions. For example, contact tracing alone can quickly result in a high proportion of the population having to go into quarantine, but combining contact tracing with social distancing means that the number of individuals having to quarantine is substantially reduced without jeopardising COVID-19 control.
Haslemere social network: Graphical depiction of the network of social contact events between 468 individuals within Haslemere over the monitoring period. Individuals are shown as circles and the social connections between them are shown as the lines connecting them.The number of unique contact partners each individual has is denoted by size and colour of nodes (large & red = central) and thickness of lines show number of contacts (5 minute intervals within close proximity) between the partners. See the research paper for more details.
What is the importance of using a real social network to re-imagine how COVID-19 spreads within populations?
Detailed datasets on real-world human social interactions are relatively rare, yet research in other systems (particularly from natural animal populations) has shown that real social structures may differ in intricate ways from simulated networks, which can influence how we might expect contagions spread. Given that this is true even in species where individuals have relatively simple social behaviours compared to humans, we can certainly expect that the ‘hidden structure’ in human societies will potentially hold great influence over how we might expect COVID-19 to spread, and how effective we might predict different interventions to be. In fact, we show that even if a simulation model had the precise information on the number of individuals, the number of social connections, and the strength of these social connections between individuals, the predictions of COVID-19 outbreaks still depend on how these social ties are arranged between individuals, such as whether they occur in clusters, or whether individuals differ in how social they are.
What is the role of data and technology in COVID-19 research as it relates to your work?
Our work was only possible through a tailor-made smartphone app, and the real-world data that this provided. So, using the latest technologies, and the arising data, was really crucial here. We now aim to take this approach even further, and apply our general network-based COVID-19 transmission model to specific social contexts, such as schools and workplaces. Again, such data was only possible to collect through advances in tracking and detection technologies. We hope that this planned work will help enable identification of which COVID-19 control strategies can be most effective in different settings.
More generally, the first surge of research into modelling the COVID-19 pandemic has revolved around sophisticated modelling approaches without heavy real-world data requirements (due to the difficulty of acquiring relevant data). But, it now looks to be the case that more and more opportunities are arising to incorporate detailed real-world data. This will be very useful for advancing understanding surrounding modelling of the pandemic, and hopefully improve confidence in transferring model predictions into real-world applications.
What data are considered reliable to use and what are their primary sources?
Real-world data is going to be crucially important in assessing the spread of COVID-19 and associated control efforts. But, it is certainly important to critically examine these data sources as they arise, determine the contexts that each of them will be best suited to, and identify potential problems that the different types of data may present. A particularly relevant example of this is in terms of considering contact tracing approaches; many different sources of data are arising, each describing prior social interactions between individuals in different ways. This -for instance- ranges from using questionnaires that ask potentially-infectious individuals about their previous interactions with others, to using automated proximity sensing (e.g. Bluetooth) or positional (e.g. GPS) technology to gauge social associations between individuals. Each of these data sources will hold their own limitations as well as their unique advantages. As such, it is important that the potential opportunities, and challenges, of such data are considered in-depth by the researchers that hope to use these different data sources.
"Real-world data is going to be crucially important in assessing the spread of COVID-19 and associated control efforts."
How can the public be a part of finding a solution to the pandemic?
In terms of hoping to incorporate more data into modelling-based research and primarily shifting to using real-world data and monitored contagion patterns, the public will be the primary governing force in determining whether this is possible. For instance, going back to the example of contact tracing, it is only through public engagement that these interventions can be implemented and assessed. Similarly, if we hope to understand how real-world physical distancing efforts may relate to suppressing spread of COVID-19, researchers will again be relying on the public to continue to engage with these interventions to allow assessment of how effective these are during different phases of the contagion, and under various societal changes (e.g. as schools gradually return).
How would you advise members of the research community to engage with the public and what part does audiovisual content and the media play in this?
Many epidemiologists, and indeed those from other fields now working on the COVID-19 pandemic, are often highly experienced in explaining complex phenomena in a way which is interesting, and relevant, to the public. So, it has been great to see these skills now being put to good use for engaging with the public. In the same sense, we’re also seeing that scientists working in these areas are often well equipped to create intuitive graphical visualisation of their research, and provide audiovisual content (such as animations), some which have gained enormous attention on social media. Some research efforts have also developed and provided simple apps which allow the public to engage with the COVID-19 modelling efforts. I look forward to seeing these efforts continue and grow!
How have you seen collaboration manifest within the research community and more widely with other policy stakeholders during this time?
On a personal level, I have taken much encouragement from watching the collaborations between different groups, and even fields, manifest as researchers all over the globe hope to help with tackling the pandemic. Being from a non-epidemiological background myself, I had hoped that my experience in modelling natural social networks could be of some use, and this was made possible through The Royal Society’s interdisciplinary RAMP Scheme (Rapid Assistance in Modelling the Pandemic). Through this, myself and my long-term colleague Dr Lewis Spurgin were guided by the epidemiologists in The Centre for the Mathematical Modelling of Infectious Diseases (CMMID)-COVID-19 working group at LSHTM (particularly Dr Adam Kucharski) as we led this recent research and continue to benefit from this as we advance our social network approach in hopes of providing further insights into COVID-19 control efforts across different social contexts.
What are some of the greatest challenges the research community has been facing currently?
In regards to carrying out research, the enormous volume of current scientific investigations into COVID-19 means that it is challenging (and near impossible) to keep up with the literature in this area. Particularly because this research is coming from so many different angles and fields, meaning that it’s also difficult to navigate around these new studies as they arise. For example, while working on our most recent paper, not only was I constantly coming across new epidemic-modelling studies that were of relevance, but there was also a great deal of serological studies released that held information that could be important for parameterising models, but were difficult to rapidly interpret at the rate that these were released.
"...the enormous volume of current scientific investigations into COVID-19 means that it is challenging to keep up with the literature in this area."
In regards to barriers to research, it must be acknowledged that it is an intensively challenging time for many individuals, and this is particularly disadvantageous for promoting diversity. For instance, the school closures and lack of access to childcare during the pandemic results in a situation whereby those responsible for children find it extremely challenging to carry out intensive and speedy research during this time. Now also marks a particularly important time in terms of people working to addressing racism around the world, and it is a particularly challenging time for people who form part of the BAME community, so it is now more important than ever to dedicate efforts to promoting diversity and equality in STEM and generally, and supporting those who need it the most in every way possible.
"...it is a particularly challenging time for people who form part of the BAME community, so it is now more important than ever to dedicate efforts to promoting diversity and equality in STEM and generally, and supporting those who need it the most in every way possible."
What is the role of preprints and peer review of COVID-19 research?
The speed at which preprint systems allow research to be freely released and freely disseminated has never being more valuable, and it has been interesting to see how preprints have become the backbone of the rapid surge of scientific advancements during the pandemic. It has also changed how the scientific community, and the public, view preprint manuscripts, and we are now seeing more and more press coverage and media attention based around preprints (when previously such attention mainly revolved around journal-published papers).
Although the preprint system hopes to promote peer-review (rather than be an alternative to it), the current number of preprints being released means that – on average – most manuscripts are not receiving open reviews during the preprint stage. I think we have also seen an increase in the efforts for internal review within COVID-19 based research groups, for instance our recent manuscript was reviewed by members of the CMMID-COVID19 working group. Despite these other avenues for review, a major focus of course still remains on the formal review system provided through journals.
"... preprints have become the backbone of the rapid surge of scientific advancements during the pandemic."
What is the role of publishers during this time and where can they add the most value?
While it is true that the traditional publishing approaches aren’t always perfect, the journal-based peer-review system remains extremely valuable, and reviewers often provide a much needed and appreciated service to the research community through shaping research outputs. It has been great to see that many publishers are working hard to provide ‘fast track’ review service for COVID-19 related papers (due to the importance of timely publication of these works), and – so far – it appears that this hasn’t detracted anything away from the rigor of peer-review. Nevertheless, it is important that publishers continue to focus on prioritising rigor and reliability over and above speed of publication (particularly now that preprint services are covering the ‘speed’ side of things so well).
"It has been great to see that many publishers are working hard to provide ‘fast track’ review service for COVID-19 related papers...Nevertheless, it is important that publishers continue to focus on prioritising rigor and reliability over and above speed of publication."
In particular, publishers stand in a good position to add value by requiring authors to provide the data underpinning their analyses, and also the code to reproduce their results. If these aspects of the research can be provided as ‘open access’, then this is of even greater benefit. Sharing of the data associated code not only ensures the results are reproducible, but allows other researchers to explore the parameters of the models under different scenarios or as new knowledge arises. This also allows rapid advancements, for instance through allowing modellers to build on previous frameworks rather than starting from scratch.
"In particular, publishers stand in a good position to add value by requiring authors to provide the data underpinning their analyses, and also the code to reproduce their results. If these aspects of the research can be provided as ‘open access’, then this is of even greater benefit."
What do you think about the UK’s approach in relation to research into the pandemic so far? Are there any other efforts you would like to see?
The UK has being playing a leading role in research into the pandemic with various high profile groups, including the CMMID-COVID19 working group at London School of Hygiene & Tropical Medicine which I have been lucky enough to collaborate with during this. Along with producing an enormous amount of high quality science in a short space of time, the UK has also successfully implemented interdisciplinary initiatives, such as the Royal Society RAMP Scheme (mentioned earlier), which aimed to recruit researchers from a diverse range of fields to aid with modelling efforts. One of the benefits of national coordinated schemes like this is that it enables other researchers (who want to help with the pandemic) to do so in a manner which is overseen by epidemiological expertise, and thus avoids some of the issues associated with non-epidemiologists engaging with epidemiological research (e.g. as discussed by Prof Julia Gog in Nature Reviews Physics).
"...the UK has also successfully implemented interdisciplinary initiatives, such as the Royal Society RAMP Scheme..which aimed to recruit researchers from a diverse range of fields to aid with modelling efforts. One of the benefits of national coordinated schemes like this is that it enables other researchers (who want to help with the pandemic) to do so in a manner which is overseen by epidemiological expertise..."
Having said this, it is also the case that the UK are not alone in their efforts, and many other countries are doing providing an excellent service, with multiple groups engaging in dedicate research efforts and advancing science in this area. Indeed, many of the UKs own outputs have also involved international collaborations, which is particularly important given that this is a global pandemic.
What do you think the next three months of research into this crisis will be characterized by?
I think it is likely that the high pace of multi-dimensional research into the COVID-19 will continue for some time, leading to higher and higher quality of outputs as our collective understanding develops. In particular, the tremendous and innovative efforts currently been put into generating new data– ranging from individual-level serological details to global epidemiological trends –means we can only expect to see novel and particularly useful research arising from this. Finally, while the first phase of COVID-19 investigations has been rightly focussed on the applied outcomes for research for helping tackle the COVID-19 pandemic, it is also the case that we have never before seen such a surge of research within one specific area as we are seeing now, and never have we seen one particular topic within science so rapidly become exposed to such an interdisciplinary approach (due to the number of diverse range of research fields involved in the COVID-19 pandemic research). Given this, I think we can also except some large-scale new fundamental insights into understanding the principles of contagions and pandemics in general.
Other blogs you may find interesting:
Research in the time of a pandemic: Leading the Oxford Vaccine Trial
Research in the time of a pandemic: SENTINEL- A proactive, early warning system to pre-empt future pandemics
Research in the time of a pandemic: Korea’s COVID-19 success story
All our interviews reflect the views and opinions of the interviewees.
Josh Firth is a researcher at the Department of Zoology, University of Oxford, where he has been based since 2012, and now holds the position of Merton College Junior Research Fellow and BBSRC Discovery Fellow. Although he is currently working on the COVID-19 pandemic through the Royal Society RAMP Scheme, his primary research is based on fundamental understanding of how individual behaviour and external processes interact to shape social structure, and the consequences of this for social processes, such as contagions. He is also involved in interdisciplinary research across various topics in biology and beyond, such as using virtual systems to understand behaviour, assessing how sociality relates to health in mammalian populations, developing models of assessing the spread of conservation initiatives, and working with biomedical researchers in implementing big datasets to assess human health in relation to activity patterns. For a full description of his research, please visit his website. Dr Firth’s recent preprint “Combining fine-scale social contact data with epidemic modelling reveals interactions between contact tracing, quarantine, testing and physical distancing for controlling COVID-19” is currently in review at Nature Medicine, and he has also previously published in Nature imprints, including “Personality shapes pair bonding in a wild bird social system” and “Nonlethal predator effects on the turn-over of wild bird flocks”.
Website: www.FirthNetwork.com Twitter: @JoshAFirth
