We interviewed Coronasurveys Project leader Professor Antonio Fernandez Anta to find out more about their pioneering project that’s helping in the fight against Covid-19 and why the project has an important Liverpool connection.
Liverpool Noise: Welcome to Liverpool Noise.
Prof Antonio Fernandez Anta: Thank you for having me here!
LN: Can you tell us about the corona surveys project? how does it work? what gave you the idea?
Prof. Anta: CoronaSurveys is an academic project whose main objective is collecting data that can be used to know how the COVID-19 pandemic evolves over time. The main data collection tool is a survey though which participants can report the health status of their personal contacts. The main difference between our survey and most surveys out there is the fact that a participant is not reporting their own data. This difference results in that we get information of a larger fraction of the population with fewer responses, and that the privacy of those reported is fully preserved. (Additionally we have kept the survey very short, so it takes less than a minute to answer.)
In a nutshell, in the survey we ask every participant for the size of her circle of personal contacts, and how many of them have been affected by COVID-19 (have been sick, have shown symptoms, have been in the hospital, etc.). Aggregating these responses allows to obtain estimates of the fraction of the population that has been affected.
The project started in mid March when I realised that the official data was obviously under reporting the number of cases in most countries, due to the testing limitations. I deployed a toy experiment in Twitter with a simple survey, which revealed that the official figures of COVID-19 cases in Spain were more than ten times lower than the real numbers. This evolved into a project that covers all countries in the world and has surveys in 60 languages.
LN: The website says you work with a lot of institutions and people, can you tell us more?
Prof. Anta: Since the beginning, the CoronaSurveys project has been a collective initiative driven by volunteers from many countries and institutions. These have taken many different roles and tasks, like maintaining the server where the survey is stored, designing the website, devising the algorithms to obtain estimates, or promoting the surveys in each country. This has led to a large core of more than 30 researchers from many countries and areas (Computer Science, Statistics, Psychology, etc.) working together, and a large number of collaborators. It may be worth mentioning here the help we had from another initiative, Crowdfight COVID-19, which mobilised more than 200 volunteer translators.
LN: Can you tell us more about local surveys? Are you running one in Liverpool?
Prof. Anta:We have a main global survey that, as I mentioned, is open for all countries. However, when convenient we can deploy local surveys to get data at a smaller scale. We have done this already in two towns outside Madrid, and Liverpool will be our next local survey deployment. In the Liverpool local survey we will be collecting data at the level of local authorities. This will allow us to evaluate the percentage of the population in each area that has already been infected, and how the number of new infections evolve.
LN: What do you hope the impact of your work will be? have there been any notable achievements?
The main technique used in this project, which is the use of indirect reporting, has never been used at the geographical and time scales of this study, as far as we know. (Another novelty of the project is the intensive use of online social networks for the promotion of the surveys.) This results in a combination of techniques that is completely new and extremely promising. This project may be the start of a new methodology to probe society, as a new application of citizen science.
Of course, one of the main questions with a new methodology is how accurate the results obtained with it are. We have compared our estimates of the prevalence (fraction of the population that has been infected with COVID-19) in the different regions of Spain with the results obtained by a nation-wide serology study involving more than 60,000 people. We have found that in most regions both values are very close. This is a result that is very promising.
LN: We know the pandemic might change and evolve over time, is that something your surveys could predict, or help with?
Prof. Anta: Yes. As I mentioned, we ask participants about past cases among their contacts, but also about new cases. By now there is a significant among of knowledge about the timeline of a COVID-19 infection, like that severe cases take 3 to 10 days to be taken to the hospital after they start with symptoms. Similarly, it is known that roughly 25% of the cases require hospitalisation. Then, if we find a surge of new cases today we can forewarn hospitals that they should expect an increase in cases in the next few days, so they can take appropriate actions.
Similarly, we could help to identify if the number of cases is decreasing, possibly because the measures taken are being effective.
LN: Well, thanks for yout time and we hope the project is a big success.
Prof. Anta: Thank you for giving me the opportunity to present the CoronaSurveys project to your audience. If they want to know more, they can visit our website, which was built by a Liverpudlian! https://coronasurveys.org/