Deducing Which Pandemic Policies Work Best

Nguyen-Viet says the team identified ten stalls selling wildlife, either wild or farmed, Q&A that could have carried the virus into the market from farms in southern China. Some wild animals sold for meat, such as rabbits and ferret-badgers, are susceptible to SARS-CoV-2. Deducing which pandemic WHO team member Peter Daszak, president

EUGENIE SOPHIE EUGENIE of the non-profit research organization Eco- policies work best health Alliance in New York City, says that the farms should be investigated to see whether Since the start of the COVID-19 pandemic, were among the most effective policies there were infections in the animals or among thousands of scientists and volunteers (N. Haug et al. Nature Human Behav. 4, workers. When the team interviewed the first have been tracking the interventions 1303–1312; 2020). person known to have COVID-19, he mentioned that governments have adopted to curb But there is less agreement, when that his parents had visited a local community viral spread — from closing restaurants analysing different trackers, on how to wet market, says Daszak. to mandating face masks. They hope to rank these measures. For example, it is not deduce which policies are most effective. certain that highly restrictive measures are Did frozen wild-animal meat have a At a conference this month, scientists automatically more effective than a smart role in the early spread of the virus? involved in 50 tracking databases mix of comparatively modest restrictions The WHO team concluded that it’s most likely discussed their efforts. Peter Klimek, a implemented with better timing. the virus jumped from live animals to people, mathematical physicist at the Complexity but Ben Embarek says it is possible that the Science Hub (CSH) Vienna and the Medical Why is it hard to estimate the effects of virus entered the Huanan market through University of Vienna, who is involved in interventions? infected frozen wild animals from farms in the CSH’s tracking project, explains the It is difficult to untangle the effects of any southern China, and then sparked an outbreak. challenge. given measure from those of other policy Although researchers in China have isolated interventions. There are many statistical viral RNA from the packaging of imported fro- How much work has gone into these approaches to disentangling relationships zen fish4, Ben Embarek says the WHO team trackers? in complex systems, but none of them is concluded that these goods were not likely to In our tracker, more than 40 volunteers and perfect. And sociocultural factors can make be the route of the virus’s first arrival in Wuhan. scientists have been involved in assigning social distancing more effective in one Lipkin says there is no evidence that the codes to more than 11,000 measures in country than in another. virus entered the market through infected 57 countries. Other efforts, such as the The effects of interventions also change frozen wild animals. It could have just as easily CoronaNet consortium and the government over time. The situation has become more been brought in by infected people who had response tracker run by the University complicated as government interventions handled wild animals, he says. of Oxford, UK, each have hundreds of have become more diverse, and as people volunteers and researchers. Some tracking adhere less willingly to restrictions. Was the virus circulating in animals efforts have received funding, but most are in China before the pandemic? struggling owing to a lack of money, and Why not combine the trackers? To establish which animal passed the virus to some have had to stop. It’s also a challenge Each tracker has its own perspectives. Some people, researchers need to find evidence of to keep volunteers motivated. do integrate data from different databases, the virus in that species. Researchers in China including one maintained by the World tested some 30,000 wild, farmed and domes- How have the trackers been used? Health Organization. But this comes at the tic animals in 2019 and 2020 but found no evi- We advise the Austrian government on expense of some of the granularity that dence of active or past SARS-CoV-2 infection, policy measures to contain the spread the original databases might have had. except in some cats in Wuhan in March 20205. of coronavirus and avoid health-system From the perspective of data quality and However, Ben Embarek says that these sur- overload. When we’re asked questions, such reproducibility of results, merging trackers veys were not representative of China’s overall as why some countries have much lower into a super-database isn’t a good idea. animal population, and that many more ani- case numbers than others, the first places we mals need to be tested for traces of infection, look are the databases tracking government How might this sort of work change particularly on wildlife farms. “The amount interventions. We still don’t know what is the in the future? of testing that’s been done is not sufficient to best way to plug the data from the tracking There is growing societal and political say, in any way, that wildlife farms were not systems into mathematical models. But the pressure to understand hypothetical carrying the virus,” says Daszak. trackers are a unique treasure trove that we scenarios: how not having implemented The explosive way in which the outbreak can use to make epidemiological modelling a certain measure might have changed took off in Wuhan in December suggests a data-driven science and to prepare for the the course of the pandemic. For example, that the virus was probably introduced once, next pandemic. was it really necessary to close schools? through the wildlife trade, says Daszak. He Or will the social and economic costs turn says future testing should focus on farmed What have they told us so far? out to have outweighed the health-related wild animals. When many countries applied various benefits? Without reliable tracker data, control measures simultaneously, we there will be no solid evidence to answer 1. Carrat, F. et al. Eur. J. Epidemiol. https://doi.org/10.1007/ knew very little about the effects of such questions. s10654-020-00716-2 (2021). 2. Amendola, A. et al. Emerg. Infect. Dis. 27, 648–650 (2021). government interventions. When more data 3. Apolone, G. et al. Tumori J. https://doi. became available, we found that curfews, Interview by Quirin Schiermeier org/10.1177/0300891620974755 (2020). cancellations of small gatherings, and This interview has been edited for length and 4. Liu, P. et al. Biosaf. Health 2, 199–201 (2020). 5. Zhang, Q. et al. Emerg. Microbes. Infect. 9, 2013–2019 closures of schools, shops and restaurants clarity. (2020).