The outbreak science or “the use of infectious disease modeling to support public health decision making” is not something new, yet it has never been so influential before. For example, Jacco Wallinga, Mathematician and Chief Epidemic Modeler at the Netherland’s National Institute for Public Health and the Environment (RIVM) has been advising the government on whether actions such as shutting schools and businesses will dampen the Covid-19 pandemic in the country.

What Wallinga did was to forecast the number of infected patients needing acute hospital services. His work had resulted in the Dutch Prime Minister Mark Rutte to take a less adverse approach of keeping the virus within groups that are not at risk of becoming seriously ill and ensuring these individuals do not overwhelm the healthcare system. In the UK, the government has also taken a soft approach at the beginning. The model devised by researchers at the Imperial College London warned that drastic measures of locking cities and closing entities may “result in a large second epidemic once measures were lifted”.

However, within a fortnight, the same group of researchers published a heavily revised model, conjured based on the latest data collected within the country and Italy. It suggested letting loose of the present control may not be ideal as it will still create a spike and fill up to twice as many hospital beds in intensive care units. Thus, the UK should have stricter control and this led to the government announcing a lockdown within days.

A fine line between estimates and controversies

Building a computational model forecasting how the pandemic will spread in an area requires estimates and assumptions. In the case of Wallinga, he employed a compartment model; by assuming the Dutch population is homogeneously mixed; he spilt them into four groups: The healthy, the sick, the recovered and the deceased. He came up with equations that simulate how people will move from one category to another as the pandemic lingers over the weeks and months.

Wallinga told Science, there is no sophisticated mathematics here, but since Covid-19 is a novel coronavirus, so he has to make many approximations before the actual modeling. Often, there is just a fine line between these estimates and controversies because a tiny tweak in the pathogen’s features and which population it’s going to affect, will change the model outcome is a very different manner. Wallinga’s nightmare will be an error in his model which causes relentless demands in hospital beds and depleting the nation’s healthcare resources.

To prevent that, Wallinga and his team had spent much time around R0, a parameter first published in late January by other epidemiologists, which approximate the number of new infections brought about by each infected individual when no control measure is put in place. Building on this, Wallinga and his team believe it will take three to six days before an infected person begins infecting others. What Wallinga does may only apply to a relatively small country like the Netherlands. Some computational modelers, especially those targeting at a heterogeneous population as in the US or Europe, will choose to trace the daily routines and interactions of thousands and millions of people.

Policymakers’ reliance on modeling

Regardless of which, as mentioned, modeling can be controversial, depending on how it’s used and explained. For example, a group of modelers from the University of Oxford concluded the death toll in the UK does not explain the severity of the disease. Rather, the SARS-CoV-2 might have been spreading widely in the population since January and infected over half of the population; making some of them immune while others become seriously ill.

Besides, policymakers’ reliance on modeling also made some questioned if that’s the right thing to do. After all, most modelers are academia who don’t usually play a part in decision making or policy. Furthermore, the virus has not been thoroughly studied and theoretical models based mostly on assumptions may not usually play out as it is in real life. Nevertheless, policymakers can’t make decisions on the basis of nothing, particularly in times of crisis. As such, it’s important for them to know the limits of models.

Models are best for identifying things that are less noticeable like mass temperature screening in public areas may not successfully singling out individuals infected with the coronavirus. Likewise, models cannot tackle “response fatigue” or the failure to keep up with extreme quarantine measures and social distancing over a prolong period of time on a vast number of people.

Last but not least, models are enacted into policies without considerations made to long term impact. Most countries have neglected the fact although massive, long lockdowns may bring a halt to the spread of the virus but it will also have an enormous economic impact that may even be more damaging than the pandemic itself.


Author Bio

Hazel Tang A science writer with data background and an interest in the current affair, culture, and arts; a no-med from an (almost) all-med family. Follow on Twitter.