Putting epidemiology at the service of the economy during the COVID-19 (Coronavirus) pandemic

It has been almost a month since my last blog on COVID-19, but it feels like a lifetime. So much ink has been poured into the ever-growing literature, both serious and peer-reviewed, and elsewhere in the news and social media. We do know much more now about SARS-CoV-2 (the virus), COVID-19 (the disease itself), and the making of the pandemic.

We have also begun to learn more about the health – both physical and mental – and socioeconomic consequences. We have come to realize that except for a few countries like Republic of Korea and Singapore the global community was caught unprepared, and as such, had to resort to drastic social distancing and lockdown of non-essential economic activity to contain the contagion.   

Many countries, rich and poor alike, are grappling with severe challenges: shortages in personal protective equipment, insufficient test stocks to detect all suspected cases, and limited capacity for treating the severely sick in intensive care units.

What will be next?

All countries will now have to ride this epidemic wave with whatever public health and medical care measures they have at their disposal. A few may also benefit from the largesse of others. There is not much time to question the courses of action taken by countries around the globe, as they all are busy extinguishing the fire and learning by trial and error. Yet, there is a lingering dilemma, a yet-to-be-explicit debate on the tradeoffs.

Are the drastic measures are going too far, to the point of risking a total economic meltdown? Could there be a middle ground, a tradeoff between optimal health security and economic shutdown?  Where would we draw the line?  Many economists in academic circles and at the World Bank have been modeling different scenarios, which are as credible as the assumptions their models are built on.

A wealth of data

I believe epidemiology has a lot to contribute to the debate, perhaps more so now than it could have a few weeks back. First, we have now reached almost two million cases globally. Equally important are about 500,000 recovered cases, in addition to perhaps another million or so asymptomatic cases that have not been diagnosed.

With such large numbers, cases could be broken down by location, age, sex, and other critical parameters to dive deeper into the specifics of the pandemic. 

Second, not a day passes without a new test being readied for rapid on-site diagnosis. This means we can increasingly diagnose cases with the now well-known Polymerase Chain Reaction (PCR) tests, which detect the virus during the period of infection (usually lasting around three weeks), but not following recovery.

But even more importantly, we are seeing the arrival of tests that detect the presence of antibodies that the immune system produces to fight infection. There are two types of antibodies: the Immunoglobulin M (IgM) antibody becomes detectable sometime during the onset of symptoms and disappears as the patient recovers, and the other, Immunoglobulin G (IgG), appears sometime during recovery but remains detectable beyond recovery – perhaps permanently and rendering the recovered patient immune to re-infection. 

Third, we are operating in a novel context of “natural experiment,” where mobility is mostly –and in some places, totally – restrained by social distancing and quarantine measures, preventing international and even domestic travel.

We have an unprecedented opportunity to use both tests to precisely estimate the distributional patterns and trends of COVID-19 cases through mass screening or random surveys and sentinel site surveillance. These tools can provide us with a thorough analysis of cases broken down by key epidemiologic characteristics of the infected, asymptomatic and recovered by location. 

This data could help us determine whether we really must resort to a total lockdown, for example, in a country of islands where most of the cases are located on one. We needn’t have the same drastic measures in all provinces if patterns in case distribution could allow authorities to quickly circumscribe the “hot spots,” alleviating the socioeconomic consequences of the pandemic.

Tailoring quarantine measures to specific locations

It is time to gradually move away from aggregate numbers and rates. We need to take full advantage of the large and growing case numbers to pinpoint specific geographies for containment efforts – putting the finger where it hurts and relieving pressure from where it does not – making full use of the notion of a cordon sanitaire.

I would defer to modelists on whether location-specific containment efforts might have consequences for the evolution of herd immunity and the flattening of the curve. Many location-specific curves might be more relevant to forecasting unmet needs in the health workforce, ICU beds, and ventilators, and enabling better-informed decisions on social distancing.

I bet the costs of scaled-up testing would be dwarfed by the potential economic benefits of tailored lockdowns, not to mention the added benefit of being able to determine where to best deploy scarce health care resources. 

Way back in 1854, John Snow, an inquisitive physician living before the era of germ theory, was able to identify a source of the cholera outbreak in London by mapping out all the cases with a piece of paper and a pencil.  This allowed authorities to remove the handle of the culprit water pump on Broad Street instead of those of all water pumps in the city. In an age of spatio-temporal modeling and small area estimation, we should be able to do the same as Snow, mapping COVID-19 across diverse geographies by drawing on big data and massive computational power.

I think it is worth a try, given the alternative of total nationwide lockdowns and the threat of new epidemic waves on the horizon.

What do you think?