By Veronica Perry
As America copes with the unfolding ramifications of COVID-19 prevention measures, information about the pandemic’s effect on the economy is taking center stage. Responding to the urgent need for insight, on March 27 the USC Dornsife Department of Economics hosted a panel on the economic consequences and implications of the COVID- 19 health crisis. USC economists Romain Ranciere and Simone Schaner moderated a discussion of expert professionals, academics and fellow economists from USC, UC Berkeley, UCLA, NYU, The International Monetary Fund, Geneva Graduate Institute of International and Development Studies, and Ecole Polytechnique. The event was viewed by 300 people, the maximum allowed for a Zoom meeting.
The various presentations outlined the severity of the crisis on emerging economies, lasting effects on the financial market and trade, implications for mental health, social consequences across the globe, U.S. attitudes and behaviors towards the virus, urban and environmental effects as well as effective testing strategies. Neeraj Sood, professor and vice dean for research at the USC Price School of Public Policy and Senior Fellow at the USC Schaeffer Center, discussed COVID-19’s fatality rate estimates.
Currently, reports share that the mortality rate from the virus ranges from 1-4% (i.e. one to four people out of 100 who are confirmed to have the disease will die from it). In these calculations, Sood explains that the numerator includes those who have died from COVID-19, while the denominator represents every person who has a confirmed case. This is where he expresses the crux of the issue: America has only been testing people who are severely symptomatic or meet other limited criteria. In doing so, they are missing several people from the denominator: those who had no or mild symptoms from COVID-19 or those who did not receive a test despite symptoms. According to Sood, the confirmed cases of COVID-19 suffer from major selection bias.
The modelers from Imperial College assume that for every person that is a confirmed case it is likely that two additional people in the population also have the virus – but are not represented in the denominator. Sood thinks this is a vast underestimate. To put that in perspective, the Center for Disease Control (CDC) uses as a multiplier of 80 when examining the flu. This means that for every confirmed case of the flu reported to the CDC, the CDC estimates that 80 more people had the flu but did not report it.
To find the reasonable multiplier for COVID-19, Sood asserts the way forward should involve random sampling. “You take a random sample of the population and test everyone in the random sample to determine whether or not they’ve had COVID-19,” Sood said to the Zoom audience. “Once you know that, you have a better understanding of the total number of people infected in the population and that becomes your true denominator.” Then you can look at the number of deaths from the virus and compute the true infected mortality rate.
There is some evidence from testing in other places, where extensive testing has tried to determine the appropriate multiplier for COVID-19. For example, every person who left Wuhan on flights to different countries were tested. When they arrived at their destination, they were tested immediately. Based on those tests, it’s estimated that since January 31, the COVID-19 infection prevalence in Wuhan was about 1%. “What this means is that there were about 200,000 people in Wuhan who had COVID-19, but the number of confirmed cases in Wuhan was 30 times less. This basically says that in Wuhan, with all the testing that was done, the multiplier was around 30,” Sood said.
Sood suggests that random sampling should begin in the United States as soon as possible to figure out exactly how many people are infected and determine the true mortality rate. Sood explained further than random sampling will show whether or not social distancing is effective in limiting the spread of the virus. “If 30% of the population already had COVID-19 and has recovered, that has different implications for social distancing,” he concluded. Sood is now launching a study to conduct such random sampling in Los Angeles county.
Read more about Sood’s recommendation to implement random sampling in his op-ed, featured in The Wall Street Journal.
Professor