One of the scarier aspects of COVID-19 is that many people who are infected don't know they are carrying the virus, because the symptoms typically range from mild to nonexistent. At the same time, that fact means raw case fatality rates (CFRs), calculated based on the number of known infections, make the disease look much more lethal than it actually is. How big is the difference between the actual CFR and the CFRs suggested by the official numbers? That remains unclear, but recent research and expert estimates suggest some possibilities.
Speaking to reporters earlier this month, Brett Giroir, assistant secretary for health at the U.S. Department of Health and Human Services, called attention to the "denominator problem"—i.e., the exclusion of many people with mild or nonexistent symptoms from official counts of confirmed COVID-19 infections. Giroir noted that "the typical mortality rate for seasonal flu is about 0.1 percent or 0.15 percent." By contrast, "the best estimates now for the overall mortality rate for COVID-19 is somewhere between 0.1 percent and 1 percent."
That range is "lower than you heard, probably, in many reports," Giroir said. "Why is this? Number one, it's because many people don't get sick and don't get tested…so probably for every case, there are at least two or three cases that are not even in the denominator." The CFR for COVID-19 "certainly could be higher" than the CFR for "normal flu," and "it probably is," Giroir said, but it is "not likely to be in the range of 2 to 3 percent," as the crude CFRs for some countries suggest. While the COVID-19 pandemic "is likely more severe in its mortality rate than the typical flu season," he added, "it's certainly within the range."
Three federal public health officials—Anthony Fauci, director of the National Institute of Allergy and Infectious Diseases; H. Clifford Lane, the institute's deputy director for clinical research and special projects; and Robert Redfield, director of the U.S. Centers for Disease Control and Prevention (CDC)—struck a similar note in a New England Journal of Medicine commentary last month. "If one assumes that the number of asymptomatic or minimally symptomatic cases is several times as high as the number of reported cases, the case fatality rate may be considerably less than 1 percent," they wrote. "This suggests that the overall clinical consequences of COVID-19 may ultimately be more akin to those of a severe seasonal influenza (which has a case fatality rate of approximately 0.1 percent) or a pandemic influenza (similar to those in 1957 and 1968) rather than a disease similar to SARS or MERS, which have had case fatality rates of 9 to 10% and 36%, respectively."
A study in Science last week estimated that 86 percent of all infections were undocumented in the early stages of China's epidemic, before the travel restrictions that the government imposed on January 23. In other words, the actual number of infections was roughly six times as high as the official number. If that holds true at this point in the United States, where about 41,000 cases have been confirmed so far, the actual number of infections right now would be nearly 250,000. The actual CFR would be 0.2 percent, compared to the current crude CFR of 1.2 percent.
That makes a huge difference in terms of projected deaths. The CDC's worst-case scenario, which assumes that containment and suppression efforts are largely unsuccessful, imagines 214 million infections (65 percent of the population) and 1.7 million deaths, which implies a CFR of 0.8 percent. If the true CFR were 0.2 percent, the number of deaths would instead be 428,000.
How likely is it that 65 percent of the U.S. population will be infected? The Science study estimates that the basic reproduction number for COVID-19—the average number of people a carrier could be expected to infect—was 2.4 prior to January 23, "indicating a high capacity for sustained transmission." During the period from January 24 through February 8, the researchers estimate, that number dropped to 0.99.
One important question for the United States is the extent to which that reduction depended on government-imposed restrictions, as opposed to voluntary changes in behavior as people took more precautions to avoid infection. While lockdown enthusiasts give the lion's share of credit to coercive measures, it is not reasonable to assume that voluntary steps—avoiding crowds, limiting social interactions, paying extra attention to hygiene, and so on—would have no impact on the reproduction number. It is notable that jurisdictions such as South Korea, Singapore, Taiwan, and Hong Kong seem to be having similar success at controlling their epidemics without imposing restrictions as severe as China's.
Even if the CFR for COVID-19 turns out to be something like 0.2 percent, it would still be deadlier than the seasonal flu, which has an estimated CFR of about 0.1 percent. Furthermore, COVID-19 seems to be more readily transmissible, with an estimated basic reproduction number between 2 and 2.5, compared to around 1.3 for the flu. COVID-19 also has a much longer incubation period than the seasonal flu: up to 14 days vs. up to four days.
"In theory, COVID-19 could spread further and faster than flu," notes microbiologist Alex Berezow, vice president of scientific communications at the American Council on Science and Health. Yet so far, he says, "it doesn't appear to be doing that….Every year, up to 1 billion people around the world get influenza and about 300,000 to 500,000 will die. The average seasonal flu case-fatality rate is 0.1%. But the sheer volume of cases means that, in the U.S. alone, 22,000 to 55,000 Americans have already died of flu during the 2019-20 season."
By comparison, the current official tally of COVID-19 cases is about 367,000 worldwide. Even if that number is off by a factor of six, the true number of infections would be around 2.2 million, a tiny fraction of annual flu cases. That certainly is not cause for complacency, since the number is bound to grow and since existing precautions help explain why the number of COVID-19 cases so far seems relatively small. But the comparison does suggest that the fears underlying the worst-case scenarios are overblown.
We do not have a clear picture of the infection rate in the United States because there has been no testing of nationally representative samples. Iceland, the one country that is screening its general population, recently reported the results of 1,800 tests involving asymptomatic people, of which 1 percent were positive for COVID-19. In a larger sample that included people with symptoms, about 6 percent tested positive.
It would be risky to extrapolate from those results, which involve a relatively isolated country with a total population of just 364,000. But unless something similar is attempted in the United States, policy makers will continue to make decisions in the dark, with potentially devastating consequences.