As New Data Improve Our Understanding of COVID-19, Our Responses to It Could Become More Effective

U ntil recently, the most commonly accepted timeline suggested that the COVID-19 first appeared in Wuhan, China in late December 2019, first hit the U.S. in mid January, and caused its first American death at the end of February. In total, over 1 million individuals have been infected with the coronavirus in the U.S., and 58,000 have died, for a rough American case-fatality rate of 5.8 percent.

Unfortunately, this epidemiological profile comes with several caveats: There are many Americans who have caught the virus but don’t show up in the official infected count because they haven’t been tested. And at least some have died from the virus without being counted toward the official death toll, so the true case-fatality rate remains unknown. The noise inherent in the data makes it all the more essential that we continue to develop our epidemiological understanding of COVID-19 so that we can make sound policy decisions in the fight against it.

Fortunately, over the past few weeks, new information has emerged that has helped that effort along. It is now apparent that COVID-19 landed on the shores of the U.S. earlier than we’d thought. After an autopsy, it was determined that a 57-year-old California woman, who was diagnosed with the flu and died on February 6 of what was thought to be a heart attack, had actually died of complications from COVID-19. February 6 was three weeks earlier than the COVID-19 death previously thought to be the first in the U.S., meaning the virus had three crucial extra weeks to spread through our country.

There is also new word coming from several recent regional antibody surveys on the percentage of people who were infected during the early months of the pandemic. It is assumed that individuals who have been infected develop antibodies to COVID-19. While it is still unknown whether the presence of antibodies can provide some degree of immunity, the percentage of people with antibodies gives a good indication of how many people were infected in the past. A study of New York State found that among 3,000 individuals who were out shopping, 13.9 percent had antibodies to the virus. (In New York City, the rate was 21.2 percent.) A study of 3,000 Facebook users in Santa Clara County, Calif. showed antibodies in between 2.2 percent and 3.4 percent of those tested. And in Los Angeles, 4.1 percent of a random sample of 863 people were shown to have antibodies. Each of these studies can only tell part of the story, because of how their subjects were recruited. But they still provide a helpful idea of the extent of the virus’s spread in these areas early in the outbreak.

We have also learned much about the distribution of the disease across the country and within cities and states. We know that large metropolitan areas are, in general, more heavily affected. We know that the number of COVID-19 cases varies by neighborhood in these cities, and that cases and deaths have been higher in African-American and Latino neighborhoods, which often have high population density and limited community resources. American-Indian reservations with similar environmental conditions seem to have an unusually high number of cases as well.

There is also emerging data that suggest we still do not know COVID-19’s true case-fatality rate. It has long been assumed that the case-fatality rate would drop to below one as more asymptomatic individuals were tested, leading to an increase in the equation’s denominator. But we are now learning that more reported deaths may need to be attributed to coronavirus than previously were, increasing the equation’s numerator and potentially keeping the case-fatality rate higher than anticipated.

Recent changes to how the CDC reports COVID-19 deaths could also increase the official death toll. The CDC has changed the COVID-19 case definition that is used to standardize what is considered a true reportable case of infection. Cases are now defined as confirmed if there is laboratory evidence of infection and probable if there is insufficient laboratory evidence but the individual was in close contact with a confirmed case or had certain clinical indicators. As of April 14, CDC case and death numbers included both categories. In the past few weeks, the CDC has also expanded the number of clinical features included in the case definition.

An analysis by the Washington Post and Yale University researchers further validates concerns about the case-fatality rate. It showed that there were more total deaths from all causes observed during March and early April than were observed on average in past years. It found that there were over 15,000 “excess” deaths nationwide — that is, 15,000 more than we’d have expected from historical data — during this time period. One obvious interpretation is that these excess deaths are unreported cases of COVID-19. But of course, it’s also possible that because of COVID-19’s drain on health-care resources, people with other mortal conditions did not receive the care they normally would have and died.

All of this new information helps fill in the gaps in our understanding of the virus and its spread. We can now assume given that by the time travel from China (January 31) and then Europe (March 11) had been banned by the Trump administration, the virus had already seeded New York City and large metro areas on the West Coast, and that it spread widely throughout the country thereafter. It’s also clear that New York City and other such metropolises were particularly hard hit, in all likelihood because of their population density.

Grim as it may be to read, this is good news: The better our understanding of the virus, the more effective we can be in responding to it. But now is not the time for congratulations. We must continue to produce more data to assist government leaders in developing their plans for when, where, and how to scale back on extreme-mitigation measures and move to a more measured containment strategy of wide-scale testing, masks, social distancing, good hand hygiene, and contact tracing. Countless lives and the health of our economy depend on it.