As the pandemic stretches into fall, some COVID-19 trackers have reached a harrowing milestone: 200,000 deaths. But The COVID Tracking Project’s tracker hasn’t yet reached that number, and we want to explain why.
Just as there’s no federal mandate requiring states and territories to compile COVID-19 data the same way, there’s also no universal approach to reporting available numbers. Various news outlets and research institutions are handling the process differently, and that’s not necessarily cause for concern.
Some large cities—New York City, for example—have their own health departments, and sometimes these health departments have a process of counting COVID-19 deaths that’s different from their state health department’s approach. At The COVID Tracking Project, our methodology relies on state-level data. What that means is our numbers can run behind the trackers that prioritize data from counties or cities.
For example, as of September 22, New York City has reported 4,630 probable COVID-19 deaths. State health officials in New York, however, have decided to report only confirmed deaths. That discrepancy helps explain why the state’s official death toll for the city is far lower than the city’s own number. The COVID Tracking Project pulls its numbers from the state’s account.
We want to be clear—just because our data doesn’t include some probable deaths doesn’t make them any less worthy of being counted. This is simply a matter of methodology differing between jurisdictions. Our process was designed to collect the most comprehensive standardized data we can at the state level, given our resources. Though the CDC offers guidance about reporting probable deaths, not all states follow it, so the reality is that some deaths won’t be reflected in our count.
It’s also worth remembering that 200,000 COVID-19 deaths in any outlet’s data is almost certainly a steep undercount. Given the country’s limited access to testing, especially but not exclusively in the pandemic’s early months, a more accurate COVID-19 death toll is impossible to calculate. In epidemiological terms, “excess deaths”—numbers of deaths that exceed what’s expected in a given period—suggest even higher levels of mortality from COVID-19 than state trackers may ever report.
As of September 22, The COVID Tracking Project is reporting 192,741 deaths. The New York Times, by contrast, is reporting 200,443, Johns Hopkins is reporting 200,541, and the CDC is at 199,462. The death toll has doubled in the past four months; The New York Times reported 100,000 deaths on May 24.
More “Hospitalization and Death Data” posts
We're up to 24 states publishing both confirmed and probable COVID-19 deaths, and we're adding those data points into our API. But states are also using two different ways of deciding which deaths to count as COVID-19 deaths.
States across the US use two primary methods for announcing COVID-19 deaths: date of death (reported) and date of death (actual). To analyze how the pandemic is trending across the country, understanding the relationship between these two data points is crucial. Here's what we've learned from investigating both methods in three of our largest hotspots: Arizona, Florida, and Texas.
A long holiday weekend makes ambiguous testing data even harder to understand, but hospitalizations are dropping, which is good.