Deborah Birx, the White House’s coronavirus response coordinator, has earned the public’s confidence through a career focused on preventing the spread of AIDS in the United States and around the world. Her understanding of the data and systems that can be used to combat epidemics is a critical asset in the effort to halt the spread of the novel coronavirus that has infected at least 85,000 Americans.

It’s therefore worth pointing out when Birx’s focus on the data doesn’t actually support her argument. As with an assertion she made during a briefing Thursday.

“We do have 19 out of our 50 states, to be reminded, that had early cases, but have persistently low level of cases,” Birx said. “And at this point have less than 200 cases. So that’s almost 40 percent of the country with extraordinarily low numbers.”

She was making this case — that much of the country has effectively contained the coronavirus — in apparent service to President Trump’s focus: scaling back restrictions on businesses and public events, which were aimed at slowing or stopping the virus’s spread. But the publicly available data, compiled by Johns Hopkins University, don’t support the optimism Birx conveyed.

First, there’s the simple fact that 19 states is not necessarily equivalent to “almost 40 percent of the country.” As of Thursday, Johns Hopkins’ data had 17 states which had fewer than 200 confirmed cases. Those 17 states are one-third of the 50 states and Washington, D.C. But they are also home to just over 7 percent of the country’s population.

There is a correlation between population and case totals, for understandable reasons. The epicenter of the outbreak in the United States is the nation’s largest city, New York. It has driven the state’s total confirmed case count to nearly 40,000 — more than in all but four countries excluding the United States. That total has risen rapidly; the state’s first confirmed case came later than a number of others but the total rose quickly. (The circles on the graphs below are scaled to the state’s total population.)

You can see on that graph all of the states with fewer than 200 cases, in blue. You’ll notice that most of them saw their first cases relatively recently as well. More on that in a bit.

If we exclude New York from the graph and control for population — that is, assess the number of confirmed cases per 100,000 residents — the picture shifts a bit. Those states with fewer than 200 cases overlap with a bigger cluster of states in which cases emerged at about the same time. States such as Vermont stand out as having quickly added cases relative to population.

It’s the rate at which cases are confirmed that’s the concern. Looking at the rate at which new cases were added relative to each state’s 10th confirmed case gives us a picture of how the rates of new cases vary by state. New York has added a lot of cases quickly. Washington state, the site of the first confirmed case in the United States, has added them more slowly.

If we again control for population, those states with fewer than 200 cases offer less cause for optimism. Vermont, for example, is adding new cases even more quickly than New York or New Jersey, relative to total population. There’s not much obvious difference in the rates in those states with fewer than 200 cases and many of the states with more than 200 cases.

That’s true in terms of actual counts, too. Many of the states with fewer than 200 cases will obviously soon surpass that (arbitrary) benchmark. Some states do stand out as having slower-than-average increases in their totals, including Nebraska.

As we noted Thursday, though, Nebraska has only conducted about 71 tests for every 100,000 residents, about half the average of 144 tests per 100,000 residents across the country. (Birx followed her comments about the states with relatively few cases by insisting: “And they are testing.”)

Again, many of the states with relatively few cases saw their first cases more recently. If we compare the total number of confirmed cases to the day of the month, we see both that the rate of increase in those states generally matches the rest of the country and that many of them saw their first confirmations later in the month.

Birx went on to downplay models that suggest a much larger death toll than the available data would suggest is likely, which, if accurate, is one bit of good news. She noted a significant reduction in one influential model, a reduction the study’s authors attribute to countries enacting the sorts of distancing measures meant to limit the spread of the virus — and the sorts of measures that Trump wants to scale back.

The argument that 40 percent of the country is relatively unaffected by the virus doesn’t match the data. But even if it did, that’s still not necessarily good news.

China’s public data suggest that it managed to limit the spread of the virus to one region of the country. The data in the United States show that we’ve been less successful. Even if it were the case that there was little to no spread in 20 states, that’s hardly good news for the other 30 states and Washington, D.C.

It’s certainly little consolation for New York City.