And perhaps more importantly, a consortium of public health researchers, including Harvard, finds that only a handful are doing enough testing to effectively suppress the virus, that is, to bring new cases down to a low enough level to allow everyday life to return to some semblance of normalcy.

When NPR and Harvard first analyzed testing on a state-by-state basis May 7, the nation was conducting about 250,000 tests every day, and according to the analysis, only nine states were doing enough testing to keep their outbreaks under control.

But the national totals are far below what the Harvard group says is needed to contain the current outbreaks — 1 million tests daily — and yet farther from a level that could truly beat back the pandemic in this country.

The less ambitious target is intended to achieve what public health experts call "mitigation." This means keeping the size of current outbreaks from growing; this approach requires enough testing to get to 10% or fewer positive tests.

To achieve suppression, states and institutions need to take a proactive approach to testing and "cast a much wider net," Jha says.

So how much testing would be needed to go beyond mitigation and actually achieve suppression.

For that, the Harvard Global Health group developed a new metric in collaboration with other academic groups across the country: They estimate that communities need to test widely enough so that the number of people testing positive for the virus is 3% or lower.

overall would need 4.3 million tests per day, with the amount of testing in each state varying depending on the current size of their outbreaks, to achieve that goal.

It could bring communities to the point "where people will have confidence going out to restaurants and bars and opening up schools without having large outbreaks and without having to shut down again," Jha says.

According to the new analysis, only four states are doing enough testing for suppression: Alaska, Hawaii, Montana and Vermont — all relatively small populations with small outbreaks.

Plescia says the main bottleneck to getting enough testing continues to be chronic shortages of the test kits themselves as well as the supplies needed to perform the tests.

Jennifer Nuzzo, an epidemiologist at Johns Hopkins tracking testing who is not involved in the Harvard analysis, questions the usefulness of projecting specific numbers of tests needed to control the virus

Instead, Nuzzo says her group has long advocated focusing on the percentage of people testing positive as a more reliable metric

The lower the percentage, the greater the chance testing is catching enough new infections to control the virus, Nuzzo says

I don't envision that we are ever going to be testing so broadly," she says