by James A. Bacon

Later today Governor Ralph Northam will issue his mask-wearing mandate for Virginians. I’ll reserve comment about the details of the plan for when I see them. In the meantime, it is worth considering what data he might draw upon to justify his measures.

The Virginia Department of Health COVID-19 dashboard highlights several “key measures.” One, seen above, tracks the total number of COVID-19 cases. This particular graph strongly suggests that the COVID-19 epidemic is spreading as rapidly as it ever was — indeed it has ticked back up in recent days.

Reinforcing this impression, today’s VDH dashboard reports a record increase of new confirmed COVID-19 cases — 1,615. Data compiled over the weekend is subject to jinky behavior due to reporting delays, but that number follows 1,483 cases the day before. So, if the Governor wants to make the case that the coronavirus is still spreading, there is data to support him. The Washington Post, which can scarcely keep its pro-shutdown sentiments out of its news articles, picked up on the latter number in its news headline this morning — and we all know that Northam (or his people) pay close attention to the WaPo’s coverage.

However…

What the Post neglected to report (I won’t speculate upon the reason) is that Virginia is finally experiencing its long-promised surge in testing. The VDH website reports more than 36,000 tests over the past three days, compared to 20,000 over the previous three-day period and 22,000 the three days before that.

All other things being equal, more tests = more confirmed cases. It is difficult to know if the increased number of confirmed cases should be attributed to actual spread of the disease or the increased frequency of testing subject to VDH protocols and priorities.

That’s why the VDH key measures also includes a graph showing the percentage of tests that are positive. The idea is that if the percentage of positive tests is declining, the virus may be retreating.

But there are problems with this graph, too. To some degree, it compares apples to oranges. In the early days of the epidemic, tests were scarce, and hospitals and public health authorities reserved tests to confirm what, based on symptoms and severity, seemed to be the most likely and most acute cases. Some days the percentage of positive tests exceeded 30%, and the seven-day moving average, as shown above, exceeded 25%. As testing became more available, especially by private labs, the tests came to reflect the prevalence of the disease in the broader population where, not surprisingly, the incidence was lower than in hospital sick wards.

A different measure, not subject to such testing biases, suggests that the spread is receding markedly. When people get extremely sick, they check into the hospital. VDH data captures the full universe of COVID-19 patients admitted to the hospital, not a small and potentially biased sample. The trend of hospitalizations is clearly down.

While the epidemic may be losing momentum, the COVID-19 virus is still out there. And it is still spreading in the community. Although deaths have been concentrated in nursing homes — almost three out of five statewide — the cases have been broadly diffused overall. Outbreaks of 10 or more cases in a single setting account for only 18% of all confirmed cases.

Clearly, some caution is still called for. I, for one, eagerly await the Governor’s announcement.