A Closer Look at Coronavirus Numbers in Pa.

Much of the coverage of coronavirus in Pennsylvania and nationally—especially the headlines—is about the number of cases, both new cases and total cases. This is especially true whenever there is a big number, like 100,000 cases. Those big numbers make for scary headlines and good clickbait, but fail to give a broader context.

To help provide that, I’ve been compiling charts with Pa. specific data for a few months. Here are several reasons why cases, and cases alone, are a poor metric for headlines.

Recoveries

 Since May 22, the Pennsylvania Department of Health has been providing the percentage of cases deemed “recovered”. They define recoveries as individuals who tested positive but have not had symptoms for 30 days. As of Aug. 15, 78% of “total cases” are deemed recovered.

That is to say, while news headlines will say “123,000 cases”—nearly 100,000 of cases are recovered, while only 27,000 are active.

Testing

 Pennsylvania reported a drop in the number of daily cases from early April through late June, then a rise from late June to the end of July. Part of this rise the number of cases was due to the increase in testing. Even though Pennsylvania ranks 49th in the nation in tests per capita, the number of tests given per day has risen dramatically.

In mid-April, more than one-fourth of all individuals getting tested have tested positive. Since early June, the percentage testing positive has remained between 3.8% and 6%.

Hospitalization and Flattening the Curve

The original purpose behind social distancing and government mandated business shutdowns was to “flatten the curve”—that is to prevent hospitals from being overwhelmed. Yet throughout this entire time, Pennsylvania hospitals have never come close to being overwhelmed. Even at the peak in mid-April, COVID-19 patient were only occupying about 7% of Pa. hospital beds and about 13% of invasive ventilators.

That usage rate has hovered around 2 percent of each since late June.

Cases vs Hospitalizations

As the chart above shows, the rise in the number of cases in late June to late July did not lead to an equivalent rise in the number of hospitalizations, or the number of deaths or ventilators in use. In fact, the hospitalization rate has fallen dramatically.

In mid-April, more than 10% of COVID cases were hospitalized; by August, the percentage has fallen to around 2.5%. This may be because more individuals are being tested—including those with mild or no symptoms—or because the virus has become less severe, or because we’ve gotten better at treating it without hospitalization.

But regardless of the reason—hospitalization trends are far more important in measuring our capacity to deal with coronavirus than merely the “number of positive tests.”

Good States and Bad States

Gov. Wolf has taken to pointing to other states, especially Florida and Texas, for “reopening too soon” and seeing a rise in cases, while arguing Pennsylvania and states like New York and New Jersey have done a better job managing.

It is true that warm-weather sunbelt states saw a rise in the number of cases during the summer, and Florida and Texas now have more total cases than Pennsylvania and even New York. Never mind that these states actually had a flatter curve than Pennsylvania and New York, or that residents fled from the New York metro area to other parts of the country and haven’t returned.

Perhaps the most telling trend in these state comparisons is that Florida and Texas have far fewer deaths per capita than Pennsylvania, and dramatically less than New York and New Jersey. In fact, as a percentage of identified cases, fewer than 2% of those testing positive in Texas and Florida have died. Nearly 6% of those testing positive in Pennsylvania have died, and New York and New Jersey have see more than 7% and 8% of cases dies, respectively.

To the extent state policy can protect people from coronavirus, the media narrative about who has done well has it exactly wrong.

Ages and Impact

It has been well noted that coronavirus disproportionately affects older populations. Nearly 68% of the deaths in Pennsylvania have occurred in nursing homes—a topic that deserves far greater discussion than I’ll give it here.

While younger populations have contracted and tested positive for COVID-19, the contrast between the survival rate of older and younger populations is dramatic.

Everything Else

While there are daily updates on the number of cases and deaths from COVID-19, there’s a lack of reporting on other causes of death. Anecdotally, we know things like suicides and drug abuse are on the rise, and screening for cancers and preventative care for heart disease are done—particularly when hospitals were banned from doing elective procedures—but it will be one to two years until we have official numbers on those conditions.

You might think that this is because coronavirus is the leading cause of death. The reality? Far from it. Coronavirus-related deaths (even though coronavirus may not be the cause of death in many of those cases), lags far behind heart disease and cancer, the two leading causes of death in Pennsylvania.

On a per-day basis, non-nursing home COVID related deaths, lags behind causes such as accidents and stroke, and is in line with the rate of drug overdoses in 2018.

The preliminary data CDC does report indicates than from February 1 to August 8, 76,234 Pennsylvania residents have died. More than 90 percent did not have COVID.

Shutting down the economy

 Lastly, the impact of shutting down the economy has to be noted. Pennsylvania’s unemployment in the last three months has been higher that at any prior point since we started measuring that—and the official unemployment figure underestimates the amount of jobs and income lost.

Gov. Wolf claims that his most recent order to close bars and limit restaurant capacity was based on data that restaurants and bars were places coronavirus was being spread. Yet his administration’s own data—data that he didn’t want to release—belies that. Fewer than 3% of people testing positive told contract tracers they had visited a restaurant in the prior two weeks.