One notable feature of the current coronavirus epidemic is a pattern of sharp, but relatively short, waves of cases. The U.S. has previously endured two of these waves. One, in the spring, was centered in the Northern states, particularly the Northeast, and another was centered in the South and Southwest in summer. The current autumn case wave has been particularly intense in the Midwest and Rocky Mountain states.

The attributes of these case waves, here and abroad, have been very sharp case rises, but a relatively short duration measured in weeks, followed by a quick decline in cases, falling off to a long tail.

Policymakers and their advisers might have taken note of this pattern and adjusted decisions accordingly.

The two most critical parameters to monitor for management of any epidemic are the number of active cases and the trend in those active cases. Active cases represent those persons who may be infectious, capable of transmitting the pathogen to others, who might then become infected and infectious. A number of states report this important statistic; Minnesota for unknown reasons does not.

Not only is it not reported, but with the information we are provided it can be difficult to ascertain.

The best way to calculate it with the available data is to take the total PCR-confirmed cases for each day as reported on the table of cases by specimen collection date. Subtract from this the “no longer needing isolation” number, which you must record every day from the situation report, as the state does not make a daily table available. These are persons whose period of infectiousness has passed. Then you further subtract the total number of deaths for the day.

There are a variety of factors which might suggest this method produces an overestimate of actual likely infectious cases, including an inability to adjust the “no longer isolating” number for total cases added from 10 days or more earlier. On the other hand, there also are likely many undetected cases. But adding an estimate of those would not change the direction of the trend analysis.

Using these methods, I have tracked the number of active cases, the daily change in that number, and the percent change since Aug. 1. The last week or so of data is incomplete due to reporting lags.

My calculations show that total active cases in Minnesota remained below 10,000 through most of September. Active cases then rose above 20,000 by the third week of October and surged to around 70,000 in early and mid-November. At that point active cases plateaued and began to decline.

The daily percent change in active cases, which is the most important parameter for evaluating the trend, peaked at around 14% on Nov. 4 and has since declined to under 5%.

As overall cases increase, active cases grow steeply. As cases begin to decline, active cases fall rapidly. This reflects the reality that when cases are increasing, there are more people who can be infecting others. When cases begin to decline, more older cases are being rolled into the “no longer needing isolation” bucket than new ones are being added, which means fewer people are able to infect others.

The shape of the active case and active trend curves are due to reaching a level of infections in which rapid transmission to susceptible persons is no longer possible, and they reflect classic unmitigated epidemic curves.

According to my analysis the trend in active cases had plateaued or peaked in the first week of November. With each succeeding day’s case report, that peak is being confirmed. This peak came before Gov. Tim Walz issued any additional orders, in particular the infamous ban of social gatherings in one’s home, a type of restriction an expert referred to as “bizarre and unscientific” in a New York Times article published in the Star Tribune. (“Are small gatherings causing the surge?” Nov. 24).

The state surely must be doing a similar analysis; if not, one wonders how they possibly would be accurately tracking the course of the epidemic.

Given this analysis, the known pattern of other waves, and the fact that the state now tells us that it takes 28 days for a mitigation-of-spread measure to take full effect, patience rather than panic would have been advisable. And we should remember that it was no action of the governor’s that caused the decline in active cases, if he attempts to take credit.


Kevin Roche is a health care investor and writes a daily blog on health care research and policy at