[See addendum at end for an update.]
It has been 46 days since the first case of Covid-19 infection was reported in Kentucky and 36 days since the first death– not as long as it seems for those of us riding out the storm at home or still on the job.! Nonetheless, we are hearing increasingly broad demands to walk away from the non-medical public health approaches we are using to mitigate the impact of this highly infectious agent. However, given the very limited availability of viral testing, of what is at best a decrease in the exponential growth rate of new cases, and continuing sporadic jumps in the number of new deaths daily; it is not at all clear that we have broken the back of Kentucky’s part of this pandemic. It does appear that our personal and other community sacrifices have awarded us success compared to other states! We have avoided a disabling flood of very sick Covid-19 patients on the capacity of our hospitals– one of our most important goals. However, in my opinion and as based on the raw numbers available to me, we do not have the evidence in-hand to declare that we have reached the plateau needed to justify anything more than thoughtful planning for progressive gradual stand-downs. The lack of a fully functioning viral testing and reporting system has not reached anyone’s minimal expectations. We are flying blind. Governor Beshear’s reports over the weekend through Tuesday evening show continuing substantial volatility in the counts.
The New Numbers as of April 21.
I have updated the assortment of data visualizations on the Institute’s interactive Tableau Public website. Click through the several tabs for data and explanations of a variety of reporting metrics. I will update these daily including tonight in advance of an eventual additional article The discussion below provides a snapshot of our status as of Tuesday evening, April 21.
The number of (unconfirmed) new cases per day jumped from a new high of 253 and to as few as 90 on Monday. Even when calculated as a running three-day average to smooth out short-term bumps in reporting, (see above figure) it cannot be said we have reached a stable plateau, and certainly not a decline in the rate of new cases. Although a curve was displayed yesterday that showed a downward drift from the curve of a semi-log plot of new cases, this represents at best only a decrease in the exponential spread of the disease. Exponential implies compound growth and is still dangerously high.
The number of new deaths continues unabated with a new daily high of 17 reported Tuesday. Continuing volatility in the number of deaths is likely to continue as outbreaks in long-term living facilities and other closed-shop situations punctuate our concern. Although no less tragic, the number of our deaths seems low compared to many other states– but this is also one of our important goals. I have a new data source of national state-specific experience and will attempt to document our success in this regard! The demonstrated ability of Covid-19 to spread in the enclosed environments of group living, religious revivals, or cruise ships (or aircraft carriers) does not require statistical analysis to make the point that Covid-19 is a bad actor. Think of what would have happened if our schools and colleges had not sent their students home. Six feet of separation is meaningless in a room, store, restaurant, or church with fans, circulated heat, or air conditioning!
Viral testing, the weakest element in the numbers.
The reported number of tests performed fell to a recent new low of 94 rising only to a maximum of 511 on Monday. This is vastly far from the thousands of tests needed to get a handle on matters and to plan for- and monitor in the future. Certainly, much of daily volatility is related to artifacts of data collection and analysis. Tuesday’s number of new tests at 498 is hardly reassuring. Without valid data, our ability to predict what lies ahead is compromised.
Despite heroic efforts, both the Governor and Commissioner Stack expressed concern and disappointment with our ability to capture the number, distribution, and clinical status of Kentuckians with viral infection throughout the state. There are many pending test results out there we do not even know about. Some of the larger commercial labs are slow to report their results. The more labs that test, the more complicated to collect their results. The availability of sample swabs and other collection materials, and of the chemical reagents necessary to actually run the tests is extremely limited both in Kentucky and nationwide. Our testing problem is the nation’s. We need even more than a national man-on-the-moon initiative! Few if any state can do it alone.
Lack of testing distorts other metrics.
We have been offering viral testing in a targeted manner. That is to say we are prioritizing tests for those with the most severe illness, and for healthcare workers and first responders. As a result, we have no idea what the background incidence of Covid-19 infection is in our communities until it appears at a hospital door or in the obituaries. For example, over the past weeks, the percent of viral tests that have been reported as positive has been rising to a maximum yesterday of 9.6%. Of course that is not the background incidence in Kentucky– or one would hope. [New Jersey has much higher incidence!] The fact that this percent has risen continuously tells me that testing has not even been able to catch up with new cases. Surely there are people having or dying of Coronavirus that we do not even know about. We are falling behind.
Similarly, the observed mortality rate in Kentuckians as defined by the number of deaths divided by the number of confirmed cases also continues to rise. It is 5.4% at present. This higher than virtually all current estimates of the expected actual mortally of Covid-19 in the population at large. The over-representation of the very elderly in the case reports of Kentucky’s deaths is powerful evidence that mortality will be even higher depending on age, other coexisting health problems, and certainly on any number of non-biological health outcome determinants.
Our goal must be more than “flattening” the curve which may only lengthen the duration of our current outbreak of Covid-19 or to make it more manageable. We want more importantly to decrease the “area under the curves” to limit the number of individuals who get sick or die, and to diminish the impact of what we must assume will be recurrent outbreaks in the future. Surely this will require time do find effective treatments for the sick, and to develop and test vaccines to prevent illness in the first place. We must and will plot a course to keep us functioning as a community both locally and as a nation. There is much that is unknown and scary for that reason, but the thing I am most sure about is that we will either stand or fall together as Kentuckians and Americans . As it has always been, no one can be healthier than the person next to them.
[Addendum Thursday April 23, 2020:
Wednesday evening’s new data release was marked with a jump of new daily reported tests from 498 to 2747. This is good. Some new testing locations were announced and for a few of these, there were no restrictions on eligibility for viral testing. Asymptomatic individuals can be tested within limits of availability.
Not unexpectedly, the daily number of new confirmed cases rose to 181 from 142 the day before. A 7-Day running average of daily new confirmed cases continues to rise a little. Using a running average smooths out some of the volatility in data collection and reporting. It will likely take some more time to determine what the overall direction of new case discovery will be. (See figures below which add new the case numbers reported the evening of April 22.)]
Peter Hasselbacher, MD
Emeritus Professor of Medicine, UofL
22 April 2020