New daily cases of Covid-19 in Kentucky have been dropping sharply since mid-September in close parallel with the Test Positivity Rate. Hospital utilization is falling as well but remains high with ICU utilization even higher than it was last winter as measured by number of beds currently filled. Daily reported deaths (expected to lag new cases) have not yet begun to fall with a current 7-Day average of 35 deaths per day. A complete tally of Covid-19 deaths will require weeks to months to be compiled. This welcome turn-around in what was an even worse flair of the Kentucky epidemic is the result of more rigorous attention to the same proven public health measures that reversed our three previous ones and the increasing rates of effective vaccination. The fact that case and death rates match or exceed those of last mid-winter should be sobering as the pool of susceptible individuals of any age by virtue of vaccinated status has shrunken! Raw incidence rates of new cases show that all but two of Kentucky’s counties are still in the in the high or moderate range of risk. We need to remain humble and focused until we can reliably consider ourselves to be approaching “normal.”
There is no longer any doubt whatsoever that many fewer people are getting sick from Covid-19. Even though considerably volatility in daily case counts throughout each week remains, the counts for each of the seven individual weekdays are way down. The 7-Day average of new cases is falling much faster than that of the 14-Day average. Each of the last three full weeks has had dramatically fewer cases than the week before. The semi-log plot of new cases is also decidedly falling. However, the open question remains, “how low can we go” as we enter the indoor winter season with its sports and holidays and the delta variant still alive and breathing down our necks throughout the state and country. In retrospect, it is somewhat unnerving to see the relentless aggregate rise in both cases and deaths over the last twelve months. As reassuring as things seem now, we have been here before. In my thinking, our best public health measure is to both believe and act on our state motto: “United we stand, divided we fall.”
Here are some representative data visualizations. You can see the full interactive profiles on KHPI Tableau Public Website.
The following is a simple plot of aggregate cases and deaths from October 1, 2020 to the present. While we have has spurts and remissions since our epidemic began in earnest last Fall, the upward thrust has been relentless. This is what we must turn around!
Peter Hasselbacher, MD Emeritus Professor of Medicine, UofL Kentucky Health Policy Institute 12 October 2021
Hospital utilization in crisis state. Deaths rising as expected. Community struggling to find the right path forward.
[Addendum Sept 25, 2021: Data on Tableau Public updated as of Friday. New cases and Test Positivity Rates clearly going down. Deaths not yet showing meaningful declines. Hospital, ICU and Ventilator utilization still at record highs. New cases for ages 18-and-under averaging 27% but declining from high of 31% on September 1. See additional notes at the bottom.]
Since the last article in this Covid-19 tracking series, I have been waiting to get through the low reporting-days of a weekend to allow time for the necessary 7-Day and 14-Day averages to reveal a reliable direction of our Kentucky epidemic of Covid-19. A number of indicators show new daily cases to be holding at a steady state with indications of a possible decline. The concern is that the 14-Day average of some 4000 daily case is still higher than any time since March 2020. With three reporting cases left this week, this September will have given us the largest monthly number so far, probably outstripping even the dark months of last December and January.
On a positive note, case-counts for specific weekdays have been trending down every day for over a week. You can click through the individual days on KHPI’s Public Tableau portfolio of data visualizations. Semi-log plots of New Cases also show a stable accumulation of new cases –but not yet a definite decrease. The Kentucky Department of Public Health Test Positivity Rate has taken a sharp turn downwards since September 8th.
Deaths. It is clear though that the Coronavirus is still having its way with us. Deaths will continue to increase for a while as new cases work their way through their individual clinical courses. The sporadic reporting of deaths illustrates the delay in time built into our clinical and death certificate reporting systems.
Hospitals and ICUs. Many if not most Covid-19 deaths occur in hospital facilities. The need to be placed on a ventilator is a strong predictor of a future Covid-related death. One has to be still hiding in their Groundhog-Day hole not to be aware that our Kentucky hospitals are over-whelmed with Covid-19 patients and under-whelmed with a healthy staff to take care of them. The number of patients in the hospital, in ICU beds, and on ventilators are at the highest levels we have ever seen. The fact that inspection of the shape of a curve graphic might be said to be “flattening” is not a good sign when the number of beds is limited!
This is where we are today. It remains to be seen where we are going. We were in a similar place last December but got a lot worse in January. I would like to assume that the seeming temporary stabilization of new reported cases is a result of the good sense and commitment of Kentuckians to their communities. So far there are not enough of us.
Peter Hasselbacher, MD Emeritus Professor of Medicine, UofL Kentucky Health Policy Institute Thursday Morning, 23 September, 2021
[Additional comments to addendum, 9-25-21: September has been a tough month for Kentucky’s epidemic of Covid-19. So far, even as of only September 24, this month has already had the highest number of new cases (96,964 added to the total number confirmed). The number of cases 18 years old and under has never been higher based on data available to me. The consequences of this flair in deaths, long-term morbidity, and hospital utilization have yet to play out.
There is however some good news in case numbers as Kentuckians take the bit of disease prevention and mitigation between their teeth. (I could not resist using a Kentucky motif.) The number of weekly cases will likely be down somewhat when Saturday’s new case number becomes available on Monday. Better news is that the 7-Day average of new cases is racing the 14 day average downwards as is expected when the virus is (at least temporally) fading. The 7-Day average of the test positivity rate is sharply declining following the same path downwards in its correlation with new cases as it did when were in the home stretch of recovery from last winter’s earlier meet with the virus. The semi-log plots of new cases agains time switched from a rising epidemic trend to a flat steady-state one beginning late August 22. More recently the trend has shifted to a modestly downward slope indicating reversal of the epidemic’s expansion. I take these numbers to suggest our community efforts are advancing us as we make the final turn, but the home stretch lies ahead and we are still far from the finish line. We need to continue to work with our trainers, track authorities, and the real veterinarians to the human animal. P.H.]
It has been more than two weeks since I posted a full update of Kentucky’s Covid-19 epidemic status. As I commented on the article in following days, data collection volatility, exacerbated by the Labor Day weekend, did not give me the confidence to opine one way or the other the direction of epidemic status. Available numbers as of the evening of Thursday, September 16 are compatible with the number of new cases entering an interval of steady-state, albeit hovering between 4000 and 4500 news cases per day which exceeds the rate of the darkness of last winter. That is not a good place to be.
[Addendum: Friday evening’s new numbers do not really alter the state of affairs or conclusions outlined in the following paragraphs and graphs. Specific detail is added at the end of this article.]
Cases. The 7-Day average of new cases is playing leapfrog with the 14-Day average. When the 7-Day average is higher than the 14-Day, epidemic expansion is occurring. The numbers obviously vary greatly from day-to-day, inserting a large degree of uncertainty into predictability.
Case counts by day of the week for Thursdays, which are weekdays with reliably higher daily counts. Stable or not?
I like to rely on semi-log plots of exponentially rising counts as a way to estimate the rate and direction of epidemic expansion. Since the end of August, the plot of daily new cases is trending to a horizontal line representing a steady-state of the numbers of new cases– even if they are bouncing up against 5000 or more a day! If we had continued at the rate of the previous few weeks, we would have been at 10,000 new cases a day. Surely such a possibility demanded effective interventions, even if unpopular.
School-age people. As summer activities for children and young adults blossomed and students went back to school, the proportion of new cases 18 years old or younger reported each evening by the Kentucky Department of Public Health rose from a 7-Day average of 18% to over 30% in early September– a greater than 80% increase. This rise occurred in the midst of intense controversy about masking and vaccination of students, teachers, and school staff. Hmm.
Hospitalizations, ICUs or worse. As would be expected, the number of deaths of cases emerging since July 1( as depicted in the first graph of this series above) is entering a more rapid expansion. It will take a while for a full accounting of those deaths to be finalized. The consensus of reports from hospital and other entities is that this newest expansion of hospitalizations and deaths (and in particular ICU and ventilator use) is occurring almost entirely in unvaccinated individuals and is related to the increased virulence of the current universally present Delta variant of the virus.
We expect Covid-19 deaths and hospitalizations to move in the same direction– albeit out of phase with each other. The sequence begins when individuals are exposed and get infected. Some get symptomatically ill. A smaller proportion of these require hospitalization and an even smaller number require ventilatory support. The odds of surviving being put on a ventilator vary considerably by reporting entity, but they are not what anyone would consider good. Against this context, achieving a “plateau” of hospital, ICU, or ventilator use is not a favorable sign if those facilities are already fully in use. This is happening in many places around the country including Kentucky. Some localities are operating under a formal “crisis status” in which the most difficult life or death triage decisions must be made. (I do not ever want to hear a politician utter the words “death panel” again to scare people into not supporting reform of our non-functional healthcare system!)
To conclude. I believe Kentucky, indeed the nation, is poised at one of a sequence of tipping points in terms of the future of our Covid-19 epidemic and its consequent effects on our healthcare, social, and financial futures. I am unwilling to concede that we are doing OK. It troubles me that coordinated and informed responses to the virus do not yet exist. Our political system that should be looking our for our health and welfare interests has abandoned that responsibility to business and the private sector. I want very much to believe we can expect a rapid turn-around, but in all honesty I see things getting worse first.
Peter Hasselbacher, MD Emeritus Professor of Medicine, UofL Kentucky Health Policy Institute 17 September 2021
[Addendum 9-18-21: A new set of Kentucky Covid update numbers was reported Friday evening. Numbers for Saturday, Sunday and Monday will not be available to me until Monday evening.
New cases increased from Thursday to 5133, one of three times in the last 14 days in which counts exceeded 5000 of which twice occured in the last three daily reports. The 7- and 14-Day averages did not change appreciably. The 7-Day average still exceeds the 14-Day average compatible with epidemic expansion. The number for a Friday has been about the same for the last three weeks. A count for the past week will not be available to me until next Monday but is high. The semi-log plot of new cases remains horizontal describing a steady state of new cases. The 7-Day average of the proportion of new cases in individuals 18 and under remains high at 28.1%.
Testing: The 7-Day number of new tests reported which began rising significantly only in early August is 39,253. In June and until mid-July, daily test counts only occasionally exceeded 10,000. Friday’s 7-Day average of electronically reported tests was 12.9, down from a high of 14.6 on September 8. (There is a complicated but significant correlation of both the Test Positive Rate and number of new cases with the number of tests performed and even hospital census that deserves further consideration.)
Hospital, ICU, and ventilator utilzation for Covid-19 remain at the same high levels as they were last week.
Deaths are clearly on the rise with an additional 45 reported yesterday. This rising trend will surely continue for a while.
As of Wednesday September 1, several lines of data point to a decreasing rate of new reported cases if not a peaking– albeit at record high levels. However, hospitalizations and deaths are expected to rise further!
Cases: It’s hard to be content with 4941 new cases Wednesday. Last week delivered the highest number of cases in any week of Kentucky’s epidemic, even higher than those of last December and January. Except for one aberrant reporting day on January 6, last week contained the two highest daily case counts for the entire 544 days of known Coronavirus in our state. The 28,850 new net cases added last week (from Sunday through Saturday) were 2,423 more than the highest week of last Winter! This August had only 618 cases fewer than the previously high month of month of January 2021. Today’s 7-Day average of new cases of 4,212 is the highest to date even with the low counts over the weekend. The 14-Day average is also at a record high.
7-Day and 14-Day Averages of New Cases. When the 7-Day average of new cases is higher than the 14-Day average, the epidemic is expanding. When the 14-Day average is higher, the reverse is true. The difference between the two averages gives a measure of the the rate of expansion or contracture of epidemic spread. While the 7-Day remains higher as of 9-1-2021, the difference between the two is as small as it has been since early July. Until the last day of June when our epidemic was still shrinking, the 14-Day average was higher.
The Test Positivity Rate also has been rising rapidly to new highs, but recently appears to be rising more slowly. However, it remains at record high levels. The 7-Day average of New Tests per day is at its highest level since November 22 and still rising. More people are getting tested. Testing more asymptomatic individuals will cause the test positivity rate to decrease.
So why does this glass-half-empty analyst see a glimmer of a hope that things are slowing down? Since March of 2020 I have been slicing and dicing the daily basic reports of cases, deaths, testing, and hospital utilization data seeking a way to answer the question “are we winning yet?” This self-assigned task was made more difficult by the existing lack of an organized and uniform testing and data collection system in a Commonwealth underlain with multiple under-supported and semi-independent public health departments. Numbers varied dramatically depending on the day of the week or the presence of holidays. There were built delays in reporting to the state public health systems– sometimes dramatic ones! Large backlogs of tests, cases, and deaths were reported periodically. There were state and contractor computer system failures. No database of historical state epidemiologic data was made available to the public even when requested. I had to depend on outside organizations for things as basic as historic case and death tallies by Kentucky county. Against this background, let me outline why I think the hard-fought efforts of Kentucky’s Public Health Department and the individual citizens of Kentucky may finally be having a demonstrable effect despite well-organized political and ideological resistance.
Case counts vary rather predictably depending on the day of the week. For example, since March 1, 2021, daily counts have on average been lowest by half on Mondays, followed closely by Sundays, rising considerably on Tuesdays, reaching a weekly peak on Wednesdays, and then gradually decreasing through Saturdays. Beginning in early July, each day of the week had case counts markedly higher than the same day of the previous week. This changed beginning Monday, Tuesday and Wednesday of the current week when the weekday counts are similar to those of the preceding week. (You can click through the individual weekdays on KHPI’s Tableau Public data visualizations.)
Semi-log plots. Simply inspecting curves of aggregate case counts or even rolling averages on the usual rectangular/linear graph paper is not particularly helpful right now. The curves still ascend in a near-vertical manner. That is what exponential growth looks like. Plotting exponential or rapidly rising data in a different way can give useful insights. Graphing new cases as a function of time using a logarithmic scale for cases and a linear scale for days allows the same plot to include both small and large numbers. Furthermore, a simple linear regression on such a plot yields a straight line when the rate of exponential growth is not changing. This can allowing some degree of forecasting the future. For example, during last July, new daily cases were doubling about every 10 days! I predicted then we could hit 5000 cases per day by early August which was pretty much on target. During the weeks of August, semi-log plots of both daily cases or rolling averages have begun to fall noticeably below the predictive trend line. During August, the doubling time of daily cases lengthened to 17 days. Instead of reaching 10,000 cases per day by early September as July’s data would have allowed, and if nothing else changes, we could still reach that number by the end of September. A “plateau” or bending of the curve is simply not good enough! Hopefully Kentucky’s citizens will continue to make effective responses to our common predicament. Eventually, even if we do nothing as some would recommend, the epidemic wildfire will burn out– but as with the forest fires of our western states and elsewhere in the world, at what cost? We are asking people to do hard things. We have to support them.
Below is another view plotting new cases in a semi-log manner but using the 14-Day average of new cases to smooth out the curve. Over this two month interval, the curve first rose and then fell below the trend line.
I have nothing reassuring about hospital, ICU, or ventilator utilization to write about at this time. All three of these markers of severe infection are worse than ever. Hospitals and their staffs are being crushed. Deaths are already clearly on the rise. Given the passage of time for individuals to move thorough the stages of infection, hospitalization, and ICU to death, we have not yet seen those numbers peak. It is the unvaccinated who are filling the hospitals and dying. A leveling off of the ICU or ventilator curves at high levels would be seen when there are no ICU beds or ventilators left to use!
While I am hoping that this apparent decrease in the rate of our epidemic explosion is real, I would like to see a few more days of data to be confident. Complicating matters is the upcoming Labor Day weekend with its celebratory events and holiday-related delays in public health reporting. The fall season will present us with predictable venues for the spread of infection including schools, colleges, concerts, and sporting events which bring the vaccinated and unvaccinated alike together for education and entertainment. It could be the Fourth of July all over again! These are wild cards that upend the usual making of odds. It would be nice to win a few hands for a change. It would also be nice to feel like someone or some entity was still looking out for our benefit! Kentucky’s legislature has taken charge of day-to-day public health management and that makes me feel very much unsafe.
Peter Hasselbacher, MD Kentucky Health Policy Institute Emeritus Professor of Medicine, UofL 1 September 2021