The coming week’s Covid-19 counts in Kentucky will be of critical importance.

Since the last update in this series on July 26, the number of new weekly Covid-19 cases continued its rise into new territory with almost 4000 new cases. On none of the most recent 7 days was the daily increment fewer than 500, with a maximum of 765 last Saturday July 31. The weekly total was however less than that of the preceding week when the total exceeded 4500 new cases. I do not know if the peak days with 977 and 833 cases on July 19 and 24 included targeted testing initiatives at prisons or long-term residential facilities that would artifactually drive the case count up as it did on May 5th. While there is some hope that the interruption in the current upward climb in the 7-Day average of new cases represents a pause to a new and higher “plateau,” I believe it is too soon to tell. Governor Beshear issued an executive order to require face coverings on July 9. This met with legal and community pushback from some quarters. More recently on July 27, bars were ordered closed for two weeks and restaurant service was dialed back to 25% capacity indoors. It is too soon to determine the effectiveness of these measures as currently complied with.

While a partial report yesterday (Sunday) was lower with 463 announced new cases, this is high for a Sunday. Today (Monday) is also historically a low-count day so It is likely that we will need to see what happens during the catch-up reporting from the public health districts later this week. National reports note that the explosion of new cases in the South and West is now spreading to the Midwest– and that includes us and our neighboring states. I would like to think we will be largely spared, but not if we hide our heads in the sand or pretend it is safe to open stadiums, tracks, schools, or festivals without a demonstrable way to prove that it is safe to do so, or a social apparatus to deal with the inevitable outbreaks that will occur.

I have been updating KHPI’s online Covid-19 Tracker daily with Kentucky’s numbers and will do so again with this evening’s announcements. I invite readers to help me analyze the interactive data visualizations. Below are graphics that are current as of Saturday August 1, but will be updated again this evening.

This first is shows the aggregate cases and deaths since the first case was discovered in Kentucky. With the weight of 30,000 cases behind it, any short-term change in the trajectory of the lines will not jump out at us, but the overall rising rate of new cases since early July is well established.

When looked at on a day-to-day plot, it is apparent that daily reports vary in fits and starts, largely because of weekend interruptions in data collection and reporting by a variety of semi-independent public health districts.

Using the data-smoothing approach of rolling 7-Day averaging below, the curves become less jumpy, but the influence of an aberrantly high or low individual day will linger for the next week. Looking at the plot below lends support to hopes of a new stabilization at a higher level but does not negate the colossal increase the occurred during July.

I am liking the visualization style below in which cases, deaths, and tests are totaled by calendar week. It is easy to understand, but unfortunately only up-to-date once a week on a Saturday. What will next week’s new cases look like?

Here are deaths by calendar week. Given that deaths follow cases by weeks not days, we could be seeing a creeping up. Could they reach the levels of April? Why not?

Some are still arguing, even in today’s Courier-Journal, that our rise in new cases is simply the result of more testing. No public health scientist that I know of would agree. I refer readers again to my last article where I discuss this in more detail. I will do so again in coming days when I take a deeper look at policy implications. (Spoiler alert! No, it is not safe to open stadiums, tracks– and not without substantial risk– schools or universities.) Weekly testing in Kentucky has not increased all that much in recent weeks.

I still do not know how Kentucky or other states calculate the Percent Tests Positive (PTP). In a perfect world where tests are performed randomly within a defined population, the PTP can indicate the prevalence of a given disease in that population– but only in that defined population. Below are Kentucky’s raw numbers. I do not know how to interpret them at this time. Can you help me? Filters on KHPI’s interactive on-line site allow the viewer to vary the date interval plotted. A case can be made that the aggregate of tests performed is trending upwards, but only very slowly. In my analysis, pending confirmation of my methods, cases are rising much faster than testing. Our testing infrastructure, like that nationally, is currently under much stress. If once again the sick and high-risk are preferentially tested, the Percent-Tests-Positive rate will go up artifactually and not be representative of background disease incidence.

Enough of an update for now. I do not think anyone will disagree that the results of this coming week’s data collection and reporting will be very important. I wish again that I could be more optimistic, but I will be happy to eat my hat.

Peter Hasselbacher, MD
President, KHPI
Emeritus Professor of Medicine, UofL
Aug 3, 2020

Kentucky Added 4580 New Covid-19 Cases Last Week. New Record.

No change in rate of increase in new cases since July 1st. Simple projections are disturbing.

Yesterday’s (Saturday) release of Covid-19 tracking data by the Governor’s office capped off what is by far the largest addition to total weekly cases of Covid-19 since the first case was documented 141 days ago. The 4580 new cases marked the third week in a row with record numbers of new cases, dwarfing the 1460 new cases for the week beginning June 28. The rate of increase of the 7-Day rolling average of daily cases has been going up at the same rate since July 1st. Graphic projections of daily new cases based either on the absolute numbers of new cases or their 7-Day rolling averages agree. Both predict the same near-future estimates of 1000 new cases per day by August 3; and 2000 new cases by August 18. Based on current data alone, both would have us at 5000 new cases per day by the first week of September. The New York Times this morning rates Kentucky as having one of the most rapidly expanding epidemics in the nation. I would like to wish away such a future, but at this time I am unable to do so.

Below are some of the data visualizations that lead me to these projections and concerns. Updated interactive versions which allow different time-frames to be chosen and which include all my data are available here.

This first graphic shows the 7-Day rolling average of new cases added daily to Kentucky’s announced total. KHPI defines New Cases as the increment in aggregate total cases from the previous day. As described elsewhere, this number is usually a handful of cases fewer than announced from Frankfort or by the national Covid-data aggregators which use Frankfort’s numbers. It is impossible to ignore the abrupt shift towards the vertical beginning around July 4th with 209 cases to yesterday’s 654. With small deviations reflecting weekend delays in reporting, the plot-line is ascending relentlessly.

7-Day Rolling Average New KY Cases to 7-25-20

Another way to demonstrate the increase is to count the number of new cases in a given calendar week. Since much of the volatility in daily counts stems from weekend artifact, calendar-week counts are another way to smooth out and standardize the numbers, albeit with a built-in delay. The chart below fills out the week ending Saturday the 25th. Pretty dramatic increase!

New cases per calendar week through Saturday 7-25-20

Plotting the rolling-average numbers using a semi-log axis yields a nearly-perfect straight line indicating that the daily numbers are rising exponentially– think compound interest or credit-card debt. (You get rich or over your head in debt before you know it!) Plotting the raw daily counts yields essentially the same result. I estimate graphically that cases are doubling approximately every two weeks.

Semi-log plots of new New Cases, Daily and 7-Day Averages with trend lines.

The dashed lines are computer-generated linear projections based entirely on July numbers to date. They do not, and cannot, take into account the effect of changes in public behavior for better or worse, or the impact of public health measures employed or withdrawn to mitigate the damage of rising cases. This is what we have to work with as of today. As noted in the introduction above, both plots project that we will have 1000 new cases by August 3; 2000 daily new cases by Aug 18; and will approach 5000 cases by early September.

Numerous projections offered by other groups or individuals these past few months (including one by a high-profile official that the epidemic would fade away quickly) have proven to be inaccurate. Mine presented above, based on less than a month’s worth of data, may also come a cropper. I would like to think that this simple objective approach is free of any inappropriate bias. It is, however, subject to the same inherent properties of the Pensieve bowls featured in the Lord of the Rings or the Harry Potter movie series in that they may only foretell what the future might be.

Not just numbers.
By accounts, I believe the rise in our cases represents real sickness. Hospital and ICU utilization are rising– although I do not have great confidence in the collection and reporting of hospital data yet. New deaths have not taken a definitive turn upwards, but they can be expected to do so. Other morbidities of this new disease remain to be defined.

Hospital and ICU utilization in Kentucky statewide.
New deaths per calendar week through Saturday 7-25-20

Are increased cases due to increased testing?
What has not been tracking upwards are new tests. Below is the companion graphic showing new tests per calendar week. By reported numbers, they have not been going up at anywhere near the case rate. Testing centers, especially the big commercial ones, have been getting clobbered nationally. Supplies for testing remain limited. Delay in reporting is occurring again. Of all the data elements being reported, numbers related to tests are the ones that I least understand or have confidence in. The current rise in new cases is absolutely not explainable by increased testing. Only die-hard disease-deniers are still saying so. Ignore them.

New tests per calendar week through Saturday 7-25-20

How about Percent Test Positives? (PTP)
Another testing-related metric is getting attention in media reports and policy decisions, including in Kentucky. This is the Percent Tests Positive ratio now being announced daily from Frankfort. As discussed in my article of July 23, this ratio is the percent of all tests done in a given interval and place that are positive. It is assumed that this number provides information about the prevalence of disease in a given population. That seemingly obvious assumption is totally dependent on accurate and timely testing data, and only works if the people being tested are comparable day-to-day and place-to-place. This is demonstrably not true in our instance or elsewhere. In the beginning of the Kentucky epidemic we were testing only sick people and those at highest risk. We canvased nursing homes and prisons. The aggregate state PTP on those days was as high a 9%. This fell to a plateau of around 4% by early June and has risen lately to a little more 4.5%. (KHPI necessarily uses 7-Day rolling averages.) Given that we are still facing testing issues, and that at least some communities are preferentially testing high-risk populations again, I remain to be convinced that the PTP is a metric we should be hanging our hats on. More important in my mind would be identifying geographic hotspots where cases are rising more rapidly as a proportion of their population. I have no access to the necessary historical location data in a usable form to investigate the utility of identifying localized hotspots. I must assume the state Public Health Department has such data and I urge them again to share with other interested and qualified individuals and institutions (like me!). If the New York Times has such data, why cannot Kentuckians?

Percent Tests Positive 7-25-20. Note: Different types of tests are being added.

Can we “suggest” our way out ot this epidemic?
There may be limits to the practically or fairness of applying the same Public Health approaches to every county or community in Kentucky. According today’s New York Time analysis, the per capita distribution of Covid-19 cases ranges widely in Kentucky. Hot spot counties where new cases are popping up rapidly is not the same list as that of total county cases. I am at present not fully versed in the pros and cons in the application of universal vs. selective public health mandates. I believe that we ultimately are going to need to apply rapid case finding and contact tracking at the most granular local level possible in a surgical manner. I do not know how of this is working in Kentucky. We had a slow start-up and very little is being said about it. I would like to think that showing important positive results from case finding and tracking would be good for us to hear. (I do not mean outing people personally.) If a big soccer game (or God forbid a church service or a school) was known to spin off many new cases, perhaps we would be more willing as a community to accept stricter limits or even postponement of events! How can we say that such events are “safe” without a way to prove it? To be successful, we need enough public health investigators, a public willing to work with them, sufficient testing with rapid reporting, and a willingness of community leaders to take public health seriously. Yes, there are such things as public health laws!

What should we do today?
Waiting for Sunday’s reports seems to me to be unnecessary. Sundays are artefactually low-count days as are Mondays. We would just have to stall until next Wednesday to see if the big numbers persist. In my opinion, and based on the public information available to me, we need now to be taking positive, enforceable, evidence-based steps, or at the very least requiring commonsense efforts to mitigate the conjoined health and economic damage that this virus is doing to us. We are told we are working towards a “new normal.” I applaud the honesty of this challenge. Just as the medieval Bubonic plague changed its world in fundamental ways, from the breakdown of serfdom to opening the door for the Protestant reformation, so will our contemporary plague deliver us to a much changed world. We should not be afraid of that.

Peter Hasselbacher, MD
President, KHPI
Professor Emeritus of Medicine, UofL
American Political Science Association Congressional Fellow 1997
July 26, 2020

What if Kentucky’s New Covid-19 Cases Increase at the Current Exponential Rate?

Yesterday’s (Wednesday) evening Covid-19 update from Frankfort told of 518 new cases but from which there were an unusually high number of 38 removed as duplicates or perhaps for other reasons. KHPI reports new cases as the day-to-day difference in total cases which yields a smaller number of 480 additional new cases. This number, while higher than any new case number earlier than July 7, is less than the several 500s, 646, or 977 counts of the previous 7 days. The rise in the 7-Day average to a new high of 552 was at least temporarily blunted– but it has not begun to fall either. Time will tell us in which direction our case trend is headed. It has only been since July 9 that mandatory mask use was required in Kentucky, but its implementation was challenged in the courts and compliance has been spotty in any event. I suspect it is too early to expect a major effect on spread at this early time. End-of-week Fridays and Saturdays have historically given us higher counts in the past. The monstrously high count of 977 was reported just last Sunday. Three reporting days of the week remain just ahead of us.

[Addendum 7-24-20: Yesterday evening’s data report confirmed ongoing rapid expansion of the epidemic. KHPI noted 607 additions to total cases and 7 new deaths. With two days to go in this calendar week, the number of new cases is already nearly as many as last week. Based on the current rate of increase of the 7-Day average, new cases are doubling by at least every 12-13 days. Hospital and ICU utilization continue to trend upwards. In all this, Kentucky is mirroring the nation overall. We are all in trouble! The full panel of visualizations on-line have been updated.]

[Addendum 7-24 4:30 pm: todays new case reports adds 744 to the total. This is bad news. Even without Saturday’s report tomorrow, the number of new cases has jumped well ahead of last weeks total. At this rate, unless something changes or happens, we will be at 1000 cases per day by August 3d. Deaths in today’s report added 7 new ones. I will elaborate on the weeks finding s tomorrow evening. I place an updated graphic of 7-Day New Cases at the end of this article.]

7-Day rolling average of new cases as of 7-22-20
Continue reading “What if Kentucky’s New Covid-19 Cases Increase at the Current Exponential Rate?”

Covid-19 in Kentucky. The Bad News Gets Even Worse.

At the time of my last article on July 16, I reported the record-breaking rise in new Coronavirus cases in Kentucky. That was on a Tuesday. I amended the article twice to show that the epidemic surge was no illusion. Indeed, over the weekend the curve of new cases continued to go nearly straight up in a clear exponential manner. Yesterday’s (Monday) new case count leapfrogged over any previous high to 977. Kentucky has the dubious distinction this morning to be listed by the New York Times as among the very top states with the most rapidly increasing number of Covid-19 cases. This is not something for which recognition is desirable!

Technical note:
For these initiatives, I have been calculating daily new cases as the increase in Total New Cases from the Total of the previous day. This number is usually a few below what the Governor’s report terms “New Cases.” I understand that latter group includes duplicates, out of state cases, corrections, or perhaps other exceptions that do not ultimately show up in the Total (aggregate) Cases. This likely source of confusion can be seen in the data of the national data aggregators and probably in the data of the CDC. I did not want to include duplicates and my efforts to obtain better data from Frankfort have not been successful. The differences between KHPI’s counts of new cases and Frankfort’s are small and the shape of the curves remains essentially identical.

I have been updating the on-line interactive versions daily on KHPI’s Public Tableau website. I began this series of articles in part as a self-educational project to explore how emerging epidemiological data could be usefully portrayed and used to monitor how and where our public health efforts were succeeding (or not) in a timely way. The result was a larger number of visualizations than most readers might want to see. beginning today, I trimmed out a number of graphics that seemed less useful at this stage of the epidemic. I place below a few of the graphics that include yesterday’s new, and indeed frightening case numbers. The new link link just above opens the most up-to-date interactive versions of all the data and visualizations.

Here is a traditional Bar Chart of the number of new cases per day current up to yesterday, July 19. Click on it to enlarge in a separate screen.

Continue reading “Covid-19 in Kentucky. The Bad News Gets Even Worse.”

Cases of New Covid-19 Infection in Kentucky Soar to New Daily Records.

New cases of Covid-19 infection in Kentucky are rising dramatically and there is no way to explain away the reasons as anything less than a demonization of medical science, and a destructive politicization of public health policy. I fear that after 4 months of difficult sacrifice, we are going to have to start all over again.

The total number of new of new Kentucky Covid-19 infections jumped by 570 and 454 over the last two days. Today’s report will be available soon. The seven highest daily reports of the epidemic have occurred since July 1. Although there were 577 cases reported on Tuesday, May 5, this obviously aberrant number included a large cache of previously uncounted cases in a prison, and from several long-term care facilities. That day’s report also included catch-up cases from the immediately preceding Sunday and Monday– intrinsic low reporting days. We are in new territory.

When did the current surge begin?
Because of our multiple catch-as-catch-can reporting systems nationally and in Kentucky, it is hard to tell. Because it was marked in my mind with a rapid change in public behaviors when retail stores and restaurants opened, I believe our local expansion began with May 25 Memorial Day weekend. After all, national and some local political leaders were telling us it was safe to do more. Indeed, some political and other leaders were telling us to frankly ignore what actual public health experts were advising. (There is still a lot of that going around.)

Continue reading “Cases of New Covid-19 Infection in Kentucky Soar to New Daily Records.”

Can We Detect Local COVID-19 Surges in Time Do Anything About It?

What happens in Myrtle Beach does not stay in Myrtle Beach!

When I last wrote in these pages on June 8, Memorial Day was two weeks behind us, and retail stores and other public venues in Kentucky had recently taken an intermediate step opening up to the general public. Parking lots, roads, and sidewalks were noticeably more populated. Masks appeared to be optional. Perhaps and probably not coincidently, by June 6 both new daily-reported cases (310) and 7-Day rolling average of new cases (226) hit all-time highs for Kentucky’s part of the epidemic! This change was large enough that the New York Times included Kentucky in its daily tally of states in which the numbers of new cases were increasing fastest. Over the next 2 weeks or so, our state was regrouped by the Times with states whose numbers were decreasing, and then back into the group with “mostly the same” number of new cases. Yesterday we were back in the “increasing” cohort along with 31 other states. In only 6 states or territories were cases decreasing. This considerable volatility in case reports does not inspire confidence that our system of timely epidemiological monitoring or control is adequate. New cases are exploding across the nation. Current Kentucky numbers are rising as well. The following graphic tracks the 7-Day rolling average of new cases in Kentucky and an estimate of the Basic Reproductive Rate as of June 27, 2020. An assortment of other visualizations are presented at the end of this article. Interactive online versions of the complete portfolio of visualizations are available at the KHPI Public Tableau website.

Continue reading “Can We Detect Local COVID-19 Surges in Time Do Anything About It?”

New Covid-19 Cases on the Rise in Kentucky: Not Completely Unexpected.

How many tent revivals does it take to alter the trajectory of an epidemic?

When Kentuckians did the hard things necessary as we entered the great unknown of a new viral disease, we altered the course of the Covid-19 pandemic in our Commonwealth. If we had remained on the initial exponential curve we were experiencing, the entire state would have been infected in a matter of weeks. We have met and passed that challenge– for now. Governor Beshear, Public Health Commissioner Stack, I, and all the rest of the hopeful Kentuckians would like to have believed that the number of new cases would steady, if not decrease altogether. Looking at the reported data and given what I understand to be the definitions of the individual data elements, I believe that the number of new cases is on the rise again. Saturday’s new cases numbered 310, the highest since May 5 when an institutional outbreak was recognized and reported all at once. Indeed the 7-Day rolling average of new cases daily has never been higher than it is now. Reported deaths or hospitalizations have not been rising substantially, but those numbers take longer to show up in reports. Other states are also reporting upticks in their case counts as society opens up.

Are rising cases the simple result of more testing?
The reporting of tests done daily remains erratic. Even though “Tests” now include both viral RNA and patient antibodies, the 7-Day average of Kentucky tests reported daily has actually been decreasing since the first of June. (See below.) The Basic Reproductive Rate (the average number of people that a given case will infect in turn) has been rising. The percent of daily tests that are positive has not changed much. Based on publicly available numbers, I cannot agree yet that the increase in cases is explainable by more testing.

Continue reading “New Covid-19 Cases on the Rise in Kentucky: Not Completely Unexpected.”

Is the Covid-19 Epidemic Still Expanding in Kentucky and Its Neighborhood?

Kentucky and many other states are backing away from public health measures of varying strictness that were adopted in March or April during the exponential expansion phase of the Covid-19 pandemic in the United States. It is appropriate and even necessary to begin this process, but it needs to be done with an acceptable degree of safety. There is no doubt that measures taken so far have at least “bent” the curve, slowing down if not ultimately decreasing the mortality and morbidity of this overtly infectious disease. I believe Kentucky has benefited greatly from our collective efforts despite opposition on several fronts including armed protest. The expectation and promise is that we and the nation will be able to detect “surges” of the epidemic in a timely way and to be able to reinstate restrictions on public interactions that have proven effective. I wish I could be more confident that we can be successful in either instance.

It‘s not over yet.
In the nation as a whole, albeit to a lesser degree in Kentucky, both the number of aggregate cases and deaths continue to increase. Our ability both as a nation and Commonwealth to test for, identify, and report the presence of Covid-19 in the community and to trace exposed persons is still far behind what is needed to detect and respond to localized outbreaks before they show up two or three weeks later as increases in hospitalizations and deaths. It is from such localized hotspots that epidemic expansion can be continuously fueled. More troublesome is a background of resistance from individuals and groups which, for a variety of ideological, religious, political, or business reasons, hold the nation hostage by refusing to follow evidence-based public health initiatives that are effective only when done collectively. Unfortunately, we face these problems with a weakened and fragmented public health system and an inequitably distributed healthcare system overall.

How will we see a “surge” coming?
As the world tiptoes its way through its “reopening'” in the middle of an active pandemic that has no demonstrably effective specific treatment or vaccine yet, how can we feel comfortable that things are not getting unacceptably worse? I do not believe this is a straightforward undertaking. As testing and reporting increase, it is inevitable that new cases will continue to be discovered in new places here and around the world. The most objective indicators of epidemic expansion commonly reported are the number of deaths or hospitalizations attributed to Covid-19. However, either of these, even if consistently reported, are late markers of community epidemic status. The virus has first to find a human host, to incubate asymptomatically, to be recognized in the healthcare system as a clinical infection, to be reported to some public health entity, to be evaluated in the context of current community experience, to be recognized as a diversion from the status quo, and only then to pass some threshold to take effective public health action. Seems to me that by this time, the virus is already out of the barn, racing down the track, and harder to stop.

Who is keeping track?
Compounding the difficulty is that there appears to be no national standard for how to define the items to be reported or even to report them at all. For example, not all states have been reporting hospital or ICU admissions. I can find no federal database at the Center for Communicable Diseases that local communities of public health researchers can draw on. The challenge of aggregating a national experience has been assumed by institutions such as Johns’ Hopkins University, or the New York Times, or The Atlantic magazine that are collecting relevant data directly from individual state and international public health sources. Reporting on weekends is not common nationally and has led to large swings in daily new cases or deaths making timely detection of deviations from the expected much more difficult. Even if a single state, county, or community is doing everything right, it is at the mercy of its neighbors. What happens in Indiana, Ohio, Illinois, Missouri, Tennessee, Virginia, or West Virginia– indeed anywhere in the world– does not stay in those places. At least one of our neighboring states appears still to be in a state of active exponential growth even before “opening up.” How then can we compare our experience with that of other countries, states, or communities?

Below is a plot of aggregate cases of Covid-19 in neighboring states. I use data from the Covid-19 Tracking Project as of May 5th. I plan to work with this or similar data further in the days ahead to compare what unfolds. View the interactive series of data visualizations here.

Continue reading “Is the Covid-19 Epidemic Still Expanding in Kentucky and Its Neighborhood?”

What Can We Learn from Mini-Coronavirus Updates?

Some weeks ago, when public health experts were still visible in Washington, a reasonable-sounding set of guidelines for opening up the national economy was offered. Sadly, the White House seems now to place all responsibility on the individual states with minimal if any major Federal help. It is walking away from, if not contradicting, the advice of the best public health scientists the nation has to offer. I fear that things are going to get interesting quickly and that we will land in an uncharted place somewhere between good and disastrous.

With individual states beginning to open up their economies in different ways and to different degrees, it is apparent that our ability to identify new cases of Covid-19 infection early, to do so in unexpected places, and to be willing and able to do something about it will be critical.

What is a “Mini-Update?”
The Wall Street Journal and other publications often offer a “Coronavirus Daily Update” sidebar with a simple list of Total Cases, Total Deaths, and Recoveries for both the United States and globally. When applied to a given geographic area, these three totals are important elements for predictive epidemiologic models. The fact that the numerous models offered today differ widely (or even turn out to be wildly wrong) confirms the truism that any model is no better than the assumptions it makes and the data available to it. By themselves, these high-altitude aggregate numbers are not fine-grained enough to help us predict the future for Kentucky. I do suggest there are some insights to be gained by examining them. In any event, the numbers are sobering.

What might we learn?
Readers of these articles may notice that I have been educating myself (and I hope some of you) about how we can best use the limited and imperfect epidemiologic data available to us to monitor the opening our economy. Are we are on the right path– or are we falling off the wagon? This is today the major healthcare challenge facing us as a nation. What might we learn from countries where the epidemic started earlier? How are we similar or dissimilar?

Continue reading “What Can We Learn from Mini-Coronavirus Updates?”

How Can We Tell If Our Covid-19 Epidemic Status is Under Control?

What would a recurrent surge of infections look like?

As fifty states with varying intensity of public health approaches to decrease the impact of this highly contagious disease begin to loosen their restrictions, how will we be able to recognize the very real threat of a “second-peak” surge of infections?

Most real experts agree that aggressive testing, new-case finding, and tracking of contacts (backwards and forwards) will be important– indeed critically essential. Small local micro-outbreaks need to be identified quickly and dealt with aggressively. This is going to be a challenge for a number of reasons!

We have become accustomed to seeing a variety of graphs and tables in our public media used to show the status of the epidemic and the hoped-for success in dealing with it. Such macro-presentations will continue– including by me! The problem is that by the time the significance of a given graph becomes evident, the horse may already be out of the barn and running. Nonetheless, what we can’t count, we can’t control. One important metric thought to justify a loosening of restrictions is a sustained two-week decrease in the number of new cases in a given locality. What would this look like in the different possible data visualizations?

Various reports and models use different valid analytical approaches, but care is needed not to unintentionally misinterpret the results. Such graphic representations are used not just to see how things are today, but to predict where we will be in the future. For example, when we visualize positive cases, we can look at daily counts or cumulative counts. Because daily counts vary widely depending on the timeliness of reports and weekend interruptions, using weekly averages is common. When we try to predict where the graphs are going, do we start with the first case, or begin with the 100th to allow matters to settle down? To deal with the difficulty of comparing small numbers that rise exponentially to big ones at the same time, it is common to use logarithmic scales in graphs. What does a “plateau” of cases look like in such different visualizations? Based on current status, how can we tell if things are really getting better– or how much worse they might be?

I prepared the following data visualizations to educate myself what a plateau of the counts of cases, deaths, or tests would look like. What would a 14-day decrease look like? I invite the reader or viewer to step through the seven different graphic representations of the number of current Covid-19 cases in Kentucky and two different futures. I hope the annotated figures are self-explanatory. In these hypothetical scenarios, I used the actual Kentucky case counts from the first reported case through May 4th. I then assumed that the number of new cases would plateau at 170 per day for the next 14 days, and that thereafter the number of new cases daily would decrease by 10 each day until there were no new ones. I was not 100 percent certain in advance what they would look like! I hope you find them useful too. The fully interactive versions can be accessed on the Institute’s Tableau Public website.

Warning: The rest of this article gets quite technical. I am asking for advice from other data nerds about how to monitor the nations’s easing up on its social distancing. Even if you do not look at the 7 figures in detail, at least notice how the same data can be looked at in different ways and the nature of the numbers we have to work with. Trends will be magnified or minimized by the choice of axes or other data transformations!

Continue reading “How Can We Tell If Our Covid-19 Epidemic Status is Under Control?”