Title

We are all about marketing, data, analysis, innovation and technology

Tuesday, June 23, 2020

What Do The Latest Coronavirus (COVID) Data Reveal About Testing, Positive Cases & Deaths

Every morning we wake up to cable news channels like MSNBC, Fox, CNN or our local affiliates to reporters exclaiming "spiking" Coronavirus cases based on yesterday's test numbers.  They give us nothing to compare our numbers to or to baseline with.  As a result we are making decisions based on incomplete data.  As a statistician, I find this very frustrating.

I simply want all the data presented to me about the trends so I can make my own informed decision.  As any statistician will agree, more data points are always better than less.  So, Rhonda and I decided to do our own research as we did back in March at the start of this horrible pandemic.  Data we searched for to include in this report is comprised of :
  • COVID testing numbers by date
  • Positive COVID cases by date
  • Hospitalizations due to COVID by date
  • Emergency room admissions for flu like symptoms by date
  • Deaths due to COVID by date
  • Deaths due to cars accidents and the flu/pneumonia annually
So, lets look at the facts.

COVID Testing Trends
We will begin with a look at the COVID testing numbers which have been a hot issue as of late on the various news channels.  The data below is from John Hopkins University and The COVID Tracking Project. 



This chart is showing incremental and NOT cumulative testing occurring on a daily basis.  So yes, testing is way up!  And, we also see the percent of positive cases per day has decreased significantly since late March and early April (the line chart).

Does this make sense?  Yes of course.  Initially, only those tested were those that were highly likely ill.  As a matter of fact, they were the only ones allowed to be tested.  Today, testing is open to anyone and testing centers are finally, thank goodness, everywhere including less affluent and predominately black communities (don't get me started on this).  As a result, many being tested today show no symptoms because they were either not sick or were asymptomatic.

But one concern we have based on this data is the recent uptick in the percent of positives -- since the opening up of America (around June 1).  This is something we do need to keep a close eye on and hope we get this to flatten quickly.  But all in all, good news with some caution.

Tracking Hospital Outpatient Visits for Mild COVID Symptoms

The U.S. Outpatient Influenza-like Illness Surveillance Network (ILINet) provides data on visits for influenza-like illness (ILI).  Mild COVID-19 illness presents with symptoms similar to ILI.  So, ILINet is being used to track trends of mild to moderate COVID-19 illness and allows for comparison with prior influenza seasons.

The graph below shows the trend of this data.  The time line is reported as "year + week."  So, "202003" represents the third week of 2020, for example.




As we can see, the height of hospital visits for these Mild COVID symptoms occurred during the last week of December, 2019 through the 12th week of 2020 (end of March).  This is represented by the overall solid black line.

Today our percent of emergency room visits for such symptoms is well below the national baseline of 2.4% for the flu.  So again another good sign that we are at least moving in the right direction... assuming we continue to observe social distancing, hand washing, etc.

Tracking Emergency Room Visits for COVID Like Illnesses
NSSP is a collaboration among CDC, federal partners, local and state health departments and academic and private sector partners to collect, analyze and share electronic patient encounter data received from multiple healthcare settings. To track trends of potential COVID-19 visits, visits for COVID-19-like illness (CLI) and ILI to a subset of emergency departments in 47 states are being monitored.



Similar to the chart produced by ILINet (in the prior section), trends began on a downward slope around April of 2020.   And, currently for all emergency room visits, we are at a rate of around 1.9% for COVID and .8% for the flu.  Once again, we are heading in a good direction.

COVID Hospitalizations Over Time
Again, using data from the CDC we can track the number of hospitalizations for COVID patients.  To date we have over 23,000 such occurrences in the U.S.



As this graph shows we are currently on a downward trend regarding hospitalizations.  Again this assumes we maintain good social distancing and hygiene practices.  But all in all, good news.

One caution with respect to statistics on hospitalizations:  the data lags in reporting and is subject to review by many doctors prior to release to the CDC.  Thus, any patients with missing data are not reported until data is obtained so the absolute numbers understate the hospitalizations, but the trend is still meaningful.

Deaths Due to COVID
So, what do the real COVID mortality numbers look like?  What is the trend?  Again, a chart that the media outlets are just not showing us whether FOX, CNN, MSNBC or your local news...  which I personally find very frustrating and upsetting! 

Below is a chart showing the number of deaths by day for the US, Italy, Germany and the United Kingdom.  Please remember that the US population is anywhere from 4 to 5 times that of these other countries when comparing charts below.



To date we are still at around 500-800 deaths per day.  This is still much higher than the number of deaths due to the flu and pneumonia combined which stands at about 200-300 per day.  Flu deaths per year range from 30,000 to 60,000 while pneumonia deaths per year average around 49,000.  

Yes, believe it or not, each year we lose about 100,000 citizen due to the flu and phenomena.  A horrible figure.  Not that any loss is good, but the COVID figures above do appear to be moving in line with these other illnesses that we cope with yearly.

Weighing the Risks:  Car Accidents vs. COVID
Believe it or not, for most under the age of 45, your odds of dying in a car accident are significantly higher than contracting and dying of COVID.  Bloomberg New Media, recently did an in-depth analysis of this data.



So, for example, if you are 35-44 years old, you are 35% less likely to die of COVID than to die in a transport accident (0.008626% vs 0.0137%).   However, if you are 65-74, you are 9-10 times more likely to die of COVID than die of a transport accident (0.135047% vs. 0.01398%).  

NOTE:  0.135047% is not "13.5 percent" but rather ".135 percent" -- slightly more than one-tenth of one percent.

Please note that this complete analysis is only showing and analyzing national data.  We do realize that for various states or cities, other things might be going on due to lack of social distancing or lack of wearing face masks in public.  But on a national, bigger picture level, we do appear to be showing some light at the end of this very long tunnel.

Rhonda and I both hope this helps you put this terrible virus and pandemic in perspective.  We are just stating the facts and do apologize in advance for the lack of empathy that might appear in this report.  That was not our intent at all.  We are simply stating the numbers.

And as always, those at risk like the elderly or those with preexisting conditions, caution must always be exercised as the data also shows. 

Warm regards and stay healthy!
Perry & Rhonda Drake

Sources:
https://www.health.com/condition/cold-flu-sinus/how-many-people-die-of-the-flu-every-year
https://www.cdc.gov/flu/vaccines-work/burden-averted.htm
https://www.cdc.gov/nchs/fastats/pneumonia.htm
https://www.cdc.gov/coronavirus/2019-ncov/covid-data/covidview/index.html
https://covidtracking.com/
https://coronavirus.jhu.edu/
https://ourworldindata.org/grapher/daily-covid-deaths-3-day-average?country=DEU~ITA~KOR~USA
https://www.bloomberg.com/opinion/articles/2020-05-07/comparing-coronavirus-deaths-by-age-with-flu-driving-fatalities
https://ourworldindata.org/coronavirus
https://ourworldindata.org/grapher/daily-covid-deaths-3-day-average
https://gis.cdc.gov/grasp/COVIDNet/COVID19_5.html



No comments:

Post a Comment