COVID-19, cases in Belgium March 15, Sunday

From the Public Health

SPF Santé publique - 197 cas 15 mars, 2020

It gives 197 new infections today, 137 in Flanders, 27 in Brussels, and 28 in the Walloon region.

A total of 886 positive tests since this all began.



My additions:

My little map ...


On this map, click on one of the four icons to reveal more details: confirmed cases, and percentage of cases in Belgium. This by Region, plus unknowns (missing data).


Regional distribution:

Distribution régionale - 15 mars 2020



A summary of the total tests analyzed to date: This is no longer possible.

These data are no longer shared with the public by SPF Public Health, nor by their replacement Sciensano, a group of epidemiologists.


Pity. Its very important. Other countries in the world would say so.


Nombre de nouveau cas - 15 mars 2020


Total Cases (i.e;, total positive tests reported) 

Cas confirmés en Belgique - 15 mars 2020

Q: "Still a linear progression, or has it become exponential?"

(click to enlarge)

Linear or exponential - 15 March, 2020


  • Left, the data from the Public Health to date.
  • In the middle (n ° 1) the linear relation: day of observation (we are the 15th). The current cases 886, and those predicted: 655. An error of estimation of -26.1%. A level of correlation, R² = 0.86639 (yesterday, 0.8912 the day before yesterday: 0.91334).
  • On the right, a relation proposed to be exponential: R² = 0.94691. (yesterday: 0.92751; day before yesterday, R² = 0.90019). Predicted cases for this day: 1,550. Estimation error 75.0%.


Conclusion today: The linear model is now weaker than the exponential, even though the latter has a significant error of estimation in evidence. How to explain the two together?


Data principle #1 : Garbage In, Garbage Out.

When a correlation becomes stronger but the error of estimation is incresing at the same time, it's time to go back to the raw data.


The definition of the above correlations (which I know many find too complicated. Sincerely sorry for that) assumes a normal distribution of the sets of incoming data. But are they normally distributed? If this is not the case, the estimates of errors which seem significant, are perhaps simply hinting at this fact. So let's take a look. Time for a U-turn.


Q: "Is the distribution of test data normal?"


A: "No, not at all. It is actually quite skewed in one direction."


And here is the reason.

Here is a second little map.


Look carefully to find this skewed distribution ...


This is a map of the laboratories responsible for analyzing the samples for SARS-CoV-2. (That is the current name of the virus that actually causes the disease, COVID-19).


On this map, click on an icon to reveal more details: the name of the laboratory, for example. Those with a yellow "L" are in the Flemish region. A blue "L" = Bruxelles-Capitale. And the green "L" icon indicates the position of a laboratory in this network in the Walloon region.



I'll help you count them:


Labs: Total

Région Flamande Rég. Bruxelles-Capitale Région Wallonne
22 16 2 4
100% 72,73% 9,09% 18,18%


I can already hear someone say, "It's like that, and it won't be any different."


But we were looking for an answer to a question about the regional distribution of test results in Belgium. Is that what we've got, an answer?  Let's add a line to our table.




Région Flamande Rég. Bruxelles-Capitale Région Wallonne
  22 16 2 4
  100% 72,73% 9,09% 18,18%

Cases COVID-19

886 570 98 202





A corelation with just three data pairs is pretty weak research. Nevertheless this doesn't look like something caused by a coronavirus.

Correlation of Labora tories and COVID-19 cases


A coincidence? Not at all. Probability that it's due to chance? (p <0.026) Only 26 chances in 1000 that it is random, and with no known explanation. So I doubt it.


Where are the cases? Where are the laboratories?


Liste Labo - 15 mar 2020


A newspaper picked up on Friday, this map from Sciensano (the epidemiology group used by Public Health) to loudly say in an article: "


"Coronavirus: the most affected region in Wallonia is Mons"

Their conclusion based on this map of positive cases of COVID-19:


Distribution des cas confirmés en Belgique - 13 mars, 2020


Their conclusion may be wrong. They should have titled the article:


"Coronavirus: the most tested and diagnosed region in Wallonia is Mons"


Go back for a moment to the map of labs above. Mons is very close to several laboratories. Mons is therefore very close to the place of diagnosis when someone has a fever, cough and shortness of breath, and someone sticks a cotton swab deeply into their nose to obtain a sample and send to lab. Almost the lab around the corner.


Let me continue

The government, their experts, Public Health, Sciensano, the News on TV and everyone, except a lady I know who is blind and deaf but also very nice, says that COVID-19 is increasing in Belgium and will start to increase further.


The above correlations studied here, strongly suggest the same thing. In fact, we are probably already in this exponential or geometric phase this day. We just don't have the data to prove it. Why? Because the figures for Wallonia do not reflect the real situation. There must  already be more cases.


Let me digress a little (although it really isn't).

In the United States where I practiced surgery and intensive care, lawyers were very present in the system, perhaps considering themselves as caregivers.


Compared to European medicine as a whole, lawsuits against doctors and hospitals are there, ridiculously frequent and for ridiculously high settlements.


By far the most common reason cited for successfully prosecuting a doctor is for "failure to diagnose". This is the main task of any doctor: to name the patient's illness. Without that, how can treatment follow? This is true for all treatments. No diagnosis ... so, "What are you treating, doctor?"


This is as true in surgery as it is in all the other specialties, from pediatrics to cancer care to psychiatry.


Now COVID-19 poses a diagnostic challenge: perhaps minimal signs and symptoms if detected early, a patient who will make you sick in your office if he has it, and a history obtained now by phone: rather than in person, and without a good physical exam. In the United States, I assume lawyers are already packing their briiefcase in anticipation of proving "diagnostic failure" before the judge.


We will have to accept this challenge. The best diagnostic test has always been a good history of the patient's illness followed by a physical exam, performed by a competent doctor. Then again, following Father Damien's example and staying amont the lepers (or COVID-19 patients) until you get it is not for everyone.


Now we can only do a laboratory test, and indicated by a basic set of screening questions that don't always lead to a diagnosis.

I have said "we" here a few times, but I'm retired. That makes one aspect of all this more stress-free.


I suppose no one in Belgium will set up portable CT scanners to do scans and quickly examine the patient's lungs like they did in China. 200 per day and per center.


And we are even more stuck, given the directives of Public Health 3 days ago, useful for underdiagnosing a whole segment of the population in all the regions of Belgium.


Now add to that there is a bias on who is tested, depending on the number of laboratories in a region.


It is actually, really shameful.

And if those who decide all these things think that the Walloon is just going to sit idly by and drink his beer to swallow hiw fries with, while waiting to get sick, while not having a diagnosis, ... I bet it won't fly like that. It's time to bring down the drones over Wallonia, and get a little serious up close and personal.


What is the worst in all this?

You've heard enough bad news in the past few days and you don't want to hear more. I know. I'm almost done.


Here's the worst ...


The sooner a disease is diagnosed, the sooner treatment can start where it is needed.

Reducing delay, invariably obtains a better outcome from the disease. I therefore suppose that when history examines the cases of COVID-19 that arose in the Walloon Region, in March and April 2020, it will find that this is where the most serious cases, with the most morbidity and mortality, were eventually observed.


That would be inexcusable.



"Pardon me, Sir ..."

"Could you stop belly-aching just long enough to suggest a positive solution to this problem as you see it?"


Increase testing presence and capability in the Walloon Region.

That's all. Diagnoses of COVID-19 will increase.


I find that preferabe to saying: "Let's let the whole herd get COVID-19, then we won't have to sort those who have it from those who don't. They'll sort themselves. Kind of like in Italy.

More testing capability in the Walloon Region. ASAP!

>>>>>> to world data for today, March 15.


<<<<<<<<<<<  Home






0 Poster un commentaire

Inscrivez-vous au blog

Soyez prévenu par email des prochaines mises à jour

Rejoignez les 13 autres membres