Read on Twitter
Debunking the study that sank the #BigTEN
@ProfDFrancis does so in this thread - please read
I’m going to simplify his thread even further to explain how he debunks a recent study in Jama that shows heart inflammation after Covid infection in a high % of patients https://twitter.com/profdfrancis/status/1294370975702822914
@ProfDFrancis does so in this thread - please read
I’m going to simplify his thread even further to explain how he debunks a recent study in Jama that shows heart inflammation after Covid infection in a high % of patients https://twitter.com/profdfrancis/status/1294370975702822914
First, here is the study https://jamanetwork.com/journals/jamacardiology/fullarticle/2768916
Here is the entire table and I’ve circled the one figure he uses to disqualify the study
Left Ventricular Ejection Fraction (LVEF) for Covid Patients was 56 (54-58)
56 is the median
54-58 is Interquartile range (IQR) meaning 50% of all patients fall into that range
Left Ventricular Ejection Fraction (LVEF) for Covid Patients was 56 (54-58)
56 is the median
54-58 is Interquartile range (IQR) meaning 50% of all patients fall into that range
First, he shows that in normal population, the mean LVEF% is 70.3%
And the Standard Deviation (SD) is +/- 6.9%
SD is different than IQR so next he explains this
And the Standard Deviation (SD) is +/- 6.9%
SD is different than IQR so next he explains this
1 standard deviation encompasses 68.3% which he rounds down to 2/3
So, 7% +/- =equals 1 SD makes up 2/3 of the total sample
Next he puts the +/- 2% LVEF from the study into the standard deviation model
He calculate 2% LVEF would be 2/7 of 1 Standard Deviation
So, 7% +/- =equals 1 SD makes up 2/3 of the total sample
Next he puts the +/- 2% LVEF from the study into the standard deviation model
He calculate 2% LVEF would be 2/7 of 1 Standard Deviation
Next he multiplies 2/7 (% of SD) x 2/3 (% of total represented by 1 SD)
4/21 or about 20%, but this is a little low because the middle bit of the normal distribution curve is higher
He raises it to 25%
4/21 or about 20%, but this is a little low because the middle bit of the normal distribution curve is higher
He raises it to 25%
Finally he concludes:
“So the chance of a random person's EF,
drawn from a population which had the most favourable mean, and typical SD (7 units),
having a value within ±2%of the mean,
is about 25%”
Throws this in for a little common sense
#wewantaseason #wewanttocoach
“So the chance of a random person's EF,
drawn from a population which had the most favourable mean, and typical SD (7 units),
having a value within ±2%of the mean,
is about 25%”
Throws this in for a little common sense
#wewantaseason #wewanttocoach
