Harlan Krumholz

And I thank all the people who point them out to me.

And Simon pointed out last week I made a major error in terms of test sensitivity and specificity.

Not difficult because often within these terms you make slight mental errors and get certain aspects wrong.

And what I got wrong was just the result from sensitivity.

We talked about taking 1,000 people with a prevalence of antibodies of 5%, so 50 people in 1,000 have antibodies, 950 do not.

The sensitivity of the test was 70%, therefore 35 people

out of the 50 will have the disease and test positive.

And importantly, 15 people with the disease will test negative under that seroprevalence and that result.

And I'm illustrating one of my major teaching points there is to not use the languages of false positives and false negatives and throw that in the bin and actually say what it means.

It is the people without the disease who test positive.

And if I'd have said it that way, it would have been correct.

But by using the gobbledygook sometimes of EBM and test accuracy, I made an error.

So sorry, folks.

Gobbledygook of test accuracy.

Goodbye from me.

Yes, hi, I am Editor-in-Chief of BMJ Evidence-Based Medicine, Professor of EBM at the University of Oxford and a GP.

Yes, hi, I am Editor-in-Chief of BMJ Evidence-Based Medicine, Professor of EBM at the University of Oxford and a GP.

I agree.

This is a very interesting time in the world of evidence-based medicine.

As we're going to hear on the programme, you're going to have evidence from case reports.