Thread by @ProfDFrancis

In a spirit of solidarity with actual scientists, I did an interview with Eike Nagel, the senior author of the famous JAMA Cardio paper that accidentally shut down some US Sports.

I apologised for the unintentional hurt, explaining again that I thought I was just teasing Valentina (Mrs Eike) about what I thought was a typo in the abstract.

And only looked at Table 1 when after 12 hours people didn't seem to pick up the query in the abstract.

I reiterated how impressed I was with the rapidity of the update of the numbers.

I have NEVER seen anyone move so quickly to fix errors.

*THAT* is the mark of a true scientist.

I explained that I had carefully read the correction document, and had considered the explanation given, namely that the raw data had been correct, but the incorrect statistics originally presented (means instead of the stated medians).

I asked what the originally presented ranges were, because obviously they were not IQRs, and even with the explanation provided I could not see how they could be SDs, SEs, CIs, or anything else I could think of.

He was not willing to speculate on what they were, other than that they were a mistake.

I moved on to a few other questions which I will not related here, which again he was not willing to speculate on.

At this stage, he became apprehensive and withdrew his permission to broadcast the video. I accept his right to do so, as this was only a scholarly discussion.

Unfortunately I am a bit numerically biased, so I can't help being curious about the origin of the original numbers, as I don't think the explanation in the Correction quite fits.

In the end we agreed that nobody would ever be able to find out how the original numbers arose.

It will be a permanent mystery.

To me, this means the Correction is not quite as perfect as it could be, because (to me) it implies that the original numbers were correctly calculated wrong statistics.

But I think it is only me that is interested in that, so don't worry.

We moved on to the more congenial part where we discussed what it all meant.

This was much easier to get agreement on.

I really want to show you the video so you see the whole picture and context, but after a day to think about it, he still hasn't agreed to show the video (I sent him a copy to review), so he HAS given me permission to convey his interpretation.

(Remember, these are not quotes, they are my invented summary)

Q. Why don't you explain on Twitter?

A. Because on Twitter, if you say something, and reply, and you don't reply back quickly, people think you are being shady or hiding something, when I am just doing my actual job

(I accepted that this can happen. Of course people realise you have a job, and can't answer everything, but they would be surprised if you didn't answer key questions. Nevertheless his point is important: not all scientists want to spend time on Twitter.)

Q. Should people use your study to shut down sport in USA?

A. I have no comment on that. We DEFINITELY didn't say to do that in the paper. I don't know why they would use our paper to say that.

(agreed!)

Q. Aren't you reassured that the CRP and Troponin is pretty normal in the revised data? Would you stop such patients doing sport?

A. After a bad infection, I would encourage them to only gradually ease themselves back into heavy activity.

(Fair enough)

Q. What about ordinary people wondering if Coronavirus has infected them without their knowledge and is causing heart disease?

A. I don't think ordinary people should worry about this at all.

(Excellent!)

Q. Don't most severe virus infections hospitalising people cause problems all over the body including the heart, which then fades away all over the body, including the heart?

A. Yes but people haven't studied it comprehensively for other viruses. Ours is the first for Covid.

Q. What is the main message you want to convey to people?

A. After coronavirus, there are subtle but definitely visible abnormalities in the scans (but not the blood tests) that are beyond what are found in the risk factor matched controls... (cont)

... We don't know whether these are harmful in the long term, or whether they will just be there like a little scar in the skin after a cut.

In other conditions these features can be the precursor of heart problems many years later.
...

...

We will be studying these patients in the long term to see what impact these actually have.

We did try to make this clear in the paper: the need for more research.

------------------------------------

He did give me permission to convey a summary like this of his view, and I hope it is useful.

Finally, I must reiterate that please do not joke about people needing to resign etc, as his institution has no sense of humour unfortunately.

You can tell he is a top-class scientist because he raced to correct the slipups, and was happy to talk to me despite worrying that i might be out to get him.

I also spoke to Francois Balloux (whom I did not know before) yesterday about his joking tweet about resigning - he agreed to clarify that he was just joking, for the benefit of humourless people.

Francois is also unsatisfied with the Correction, for different reasons from me, but he is happy to reiterate that he was only speaking in jest.

Francois thinks it is unwise to say the message hasn't changed, when the press etc has misinterpreted the message. People may think the paper "still" supports shutting of sports, when in fact it never did.

However I hope in a few days Eike will OK me to show the video.