Martin Mayer (@MartinGMayer) header
Martin Mayer (@MartinGMayer) profile picture

Practicing clinician-scientist ▪ Internal medicine & cardiology ▪ Research methods, biostatistics, & knowledge synthesis ▪ Expertise in R ▪ Website has details

(I'm not active on Twitter)
Joined December 2014
Photo by Martin Mayer (@MartinGMayer)Photo by Martin Mayer (@MartinGMayer)Photo by Martin Mayer (@MartinGMayer)Photo by Martin Mayer (@MartinGMayer)Photo by Martin Mayer (@MartinGMayer)
Thread on the reporting of hierarchical outcomes inspired by 1 of the only 2 #Impella RCTs.
7
65
20
125
Quoi? Another chapter in what is starting to feel like a kafkaesque saga...
This Post is from an account that no longer exists. Learn more
1
This. Also relevant for people reading research. I feel like I say this more often than should be necessary, but quantity of data ≠ quality of data. Dog crap is still just dog crap, even if you have a *lot* of it in a single pile. Same goes for biased data/data at risk of bias.
This Post is from an account that no longer exists. Learn more
1
A good read, and pairs well with tandfonline.com/doi/full/10.… (among others, of course).
P values (significance tests) were first proposed as an informal heuristic to help assess how “unexpected” the observed effect size was if the true state of nature was no effect/difference. “Significance test” is confusing/misleading. “Unexpectedness test” more appropriate.
2
Sorry, @bmj_latest, but this is a bit irksome for me. It's good you're stimulating discussion/debate (e.g., this thread), but we still haven't really managed to get the basics right/optimized, so the @EricTopol review seems to be getting the cart ahead of the horse. (1/)
2
3
1
10
And surely, better training in epistemology, research methods, & statistics is much needed. Even starting w/ basic statistical concepts, e.g., link.springer.com/article/10… (@Lester_Domes, @stephensenn, @ken_rothman, Carlin, Poole, @goodmanmetrics, & Altman) would be a good start. (7/)
1
1
3
In my opinion, we need training --> mastery of the above much, much more than training in genomics, AI, etc. (8/8) @JeremySussman @ProfHayward @venkmurthy @VPrasadMDMPH @adamcifu @ADAlthousePhD @statsepi @boback @PaulGlasziou @DrMarkMurphy @f2harrell @dailyzad @Ricky_Turgeon
1
9
Finally, the @nytimes has a bio/obit on Alan Turing, brilliant mathematician and one of the fathers of modern computing. His work is central to modern life (even twitter) yet he was treated as a criminal during his lifetime. nytimes.com/2019/06/05/obitu…
1
14
1
59
This analogy is absolutely spot on. @JSurgRes, you really need to pull the paper. How can you possibly defend leaving it? Is it embarrassing? Sure. But not pulling it is shameful & will hurt your credibility more. More here (including from @ADAlthousePhD): pubpeer.com/publications/439…
Replying to @f2harrell @JSurgRes
Extending the analogy a bit further: it's the equivalent of you (with a team of statisticians) publishing an article promoting a surgical technique that was discredited 20 years ago; the journal accepting it without a surgeon reviewing the paper; and then...
5
7
Another testament (not that we really need more proof) that there's still a long way to go in actualizing Altman's charge of "less research, better research, and research done for the right reasons." teapawt.net/MartinGMayer/status/10…
1
(1/2) Glad to see some temperate voices getting published in @NEJM re: precision medicine.
Wise words: “Medicine in the molecular era will be no more precise than in prior eras - evidence synthesis, clinical judgment, and patient presences all factor in” @NEJM @kewatson @boback
1
1
8
This. You say precision, I say precaution. ...& lest we forget, care *is* personalized when using evidence-based medicine & shared decision-making. Let's focus on getting those right & est. as the true standard of care before we get carried away w/ ideas like precision medicine.
2
Now that we heard it from Prof. Tannock, I hope everybody agrees to this? #ASCO19
10
78
18
169
Is it time to retire the concept of statistical significance? Read what @MartinGMayer from @EBSCO_Health has to say: https://teapawt.net/t.co/iDjLm7utZF #statistics #EpiTwitter
7
2
This. You say precision, I say precaution. ...& lest we forget, care *is* personalized when using evidence-based medicine & shared decision-making. Let's focus on getting those right & est. as the true standard of care before we get carried away w/ ideas like precision medicine.
This Post is from an account that no longer exists. Learn more