Pandemics remain a clear and present danger. Once they get established in the human population it's often too late to stop them in their tracks but we have a real chance to prevent that happening-those who are advocacy minded should promote enhanced surveillance and data sharing.

Saw an ad encouraging you to “know your risk for prediabetes.” What does it mean if you are at risk for prediabetes? Are you pre-prediabetic?

A reanalysis by company researchers in respons to our critique of Hypotension Prediction Index (HPI): doi.org/10.1097/ALN.00…

HPI and MAP are compared, and there is no difference!

'Forward' prediction of hypotension in 5 min:
AUC for HPI: 0.923
AUC for MAP: 0.932

1/3

I think EndWokenessNow88 with the confederate flag profile pic and 50 posts a day on black crime statistics has a point. I might be letting my personal bias (existing while black) affect my views on the science.

So proud of Linda Amoafo, who did an amazing job presenting and defending her PhD dissertation. Honored to have been part of her thesis committee. #BiostatsElevated

In the next 3 episodes of Nullius In Verba we will be discussing Paul Meehl's Philosophical Psychology course. Still one of the best free online courses you watch 35 years later! meehl.umn.edu/video If you want to listen along, we will discuss lectures 1-3 two weeks from now!

When you stoke fears and anxieties, these are the sorts of draconian solutions that people cook up - 12 hour school days starting at 7am will fix absolutely nothing thetimes.co.uk/article/school…

'Weighting methods for truncation by death in cluster trials'

arxiv.org/abs/2404.10629

A beauty of an opening dissertation chapter by Penn GGEB Dept. of Biostatistics, Epidemiology & Informatics PhD Biostats student Dane Isenberg (proud cmte chair here😊).

w/ Nandita Mitra (@PennCausal) Fan Li

Stefan Feuerriegel Nature Medicine Would love to see a session on this at ICTMC 2024. We want focused sessions on AI in trials. It would be great to see some of these ideas discussed in a room that also includes experienced trialists and statisticians. ictmc.org/spotlight-sess… cc Mihaela van der Schaar

Work on AI for tissue diagnosis with tons of 🚩

🚩 no sample size calcs
🚩 no calibration
🚩 “class imbalance” nonsense
🚩 Improper scoring rules with balanced accuracy and balanced F1

I think the stat referees at Nature Medicine asleep for this one

nature.com/articles/s4159…

The difference between reference-based *multiple imputation* & reference-based *conditional mean imputation* inference should not be ignored -different trial conclusions may be obtained👇we explore this difference to help guide statisticians choice here: onlinelibrary.wiley.com/doi/full/10.10…

🚨Publication🚨
Today we publish the largest ever study of conflicts of interest and the European medical device industry
• Between 2017-2019 medical device companies disclosed €425 million in ‘education’ payments
• This is likely the tip of the iceberg
sciencedirect.com/science/articl…

Curious why statisticians recommend including the outcome in your imputation models? Check out our new paper in Statistical Methods in Medical Research! Sarah Lotspeich (She/Her), Staci Hepler, and I show with some simple mathematical derivations why this is really a requirement!

I think if you're against this kind of thing you should actually explain what you think is bad about it, rather than leaving it to innuendo and scare quotes.

**New tutorial (PREPRINT)**
We review multiple imputation approaches for handling incomplete longitudinal data (including the case with higher-level clusters), summarize available guidance and illustrate implementation (R and Stata)
#epitwitter #rstats
arxiv.org/abs/2404.06967