Not one, but two. 1. Holland vs Italy, Euro 2000 semi 2. Juventus vs Real Madrid, UCL 2002-03, second leg

One of the main points is that diagnostic cohort study designs are preferred over case-control study designs. Feasibility and study turnaround time would be challenging with cohort designs. #medicalAI #ArtificialIntelligence thelancet.com/journals/landi….

Great assist by DiMarco for Yarmolenko. Truly Inter ⭐⭐ 😒.

Programmer's day is celebrated on the 256th day of the year because the highest power of 2 which is less than 365 is 256 and 256 is the number of distinct values that can be stored/represented in the 8-bit system. #ProgrammersDay linkedin.com/posts/nevenaso…

Details of MRMCsamplesize, an R package for sizing reader studies. #rstats

"An #AI [skin cancer] algorithm that was superior in experimental studies was significantly inferior to specialists in a real-world scenario" thelancet.com/journals/landi… The Lancet Digital Health

Observational epidemiology studies have associated chocolate consumption with reduced diabetes risk. However, an RCT now shows that cocoa extract vs placebo in ~21K w/o diabetes did not reduce the risk of incident self-reported diabetes diabetesjournals.org/care/article-a… Diabetes Care, a research journal of the ADA 🌟

But, but, RCT is the gold standard of evidence. 🥹

C.H.I.E.S.A 💥💥

🚀 Introducing the Stanford AIMI Dataset Index – a diverse inventory of AI-ready datasets for machine learning in healthcare! We hope this resource will help democratize access to health data for AI development. Explore & contribute here: bit.ly/3RmEGFL #AIMIDatasetIndex

Out today! Joint project with PATH Central TB Division | #TBMuktBharat Nagpur Stop TB Partnership & Qure.ai . 16% increase in TB yield, #AI picked cases missed by radiologists. DDG - TB | #TBMuktBharat Dr Raghuram Rao Ministry of Health USAID India The Global Fund journals.plos.org/digitalhealth/…

I'm quite skeptical of this. An associate professor at Harvard needs about 40 papers. 97 papers would put one on the path to be a professor. Even if these are a lot of abstracts, it is still hard to believe that these individuals actually collected the data to write these.

RIP, Gigi Riva

For years #EndTB efforts have been negatively impacted due to lack of political commitment but it’s all changing now. Qure.ai & FUJIFILM Corporation have set up a TB x-ray van powered by AI & #StopTB’s Lucica Ditiu & @Dohgovph Health secy Teddy Herbosa MD ❎ visited the van & got tested as well.

😳

Just published Our review article of all the medical #AI randomized controlled trials The Lancet Digital Health thelancet.com/journals/landi… Ryan Han Julián Nicolás Acosta Pranav Rajpurkar

It's been a while since I saw such a dominant play by Italian teams in a major final. #Atalanta, #Lookman, #Scalvini, #Djimsiti and #Ederson schooled Leverkusen throughout the game!

My message here is: LLMs work best when combined with humans. That means the thing to be most excited (or scared) of is the power of computers+humans, not just computers on their own ...which is something Douglas Engelbart pointed out in 1962! dougengelbart.org/pubs/augment-3…

Our work on investigating the difference in performance of a CXR AI algorithm for TB detection in digital DICOM CXRs and mobile captured photos of CXR films is now published. shorturl.at/jaSmF

A simple mobile phone captured photo of an X-ray film may be enough for an AI algorithm primarily trained on digital X-ray DICOM images to "look" effectively for signs of TB. #artificalintelligence #medicalai #tuberculosis