Pulmonary Critical Care, data-science, machine-learning, and AI in critical care, health services research, University of Michigan
Joined August 2013
- Tweets468
- Following74
- Followers1.1K
- Likes184
Proposal #3: Power pulse oximeter tests to detect small but clinically important differences between racial groups
- Small performance differences translate into significant population disparities, hidden hypoxia rates of 2% versus 12% for example!
6/ It worked. We created a test dataset of chest x-rays each reviewed independently by 6 intensive care MDs and chest radiologist to validate model. LOTs of work.
1/ Our deep learning model to detect findings of ARDS on chest x-rays, now in Lancet Digital Health
thelancet.com/journals/landi…
1/ Our research letter on racial bias in pulse oximetry measurement, out today in NEJM nejm.org/doi/full/10.1056/NE…
Advice from @iwashyna: When you have to stay home to watch your kids, don’t try to work from home
Agreed, see my follow-up tweet, I suspect negative agreement is the problem driving the lower kappa, (here negative is not failing SBT)
Rates of ARDS diagnosis among physicians reviewing the same pts varied considerably, from 9-47%! 5/7
ARDS risk factors matter: pts with pna + shock + P/F<120 nearly always had consensus ARDS, patients without risk factors and P/F>180 never had ARDS 4/7
Our study: Differences between patients in whom physicians agree and disagree about the diagnosis of ARDS, with Tim Hofer, Ivan Co, @Jakob_McSparron, @iwashyna now at @AnnalsATS @davidlederer #VisualAbstract atsjournals.org/doi/abs/10.1… 1/7
We ask: Can we use detailed EHR data to more comprehensively evaluate delivery in clinical practice? (EHR flowsheets = every recorded tidal volume) 2/5
Our study, Evaluating Delivery of Low Tidal Volume Ventilation in Six ICUs Using Electronic Health Record Data, with @MichelleNgGong, Carl Haas RT, and @iwashyna is published online @CritCareMed journals.lww.com/ccmjournal/… 1/5
