We can't assume that COVID ML models developed in one hospital will perform well in another one. Data drift occurs in healthcare settings, too, due to changing patient mixes and treatments over time. https://t.co/ndyKPT8PP4 #machinelearning #healthcare #c
3,680/ Illinois, USA (adult current smoking rate was 14.5% in 2019) Of 516 hospitalized with Covid-19, 12.8% were nicotine users (not specified if current or ever) https://t.co/UT11ZThHvK
Predicting clinical outcomes among hospitalized COVID-19 patients using both local and published models https://t.co/A4S8Q8TsYd