AI for Covid? A prediction…
In June 2020 I was interviewed about how AI could assist the clinical management of Covid-19 patients. My answer in a nutshell was: It won´t.
I predicted that there will be a lack of high-quality data and that there won´t be any substantial clinical impact.
What is the status in 2021?
A living systematic review of all available clinical Covid19 models so far came to this conclusion:
“This review indicates that almost all published prediction models are poorly reported, and at high risk of bias such that their reported predictive performance is probably optimistic.”
Another systematic review of all available models for Covid-19 diagnostics in imaging came to this conclusion:
“Our review finds that none of the models identified are of potential clinical use due to methodological flaws and/or underlying biases.”
Hundreds of publications. Zero clinical impact. What a waste of time, energy, and money!
AI in healthcare is exciting and full of potential. But we need to stop the ride on the hype train. We need to do it right, based on best practices.
If you are interested in the interview watch it here:
https://www.youtube.com/watch?t=776&v=xpE3iZPS3yk&feature=youtu.be
Read the highly recommended reviews here:
https://www.nature.com/articles/s42256-021-00307-0
https://www.bmj.com/content/369/bmj.m1328
(image credit: Globe vector created by iuriimotov – www.freepik.com)
