4 ms·

It is somehow of a side comment, but I always wondered why there is no CAM - Computer Assisted Medicine. The kind of program where you would input everything yo

It is somehow of a side comment, but I always wondered why there is no CAM - Computer Assisted Medicine. The kind of program where you would input everything you checked (either physical exams, or quantitative information such as blood tests results, heart rate etc.) In winter, a typical result from this program when someone comes with a sore throat, pain in the joints and headache is "flu" - at 98%. But there is 1% left for parasitical infection of pythonus golangus, which can be extra tested by doing this and that. And then the next 10 possible illnesses, extremely low probability. What this could avoid is that an MD who is human and has zillions of patients with the same symptoms is helped to target something else either when there is something uncommon, or the typical treatment si not as effective as it would be. I am a physicist by education, and IT by work. All what we do is more or less reproducible and it is extraordinarily uncommon that a bug is due to a cosmic particle hitting the CPU rather than a forgotten semicolon. I believe however that biology is much more unpredictable, with a ton more parameters and an MD may have not heard about an illness (or have forgotten) so reminding them about the possibility could be enlightening (and direct the gut feeling they may have otherwise).

rscho 5y ago
Medicine of today is still more art than science. Although we strive towards what you describe, there are enough poorly defined variables involved in diagnosis to make CAM pretty much useless. A major cause of that is the information requirement. People will describe and interpret findings differently, making reporting useless. Trying to directly measure things will stumble across very complex issues of calibration, etc. Actually, all the current hype goes to AI/ML/automation, but the real priority would be to improve data collection. Unfortunately, it's less sexy and much harder. Old-school docs still have good times before them.

rscho 5y ago

Medicine of today is still more art than science. Although we strive towards what you describe, there are enough poorly defined variables involved in diagnosis to make CAM pretty much useless. A major cause of that is the information requirement. People will describe and interpret findings differently, making reporting useless. Trying to directly measure things will stumble across very complex issues of calibration, etc. Actually, all the current hype goes to AI/ML/automation, but the real priority would be to improve data collection. Unfortunately, it's less sexy and much harder. Old-school docs still have good times before them.