AI in healthcare.

[GGizmo]

A lot of doctors rely on diagnosis from experts.

It has the potential to scan for many conditions at once.

There are thousands of conditions in the medical books. Based off my own experience and thinking, it wouldn't surprise me if a high number of conditions aren't diagnosised and the people are struggling along in life.

 

[Superficial]

This, it would be good if it could be blindly split and recorded some way to ensure the doctors dont fall back on complacency that the AI will do the work for them.

It would be but I can see doctors protesting once it's likely that, at least their colleagues, will see the results as the only way this could work in a standard practical setting is amongst staff in the same hospital. They won't want to have their individual or collective competence tested and reported to their seniors I don't think (though I think it has or will happen anyway so they'll have to suck it up). I'm not sure how that square will be circled yet unless it's being introduced as mandatory...affecting all at onces accross the NHS. It does have the potential to be scandalous if diagnoses rates and condition prevention improve massively.

According to the below article (sorry for paywall) AI this year is fantastic at all the other stuff too..

Google AI has better bedside manner than human doctors — and makes better diagnoses

Researchers say their artificial-intelligence system could help to democratize medicine.

www.nature.com

Google AI has better bedside manner than human doctors — and makes better diagnoses​

Obviously doctor competence has been exposed when testing AI (presumably by willing participants) but I'm referring to smaller groups of colleagues seeing the results amongst themselves....it could be quite embarasssing when named doctors making errors are reported to hospital managers and senior doctors.

They'd never let patients know that AI made our dignoses and our doctors screwed it up anyway - it'll all just be reported as the doctor making the call.

 

[arknor]

There are thousands of conditions in the medical books. Based off my own experience and thinking, it wouldn't surprise me if a high number of conditions aren't diagnosised and the people are struggling along in life.

Doctors rule stuff out based on how old you are as well.

I was having breathing difficulties 2 years ago and thought it was COPD from being a smoker from about 14> 30 and 34 > 40
So i go the doctors and say I think I have COPD and describe my symptoms.

He instantly says it won't be COPD because your not over 50.
He was right in this case and it was just GERDS (acid reflux disease). excess acid had basically being burning my airways/lungs/throat and tongue as I slept.

Google says COPD is more common in people after 40, but people aged 20-50 can still have COPD, yet it was instantly ruled out as a non possibility

 

[413x]

Doctors rule stuff out based on how old you are as well.

I was having breathing difficulties 2 years ago and thought it was COPD from being a smoker from about 14> 30 and 34 > 40
So i go the doctors and say I think I have COPD and describe my symptoms.

He instantly says it won't be COPD because your not over 50.
He was right in this case and it was just GERDS (acid reflux disease). excess acid had basically being burning my airways/lungs/throat and tongue as I slept.

Google says COPD is more common in people after 40, but people aged 20-50 can still have COPD, yet it was instantly ruled out as a non possibility

So much this.
Lots of cancers (particularly colon and brain) happening younger and younger. But under 40 you won't get taken seriously vet easily.

 

[Tefal]

This is where AI could be a game changer.
Medical diagnostics.
You could have an mri and the results instantly. No waiting for weeks.

It would Truely be a game changer. Only takes half an hour for most MRIs. You could have scans basically 24/7 and only needing staff at the point of the scan.

Faster it happens the better.


Can't believe (well I can) some people didn't want thier scans looked at by the AI.
Maybe after reading the results that it picked up cancers that radiologists missed they'll regret that decision.

I wonder how the helium supply will hold up?

It's already pretty ropey

 

[HangTime]

I went to a Data conference a while back (pre-covid, far enough back that they were typically called just "Data" conferences not "Data & AI" heh) where they presented around this topic, really impressive stuff. I think at the time they said human consultants had a 91% success rate of identifying conditions compared to their ML / Computer Vision at 97% or something like that.

I've felt for a while this is a really, really good use of AI / Computer Vision. It doesn't mean "oh noes handing over our healthcare to the computer overlords", it means "why not have computers do additional vetting and flag things up for humans to take a second look that maybe got overlooked first time?".