Hacker News Reader

Comment by pedalpete | Hacker News Reader

Can you give an example of this?

Where I'm familiar with AI in healthcare is tools to help radiologists and cardiologists detect cancer early. This is a support AI, not the AI doing all of the work.

I'm also hearing about the potential for generative AI to provide a good bedside manner in "discussing" a diagnosis with a patient. Not diagnosing, but having more patience and time than a doctor has to communicate with the patient on known entities.

We use ML for measuring sleep state and finding opportunities to stimulate the brain to improve the efficiency of deep sleep.

Where are these "worst possible ways" that I'm unfamiliar with?

View on HN

The above examples you've provided sound pretty dreadful to be honest. I can't imagine how awful it would be to be told I've a serious disease and then be stuck talking about it to a chatbot of all things.

The clinicians themselves can be pretty dreadful. The uses being explored are to give coaching to them ahead of delivering news to help boost empathy and increase understanding. “It looks like you had a subdural hematoma cause a status epilepticus seizure” vs “that fall you took broke some blood vessels in your head which caused a longer seizure than we like to see”.

Doctors are using it inappropriately to diagnose and advise treatment on their phones without telling patients. It is being rolled out in all manner of automated classification, huge amounts of behind the scenes healthcare involves manual review of charts for classification purposes that is associated with various streams of funding. Low level employees are turning to general purpose AI to assist with this, leaking data, not understanding how unreliable these tools can be for this. Billing departments are "secretly" using it to "refine" coding. Pretty much all of the dumb irresponsible things you can think of.

> Billing departments are "secretly" using it to "refine" coding. What does this mean exactly?

When billing any insurance company, all procedures are assigned a "code" to standardize the payment/procedure performed across systems.

Some doctor patient visits may fall under multiple codes, so billers will optimize for the highest paying code.

Ingesting visit notes and having an LLM help determine what billing codes to use seems like a great use of AI in my opinion.

Billing codes are complicated and the rules behind them change frequently. These seem like a case where AI could present a wrong answer with authority—probably a close to correct answer, but one that could cost a patient or practice.

I'm not saying it should go without human intervention. It could do a great job at selecting the most likely codes and displaying them to the biller.

I work in an ED and I see doctors and nurses looking things up on Google and yt daily. Medications and doses are the most common thing I see. It’s not much different than how developers or mechanics diagnose a problem.

One instance I observed all the ED providers huddled around a screen watching a yt video for a pediatric patient, we are not a children’s hospital. They all watched then discussed the video before going back out to the pt.

I’m sure there should be better documentation available but so much of their jobs is judgement based already.

Medical specific trained LLMs are likely to be commonplace in near future for providers based on feedback I’ve gathered.

It's also used for visit notes generation and works really well. Doctors are really loving this one in my experience https://www.lyrebirdhealth.com/ It's also assistive of course - it basically generates a draft that you're supposed to edit.

Reply on news.ycombinator.com