AI in the Doctor's Office

[Ken]

They have had no issues to my knowledge. But they do review what AI does.

Doctors have to do the same like for line review if they are using human transcription before they sign off. So not much difference.

[Bootstrap]

So, can you point to me when adding more technology has reduced workload on doctors (and other medical staff) instead of just resulting in higher demands/expectations on them?

A friend is an office manager at a pediatric office and another friend is a doctor. They use AI to help them with their notes- instead of having to either handwrite, type or speak their notes after a visit the AI records the visit, writes the note and the doctor just has to review it. It has saved them a lot of time. The doctors all say that writing or even speaking notes after a visit just adds to their already long day so this does indeed save them time.
It is all voluntary - at least for the docs in the pediatric office and my friend.
My daughter in law- a speech therapist- uses it for her notes on her students to save time too. There are reputable AI services that help them write the notes and in deed does save time.

My doctor used to type notes as we talked. Now he sits in a chair and looks at me without a computer between us. He says he likes it that way, and so do I.

[QuietlyListening]

I have one doctor who manages to type as he talks to you yet you don't realize it and his attention is on you. He is a rare doctor- he treats the whole person and he is a specialist and reads his notes and notes others send him.

But AI has helped docs spend more attention on the patient instead of the computer - the new programs require certain questions be answered- at least many of them- and more attention is on the computer asking the questions even if they aren't typing plus a total note still needs to be written. And I speak from experience in what is required as a former home health nurse . Didn't have AI at the point in time and rather than be typing all the time I made notes and wrote an entire note later but that takes TIME. Know many docs did the same before AI. Not all use AI. But it has been a help.

[ohio jones]

They have had no issues to my knowledge. But they do review what AI does.

Doctors have to do the same like for line review if they are using human transcription before they sign off. So not much difference.

If only we could get MN posters to do the same.

[Ken]

I just asked my wife about this.

Before AI when she did full days of primary care clinics seeing 3 patients an hour in 20 min appointments she would dash off some preliminary notes in the couple minutes she had between patients and then would often be at the computer at home until 8 pm or so reviewing and finishing charts and clearing out her inbox (prescription renewals etc.). That is with the last patient seen at 4:40 pm. So leaving home at 6:15 am or so to get to the office early to review charts before the appointments that made for a 14 hour day or nearly 70 hours/week, often without time for lunch and no breaks.

Now with AI she has learned to dictate her thoughts out load during every exam and the AI recorder picks up everything she says and also every question and answer from the patient. But she has had to train herself to narrate all of her thoughts out loud to make for a productive AI generated chart. Now she only has to review all her charts at the end of the day before leaving the office, make any edits as necessary and is usually done before 6 pm. She is also a better doctor not because the AI comes up with medical answers, but because she has become more communicative during exams and so patients get a better sense of what is happening.

So it is as much the software training doctors how to narrate a good chart during an exam as anything else. They don’t normally rely on AI to generate diagnoses or prescribe treatments. They had computers doing that long before AI anyway. And most primary care isn’t like the TV series House where doctors would play Sherlock Holmes to uncover medical mysteries. 95% of the time they know exactly what is ailing the patient and it is simply a matter or figuring out the best treatment plan from a suite of possible options. And when they don’t know it is usually a process of running labs and then referrals to specialists.

[Szdfan]

I just asked my wife about this.

Before AI when she did full days of primary care clinics seeing 3 patients an hour in 20 min appointments she would dash off some preliminary notes in the couple minutes she had between patients and then would often be at the computer at home until 8 pm or so reviewing and finishing charts and clearing out her inbox (prescription renewals etc.). That is with the last patient seen at 4:40 pm. So leaving home at 6:15 am or so to get to the office early to review charts before the appointments that made for a 14 hour day or nearly 70 hours/week, often without time for lunch and no breaks.

Now with AI she has learned to dictate her thoughts out load during every exam and the AI recorder picks up everything she says and also every question and answer from the patient. But she has had to train herself to narrate all of her thoughts out loud to make for a productive AI generated chart. Now she only has to review all her charts at the end of the day before leaving the office, make any edits as necessary and is usually done before 6 pm. She is also a better doctor not because the AI comes up with medical answers, but because she has become more communicative during exams and so patients get a better sense of what is happening.

So it is as much the software training doctors how to narrate a good chart during an exam as anything else. They don’t normally rely on AI to generate diagnoses or prescribe treatments. They had computers doing that long before AI anyway. And most primary care isn’t like the TV series House where doctors would play Sherlock Holmes to uncover medical mysteries. 95% of the time they know exactly what is ailing the patient and it is simply a matter or figuring out the best treatment plan from a suite of possible options. And when they don’t know it is usually a process of running labs and then referrals to specialists.

I think this is the important point here. AI is being used as a support tool to streamline time-intensive note-taking processes. It doesn’t sound like it’s making diagnoses.

Is it possible that someday AI will be used to diagnose patients? Probably. But it doesn’t sound like from what people are saying here that we’re there yet.

[Thomas_muntzer]

More accurately, the people who employ doctors have decided to use AI as an excuse to lay off medical transcriptionists and other back-office staff and just increase the workload on doctors. We will, of course, not see a commensurate increase in pay for doctors for this - but more money will flow to venture capital-backed AI, along with greedy owners keeping more of the cash for themselves (and giving themselves executive pay raises). (A nearby hospital system pays its CEO $1.3 million; typical doctor pay is barely a tenth of that). The medical transcriptionists get a paltry $18 or so an hour, yet they're the ones targeted by AI for layoffs.

Why can't we have an AI replace CEOs?

AI is notoriously unreliable, but it will be easy to blame the AI for the inevitable medical errors that result from transcription errors due to the use of AI. "Sorry we amputated the wrong limb, but the AI mixed up 'left' and 'right'."

The dystopic side of AI more income inequality and more concentration of power in the hands of the elite

[JohnH]

I think this is the important point here. AI is being used as a support tool to streamline time-intensive note-taking processes. It doesn’t sound like it’s making diagnoses.

Is it possible that someday AI will be used to diagnose patients? Probably. But it doesn’t sound like from what people are saying here that we’re there yet.

Do you really think AI/LLM technology will be an exception to the rest of the industrial revolution and capitalism, which generally hasn't resulted in better conditions for labor, but certainly has resulted in more concentration of power and wealth in the hands of the capital class?

[Szdfan]

I think this is the important point here. AI is being used as a support tool to streamline time-intensive note-taking processes. It doesn’t sound like it’s making diagnoses.

Is it possible that someday AI will be used to diagnose patients? Probably. But it doesn’t sound like from what people are saying here that we’re there yet.

Do you really think AI/LLM technology will be an exception to the rest of the industrial revolution and capitalism, which generally hasn't resulted in better conditions for labor, but certainly has resulted in more concentration of power and wealth in the hands of the capital class?

You sound like a Communist. Welcome Comrade!

[Ken]

I think this is the important point here. AI is being used as a support tool to streamline time-intensive note-taking processes. It doesn’t sound like it’s making diagnoses.

Is it possible that someday AI will be used to diagnose patients? Probably. But it doesn’t sound like from what people are saying here that we’re there yet.

Do you really think AI/LLM technology will be an exception to the rest of the industrial revolution and capitalism, which generally hasn't resulted in better conditions for labor, but certainly has resulted in more concentration of power and wealth in the hands of the capital class?

Has the plight of workers improved or worsened in the past 150-200 years since the start of the industrial revolution?