My grandfather is old and doesn’t have a great memory, he asks every doctor if he can record their encounter and always sends me the recording after to review and help him understand. Several times it’s helped us clear confusion on what was said in a visit.
I think the only people scared of this are bad doctors who think they will be exposed doing something wrong or inadequate.
In India pretty much everyone records everyone without even asking. When I am counseling patients prior to procedures, a good number are recording the conversation, usually to go through instructions later, etc.
The last couple of appointments I’ve had in a doctors office they asked if I mind if they record - I always agree. Utilizing AI for their recorded notes has been an improvement in returning to actual conversations between caregiver & patient compared to recent years when the doctors sat at a keyboard looking back & forth between a screen & the patient while they frantically typed every possible detail. It was like having an appointment with a court reporter or if they had the budget (usually only specialists) they brought their court reporter into the room & you were having a private conversation with a third party sitting there. If I agree to my appointment being recorded then I should have full access to that recording as well. I agree with you that this would cut down on complaints about the doctors - not increase lawsuits as the AMA assumes. Case in point, my 88 yo aunt went to a convenient care for a rash that came on suddenly. She complained to me that the nurse did not tell her anything but merely handed her a script to fill for hydrocortisone. When she filled it the directions said it was to treat unicateria. She looked it up & found out that meant hives but maintains the nurse said nothing to her. So she’s telling me that she had an almost silent appointment with the nurse? I question my aunts recall about this & a recording that she could have access to would have taken care of it.
Balderdash. EPIC notes are bad enough as redundant as they are already.
I come from the golden age of medicine when the idea was to get to the answer for the patient expediously and to start treatment. Our brief notes were meant to help us formulate our plan. They were meant to help us help the patient and that was it. Now they’ve become an end in themselves. Who cares about the patient? They get in the way of the note. Video recording? How about the lazy kids come with their elderly parent? Oh, that’s too human. The machine is better. Eh?
OK we recently changed from human scribes to an AI recording/scribe. I will submit the unpopular opinion that I would not want the raw recording to be preserved or available. My human scribe (a 19-year-old college student working part time) was imperfect but helped me focus on the patient. I then went into their documentation and made extensive edits to more accurately reflect both what was discussed and my medical decision-making. I would NOT want the scribe's initial rough draft to be a part of the medical record since it was frankly inaccurate much of the time. I certainly wouldn't want it subpoenaed. I feel the same way about the AI scribe's efforts, which frequently mis-hears or misunderstands, and is also extensively edited by me after the encounter. Just as I wouldn't want my decision-making judged by my human scribe's first draft, I also don't want the same to be true with the AI scribe, which is competing with ambient noise, the kids my patient brought along, interruptions from my staff, and just garbled communication more generally.
My wife was recently asked to sign a consent to recording for the purpose of AI transcription but it included several questionable stipulations. It said the recordings can and will be deleted, will not be part of the patient's record, and the patient has no right to the recording or transcript. Even worse, they asked her to agree not to sue regarding anything related to the recording.
Also, we'd like you to place this noose around your neck for us.
Meanwhile, they also have a policy that forbids the patient from recording, and this is in a two-party consent state.
This is very wrong and is part of a person’s personal data. It should be available as a compressed transcript to ensure the official notes match the concerns raised in the appointment, as well.
I reviewed the AI notes for a visit with my Mayo internist in Scottsdale. I was horrified to see that my doctor’s comments about her own mother were attributed to me. She had not asked if she could record me. We had a conversation and I asked that she correct the notes. Be very careful what is said with your doctor. While AI can be a help and a time saver for the clinician, the recording might be garbage in garbage out.
This is one of those rare "if you have nothing to hide" moments that I agree with. If the data (recording) is already being generated, then both sides should have access to it. If it is not already being generated, then I'm OK with having a both parties must consent requirement.
It will generate some non-zero number of additional malpractice lawsuits. Some will be legit, some not. The legal and insurance industries will just have to deal and work through it. It will also generate some increase in better following through with directions by the patient. This in turn will result in some avoided health problems for patients and reduced insurance claims.
My suspicion is that it'll be a net positive. At some point you'll want some way to ensure that the recording hasn't been "adjusted" to hide/manufacture evidence, but there are already ways to do that which just require buy in and implementation (basically, this is just a cryptographically strong digital signature which is already available for text documents).
The real risk long term is when the data is exposed due to poor security practices or insecure software. The insurance company, my doctor, and I all need to know about health problems I have. Prospective employers, not so much (unless it is job relevant).
For my own reasons, I would love to have a copy of patient recordings/video after the visit to review and learn from.
I recall a particular case in my first year of independent practice that I had a family that did not want to do an x-ray on the first visit for a cough and fever. I thought the patient was sicker than the normal cough visit from history alone, but she looked great in clinic with normal oxygen saturations. I did an extensive pulmonary exam (stethoscope directly on the skin, watching her breathe, manually counting respirations for a full minute, etc.) When I called early the next day (on a Saturday) to follow up, her father (also a physician) screamed at me about what a terrible doctor I was, an idiot, useless, because they had ended up in the ED overnight with high fever and she was diagnosed with bilateral pneumonia. I hadn't heard ANYTHING abnormal on the exam, and she had skipped out of that clinic room. I believe I spent the whole morning in bed afterwards after a bit of crying because I felt so terrible that I had missed this. But what did I miss!? Had I encountered her before signs of pneumonia were there or did I truly miss something on the physical exam? Was there a rapid clinical change over the next 8 or so hours? Was there deterioration because she was recumbent while sleeping?
This is such a minor "bad outcome" but there are tons of cases I bet physicians would love to review cases where a diagnosis took along time or where there might have been clues from earlier visits.
Also, I continue to maintain that we should occasionally throw in cases to the M&M schedule where nothing bad happens but no one knows that until the end and watch physicians confidently claim the patient need a full work up at the first visit when they complained of mild abdominal pain because only the most dramatic cases end up in M&M.
In the software world you have quality assurance as a step that code (supposedly) goes through. This is supposed to find software bugs before they reach the end user. It is an accepted practice that you introduce a small number of intentional bugs into the code before QA sees it.
If QA finds 50% of your seeded bugs, then you can infer that they'll also find 50% (give or take) of the real bugs. You can then adjust things from there (accept the failure rate, do more testing and QA, whatever).
You are describing something very similar, just from a different perspective.
As a physician if I see someone elderly I will sometimes FaceTime with a family member.
I would not like being recorded every time I see a patient. It would minimize small talk with a patient and me getting to know them better. I would have to be very careful when giving medical advice to a patient. I might be misunderstood or I might not be 100% factually correct and it could potentially harm me in the future.
I consented to AI recording at my recent GP visit because 1) I like my doctor and thought it might ease his workload, and 2) we weren't discussing anything very deep
My grandfather is old and doesn’t have a great memory, he asks every doctor if he can record their encounter and always sends me the recording after to review and help him understand. Several times it’s helped us clear confusion on what was said in a visit.
I think the only people scared of this are bad doctors who think they will be exposed doing something wrong or inadequate.
In India pretty much everyone records everyone without even asking. When I am counseling patients prior to procedures, a good number are recording the conversation, usually to go through instructions later, etc.
The last couple of appointments I’ve had in a doctors office they asked if I mind if they record - I always agree. Utilizing AI for their recorded notes has been an improvement in returning to actual conversations between caregiver & patient compared to recent years when the doctors sat at a keyboard looking back & forth between a screen & the patient while they frantically typed every possible detail. It was like having an appointment with a court reporter or if they had the budget (usually only specialists) they brought their court reporter into the room & you were having a private conversation with a third party sitting there. If I agree to my appointment being recorded then I should have full access to that recording as well. I agree with you that this would cut down on complaints about the doctors - not increase lawsuits as the AMA assumes. Case in point, my 88 yo aunt went to a convenient care for a rash that came on suddenly. She complained to me that the nurse did not tell her anything but merely handed her a script to fill for hydrocortisone. When she filled it the directions said it was to treat unicateria. She looked it up & found out that meant hives but maintains the nurse said nothing to her. So she’s telling me that she had an almost silent appointment with the nurse? I question my aunts recall about this & a recording that she could have access to would have taken care of it.
Balderdash. EPIC notes are bad enough as redundant as they are already.
I come from the golden age of medicine when the idea was to get to the answer for the patient expediously and to start treatment. Our brief notes were meant to help us formulate our plan. They were meant to help us help the patient and that was it. Now they’ve become an end in themselves. Who cares about the patient? They get in the way of the note. Video recording? How about the lazy kids come with their elderly parent? Oh, that’s too human. The machine is better. Eh?
Yes our AI world makes 1984 look desirable.
OK we recently changed from human scribes to an AI recording/scribe. I will submit the unpopular opinion that I would not want the raw recording to be preserved or available. My human scribe (a 19-year-old college student working part time) was imperfect but helped me focus on the patient. I then went into their documentation and made extensive edits to more accurately reflect both what was discussed and my medical decision-making. I would NOT want the scribe's initial rough draft to be a part of the medical record since it was frankly inaccurate much of the time. I certainly wouldn't want it subpoenaed. I feel the same way about the AI scribe's efforts, which frequently mis-hears or misunderstands, and is also extensively edited by me after the encounter. Just as I wouldn't want my decision-making judged by my human scribe's first draft, I also don't want the same to be true with the AI scribe, which is competing with ambient noise, the kids my patient brought along, interruptions from my staff, and just garbled communication more generally.
My wife was recently asked to sign a consent to recording for the purpose of AI transcription but it included several questionable stipulations. It said the recordings can and will be deleted, will not be part of the patient's record, and the patient has no right to the recording or transcript. Even worse, they asked her to agree not to sue regarding anything related to the recording.
Also, we'd like you to place this noose around your neck for us.
Meanwhile, they also have a policy that forbids the patient from recording, and this is in a two-party consent state.
This is very wrong and is part of a person’s personal data. It should be available as a compressed transcript to ensure the official notes match the concerns raised in the appointment, as well.
I reviewed the AI notes for a visit with my Mayo internist in Scottsdale. I was horrified to see that my doctor’s comments about her own mother were attributed to me. She had not asked if she could record me. We had a conversation and I asked that she correct the notes. Be very careful what is said with your doctor. While AI can be a help and a time saver for the clinician, the recording might be garbage in garbage out.
This is one of those rare "if you have nothing to hide" moments that I agree with. If the data (recording) is already being generated, then both sides should have access to it. If it is not already being generated, then I'm OK with having a both parties must consent requirement.
It will generate some non-zero number of additional malpractice lawsuits. Some will be legit, some not. The legal and insurance industries will just have to deal and work through it. It will also generate some increase in better following through with directions by the patient. This in turn will result in some avoided health problems for patients and reduced insurance claims.
My suspicion is that it'll be a net positive. At some point you'll want some way to ensure that the recording hasn't been "adjusted" to hide/manufacture evidence, but there are already ways to do that which just require buy in and implementation (basically, this is just a cryptographically strong digital signature which is already available for text documents).
The real risk long term is when the data is exposed due to poor security practices or insecure software. The insurance company, my doctor, and I all need to know about health problems I have. Prospective employers, not so much (unless it is job relevant).
For my own reasons, I would love to have a copy of patient recordings/video after the visit to review and learn from.
I recall a particular case in my first year of independent practice that I had a family that did not want to do an x-ray on the first visit for a cough and fever. I thought the patient was sicker than the normal cough visit from history alone, but she looked great in clinic with normal oxygen saturations. I did an extensive pulmonary exam (stethoscope directly on the skin, watching her breathe, manually counting respirations for a full minute, etc.) When I called early the next day (on a Saturday) to follow up, her father (also a physician) screamed at me about what a terrible doctor I was, an idiot, useless, because they had ended up in the ED overnight with high fever and she was diagnosed with bilateral pneumonia. I hadn't heard ANYTHING abnormal on the exam, and she had skipped out of that clinic room. I believe I spent the whole morning in bed afterwards after a bit of crying because I felt so terrible that I had missed this. But what did I miss!? Had I encountered her before signs of pneumonia were there or did I truly miss something on the physical exam? Was there a rapid clinical change over the next 8 or so hours? Was there deterioration because she was recumbent while sleeping?
This is such a minor "bad outcome" but there are tons of cases I bet physicians would love to review cases where a diagnosis took along time or where there might have been clues from earlier visits.
Also, I continue to maintain that we should occasionally throw in cases to the M&M schedule where nothing bad happens but no one knows that until the end and watch physicians confidently claim the patient need a full work up at the first visit when they complained of mild abdominal pain because only the most dramatic cases end up in M&M.
In the software world you have quality assurance as a step that code (supposedly) goes through. This is supposed to find software bugs before they reach the end user. It is an accepted practice that you introduce a small number of intentional bugs into the code before QA sees it.
If QA finds 50% of your seeded bugs, then you can infer that they'll also find 50% (give or take) of the real bugs. You can then adjust things from there (accept the failure rate, do more testing and QA, whatever).
You are describing something very similar, just from a different perspective.
FYI: A company called ZoomRx is paying doctors to record sessions to hear how docs are discussing their therapies with patients. https://zoomrx.com/offerings/hcp-patient-conversations
I believe it can be a few hundred dollars per conversation for certain specialities.
As a physician if I see someone elderly I will sometimes FaceTime with a family member.
I would not like being recorded every time I see a patient. It would minimize small talk with a patient and me getting to know them better. I would have to be very careful when giving medical advice to a patient. I might be misunderstood or I might not be 100% factually correct and it could potentially harm me in the future.
I consented to AI recording at my recent GP visit because 1) I like my doctor and thought it might ease his workload, and 2) we weren't discussing anything very deep
I'm going to go with yes.