I don't understand your scenario where I wouldn't have to review a transcribed recording. The ED is loud and things may not be heard or the camera may not see everything. In addition, how would you prevent any audio picked up about other patients from being included in the wrong patient's recording? Never mind needing massive data warehouses to store these massive files. Spelling errors would be rampant. Things go on in my head that have to be put into the MDM. Most of my exam is done visually, so how is verbalizing every thing I am seeing efficient? I also discuss things with consultants and review records that need documentation. I then have to review any automated note for accuracy and add my thinking. I also need to document discharge instructions too which isn't going to be transcribed in real time. If the recordings are not transcribed, how will anyone be able to review it in the future without having to sift through hours of video? Finally, I will have to log in on my day off to review these transcriptions as it likely isn't going to be real time any time soon. I have used scribes and transcriptions and they both have typos and need edits. The note is not just for patient and physician communication, it is a legal record and billing document which needs to be reviewed before it is signed. Until the regulations are changed, patient privacy is not a concern, and malpractice is virtually eliminated, I don't see this being a successful reality despite a lot of startup buzz. This is why healthcare is so frustrating for startups as they don't often get the regulatory and infrastructure burdens of healthcare.
What really would speed up documentation is malpractice reform, getting rid of the various metrics I need to explain in my note, and changing documentation to only need a focused HPI, PMH which is reviewed and updated by the patient for our review rather than waiting for me/RN/tech to ask for the 1000th time in their visit, only pertinent physical exam findings are needed, and an MDM/plan. Re-documenting my EKG in MUSE and the EHR, checking the vital signs reviewed box, making sure I have 4 elements in each HPI, making sure every patient is screened for smoking even though their chief complaint is a stubbed toe, documenting all my MIPS criteria, hitting 2 findings per system in my exam, etc. all while making sure the note sounds good if ever read in court are why charting really sucks. These aren't things that can be solved by automation.
Not everyone is opposed to this idea because we are dinosaurs. I don't think anyone thinks that T-sheets are better than the EHR for patients or for us reviewing the patient's history, but they likely (paper was before my time) were easier and UI designers at the EHR companies should learn from that. As for recordings, we don't have the same protections other industries have as well as having different challenges. Our notes are not used just by us with our patients and include more than what happens in the room with the patient. These companies trying to make this a reality are so far off in the current system that I do not see it on the horizon in the next 10 years.