Dear (name),
You’re all familiar with the electronic health record (EHR)—either as a patient or a clinician. I and many others vehemently opposed the electronic health record when it was introduced despite the hoopla and promise its purveyors claimed. We worried whether it would benefit patients. Our concern was that it would be used mainly for billing purposes, disrupt the clinician-patient relationship, and create vastly increased administrative work for clinicians. But, with federal carrot and stick financial incentives, our institutions bought into it full sway. In fairness, there were a few benefits, for example, it was easier to access and view x-rays and electrocardiograms. And its potential to help patients and clinicians goes without question.
The reality, however, is that our worries were borne out. Doctors now focused on their computer rather than the patient in front of them. They then found that laboriously checking multiple boxes on numerous different screens turned out to require more time than allotted with the patient. Guess what? The clinicians now take their computers home at night so they can complete the largely-for-billing-purposes record. The nightly demand for 3-4 hours of computer work led to an unintended, unanticipated consequence: the EHR has become a major contributor to the clinician burnout you hear so much about. Notwithstanding many efforts to correct EHR problems, there has been little improvement.
Now comes artificial intelligence (AI) with the promise of some reprieve. Much as with the advent of the EHR, however, we have little data on the promised benefit or the effectiveness of implementing it. According to a recent article—The Perils of Artificial Intelligence in a Clinical Landscape—in JAMA Internal Medicine (2024; 184: 351-352) by Drs. Sakar and Bates. They warn that, lest we replicate the missteps of EHR implementation and compound the problem of patient care and clinician burnout, we must take a careful look at the following issues.
- The risk of further dehumanizing the clinician-patient relationship could occur if chatbots provide the first line of communication with patients. The authors give a chilling example of an AI chatbot receiving an electronic report of cancer from the pathology lab and then, without the clinician’s knowledge, relaying it to the (shocked) patient.
- If AI is used to record patient histories during doctor-patient interactions, it’s quite possible the volume of documentation generated may increase the doctor’s work. The latter will presumably need to read and affirm such AI documents.
- The AI is only as good as it is programmed. Critical information may be missed or misrepresented, especially when extracting it from large databases such as complicated patient hospital records or multiple records from different sources.
- A well-known shortfall of AI is its ability to fabricate information, for example, that a patient is taking insulin even though they are not diabetic.
The authors advise a different approach from that of EHR developers: “prioritizing patients and clinicians, rather than billing, profits, and the priorities of health care systems with established patterns of primary care neglect.”
How does this relate to our goal of fixing mental health care? Most care is provided in primary care settings by increasingly burdened doctors, nurse practitioners, and physician assistants. Let’s help these intrepid folks, not make their job more difficult!
I’m getting lots of great feedback on these newsletters—and being updated on what many of you are doing, how your lives are going. Thanks, I appreciate that, it helps restore the connection to so many friends I lost contact with when I retired. Keep them coming.
My book, Has Medicine Lost Its Mind?, is now with the editor at Prometheus Books (to be published in the Spring of 2025). My focus has thus shifted to generating widespread public interest in a revolutionary way to correct the mental health crisis.
Please help me spread the message by recruiting people to go to my website (Personal Website) and sign-up, either at the pop-up or at the bottom of the home page; they’ll then receive this newsletter as well as brief questionnaires for depression and anxiety and a list of shocking facts about mental health care. And they incur no obligation.
Thanks, I appreciate your help!
Take care.
Bob
Robert C. Smith, MD, MACP