In recent times, healthcare in the United States has faced significant challenges, particularly with the integration of artificial intelligence (AI) into patient care. IM Doc has been sharing insights on these developments, notably his troubling experiences with AI-generated transcription errors in patient notes. His latest report highlights disturbing changes anticipated in his hospital’s operations due to the widely-used Epic billing and records system, which aims to shift diagnostic and care decisions to AI algorithms.
This push towards automation isn’t solely fueled by AI enthusiasm; a pressing factor is the longstanding shortage of primary care physicians, which now extends to other specialties. IM Doc underscores this alarming trend, detailing its origins while acknowledging some gaps in the narrative.
The reduction in medical school class sizes began, to his knowledge, in the 1980s, possibly as a response to anticipated demographic shifts in the U.S. population. Contrary to expectations, the 2000 census revealed an increase in population, primarily due to immigration and higher birth rates among Hispanic communities. The broader expectation was that the U.S. would increasingly depend on foreign-educated physicians to fill the void.
Unfortunately, this plan backfired. Many foreign-trained doctors who came to the U.S. for work eventually returned home, deterred by the complexities and pressures of navigating insurance systems, among other challenges. Their negative experiences have potentially dissuaded future medical students in their home countries from seeking opportunities in the U.S.
A notable factor contributing to the current physician shortage is the corporatization of healthcare. Outside of metropolitan areas and concierge practices, most doctors now find themselves employed by large medical systems, which diminishes their autonomy over medical practices. This shift towards bureaucratic management has pushed many physicians to consider early retirement.
Now, regarding the changes involving the Epic system, IM Doc has shared his concerns:
I learned this week that within the next 6-18 months, our software will undergo significant updates. Oracle and Epic are investing billions into AI. The new system will generate extensive notes based on incoming lab results for each patient, detailing their health status and necessary actions. It will analyze complete patient records, including scanned documents and the national vaccine database. Additionally, audio-visual equipment will be installed to eliminate the need for keyboard use. The system will notify us of any billing discrepancies while processing patient communications automatically. Eventually, Epic, Cerner, and VA systems will interconnect, enabling automatic document retrieval. If any information is missing, it will facilitate the release of information to any doctor in the U.S.
What could possibly go wrong?
It’s been suggested that this initiative addresses the extreme shortage of primary care physicians nationwide. The hospitalist model has drastically impacted the outpatient internal medicine field, creating severe deficits as doctors shift towards concierge care or retire. For context, I see 20-25 patients daily alongside support from a nurse practitioner. The influx of complex new patients is overwhelming, often with serious health issues that have been mismanaged prior to their arrival at my office.
Moreover, nurse practitioners (NPs) and physician assistants (PAs) are increasingly filling gaps in care, but many are not trained to handle the complex cases traditionally managed by internists. The reliance on AI systems serves as a stopgap for these providers, who often lack the foundational knowledge and experience expected of physicians.
This trend has led to constant patient crises and elevated workloads. Specialists now face a surge of cases previously handled by general internists, causing a backlog in specialties, particularly in fields like endocrinology, rheumatology, and neurology.
The expectation that AI will alleviate these issues seems misplaced. My workload has become unmanageable, and the demands of insurance companies only add to the stress. I find myself longing for retirement as I grapple with these challenges.
One pressing concern is how this system will address patients’ rights regarding recording interactions. In states that require consent from all parties, will patients face treatment denial if they refuse recording? Such a policy could raise ethical and legal questions about consent and privacy, especially if patients unite against it.
_______
In a previous post, IM Doc shared a stark example of the fallibility of AI in patient recordkeeping, illustrating how inaccuracies could lead to catastrophic outcomes for patients. The technology can create fictitious medical histories and medication records that could jeopardize a patient’s chances for insurance coverage or appropriate care.
This technology frequently produces serious errors—almost two-thirds of processed charts contain inaccuracies, often misrepresenting patient issues while maintaining a façade of professionalism in medical language.
For example, one patient visit mentioned medications that were never discussed and pharmacies that don’t exist nearby. Had I not scrutinized the notes, these inaccuracies could have led to serious repercussions for the patient.
The push for a more efficient system has only resulted in increased stress for conscientious providers. Many find themselves spending more time correcting AI-generated misinformation than providing care. Notably, these AI systems track time spent with patients, potentially creating pressure for quicker consultations.
The question remains as to why healthcare leaders are so committed to this direction. There seems to be a disconnect between the benefits projected by corporate entities and the realities faced by medical professionals and patients alike.
The future of American medicine is at a crossroads, where profit-driven decisions are overshadowing the ethical practice of healthcare. If the current trajectory continues, both patients and doctors may find themselves in a precarious situation, leading to a breakdown of quality care.
In conclusion, the integration of AI in healthcare raises numerous concerns about quality, accuracy, and patient rights. As pressures mount and providers adapt to new systems, it’s crucial for stakeholders to prioritize the principles of patient care, rather than allow corporate interests to dictate the future of medicine.
