AI Is Coming For Your Doctor

IM Doc has been giving regular reports from the front, particularly on the accelerating crapification of medical care in the US. You may recall his horrorshow experience with AI, as in having it forced on him for the transcription of patient notes and having the output not simply have many important errors but even complete fabrications.¹ He has a new report below about how his hospital plans to roll out sweeping changes implemented via the widely-used billing/records system Epic, that amount to turning diagnosis and care decisions over to AI.

This development is not just due to AI and total surveillance enthusiasm. A big reason for medical systems to go this route is the long-standing primary care physician shortage, and now, IM Doc indicates, is extending into other specialities.

IM Doc recounted the history of how this came about but I am not able to find the relevant e-mail. So I am pretty confident in the general trajectory but am a bit fuzzy on the timing and the institutional forces behind it. Anyone who can provide details and any needed corrections please pipe up.

Medical schools cut the size of their classes. I believe this started in the 1980s but it might have been as late as the 1990s. This may have been the result of an expected demographic decline in the US; experts were surprised that the 2000 census showed a population increase since a fall had been widely anticipated. The rise was the result of immigration and higher birth rates among Hispanics. But my understanding is the bigger driver was the expectation (or plan?) that rather than growing its own doctors, the US would increasingly rely on foreign-educated physicians.

The latter, as IM Doc explained, did not work out. Doctors who came to the US to practice in large measure went back home. Even with the inducement of attractive pay levels, they were put off by the time, stress, and risk involved in fighting with insurance companies. Word of their negative experiences got out quickly in medical communities in their home countries, so far fewer medical students there considered, let alone tried, practicing in the US as an option.

The part that is opaque to me is that when this foreign sourcing of doctors idea fell way short of expectations, why US medical schools did not start increasing their class sizes.

A big second contributor to the primary care and now specialist shortages is the corporatization of medicine. Outside New York City and concierge practices, most doctors seem to find it necessary to work with a big medical system. That means as an employee as opposed to a solo/small firm practitioner. That in turn means that they have lost control over how they practice medicine. The loss of autonomy and dealing with a money-driven as opposed to care-driven bureaucracy has resulted in quite a few doctors who are able to retiring early. ²

Now to IM Doc’s alert on changes planned with the many medical systems using Epic:

I was informed this week that in about 6-18 months our software will be updated. Oracle is apparently investing billions in AI as is Epic. It will take all incoming lab results for each patient and craft a pages long note with details about their labs and if they are normal or abnormal and what they need to do etc. It will be able to go through all of their entire chart – the arrays we have in the system as well as every single page of scanned pdf – and the national vaccine database – and put a paragraph at the very beginning of each communication discussing each and every deficiency in their health maintenance. In each terminal station, audio visual equipment will be set up and we will supposedly never touch a keyboard again. It will immediately alert us if there are billing issues on anything we are ordering. It will analyze each incoming voicemail and computer message from patients and craft a complete response with the current standard of care for the issues via up to date. If information is requested or required from any outside source – staring in 6-12 months Epic and Cerner and VA systems will all be interlinked and the system will go and gather all documents needed. If it cannot find it – it will automatically send Release of information to any doc in the USA.

What could possibly go wrong?

I have been repeatedly told this is all for two reasons. The absolute dearth of primary care MDs all across the nation. The entire “hospitalist” movement that has now come out as the clear and total victor has decimated the general internist cohort – almost no one straight out of training is going into outpatient medicine in internal medicine. This is on top of mass retirements and transition to concierge care causing a horrific deficit in outpatient internal medicine. (FYI – I am personally seeing 20-25 people every day with an NP seeing another 12-15. I see 2-3 new people a day. Most new patients are very complicated and require much attention. My next new patient scheduled visits are in FEB or MAR – and a routine follow up is in about 6 weeks. I routinely have people calling on the phone that my nurses can tell are deathly ill and we put them in as overbooks all the time. On this past Friday – I had an acute renal failure with Creat of 5, a new onset rapid afib, and a massive leg infection – all being overbooked. This was on Friday alone. 2 of these patients had been seen in urgent cares by NPs and had been completely misdiagnosed and mishandled and had sat at home for days with festering issues that could have killed them.) The family practice doctors were simply not trained to deal with very ill patients like I and my internist colleagues do every day.

Furthermore, the system is now being backfilled with NPs [nurse practitioners] and PAs [physicians’ assistants] and these AI systems are being set up in large part because of the extreme knowledge and practice and experience deficits in these types of providers. The AI systems are thought by the MBA crowd to be a failsafe and a very reasonable aid for these NPs and PAs who were never trained to be physicians. There is no other way to put it.

What was once a way of having NPs see overflow under the direct supervision of general internist MDs has now turned into a situation where they are all over this country all on their own. On their own except for the AI systems telling them what to do and how to answer patient questions.

The system is literally imploding because of this. Routinely every day I have 1-2 patient disasters in my office because of bad care that was given days or weeks ago by people who are simply not trained. The specialists are now overwhelmed with cases of all kinds of things that in the past had been handled by general internists. There are multiple sub-specialties that were already in dire shape numbers wise – but now are just absolutely torpedoed. This is most acute in endocrinology, rheumatology and neurology. But ENT and urology are close behind.

And they think this AI is going to help? And when will the day come that I can just sit in my office and read and let the computers do everything? – This AI and the absolute crushing case load is not what I signed up for in any way shape or form. This is not the medicine of decades ago when I started. And all the while, with all this crushing load, I am on the phone every day with insurance companies begging for care for my patients. All the while, patients screaming at me and my staff because we did not get back to them in a timely manner, etc. I must say – I was trained to be strong and resilient – but this is about to break me. I look forward to retiring soon.

It is not clear to me what might happen with this scheme, which depends on recording of patients, if they refuse. Some states that require all party consent for recording.³ I suppose the bureaucrats could say, “No recording, no treatment” but would that be an arguable breach of contract? And what about the unlikely event of unionized action by patients, say >25% of patients in thees states refusing? Unlikely but one can hope. Too bad the class action bar has, by design, become so hollowed out.

_______

.¹ From a September post:

IM Doc later provided a horrorshow example of the hash it makes of transcribing patient notes. In one case, it invented multiple serious illnesses the patient had never had and even a pharmacy that did not exist. Extracted from his message:

This is happening all the time with this technology. This example is rather stark but on almost 2/3 of the charts that are being processed, there are major errors, making stuff up, incorrect statements, etc. Unfortunately – as you can see it is wickedly able to render all this in correct “doctorese” – the code and syntax we all use and can instantly tell it was written by a truly trained MD.

This patient actually came into the office for an annual visit. There was nothing ground-shaking discussed….

This patient is on no meds that are not supplements. There are no prescriptions – and yet we supposedly discussed 90 day supplies from Brewer’s Pharmacy in Bainesville. There is no pharmacy nor town anywhere around here that even remotely sounds like either one. A quick google search revealed a Bainesville MD, far away from where we are – but as far as I can tell there is no Brewer’s Pharmacy there – the only one in the country I could find was in deep rural Alabama.

The last paragraph was literally the only part of this entire write up which was accurate…

This is what I do know however

1) Had I signed this and it went in his chart, if he ever applied to anything like life insurance – it would have been denied instantly. And they do not do seconds and excuses. When you are done, you are done. If you are on XXX and have YYY – you are getting no life insurance. THE END.

2) This is yet another “time saver” that is actually taking way more time for those of us who are conscientious. I spend all kinds of time digging through these looking for mistakes so as not to goon my patient and their future. However, I can guarantee you that as hard as I try – mistakes have gotten through. Furthermore, AI will very soon be used for insurance medical chart evaluation for actuarial purpose. Just think what will be generated.

3) These systems record the exact amount of time with the patients. I am hearing from various colleagues all over the place that this timing is being used to pressure docs to get them in and get them out even faster. That has not happened to me yet – but I am sure the bell will toll very soon.

4) When I started 35 years ago – my notes were done with me stepping out of the room and recording the visit in a hand held device run by duracells. It was then transcribed by secretary on paper with a Selectric. The actual hard copy tapes were completely magnetically scrubbed at the end of every day by the transcriptionist. Compare that energy usage to what this technology is obviously requiring. Furthermore, I have occasion to revisit old notes from that era all the time – I know instantly what happened on that patient visit in 1987. There is a paragraph or two and that is that. Compare to today – the note generated from the above will be 5-6 back and front pages literally full of gobbledy gook with important data scattered all over the place. Most of the time, I completely give up trying to use these newer documents for anything useful. And again just think about the actual energy used for this.

5) This recording is going somewhere and this has never been really explained to me. Who has access? How long do they have it? Is it being erased? Have the patients truly signed off on this?

6) This is the most concerning. I have no idea at all where the system got this entire story in her chart. Because of the fake “Frank Capra movie” style names in the document I have a very unsettled feeling this is from a movie, TV show, or novel. Is it possible that this AI is pulling things “it has heard” from these kinds of sources? I have asked. This is not the first time. The IT people cannot tell me this is not happening.

I have no idea why there is such a push to do this – this is insane. Why the leaders of my profession and our Congress are all behind this is a complete mystery.

² As this extract shows, what was intended was better fleecing of patients. But the bean-counters overlooked that doctors have agency and might not embrace the MBAs’ priorities. One example via a 2013 post, Coming Corporate Control of Medicine Will Throw Patients Under the Bus:

One of the most effective scare techniques employed to preserve our grotesquely inefficient, overpriced health care system has been to invoke the red peril of “socialized medicine”….

In fact, business freedom here increasingly means the God-given right to exploit the vulnerability of the public. The example slouching into view is more corporate control over the practice of medicine. And based on the previews, it will make the horrors falsely attributed to socialized medicine look pale….

I strongly encourage you to read this post from Whole Health Chicago (hat tip Lambert) in full. It shows how the future of American medicine is to fire the ones who are unhealthy. No, I am not making that up. The writer, Dr. David Edelberg, describes a recent presentation by a large insurance company. They’ve apparently been hosting similar sessions with physicians in the Chicago area in large medical practices. Here are the key bits (emphasis original):

The speaker at these evenings is always a physician employed by the insurance company. His/her title is medical director (I begin to think there must be dozens and dozens on their payroll) and he always begins by reassuring the audience that he was in clinical practice himself so he understands something of what physicians–especially primary care physicians–are facing. I view this physician more as a “Judas steer,” the animal that leads an innocent but doomed herd of cattle through the slaughterhouse corridors to the killing floor.

• The health industry hopes that individual medical practices and small medical groups will ultimately disappear from the landscape by being financially absorbed into larger groups owned by hospital systems.

And why do the powers that be regard this as desirable? Although the article does not stress this point, doctors have an established revenue stream. So the acquirers buy them out and impose discipline on those artistic, freewheeling doctors. The “practice style,” which used to mean the independence that doctors once enjoyed, is now an Orwellianism and includes hewing to corporate guidelines as to how to operate.

And here’s what to expect:

Physicians are expected to spend a limited amount of time with each patient, and are encouraged to see as many patients as possible during a workday. The insurance companies, sometimes with the token cooperation of a few physician-employees, create vast books of patient-care guidelines to which they believe their physicians must be “accountable” (remember this word, it will crop up again). These guidelines might mean documented Pap smear and mammogram frequency, weight management and exercise, colonoscopies for patients over 50, and getting that evil LDL (bad cholesterol) below 99 by any means possible…

If the chart audit system discovers that a physician, for whatever reason, is an “outlier”–that she’s either not following the guidelines exactly or not getting the results anticipated for her patient population—she’ll be financially penalized. A quick example of what might occur: if your LDL is 115, you may be on the receiving end of a statin sales pitch from your doctor, not because bringing it down to 99 will improve your longevity, but because your refusal to do so will impact her financial bottom line.

Now of course, you might say, “Well, in fairness, medicine is too much of a cottage industry. Look at how many doctors give unnecessary annual EKGs to patients in low risk groups. How else are we going to get to evidence-based medicine?” The problem is that what we as patients will get isn’t driven by best outcomes, it’s driven by profits. Edelberg explains:

…the subtext of “standardized” always includes the unspoken “spend less money on the patient.” Thus, a doctor might be financially penalized for recommending nutritional counseling to lower cholesterol (“counseling is expensive”) instead of writing a generic statin drug (cheap). Or recommending psychotherapy (“therapy is very expensive”) instead of generic Prozac (cheaper than M&M’s). Or referring patients for massage, acupuncture, or even chiropractic (“expensive, expensive, expensive!”) instead of pushing an over-the-counter antiinflammatory (free to the insurance company, as it’s OTC).

And I shudder to think what becomes of patients who don’t hew to standard templates: the person who had a high body mass but not due to dangerous abdominal fat (which is what creates the health risk) who is pushed to take the latest, greatest diet drug. What about people who don’t buy into the religion of getting your LDL down to below 100 (one reader argued that while it may lower your risk of heart disease, it increases your all-factor death risk by reducing your ability to fight MRSA)? Will they face penalties if they fail to comply?

No, you just will find it nearly impossible to get a doctor to take you:

• Let me close with a best-as-I-recall quote from an insurance company medical director. “We can no longer afford to pay for health care under the PPO model. Our plan is to phase out all fee-for-service care during the next few years. We’ll pay you doctors a finite amount of money to take care of a defined population. We tell doctors, ‘Don’t spend much money and you can keep the difference. Period. Don’t follow guidelines, and you’ll be leaving behind some serious money on the table and we’ll just take it back.’”

In case you think I overstated the implications, Edelberg recapped the discussion that ensued:

One physician piped up…. “But what about the non-compliant patients who won’t take the meds, don’t eat well, don’t have mammograms, continue to smoke? And what about super-health-conscious patients who want their vitamin levels measured and want referrals to acupuncturists?”

Another physician answered wearily for the medical director (who didn’t disagree): “You’ve got to fire patients like that. Get the non-compliant and the super-demanding out of your system. They’ll drag your numbers down. Hit your personal bottom line.”

Hey you, patient. Yes, I mean YOU. Pink slip time! Canned! Take your medical records and don’t let the frosted glass door hit you in the…on the way out.

³ All party consent states where even notification might not suffice. California is so big that large scale patient action might throw a spanner. But it is also the most tech-forward state :-(:

California
Florida
Hawaii
Illinois
Maryland
Massachusetts
New Hampshire
Oregon
Pennsylvania