1. Podcasts
  2. Fixing Healthcare Podcast
  3. FHC #234: The risks of moving too fast with AI & the dangers of standing still
Fixing Healthcare Podcast

11 Nov 2026 · 38m · 13 chapters

Episode artwork

FHC #234: The risks of moving too fast with AI & the dangers of standing still

In short

The episode argues that while frontier generative AI poses real safety risks, healthcare should not ignore near-term benefits of AI agents for continuous monitoring and earlier intervention, because “standing still” is also dangerous. It also links healthcare unaffordability to national debt and borrowing costs, claiming the system needs strategic pivots (capitation, site-neutral payments, drug price negotiation) rather than incremental tweaks.

Guests

Dr. Robert Pearl, former CEO of the Permanente Group (18 years), Forbes contributor, Stanford professor, and author of Mistreated, Uncaring, and ChatGPTMD. Host: Jeremy Corr (Fixing Healthcare; Pop Union Books and Medicine; CEO of Executive Podcast Solutions).

Key claims

AI can augment clinicians via Bluetooth/video monitoring; it won’t replace doctors; regulatory and collaboration barriers slow adoption; affordability pressures will force business-model change.

Notable examples

postoperative wound infection detection; home BP/glucose monitoring; chronic heart failure daily monitoring (weight, swelling, endurance, lying-flat); AI-enabled wound photo triage; Utah pilots for AI medication renewals and acne photos; Kaiser Permanente outcomes (better mortality and hypertension control before ChatGPT).

Written by AI. May contain mistakes.

Chapters

0:46

Risks of Generative AI

Discussion on the growing concerns about the risks of generative AI in healthcare.

“broad range of healthcare topics, you can visit his website, robertperlemd.com.”
2:59

Opportunities in AI for Patient Care

Exploration of how generative AI can improve patient monitoring and outcomes.

“Harnessing older and less powerful degenerative AI applications, I think they could improve patient care and save hundreds of thousands of lives.”
6:15

Implementing AI in Chronic Disease Management

Insights on the implementation of AI tools for managing chronic diseases like diabetes and hypertension.

“When you were the CEO in Kaiser Permanente, how did you approach this opportunity?”
8:34

Impact of AI on Hospitalizations

Discussion on how continuous monitoring through AI can reduce hospitalizations for heart failure.

“For us, in Kaiser Permanente, keeping patients healthy, that was financially positive, since almost all of our members recovered under their pay-for-value, a capitated arrangement.”
12:39

Surgical Recovery and AI Monitoring

How AI can assist in monitoring surgical recovery to prevent complications.

“Agents could make that kind of continuous monitoring easier and more affordable for all Americans.”
14:00

The Role of AI in Healthcare

Explore the applications and barriers of generative AI in medical care.

“analyzing the progression for signs of improvement or deterioration, and asking about pain and drainage.”
18:28

Healthcare Affordability Crisis

Discuss the economic implications of healthcare unaffordability in the U.S.

“Robbie, I'd like to go back to the point you made about why there has been so little action relative to applying these generative AI tools for patients as a means to improve health and lower total medical costs.”
24:34

Responses to Financial Pressures in Healthcare

Analyze the strategies healthcare organizations are using to cope with rising costs.

“How have healthcare organizations and elected officials responded so far?”
28:01

Hospital Revenue and Physician Compensation

Learn how hospitals are adjusting physician compensation amidst scrutiny and cost pressures.

“have increased prices by around 15 percent without improvement in medical care or clinical outcomes.”
29:18

Hospital Spending and Congressional Scrutiny

Understand the financial pressures hospitals face from government programs and self-funded businesses.

“How do you see hospitals responding to this downward pressure on medical costs?”
Show all 13 chapters
31:05

Outpatient Services as Revenue Generators

Explore how outpatient services are becoming crucial for hospital revenue and cost management.

“Under this program, eligible hospitals and clinics purchase outpatient medications at discounted prices, and then they sell them at much higher prices.”
33:12

Pricing Pressures on Drug Manufacturers

Examine the challenges drug manufacturers face regarding pricing and negotiations.

“Robbie, last but not least, how about the drug manufacturers?”
35:27

The Strategic Inflection Point in Healthcare

Discuss the potential strategic shifts necessary for the healthcare industry amidst rising pressures.

“And without a doubt, all the current healthcare incumbents are likely to face pressure in the future to lower costs while improving clinical outcomes.”
A man walking through an airport, asking his phone a question

Ask about this episode

Get the gist, explore an idea or ask about something said.

Claude

On Fixing Healthcare Podcast, what are the key points of this episode?

Searching your listening

Transcript

Automatic transcript. May contain errors.

0:11Dr. Robert Pearl: Hello and welcome to Diving Deep, part of the Fixing Healthcare podcast series. I'm your host, Jeremy Corr, also host of the Pop Union Books and Medicine podcast and CEO of Executive Podcast Solutions. With me is Dr. Robert Pearl. For 18 years, Robert was the CEO of the Permanente Group, the nation's largest physician group. He is currently a Forbes contributor, a professor at both the Stanford University School of Medicine and Business, and author of the best-selling books, Mistreated, Why We Think We're Getting Good Healthcare and Why We're Usually Wrong, and Uncaring, How the Culture of Medicine Kills Doctors and Patients.

0:39Dr. Robert Pearl: His new book, ChatGPTMD, How AI-Empowered Patients and Doctors Can Take Back Control of American Medicine, remains a bestseller. If you want information on a broad range of healthcare topics, you can visit his website, robertperlemd.com. Robbie, listeners are increasingly concerned about the dire warnings that companies and computer scientists are voicing about the risks of generative AI. They're wanting to know your thoughts, given your deep knowledge on the technology.

1:02Jeremy Corr: Jeremy, listeners are correct that the warnings about artificial intelligence do keep getting louder. And in fact, we've discussed in previous episodes, even the companies building the world's most powerful AI systems are reporting troubling behavior from their own creations. We recently talked about the hugging face break-in when hundreds of generative AI agents escaped the laboratory in which they were being tested for cybersecurity risks, broke into the Internet and hacked into the homepage of the Hugging Face Company, a major technology business. And if that's not problematic enough, some of the researchers closest to the development of the frontier AI tools and applications warn that powerful new systems could cause human extinction.

1:48Jeremy Corr: What we've seen are scientists from companies like Anthropic and OpenAI expressing fear about their own organization's ability to control this new generation, this next generation of frontier models. And elected officials are becoming more concerned as well. Government leaders in the United States and abroad are engaged in intense debate about how to keep increasingly capable models from causing harm. In fact, Inanthropic's recently released S1 IPO filing, 97 pages. We're talking 30 % of the document is on risks, although not all of them are specific to AI safety and societal danger. Some of the topics disclosed included ones on loss of human control, cyber attacks, autonomous behavior, weaponization, misinformation, and the growing existential threats to humanity.

2:41Jeremy Corr: I believe that these and other concerns have to be taken seriously, and I in no way downplayed them. But as a physician, I also worry that fear of AI is taken sometimes too far, and that doing so will cause our nation to overlook extraordinary opportunities. Harnessing older and less powerful degenerative AI applications, I think they could improve patient care and save hundreds of thousands of lives. We've seen across time that inventions that can harm people, they're often the same ones that can save lives. Radiation is a clear example. Get exposed to too much radiation, it can take your life.

3:20Jeremy Corr: But if you have a cancer, radiation therapy can kill the tumor and save your life and eliminate the need for deforming surgery.

3:29Dr. Robert Pearl: Robbie, this makes sense. Could you provide listeners with a sense of how a generative AI tool could make a positive difference for patients?

3:35Jeremy Corr: Jeremy, let me ask you to consider a question that I often pose to medical professionals. I ask them if they had the time to check on patients every day. Would they be able to detect complications earlier and improve clinical outcomes? Almost always the answer is yes. I mean, think about a patient who has surgery. The physician tells the family to call the doctor's office if they suspect an infection. The handout they are given says that they should look to see if the wound is red, draining, or painful. But for the patient and the family, identifying a problem at its initial stage is challenging.

4:14Jeremy Corr: Most wounds are red due to inflammation. It's just part of the natural, normal healing process. The majority of wounds have some drainage, and nearly all are uncomfortable and even sometimes painful. The line between healing and infection, that can be very hard to identify in the first hours or even the first day. But for a postoperative patient, recognizing an infection at the first signs, often a couple of days before it becomes clear as to what's happening, that could be the difference between solving the problem with antibiotics versus having to return to the operating room to drain an abscess.

4:53For patients

4:54Jeremy Corr: with diabetes or hypertension, the daily variation in readings, that makes it very challenging for patients to identify when their chronic condition is improving versus not responding to the medication that their doctor has prescribed. High blood pressure, it's responsible for 40 % of strokes. When a physician makes the diagnosis, the individual usually is started on a drug. Using an inexpensive electronic blood pressure cuff, the patient could obtain, let's say, three readings a day. By the end of the month, they would have 100 data points. But what does it all mean? Few patients have the expertise to assess the progress, and no doctor wants to receive 100 blood pressure readings from hundreds of patients every month.

5:37Jeremy Corr: But a generative AI application, one connected through Bluetooth to an electronic blood pressure cuff, the application could. And the same for blood glucose reading in patients with diabetes. And diabetes is the number one cause of kidney failure, 80 % of lower limb amputations, and a major contributor to heart disease. Rather than waiting months to have these medication doses changed or a new prescription provided, doctors could now make the changes quickly, allowing these chronic conditions to become controlled in a matter of a few months rather than a year or longer.

6:15Dr. Robert Pearl: When you were the CEO in Kaiser Permanente, how did you approach this opportunity?

6:20Jeremy Corr: I saw the value of more frequent outreach and evaluation, as you say, when I served as CEO and was responsible for the health care of 5 million Kaiser Permanente members. We put in place a variety of programs to monitor the status of individuals with chronic disease on a monthly, if not weekly basis. And the frequency would depend upon the patient's level of control. As a result, over a 15-year time period, heart disease and stroke mortality amongst Kaiser Permanente members declined twice as fast as it did nationally. It wasn't that patients in the community didn't have excellent doctors or that these physicians didn't know what to do.

6:58Jeremy Corr: It's just difficult to obtain effective control of a problem that exists every day of your life when the physician only sees the person every three or four months. As a result of our success, the National Committee on Quality Assurance recognized our organization for reducing the mortality of heart attacks among our members, 30 % lower than for individuals in the communities around us. Overall, our medical group, TPMG, controlled hypertension in 85 % of patients, and that's compared to only 54 % nationally. But these positive results were accomplished prior to the availability of generative AI tools.

7:37Jeremy Corr: As listeners know, the first version of ChatTBT was only released by OpenAI on November the 30th of 2022. As such, rather than using technology in the home, we had to accomplish improved management of patients with chronic conditions by having hundreds of clinicians check on patients every week or so over the phone. During these virtual visits, the clinicians would assess how's the patient doing and could make the necessary medication changes or notify the patient's physician at that time. In retrospect, that approach produced remarkable outcomes, but the level of clinical control, it wasn't as excellent as it would be possibly now because now we could do continuous monitoring through a generative AI application.

8:24Jeremy Corr: And of course, the expense of the vast number of clinicians was substantial and was beyond what most doctors and health systems in the United States could spend today to make available to their patients. For us, in Kaiser Permanente, keeping patients healthy, that was financially positive, since almost all of our members recovered under their pay-for-value, a capitated arrangement. With the rapid advances in generative AI, doctors and patients now have the opportunity to do even better.

8:53Dr. Robert Pearl: Robbie, can you provide more details about how the process would work?

8:58Jeremy Corr: Jeremy, as I just mentioned, most medical groups cannot afford to assign hundreds of clinicians to provide that level of attention consistently. And even if it were funded, there are not enough doctors or other clinicians in the United States to do it nationwide. But the availability of ubiquitous, low-cost generative AI agents, that would change the calculus. Those applications could be made available on home computers or smartphones. And using Bluetooth technology, they could monitor data, respond to health changes, and take predetermined steps without direct clinician oversight on a day-to-day basis.

9:33Jeremy Corr: Of course, when a problem is identified, the doctor would immediately be brought into the loop to confirm the findings and make the recommended medication adjustments. Consider, for example, how AI might help one of the 6.7 million Americans living with chronic heart failure. If you asked cardiologists whether they could improve clinical outcomes if they were able to assess all their patients with this problem on a daily basis, they would universally say yes. They'd be able to more accurately identify when patients with chronic heart disease were deteriorating and progressing to an acute crisis, one that would require admission to the ICU and put the patient's life at significantly greater risk.

10:14Jeremy Corr: This would allow physicians to intervene sooner, assuming there was continuous monitoring, and they could make minor adjustments in medications or diet in order to avoid the emergency that would happen a day or two later. A cardiologist monitoring a patient in their home would look for increases in weight, ankle swelling, and fluid in the lungs. Combine those signs with reduced endurance and difficulty lying flat comfortably would indicate worsening heart failure. These assessments are neither especially complex nor beyond what AI could do today. Bluetooth can connect agents to electronic scales and stethoscopes.

10:51Jeremy Corr: Video monitoring can allow agents to evaluate ankle swelling, measure stair climbing ability to assess whether the patient can lie flat as comfortably today as they could yesterday. And since the application would have all the information and the visual findings from the day before, the assessment would be objective and clinically helpful.

11:13Dr. Robert Pearl: Does this mean that generative AI would replace doctors?

11:16Jeremy Corr: No. In nearly all applications, the generative AI augments rather than replacing clinicians. The idea that big cost reduction with generative AI in medicine would be by replacing people with technology, that's just inaccurate. Unlike in most other industries, in healthcare, the cost reduction for each of these examples doesn't come from eliminating people. It happens by improving patient health and avoiding costly, life-threatening medical problems, which replicates what doctors would do if they had a time which they don't. In each case, these agents would augment what the physician already does.

11:55Jeremy Corr: It wouldn't replace it. If a clinician would see a patient every three months, the doctor would continue to do so. But in the other 90 days, the ones between the office appointments, the genitive AI agent would assess the situation. The question isn't whether these technological applications are as good as a physician. It's whether they're better than nothing, which is today's reality. With hypertension controlled less than half of the time, and diabetes even less often, American medicine has a huge opportunity to improve clinical outcomes. And the potential benefits are significant. Of course, dozens of randomized trials, remote monitoring of patients with heart failure, has reduced hospitalizations by roughly 20%.

12:39Jeremy Corr: Agents could make that kind of continuous monitoring easier and more affordable for all Americans. The same logic applies to chronic diseases, such as hypertension and diabetes. Nearly 120 million American adults have high blood pressure and 38 million live with diabetes. Both illnesses can damage the heart, the kidneys, and blood vessels when poorly controlled. considering the upside for patients with hypertension. The CDC estimates that better control of high blood pressure could prevent 92 ,000 heart attacks, 139 ,000 strokes, and 115 ,000 cardiovascular deaths over a five-year time period.

13:20Jeremy Corr: And agents could support that care by making monitoring and evaluation easier to accomplish. Another opportunity that I mentioned earlier comes from surgery. As I pointed out, once home, patients are instructed to watch for signs of infection at the incision site, redness, drainage, and pain. But because these changes are part of the healing process under usual circumstances and following most surgeries, patients struggle to recognize when they signal a complication. If surgeons can see the patient daily, they compare the wound's appearance with how it looked 24 hours before. AI agents could perform a similar evaluation by collecting photos of video from the patient's phone, analyzing the progression for signs of improvement or deterioration, and asking about pain and drainage.

14:08Jeremy Corr: If the application identified a likely problem, it would send the images and findings to a clinician. The individual would quickly determine whether the patient needed to be seen. Patients with concerns could, of course, still contact the surgeon directly. And Jeremy, surgical site infections, they're not trivial. They add about$3.3 billion to U.S. healthcare costs every year. Earlier recognition would bring patients to medical attention sooner, when antibiotics often prove sufficient rather than requiring a surgical intervention. The application of generative AI to the direct delivery of medical care is slowly becoming more common.

14:45In Utah, a state-authorized pilot allows a generative AI application to process

14:51Jeremy Corr: routine renewals of previously prescribed medications. The state has also authorized a pilot in which the Genitive AI tool analyzes patient-submitted facial photos and recommends topical treatments for mild to moderate acne. In a recent randomized feasibility trial, cardiac surgery patients use an AI-enabled platform to submit wound images from home, allowing the system to prioritize potentially concerning changes for clinician review.

15:21Dr. Robert Pearl: Given that embracing these opportunities would improve quality, save lives at lower cost, and that Bluetooth video and conversational AI already exist, why aren't they currently being used by doctors?

15:33Jeremy Corr: Jeremy, the biggest barrier is the complexity and expense of obtaining regulatory approval for medical tools, ones that are used by patients. Rather than being treated like a new type of stethoscope, they're approached by the FDA as one might a new medication. And the cost and time required to gain that approval, it's massive. But Medicare is now offering a place to begin. Its new access model supports technology-enabled chronic care and rewards providers for improving clinical outcomes rather than simply delivering more services. Working with the FDA and NIH, access could provide a structured setting to test these agents, determine where they improve medical outcomes, and create a quicker and more rapid path toward approval.

16:16Jeremy Corr: And a second barrier is collaboration. Safely developing these agents for patient use at a minimal cost now will require clinicians, medical societies, and health systems to work with AI companies such as OpenAI, Anthropic, and Google. Empowered independent researchers will need to test reliability and ensure that false alarms do not overwhelm clinicians. How would you put all these pieces together? The existential risks posed by increasingly powerful AI systems, they demand serious safeguards. There's no question that these frontier models are becoming incredibly powerful and appear to be insufficiently regulated.

16:56Jeremy Corr: But before assuming that it's risky to use generative AI to help patients improve their health and avoid medical complications, let me offer two thoughts. The first is the generative AI applications, the ones currently available, are different from a risk perspective than the frontier models and what they're capable of doing. And even if we were to use some of the most advanced generative AI agents, we still need to ask whether the opportunities to apply them to save lives and make care affordable, whether this would justify and offset the theoretical existential threats. And my view is that the broader application than for American medicine, opportunities to save hundreds of thousands of lives annually would be worth it.

17:36Jeremy Corr: Obviously, there's a balance, and others may disagree, but before people assume that the currently available generative AI large language models can't be trusted and must be in the direct physician oversight, they should remember that hundreds of thousands of Americans die annually for misdiagnoses and poorly controlled chronic diseases, and that these problems could be solved, but currently aren't using a generative AI tool. I believe as a nation, we shouldn't treat every use of generative AI as equally dangerous, and we need to recognize that when we choose to apply them to control chronic disease, diagnose a postoperative infection, or prevent acute heart failure, that we won't be increasing the existential dangers that they pose.

18:18Jeremy Corr: That risk, it's independent of what medicine does. If I had to summarize my conclusion in one sentence, it would be that the risks of moving too fast are real, but so are the dangers of standing still.

18:33Dr. Robert Pearl: Robbie, I'd like to go back to the point you made about why there has been so little action relative to applying these generative AI tools for patients as a means to improve health and lower total medical costs. How big is the problem of growing health care unaffordability for our nation?

18:49Jeremy Corr: Jeremy, I see health care affordability as a problem that poses massive economic threats, not only to the physical and psychological health of our nation, but also its financial health. And the most recent midterm election demonstrates that unaffordability of medical care is a huge issue across the country. But don't take it from me. Instead, let's consider what the people who analyze the economics of the country and its future direction are saying. Jeremy, as you know, I teach in the business school, and there's a saying on Wall Street, that bond traders are the smartest people in the room. Unlike stock investors, who can be swayed by promises and hype, bond investors, they focus primarily on objective data, and they look at future financial risks.

19:39Jeremy Corr: When inflation concerns rise alongside mounting government debt, they respond by demanding higher long-term interest rates in return for lending their money. And last month, we saw the yield in the 30-year U.S. Treasury bond reach an intraday high of 5.73%. This is the highest level in more than two decades. Inflation concerns and the nation's growing debt were the two major factors exerting upward pressure on borrowing costs. Robert, can you provide listeners with some data to demonstrate this point? Jeremy, I'm happy to do so. The United States recently surpassed$40 trillion in national debt.

20:20Jeremy Corr: That's more than twice the total from less than a decade ago. The Congressional Budget Office projected a$1.9 trillion federal deficit for 2026, and estimated that the debt held by the public would rise from 101 % of gross domestic product in 2026 to 120 % for 2036. Annual net interest payments are expected to more than double, from the current approximately$1 trillion to a projected$2.1 trillion. Together, these trends will increase pressure on elected officials to limit federal health care spending. In my strategy course at the Stanford Graduate School of Business, I encourage students to analyze trends, including the direction of the trend and the rate of change.

21:08Jeremy Corr: They need to determine the probability that the trend will continue and act before the most damaging outcomes become unavoidable. I would offer the same recommendations for listeners and elected officials as they think about their finances and the economic health of the nation. Recognizing the changing financial environment, making appropriate strategic decisions, and acting decisively, these are what separate success from failure.

21:36Dr. Robert Pearl: If our debt has been rising rapidly for a decade, and nothing much has been done to flatten the slope of the line, why is now different?

21:44Jeremy Corr: As you note, healthcare affordability has been described as a crisis for more than a decade. The bond market's warnings about mounting financial risk, they add urgency to addressing it. Nowhere is the pressure more visible. Nowhere is the data more undeniable than the federal government. CBO projects federal spending on Medicare, Medicaid, and other major healthcare programs will total an estimated$1.9 trillion this year. Medicare alone will cost the country about a trillion dollars. If we exclude the legally mandated payments, ones that Congress has authorized for Social Security and the interest payments to the national debt, Medicare and other federal programs are expected to account for nearly half.

22:32Jeremy Corr: We're talking 45 % of the remaining federal budget. Reduce federal spending without addressing health care payments. It just won't be possible. We'd like it to be. It's just not possible. Talk to clinicians, and many will tell you that physicians deserve higher payments for the work they do. I'll talk about the decade of training required and the stress that the profession produces. And although from a medical perspective, they're absolutely right. Through a strategic lens, if the odds of the government reducing doctor pay rather than augmenting it are high, then assuming that it won't happen, that's a major strategic error.

23:10Jeremy Corr: And strategic errors always produce negative consequences. But this pressure on the government isn't the only pressure driving up the bond rates and causing families to experience greater unaffordability in the health care they receive. In parallel, businesses which provide health coverage to approximately half of Americans, they face similar pressures as the federal government. Year after year, insurers have raised premiums, forcing employers and employees to pay more. Employer health care costs have risen more than 7 % annually for the past three years, with the benefits consultant Aon projecting a 9.5 % increase for 2027.

23:52Jeremy Corr: According to the Kaiser Family Foundation, 62 % of insured adults worry about affording their deductibles. The consequences extend beyond the household. A second Kaiser Family Foundation survey found that 36 % of adults receiving treatment for three or more conditions reported skipping or delaying care because of cost. Among adults with multiple conditions who delayed or skipped care for any reason, 27 % said that their health worsened as a result.

24:28Dr. Robert Pearl: Robbie, when it comes to the problems people are experiencing, families clearly are reaching that financial cliff. How have healthcare organizations and elected officials responded so far?

24:38Jeremy Corr: Jeremy, for decades, healthcare organizations have responded to financial pressures through small tweaks and incremental changes. They've relied on higher premiums, narrower networks, reduced benefits, and tougher reimbursement negotiations. Unfortunately, rather than succeeding, these measures have failed to resolve the affordability crisis. Through a strategic lens of the U.S. healthcare system has remained remarkably stable. Insurers continue to restrict utilization. Hospitals continue to consolidate to strengthen negotiating leverage. Drug manufacturers continue to rely on patent protection to maintain high prices, and physicians continue to join hospitals and health systems to improve their bargaining position.

25:24Jeremy Corr: So far, these traditional approaches, they've been sufficient to ward off transformational change in the healthcare industry. But the combination of increasing government debt, higher borrowing costs, and financial pressures on employers make it less likely that these types of approaches will be adequate in the future. Four major healthcare players, insurers, physicians, hospitals, and pharmaceutical manufacturers. Now each must decide to maintain its current business model or pursue a fundamentally different strategy.

26:00Dr. Robert Pearl: Kirby, that's a list of who's who in healthcare. Let's take them one at a time, starting with the insurers, both what they've done and what you believe they could do to break the current vicious cycle.

26:10Jeremy Corr: Jeremy, insurers have tried to ramp down costs by avoiding expensive diagnostic evaluations, treatments, and medications. Their primary tools have been prior authorization and claims adjudication. But they've done little to improve the efficacy and efficiency of medical care. The result is that costs have risen to a level almost double that of other nations. We're talking on a per capita basis. But U.S. outcomes lag on nearly all key measures from life expectancy to maternal deaths and childhood mortality. With the midterms now over, the Democrats having wide margins in Congress, they're likely to push for lower prices and fewer restrictions on the care that doctors provide.

26:53Jeremy Corr: One option to avoid heavy-handed regulation would be to pivot toward long-term population-based payment arrangements with providers sharing the financial risk. With decapitation, a physician group or health organization, it receives a single predetermined payment to manage the care for a defined population of patients rather than billing separately for every service provided. Rather than trying to restrict access, insurers would negotiate multi-year agreements that reward providers when they improve clinical outcomes, when they reduce heart attacks, strokes, and kidney failures, or when they lower total medical spending.

27:33Dr. Robert Pearl: Robbie, that makes sense. What about for doctors?

27:36Jeremy Corr: Jeremy, physician reimbursement remains under massive pressure. According to the American Medical Association, Medicare physician payment rates adjusted for rising costs of operating a medical practice have declined 33 percent between 2001 and 2025. In response, physicians have joined hospital systems and they've signed up with private equity-backed groups to gain negotiating leverage. The data show that both of these arrangements have increased prices by around 15 percent without improvement in medical care or clinical outcomes. But with hospitals coming into greater scrutiny for the monopolistic market control and facing greater pressure to reduce spending, preserving hospital revenue we can expect will be a higher priority for administrators in increasing physician compensation.

28:25Jeremy Corr: As such, this might be the time when doctors, rather than seeking protection, decide to take the risk and contract directly with insurers under a capitated contract. Under this arrangement, as insurers shift towards value-based care, physicians would accept greater financial accountability, they would sign multi-year agreements, and they'd benefit significantly when they could prevent chronic disease, control chronic conditions, and avoid the expensive life-threatening complications that have driven up American medical care so high. Once signed, they would likely embrace generative AI to improve chronic disease management.

29:03Jeremy Corr: And by reducing life-threatening complications that require hospitalization and demand costly procedures that earn more than by simply seeing additional patients each day at progressively lower reimbursement rates.

29:18Dr. Robert Pearl: Barry, that makes sense. How do you see hospitals responding to this downward pressure on medical costs?

29:23Jeremy Corr: Jeremy, hospitals are the largest single category of U.S. health care spending. According to the Centers for Medicare and Medicaid Services, hospital care accounted for approximately 31 % of the nation's$5.3 trillion of healthcare expenditures in 2024. But when you exclude items like interest in the national debt and social security payments, ones that are legally required, hospital costs rise to over 40 % of all medical expenses, as we mentioned earlier. That makes hospital spending a prime target for government programs and self-funded businesses seeking major savings. Two major sources of hospital revenue are already facing intense congressional scrutiny.

30:05Jeremy Corr: First, Medicare pays more for certain services delivered in hospital outpatient departments than for comparable services in community physicians' offices. In the past, when hospitals treated unusually sick patients at their inpatient sites, individuals who required medical care that was beyond the skill of doctors in the community, that made sense. But critics point out that now hospitals are acquiring doctors' practices in the community. And by putting them under the hospital license, they're billing significantly more and getting paid more for the same services as the physicians had provided in the past.

30:41Jeremy Corr: And they're able to do this even when the physician's offices are located miles away from the inpatient site and the care is identical as it was before. The Medicare Payment Advisory Commission has examined these differences and it's recommended site-neutral payments for selected services. Implementation would significantly decrease hospital revenue. A second threat to the hospital industry is related to the federal 340B drug pricing program. Under this program, eligible hospitals and clinics purchase outpatient medications at discounted prices, and then they sell them at much higher prices. A study published in the Indian Journal of Medicine found that private insurers were billed and reimbursed 340B eligible hospitals a median of three times their estimated acquisition costs, and this is for the physician and minister drugs.

31:34Jeremy Corr: Capping what hospitals can charge insurance companies and charge patients for these medications based on acquisition costs, that would lower total health care expenses. As such, Congress could now enact legislation to restrict profits. If hospitals want to protect themselves from watching the revenue drop, one place to look would be at the outpatient services arena. This is the fastest growing revenue generator for many inpatient facilities. In fact, a 2026 American Hospital Association report calculates that an average outpatient services account for 57 % of hospital revenue, and that was in 2024.

32:13Jeremy Corr: Instead of trying to maximize inpatient bed utilization and raising prices to pay for higher care delivery costs, integrated healthcare systems, ones with their own hospital network, could build on their outpatient capabilities by negotiating bundled payments for managing high-cost chronic conditions such as heart failure. This would leverage the combination of outpatient and patient capability. Instead of being paid a separate fee for every visit, test, and hospitalization, it assumed responsibility for the totality of their patient's care, using their size and scale to improve outcomes and thereby reducing spending.

32:49Jeremy Corr: Genitive AI, remote monitoring and coordinated outpatient teams would help identify clinical deterioration and deliver treatment before patients require emergency care or ICU admission. And when inpatient admission was required, the health system would use its inpatient facility and capability and expertise to provide care quickly and efficiently. Robbie, last but not least, how about the drug manufacturers? Chairman, Americans pay substantially more for brand-name prescription medications than patients in other wealthy countries. A RAND Corporation analysis commissioned by the U.S. Department of Health and Human Services found that U.S.

33:27Jeremy Corr: brand-name drug prices remain more than three times as high as those in peer countries, even after accounting for estimated drug manufacturer rebates. As Americans struggle to afford the medications they need to improve their health, the question Congress is asking is why the United States should fund the cost of development for Europe and Asia and generate the majority of a drug company's profits. Today, Congress has restricted the government from negotiating drug prices, even for patients enrolled in Medicare and Medicaid. Nearly all pure nations rely on these types of negotiations to prevent price gouging by drug manufacturers.

34:03Jeremy Corr: With Democrats now controlling both the House and Senate, the pressure is going to rise for this prohibition to be repealed. Already, the executive branch is imposing growing pressure on manufacturers to link U.S. prices to international benchmarks. There's increasing talk about legislating a most favored nation provision, which the United States wouldn't pay more than other wealthy, pure nations for identical medications. But rather than waiting for the government-imposed drug pricing restrictions, not only on Medicare and Medicaid, but also for people covered by their employers, manufacturers today could negotiate lower prices with self-funded businesses and pharmacy benefit managers in exchange for expanded coverage and greater sales volume.

34:45Jeremy Corr: Consider GLP-1 Medications, a company that currently excludes obesity drugs from its health plan, but be willing to offer coverage if a manufacturer agreed to lower prices substantially to levels comparable to those in other wealthy countries. In exchange, the manufacturers to receive preferred formulary placement and access to a much larger population of eligible patients. Although profit per prescription, that would decline, the manufacturer would gain additional sales. Even if total revenue went down under this agreement, manufacturers would continue to earn higher profits than under a government-imposed most favored nation pricing policy.

35:26Dr. Robert Pearl: Rabbi, it's clear that the American healthcare system isn't meeting the needs of most people. And without a doubt, all the current healthcare incumbents are likely to face pressure in the future to lower costs while improving clinical outcomes. How do you see the process evolving?

35:39Jeremy Corr: Jeremy, as we discussed, federal deficits, rising borrowing costs are putting immense pressure on public budgets and American families. Employers are struggling to absorb healthcare inflation. Patients are finding medical care increasingly difficult to access and afford. For insurers, physicians, health systems, and the pharmaceutical manufacturers, the magnitude of these pressures indicates that a potential strategic inflection point may have been reached. The strategic inflection point is a moment in time when the world changes and never goes back to what was before. You can think of the introduction of the smartphone.

36:16Jeremy Corr: It's just such a point from the past. And once that moment happens, it becomes unimaginable that we could ever have done it the way we did it before. Already the bond market is flashing warning signs, warning lights. Rather than waiting for the crisis to erupt, health care leaders must now recognize the dangers ahead. And once they do, they can decide whether to cling to the established business models of the past, or whether they want to pivot and embrace fundamentally different approaches to financing and the provision of medical care. Companies in a variety of industries, including computers, automobiles, and steel production, They've made this type of shift in the past, and many have become even more dominant than before.

36:58In contrast, others have clung tightly to their past way of doing business, their past strategy, and they failed as a result.

37:07Jeremy Corr: Figuring out the best approach in a dynamic environment, that's challenging. The right path is almost always clear, of course, in retrospect, but leaders don't have the luxury of viewing the world through the rearview mirror. They must decide and act before the data is 100 % clear. When it comes to American medicine, there are strong reasons for hospitals, insurers, drug companies, and physicians to make a shift. But there are equally strong reasons to hold on to the current path. There just isn't a clear answer. Bond traders aren't always right. But before healthcare's leaders decide that making a shift is too risky, they should ask themselves, what do they know?

37:49Jeremy Corr: that the smartest people in the room don't.

37:53Dr. Robert Pearl: We hope you enjoyed this podcast. We'll tell your friends and colleagues about it. Please follow Fixing Healthcare on Apple Podcasts, Spotify, or your favorite podcast app. If you like the show, please rate it five stars and leave a review. Visit our website at fixinghealthcarepodcast.com and follow us on LinkedIn, Facebook, or Twitter at FixingHCPodcast. Thank you for listening and have a great day.

From the publisher

Artificial intelligence is advancing at an extraordinary pace, and warnings about its dangers are becoming increasingly alarming. Technology companies are reporting troubling behavior from powerful AI systems, while researchers and government officials debate how to prevent potentially catastrophic consequences.

But Dr. Robert Pearl sees another danger: What if fear of AI prevents medicine from using the technology to save lives?

In this episode of Diving Deep, Pearl and co-host Jeremy Corr begin by exploring the risks and opportunities of generative AI. Later in the episode, they offer another warning that healthcare leaders can no longer afford to ignore: America’s mounting economic pressures.

The medical dangers of standing still

Dr. Pearl begins by acknowledging the existential risks posed by increasingly powerful AI systems. Those threats demand serious safeguards, he says, but not every application carries the same level of danger.

Consider patients recovering from surgery, living with heart failure or managing chronic diseases such as hypertension and diabetes. Doctors know that more frequent monitoring could help identify complications earlier and improve outcomes. But most physicians lack the time to review daily blood-pressure readings, evaluate surgical wounds or track subtle changes in patients’ health between office visits.

Generative AI agents could help fill those gaps. Pearl draws on his experience leading The Permanente Medical Group, where more frequent monitoring contributed to substantial reductions in cardiovascular mortality and better control of hypertension. AI, he argues, could make comparable levels of attention more affordable and accessible nationwide.

What the bond market is telling healthcare

The second half of the episode shifts to another source of concern: America’s mounting debt and the rising cost of borrowing.

Pearl explains why bond-market trends should matter to physicians, hospitals, insurers and drug manufacturers. With federal debt exceeding $40 trillion, healthcare spending continuing to rise and employers struggling with escalating premiums, the financial pressures on American medicine are becoming increasingly difficult to ignore.

Yet healthcare organizations continue relying on familiar strategies: higher prices, tighter insurance restrictions, hospital consolidation and greater negotiating leverage.

Pearl argues that these approaches cannot remain effective indefinitely. Instead, he challenges four major sectors of healthcare to consider fundamentally different strategies:

  • Insurers: Move beyond restricting access and embrace payment arrangements that reward better health and lower total costs.
  • Physicians: Consider greater financial accountability through value-based contracts that reward prevention and effective chronic disease management.
  • Hospitals: Shift their focus from maximizing inpatient revenue toward coordinated outpatient care that prevents avoidable hospitalizations.
  • Drug manufacturers: Explore lower prices in exchange for expanded coverage and greater prescription volume.

Each approach involves risk. But Pearl warns that preserving the status quo may ultimately prove more dangerous than embracing change. He closes by describing what business strategists call an inflection point: a moment when existing assumptions no longer hold and leaders must make consequential decisions before the outcome becomes obvious.

The two discussions offer a common lesson for American healthcare. Whether confronting powerful new technology or mounting financial pressures, leaders cannot assume that avoiding change will protect them from risk. Listen to this episode of Diving Deep to explore how AI, economics and strategic decision-making could reshape the future of medicine.

Helpful links:

* * *

Dr. Robert Pearl is the author of ChatGPT, MD: How AI-Empowered Patients & Doctors Can Take Back Control of American Medicine.

Fixing Healthcare is a co-production of Dr. Robert Pearl and Jeremy Corr. Subscribe to the show via Apple, Spotify or wherever you find podcasts. Join the conversation or suggest a guest by following the show on X and LinkedIn.

The post FHC #234: The risks of moving too fast with AI & the dangers of standing still appeared first on Fixing Healthcare.

More from Fixing Healthcare Podcast

All episodes