Mailbag: Virtual Nurses, Bundled Payments, and Single Payer

Welcome to Healthcare Rounds, where we give you smart conversations with the people moving healthcare forward. I'm John Marchica, CEO of Darwin Research Group. Today's a mailbag episode, and I'll be covering technologies to combat nursing shortages, bundled payments, and the likelihood of the U.S. going to a single payer model. If you have a question or comment from me, please send it to healthcare rounds at darwinresearch.com. And if you're not receiving our weekly newsletter, you can subscribe at darwinresearch.com. Okay, let's get to your questions. This first one comes from a commercial life sciences executive who asks, How are health systems or IDNs applying technology to help address the nursing and nursing and staffing shortage? What are they investing in? Well, health systems are putting real money behind three things right now. In our ongoing research at Darwin, addressing staffing issues has been a top strategic priority, honestly, for as long as I can remember. And of course, COVID put a sharper focus on employee retention, staffing, and burnout. But the first and largest bucket or category, I would say, is virtual nursing. And here are some specific examples that I pulled from our own IDN research as well as a couple of articles from Becker's. Advocate Health saved over 43,000 hours across 25 hospitals in a single year using virtual nurses for admissions, discharges, and transfers, and cut enough nursing turnover to avoid $6.3 million in costs. Piedmont scaled virtual nursing across 17 hospitals and 2,700 beds in under 18 months. Prime Healthcare, a 55 hospital system, expanded from virtual observation into full virtual nursing after a pilot that cut patient falls on medical surgical units by 84%. Advent Health started with a pilot in Florida using off-site RNs to communicate with patients in the ED and inpatient units, and one hospital in that rollout saw its RN turnover drop from 46% to 16% year over year. H-Rim Health assigns virtual nurses to about 10 patients at a time, handling safety monitoring, charting support, and admissions and discharges. And finally, Baylor St. Luke's hospital pairs virtual RNs with bedside nurses and patient care assistants, and they've expanded from an initial 32-room pilot to 64 rooms since 2023. So, how do actual nurses feel about virtual nurses? Well, a recent University of Pennsylvania survey of nearly 900 nurses across 10 states found 57% said virtual nursing didn't reduce their workload, even though 53% said care quality improved. And the researchers were frank. They said, well, virtual nursing doesn't help much without adequate bedside staffing and clearly defined roles to begin with. And there's a real undercurrent of nurses suspecting health systems are using it to avoid hiring or replace existing nurses rather than to genuinely support the floor. So the second area that I should mention is ambient documentation. And we know that that's been the case in the physician world for some time, but it's been moving into nursing. And again, I found a few examples. Parkview Health deployed ABRIGE for nursing documentation and went from 60,000 AI process tokens a month to 3 million in just a matter of months. Mayo Clinic is co-building ambient AI tools specifically designed around nursing workflows, not just repurposing the physician version, because obviously the two roles document completely differently. Providence launched a dozen epic AI tools in a single month this spring, covering ambient documentation, predictive analytics, and administrative automation. And finally, Jefferson Health is testing whether ambient tools built for physicians can be adapted for nursing workflows at all. So in general, ambient documentation is becoming commonplace. According to Epic, two-thirds of U.S. hospitals on the Epic EHR were already using some form of ambient AI documentation as of 2025. And the last category or bucket is predictive staffing and scheduling. These tools adjust staffing in real time based on patient acuity and census trends, specifically to cut overtime and agency labor costs, which is where a lot of the real money is bleeding out of hospital budgets. Vendors like UKG, Oracle Health, and Workday are building healthcare-specific workforce analytics that link hiring, scheduling, and retention data into a single feedback loop. So a hospital can see, for example, whether new graduate nurses hired through one channel are turning over faster than others, and then they can adjust sourcing accordingly. So to wrap up this question, we are seeing widespread adoption of virtual nursing, ambient documentation, and predictive staffing. Each of these tools has the potential to increase productivity, reduce costs, and improve the patient experience. But honestly, as far as I can tell, none of it fixes the underlying shortage. And that problem has a lot to do with accredited nursing programs themselves, and a long-standing shortage of instructors who see higher salaries and career opportunities outside the classroom. I've mentioned that on the show several times. These tools shift work around in my hospital's financial breathing room. Whether it reduces nurse burnout or just relocates it somewhere else in the system, that's still an open question, and I don't think that we have a clean answer yet. Second question comes from an IDN clinical leader who asks, Do you see bundled payments for medical conditions becoming standard for all traditional Medicare patients? Well, the short answer is no, and certainly not for all conditions. For the foreseeable future, I think that CMS will continue to test models and commercial payers are going to follow suit, learning lessons from the CMS experiments. And that means that bundles, ACOs, global payments, Medicare Advantage, other models that we call alternative payment models or value-based payment models, those are going to continue. That means that bundles, ACOs, global payments, Medicare Advantage, and other models that we call alternative payment models or value-based payment models will continue alongside the bundles. As a reminder, a bundle payment model pays a single predetermined amount for all the services involved in a patient's episode of care. For example, a hip replacement, including the surgery, hospitals stay and follow-up rehab, rather than paying separately for each individual service. For years, CMS tested bundled payment models voluntarily, and the problem with all voluntary payment models is you're going to run into selection bias. The hospitals that sign up are the ones who think they'll do the best. While the concept of bundling payments has been around for decades, most of us are familiar with the Comprehensive Care for Joint Replacement, or CJR, Bundled Payments for Care Improvement, or BPCI, and an advanced version of that program, with a goal of reducing total cost of care while improving quality for procedures like hip replacement surgery. CJR ended in 2024, while BPCI and BPCIA ended in 2025. Some participants from both these programs transitioned to TEAM. The latest from CMS is the Team Bundle Payment Model, which includes five surgical episodes and is now underway in 750 hospitals in about a quarter of the country. The most important thing to know about TEAM is that it's mandatory. So mandatory fixes the selection bias problem. There's no getting out of it. But it also means that CMS is really getting serious about bundle payment models going forward. CMS has already finalized a physician-focused mandatory model starting in 2027 for heart failure and low back pain called the ambulatory specialty model or ASM. Where previously most bundle payment models were for surgical procedures like hip and knee replacements, ASM is focused on chronic disease management. So now we're looking at episode-based payments focused on exactly the kind of long-term condition management that drives most of Medicare spending. And just a couple of weeks ago, CMS announced that it's expanding or resurrecting the Comprehensive Care for Joint Replacement Model. Starting in January 2028, most hospitals will be required to participate in the expanded model, referred to as CJRX, making it the first expanded mandatory test of an episode-based payment model. Under the expanded model, an episode will span the joint replacement surgery, hip, knee, or ankle, hospitalization, and the first 90 days of recovery after discharge from the hospital, including follow-up treatments such as physical therapy. Hospitals that are participating in TEAM will not be required to participate in CJRX. For health systems, the implication is straightforward. This means that hospitals and IDNs need to shore up discharge planning and tighten relationships with high-performing skilled nursing facilities and home health agencies, or they risk losing money on the bundle. And because ASM extends the same logic to outpatient specialists starting in 2027, health systems need to think about the physicians treating these patients today and whether they're employed, affiliated, or entirely outside the system. And if it's the latter, ASM gives the system no direct lever over how these physicians manage the episode, which means the relationships and data sharing agreements need to get built now, well before 2027 rather than after the model goes live. And if you're in pharma, you have other considerations, and it's not as simple as, quote, bundles encourage the use of lower cost generic drugs, which is kind of the typical wisdom. We'll start that sentence again. If you're in pharma, you have other considerations, and it isn't as simple as bundles encourage the use of lower cost generic drugs, which is sort of the typical wisdom, but the moment you put a chronic condition inside an episode payment, you change how drugs get used inside that episode. Clearest example of this is dialysis. Back in 2011, Medicare rolled anemia drugs like EPGEN into the ESRD bundle instead of paying for them separately. Utilization dropped almost overnight, more than 20% below prior levels. When researchers look closer, the drop was concentrated almost entirely among patients who didn't clinically need the drug in the first place. And that's the bundle working exactly as it's supposed to, squeezing out low-value use without touching necessary care. Now, if you're bringing a new drug to market and you want your drug to be part of, say, a heart failure bundle, pay attention. Because bundle payment benchmarks are built off historical spending data. A genuinely new, more expensive drug can blow up a practice's target price the moment it hits the market through no fault of the practice of the patient. And Medicare actually ran into this with the oncology care model, where drug spending makes up close to two-thirds of total episode cost. They built in something called a novel therapy adjustment, which is essentially a patch designed to keep practices from getting penalized financially for using new cancer drugs. But even that patch had a strange, if you will, side effect. CMS's own data shows that in some cases, using the more expensive brand name version of a drug instead of an available generic qualifies a practice for that adjustment, because the adjustment is based on share of spending going to newer therapies. So a mechanism built to protect innovation can end up nudging practices toward the pricier option instead of a cheaper one. The opposite of what you'd expect a bundle to do. So the bottom line is bundling a drug into a payment doesn't just push total costs down, it reshapes which drugs get used and when, in ways that take real design work to get right, and Medicare is still tuning that formula as it goes along. If ASM extends the same logic into heart failure and low back pain, expect the same growing pains. Real savings on discretionary use, real friction for anything new coming to market, and CMS patching the model in real time to keep from punishing innovation. It doesn't mean to punish. In my experience, most people, when they say single payer like Canada, assume every other developed country runs the same system, and they don't. But there's a well-established framework health policy people use to sort this out. And there's three basic models, and it's worth being specific about which one people mean because single payer just gets used as a catch-all. The first is called the Beveridge model, which is named after a British economist named William Beveridge. In this model, the government doesn't just pay for care, it owns the hospitals and employs the doctors directly. That's Britain's National Health Service. And it's the purest version of what people picture when they use the term socialized medicine. The second is the Bismarck model, named after Otta von Bismarck, the Prussian chancellor who invented it in the 1880s. And what he built was an insurance system funded jointly by employers and employees through payroll deductions, paid into nonprofit so-called sickness funds. In this model, doctors and hospitals stay private. And it's the model that would look most familiar to Americans because it's built on employer-linked insurance. It's just heavily regulated, and it's universal, unlike our model. The third is the national health insurance model, and this is the Canadian model. The government is the insurer, financed through general taxes, but doctors and hospitals remain private. So Canada, publicly financed, privately delivered, free at the point of use. Technically, it's 13 separate provincial single-payer systems, one per province and territory, operating under shared national rules rather than one national program. There's a private market in Canada too, but it's it's mostly for what the public plan doesn't cover, which is dental vision and certain prescription drugs. Now let's look at a few other countries and see how they finance and deliver health care. After a few of these examples, you see where I'm going with this. The UK is the textbook beverage case, government-run hospitals, salaried public employee doctors, funded through general taxation, and free at the point of use. Germany, the original, still runs through nonprofit sickness funds, financed jointly by employers and employees, but it's genuinely multipayer, with over 200 competing funds, all privately delivered. France runs the same public financing structure as Canada, but with a key difference. Patients often pay out of pocket at the time of care and get reimbursed by the public insurer afterward. And nearly everyone also carries a supplemental private policy called a mutuel, please forgive my French pronunciation, to cover what the public reimbursement doesn't reach. So it's genuinely a public-private hybrid system dressed up as single payer. And by the way, it's frequently ranked among the best performing systems in the world. Here's a country that's universal but not single payer at all. Brazil guarantees healthcare as a constitutional right through a public system called SUS or SUS, an acronym for unified health system, which is free at the point of use for everyone. But over half of all health spending in Brazil is private, and roughly a quarter of the population also buys private insurance because the public system is stretched thin and underfunded. Switzerland has no government insurer whatsoever. Everyone is legally required to buy private insurance from competing, tightly regulated nonprofit insurers, with income-based subsidies for people who need help affording premiums. It's private coverage made universal entirely through mandate and regulation, and it's the system that Americans debating an individual mandate usually point to. And finally, Australia runs a public insurance program, also called Medicare, that covers everyone for basic hospital and physician care, but the government actively pushes higher earners into private supplemental insurance through tax penalties if they don't buy it. A deliberate two-tier design, and that's not an accident of history. And here's a domestic example worth throwing into a mix. Maryland has run an all-payer rate setting model since the 1970s under a state agency called Health Services Cost Review Commission. Every payer, Medicaid, Medicare, commercial insurers, even self-insured employers, they pay hospitals the same rate for the same service. Since 2014, Maryland has taken it further with global budgets. Each hospital gets a fixed annual revenue set in advance, so the incentive to run more volume through the building mostly disappears. It's not single payer. Insurers stay private, and coverage still runs through employers the same way it does everywhere else in the country. But it proves you can get Canadian-style price discipline without Canadian-style financing. And CMS is now trying to export the model to other states. It's worth remembering the next time someone tells you single payer is the only way to control hospital costs. And it's a clean illustration of the split I mentioned earlier. Who pays and who delivers are separate questions. And Maryland answers them differently than Canada or even the rest of the United States. One last point. Our system is plenty socialized already. Now, see the actual question of whether we're headed to Canadian-style health care, whether you mean universal coverage or single payer or any system like the ones I talked about earlier, no. Not in my lifetime at least. It's toxic politically. Just look at what we went through with the adoption of the Affordable Care Act and what politicians are still arguing over today, or look to the failed Clinton health care reforms of the 1990s. After the Medicaid cuts and the ACA subsidies expiring last year, you're seeing a real wing of the Democratic Party argue that just restoring the old system isn't enough, and Medicare for all is showing up again in 2026 primary messaging. And actually, if you poll it with a fair description, one that's upfront about the fact that some people's taxes would go up, you still get majority support, somewhere just north of 50%. But here's what hasn't changed. Roughly 150 million Americans get their coverage through an employer, and a large share of those people vote. And every time single payer gets tested at the state level, even in progressive states, it tends to hit a wall the moment it's attached to an actual tax bill. So the question isn't whether the rhetoric is louder. It's whether that rhetoric translates into legislation, and that's where I end up just saying no. And what I also don't think changes, regardless of how loud the 2026 and 2028 conversations get, is the pattern this country has followed for 60 years. We move on health coverage in increments, not conversions. Medicare, Medicaid, ACA, none of them were single payer, all of them were incremental. And I'd bet on that pattern holding, even as the rhetoric around Medicare for All gets loud. That's it for this week's mailbag episode. Please send your questions and comments to healthcare rounds at darwinresearch.com. If you enjoyed this episode, please forward it to a friend and rate reviews subscribe on your favorite podcast platform. I'm John Marchica. See you next time.