• Pay for Health Outcomes: Talking to Vivian S. Lee

    How does pay-for-action US healthcare not just waste money, but impede public health — and even harm patients? How might applying best practices from industrial production actually make our hospitals more patient-friendly? When I want to ask such questions, I pose them to Dr. Vivian S. Lee. This present conversation focuses on Lee’s book The Long Fix: Solving America’s Health Care Crisis with Strategies that Work for Everyone. Lee is President of Health Platforms at Verily Life Sciences, and former CEO of University of Utah Health (at a time when University of Utah ranked first among the nation’s university hospitals in quality and safety). An MR radiologist who developed novel methods for measuring kidney function and vascular disease, Lee has been elected to the National Academy of Medicine, and has received the International Society for Magnetic Resonance in Medicine’s highest award for scientific contributions and leadership. She has authored over 200 peer-reviewed research publications, and presently serves on the Defense Health Board (advisory to the Department of Defense). She earned a PhD in medical engineering from Oxford University, an MD from Harvard Medical School, and an Executive MBA from NYU’s Stern School of Business.


    ANDY FITCH: First, with demographic- and technology-driven increases to healthcare expenditures worldwide, with the US already spending two to three times more per capita than comparator countries, with the US standing out among such nations for the strikingly uneven outcomes of this care (and for its 44 million uninsured citizens), how should we be prioritizing eliminating waste, foregoing ineffective treatments, and better coordinating care delivery — instead of simply “cutting back” on care?

    VIVIAN S. LEE: The US spends, as you suggested, much more per capita than our counterparts in Europe, Canada, Australia, Japan — yet we get significantly poorer outcomes, in large part because we waste so much money. In 2012, as the Affordable Care Act became law, the National Academy of Medicine issued a report estimating that we wasted about 30 percent of healthcare dollars. A 2019 study showed that we still waste at least 25 percent, close to a trillion dollars each year.

    To cut healthcare costs, The Long Fix shows how, rather than making an across-the-board cut, we need to target waste. To achieve that, we need to turn healthcare’s economic model upside down. Today, most payors pay doctors to do things to patients, to perform procedures, to operate, to recommend imaging studies. Payors pay regardless of whether these actions improve health. Efforts to reduce the overall costs of care fly straight into the headwinds of capitalism. We really need to turn this plane around. Instead of paying for action, we should pay for care that makes people healthier. We need the tailwinds of capitalism to push us forward: to push towards improved quality and efficiency through innovation, and competition.

    Could you make concrete some most problematic ways in which this prevailing fee-for-service model deemphasizes preventative care, distorts market incentives, and even may discourage successful treatment of a patient (as opposed to extracting further profits from patients’ ongoing conditions)?

    Where do I begin? Well first, even before offering some statistics, I want to make the broader point that over-treatment in our present system comes at the expense of safety. Consider the classic first principle of medicine: Primum non nocere (first, do no harm). Today, medical mistakes have become our third-leading cause of death in the US, right behind cardiovascular disease and cancer. It doesn’t have to be so. Other industries have embraced total quality management: consistently raising their standards of production, reducing variation, improving outcomes, and decreasing error rates (to nearly zero). That kind of management model still remains in its infancy in US healthcare.

    When I ran University of Utah Health, we studied one of these total quality management systems. We took a trip to Ogden, Utah, just up the road from Salt Lake City, to tour the Autoliv airbag manufacturing plant. This plant has had a special relationship with Toyota since they were first contracted to manufacture Toyota airbags in the 1990s. One of Toyota’s experts came over and spent several years working with everybody at the plant, and since then, production quality has steadily improved. When we took our tour, it just put us to shame. They weren’t talking about one-part-per million defects. They were talking about zero defects. In healthcare, by contrast, we regularly see error rates closer to 20-25 percent.

    The fee-for-service model also undervalues preventative care. We have clear examples of how moving towards a payment-for-health model can dramatically improve prevention and reduce avoidable hospitalizations. ChenMed (a medical group originally from Miami) and Iora Health (started in Boston) both have developed innovative multi-site medical practices. Both groups have embraced new Medicare Advantage contracting models, in which they receive a predetermined payment per year for each Medicare member on their rosters. The precise amount is determined by the underlying health of each patient, higher for those with a greater number of conditions. If the medical group (the physicians and the nurses and the team) can keep this person healthy and out of the hospital, then they keep the savings. If the patient needs more care than expected, then the medical group loses money. If these practices keep patients satisfied, then the patients likely will stay with them the next year.

    The results have been impressive. Instead of 10 minutes per visit, doctors in these groups might spend 30 minutes to an hour with a patient. Frail patients need to get seen regularly, so these practices might, for example, offer shuttle service to the clinic, or even make home visits. They don’t just prescribe medicine. They make sure their sometimes impoverished patients can get prescriptions filled, often at pharmacies within the clinics themselves (or refilled, perhaps again through home deliveries). As part of a falls-prevention strategy, these clinics offer yoga classes, Zumba classes, or Tai Chi. They put a big focus on reducing social isolation. Some have community rooms in the clinic, functioning as a kind of social drop-in space. As a result, even though these facilities may cost more to run than a typical primary-care clinic, they provide significant long-term savings. Best of all, by avoiding hospitalizations, their patients enjoy independence and better health.

    Then more broadly in terms of restructuring market incentives, Big Pharma firms have enjoyed a long run of staggering patent-protected profit margins. Hospital systems routinely exploit their regional-market dominance to the detriment of patients. An entire sector of lucrative insurance-related administrative work only exists at this scale in the US. So could you make concrete a few baseline hypocrisies in these various industry players arguing that structural interventions to address US healthcare’s waste would compromise some supposed free-market dynamism?

    We do hear a lot about the pros and cons of free-market capitalism driving American healthcare. Personally, I think we should consider this market from two perspectives — first focusing on individuals as consumers of health. For patients, it’s pretty hard to think of American healthcare today operating as a functional market, given the negative externalities. Few people understand medicine, leading to significant asymmetries of information. Few people understand how health insurance really operates, leading some patients themselves to demand unnecessary treatment (such as antibiotics for the common cough), an example of moral hazard. And even with transparent pricing, most people would be unable to shop or negotiate effectively for healthcare — especially if they are experiencing crushing chest pain, for example.

    On the other hand, the market in which healthcare payors (insurance companies and the government) buy goods and services (from health systems, doctors, pharmaceutical companies, or device manufacturers) could function better, with more transparency and more data on patient outcomes, quality of care, and cost-effectiveness of medications or devices. These could all improve competition and potentially drive innovation, leading to better value and to lower costs.

    On this topic of transparency, US healthcare’s lack of transparent pricing definitely stands out. Could you sketch some most drastic examples of healthcare providers and institutions themselves having no idea what they pay or get paid for various supplies and services? Could you point to some outrageous unfathomables of medical billing (with which most Americans already will be all too familiar)? And again, could you make the broader case for taking our most fundamental step towards pricing transparency by shifting from a pay-for-action to a pay-for-results model?

    Just by shifting from pay-for-action to pay-for-results, we can reshape incentives, demanding that every successful player in the market demonstrate they make people healthier at a comparatively low cost.

    During my time as CEO of University of Utah Health, two things happened almost back to back that dramatically affected our thinking. First, the state of Utah announced that its Medicaid program would change the payment model. Instead of paying each time a patient visited us, the state would pay a fixed amount per year for each Medicaid patient. Now all the sudden, the woman who had visited our emergency room 52 times the previous year became a major concern. Those costs would become our expenses (instead of revenue). We immediately assigned her a primary-care provider and a care manager, who ensured that she had transportation to the clinic, that she had her prescriptions filled, that she stayed healthy and didn’t wind up back in the ER.

    Secondly, that same year, Medicare announced pilot projects for what they called “bundled payments.” For some of the most common procedures, like hip and knee replacements, Medicare would pay a fixed amount to hospitals for the pre-operative care, the surgery, the follow-up needed after an operation. If a patient had to stay in the hospital for a long time, or had to return to the hospital due to an infection, then the hospital would provide that additional care with no further charges — without passing on any additional costs to Medicare.

    The hospital’s ability to make money (or not lose money) now depended on whether we could keep our costs below these Medicaid annual rates and these bundled Medicare payments. Our hospital recognized the need to better understand how much it cost to provide care for each individual patient. If we understood our costs, we could identify opportunities to reduce those costs, which would improve our profitability.

    We’d never considered costs at this level of detail. But when we did, our doctors (myself included) were fascinated to see how much variability existed. For example, the types of artificial hips our doctors used varied in cost by three-fold — even though the different brands seemed to work similarly well for patients. These kinds of discoveries we made about our own institution landed us on the front page of the New York Times, and motivated me to write The Long Fix. You can’t manage what you can’t measure. And in this case, it’s just remarkable that our national healthcare system, so concerned about costs, has such little insight into these costs of care at the individual patient level. I consider that one of our greatest opportunities for improvement.

    Then alongside patient outcomes, and patient satisfaction, your book also makes clear medical providers’ own workplace dissatisfactions — say with doctors’ increasing relegation to rote data-entry duties. Hospital systems themselves have become notorious among economists for exploiting labor-market dominance (even if I can quit my nursing job at one hospital and get hired at another 5 miles away, I might still have the same employer). So before we continue with what makes our present healthcare system problematic for patients, for business owners, for taxpayers, could you flesh out what makes it sometimes so demoralizing (even pre-COVID) for its workers?

    Here again, in a fee-for-service environment, doctors face constant pressure from their health systems to generate more “revenue” by seeing more patients, and documenting all the procedures or other actions they’ve taken. This record-keeping consumes an incredible amount of a physician’s work day — and even home life. Physicians routinely point to all the hours spent at night and early in the morning just updating records, documenting every single thing they’ve done, in order for the health system to bill for them. It’s not surprising that nearly half of physicians surveyed in the US have described themselves as burnt out in recent years, even before COVID-19.

    Yeah, you’ve already offered some pretty disturbing statistics on medical mistakes — which, as The Long Fix points out, also include 10 instances of serious harm for every reported death. Your book gives an especially telling account of drug delivery. Could you describe some of the institutional challenges here, and also the stand-out successful interventions that you have seen? And again, beyond any one single care facility, could you point to why we need redirected market incentives to address these nation-wide failings?

    Studies have shown that every time a patient receives medication in the hospital, there’s a 20-25 percent chance that a medical error happens. Now, sometimes these errors cause no real harm. Maybe, for example, the timing of medication delivery didn’t fully optimize a particular drug’s potential benefits. But we also see more significant mistakes, such as the wrong dose or the wrong injection rate.

    We even see patients receive the completely wrong drug. In the book, for example, I address the seemingly simple topic of sound-alike names for very different drugs. Take Celebrex (an arthritis drug), Cerebyx (a seizure medication), and Celexa (an anti-depressant). Very different medications, for very different conditions. You certainly don’t want to get these confused. But that’s quite common. The FDA in fact regularly publishes online lists for all these sound-alike drugs. Instead of requiring more distinct names for them, the FDA recommends just putting three letters in the upper case — so that medical providers can physically see the difference between, say, clonazePAM and cloZAPine. There must be a better way.

    Electronic medical records and e-prescribing already have significantly reduced the error rate. But we still could do much better with medications. The Long Fix describes, in a real-world hospital setting, the many interruptions that doctors, pharmacists, and nurses can face when writing prescriptions or dispensing medications. This goes back to what I said about how hospitals haven’t yet benefited enough from the kinds of operational advancements that we see in well-run manufacturing plants. We certainly could embrace the same basic lessons of industrial engineering, and adopt best-in-class processes for running healthcare facilities with a degree of coordination more like airbag manufacturing plants or piano factories.

    Many healthcare processes could likewise benefit from better human-factors design. Right now, we often have harried clinicians navigating complex, distraction-filled environments, with their cell phones ringing continuously. It can be better. For one example, at the University of Utah we assigned a team to focus on reducing hospital-acquired infections, particularly catheter-related infections. Catheters are these long tubes, like IVs, but they extend more deeply into the body, delivering antibiotics or nutritional supplementation for days, and sometimes weeks or even months. The catheters can easily become sources of infection if not inserted correctly, or if not cleaned effectively — especially in vulnerable patients, like those with burns.

    Our nurses worked closely with a human-factors design expert to create a simple central-line tool kit for dressing changes. This kit holds all of the necessary supplies, so nurses don’t need to run around trying to find some gauze here, or pick up some iodine swabs (for cleansing) there. It’s all in one kit. With every step, the nurse unfolds another part of the pouch-like kit, so that each of the supplies gets revealed in its proper sequence. With these dressing-change kits and a few other adjustments, the burn unit’s infection rate declined almost immediately. Over the next four years, this ward had zero central-line infections — basically unheard of in the US.

    By extension, could you point to drawbacks (at least for many patients and providers) in our extravagant medical-malpractice culture — with US safety records often much worse than, say, countries that have implemented a “no-fault” system for reporting mistakes?

    New Zealand started using a medical-malpractice no-fault system as early as 1974. This no-fault system makes medical malpractice more like workers’ compensation. Patients file claims with the government, and are compensated from a central pool of public funds. While researching The Long Fix, I discovered that Virginia and Florida actually have had no-fault programs for decades, covering the compensation costs when infants suffer brain injuries during birth. These programs, which don’t involve the courts, tend to result in comparable compensation for injuries, but with much lower administrative costs for families.

    You’ve also mentioned the need for establishing empirical measurements of success, alongside implementing procedural best practices. Could we pivot towards a broader discussion of the predictable snags (staff skepticism, perverse gaming-the-system scenarios) that come with prioritizing and sometimes even publicizing quantifiable End Results and patient-satisfaction metrics? Could you give some personal experience on how to most constructively track and report such data?

    One of the most rewarding aspects of my career in healthcare has been the chance to work with so many highly motivated, mission-driven individuals. The physicians I worked with at the University of Utah understood well the many fundamental challenges our health system was facing. They knew why we needed to move away from the pay-for-action model (even though we had succeeded in it) to a model that would pay for health outcomes. And we all recognized that making this pivot would truly require systemic change. Our administrative team realized, for example, that not only did we need to gather and provide data about outcomes and costs, but that we also should give our medical teams significant autonomy in deciding how to implement institutional improvements.

    We kept the focus on measuring how well patients were doing under our care. We first shared information about patient satisfaction, and then later quality and costs of care. Of course, at first, some of our typically high-achieving providers may have felt a bit shocked to receive data indicating their performance was below average. They would argue that the data were incorrect, or that their patients were sicker and had more complex conditions. Sometimes they were right.

    Over time, as the data improved, we started making reports more visible to colleagues in their divisions and departments, then later more widely across the hospital. Then, at the urging of one of our surgeons, we decided to share patient-satisfaction data publicly (online) for each of our physicians and advanced-practice nurses. We actually became the first hospital in the country to post such data online. By the time we did that, most of our physicians were significantly above the national average. By first giving our providers the chance to improve, I think we encouraged a productive nurturing of best practices — allowing clinicians to learn from each other.

    We also learned the critical lesson of letting our medical teams share in the benefits that came from these improvements. When physicians reduced costs of care while improving patient outcomes, generating savings for the organization, we shared some of those savings with them. They couldn’t use these funds to boost salaries. But we gave them significant discretion to decide if they should, for example, send some of their faculty to professional meetings (to report on research), or hire more medical assistants to help manage electronic record-keeping work.

    Well still on this topic of promoting autonomy, perhaps now from a more qualitative perspective, could you also describe the importance of recognizing patients as co-producers of health — again particularly in distinction to the passivity bred by our employer-sponsored, fee-for-service system?

    That’s definitely one of my favorite topics. In medical school we actually all learn (and as the University of Utah’s dean I also taught) that while we tend to think of healthcare in terms of what happens in hospitals and clinics, it turns out that roughly 70 to 80 percent of health outcomes are determined by our own individual experiences and choices: what we eat, what we drink, what we inhale, physical activity, and so on. When it comes to improving this country’s overall health, we here again need to recognize that a shift towards paying for better results will incentivize a focus on how to promote behaviors leading to better health outcomes.

    One key influence on my thinking in The Long Fix came from Stanford economist Victor Fuchs. In the 1960s, amid the shift to a service-based economy, Fuchs observed that, unlike with industrial manufacturing, a service-based exchange relies on the producer and the consumer to co-produce outcomes. Take public safety, for example. We don’t just count on law enforcement. We put locks on our doors and perhaps install an alarm system. Or in education, we help our children with their homework and volunteer for the PTA.

    In healthcare, this kind of co-production becomes even more crucial. It centers on the relationship between a patient and his or her doctor or nurse, and acknowledges the vital role of patients (and their families) in attaining health.

    At Verily, for example, we’ve built a virtual diabetes clinic that epitomizes this principle of co-production. For someone with type 2 diabetes, a crucial element of maintaining good health is keeping blood sugars within normal range. Blood-sugar levels mostly depend on diet and exercise (and sometimes medications). With our product, called Onduo, instead of pricking a finger multiple times each day to check blood sugars, a patient can apply a device called a continuous glucose monitor (that we also develop) on the arm or abdomen. This measures blood sugars constantly, and transmits those values via Bluetooth to the patient’s smartphone. Instead of having to keep a log of everything they eat, patients can just take pictures of meals and snacks, and our algorithms can recognize most of these. With those photos and the blood-sugar tracings, a person can see exactly how different meals affect blood-sugar levels.

    Instead of lecturing people (“Vivian, you really should stop eating that extra slice of pecan pie every Thursday night”), this tool makes it clear to patients themselves exactly what’s happening, so that they can draw their own conclusions (“Here’s how my blood sugars changed when I ate that second slice”). Providing people with this more immediate and engaging type of information makes them more able to co-produce their own health. And by the way, they can lose weight too.

    During this COVID-19 pandemic, people with type 2 diabetes are more vulnerable to infection than most, and also have to fear exposures that could take place in busy clinics. Access to this virtual clinic helps them manage their own health remotely, while texting and videoconferencing with health coaches and doctors as needed. Even outside of a pandemic, these digital health technologies will have a big role to play in the future of co-producing health.

    Sticking then with future-oriented topics, The Long Fix, completed before any COVID concerns had arisen, presciently calls on the US to consider the production, procurement, and maintaining of adequate medical resources and capacities (from ongoing scientific discoveries all the way to supplies, technologies, staffing, institutions) a fundamental national-security issue. You also envision a “medically ready America” that would proactively address related concerns of housing, nutrition, education, employment, and environmental quality. What would you most hope for such agendas to look like in COVID’s aftermath?

    I have the honor of serving on the Defense Health Board, a federal committee advisory to the Secretary of Defense regarding military medicine. The Military Health System aims to provide better health, better care, lower costs, and improved readiness. That readiness requires both a military force physically and mentally prepared to deploy, and a medical force professionally prepared to support them.

    In the book I try to imagine what it would mean if all US medical personnel had a similar charge: to provide a medically ready country. If the federal government had a mandate to focus on health, wellness, and prevention (instead of mostly focusing on administering a mega health-benefits system), imagine, for example, how we’d start tackling the many underlying conditions making Americans more susceptible to COVID-19 — such as obesity, diabetes, hypertension, and cardiovascular diseases. We’d make sure that critical resources like pharmaceuticals (often in short supply) and medical equipment and devices (personal protective equipment, for example) remained adequately stocked. We’d maintain a robust public-health and preventative-medicine infrastructure. We need all of that to create a more ready and resilient America.


    Portrait of Vivian S. Lee by Aaron Wolf.