Apparently, self-insured employers hot on better managing their healthcare spend are finding truth (and dollars) in Embold Health’s mantra that “quality is the best, most sustainable way to control costs.” This health tech startup is applying the old “Centers of Excellence” framework to the individual physician level; helping identify high-performing primary care docs and specialists in local markets for employers who not only want to offer their employees better quality care, but also improve the healthcare system in the communities in which they live and work.

Daniel Stein, Embold Health’s co-Founder & CEO, explains the company’s model, which is being perfected with one of the most demanding-yet-coveted “health activist” employers out there: Walmart. In this particular case, Walmart is actually incentivizing its employees to go to the providers ranked highest by Embold’s assessment, which looks at physician performance along three categories: 1) appropriateness of care; 2) outcomes; and 3) cost-effective compared to peers in-market. Backed by the robust national BlueCross BlueShield dataset, the information Embold Health is collecting, analyzing, and doling out to employers can definitely cause some health systems to take pause — and their docs to bristle. So, how does Embold Health diffuse potential blowback? Here’s where the competitive nature of local healthcare, particularly in the world of primary care, becomes clutch. Tune in to hear the details, including some very interesting stats, as well as Embold’s latest endeavors to help docs make better referrals to specialists.

There have been disturbing reports of hospitals firing doctors and nurses for speaking up about inadequate PPE. The most famous case was at the PeaceHealth St. Joseph hospital in Washington, where Dr. Ming Lin was let go from his position as an ER physician after he used social media to publicize suggestions for protecting patients and staff.  At Northwestern Memorial Hospital in Chicago, a nurse, Lauri Mazurkiewicz warned colleagues that the hospital’s standard face masks were not safe and brought her own N95 mask. She was fired by the hospital. These examples violate a culture of safety and endanger the lives of both patients and staff. Measures that prevent healthcare workers from speaking out to protect themselves and their patients violate safety culture. Healthcare workers should be expected to voice their safety concerns, and hospital executives should be actively seeking feedback from frontline healthcare workers on how to improve their institution’s Covid-19 response.

Share power with frontline workers

According to the Institute for Healthcare Improvement, it is common for organizations facing a crisis to assume a power grab in order to maintain control. As such, it is not surprising that some hospitals are implementing draconian policies to prevent hospital staff from speaking out. While strong leadership is important in a crisis, it must be balanced by sharing and even ceding power to frontline workers. All hospitals want to provide a safe environment for their staff and high-quality care for their patients. However, in a public health emergency where resources are scarce and guidelines change daily, it’s important that hospitals have a systematic approach to keep up.

Learn from the aviation industry

The aviation industry is an example of an industry that has benefited tremendously from protocolizing safety culture. Flight recorder data from the 1970s suggested that many aviation disasters occur because communication breaks down between crew members. The KLM Flight 4805 crash in 1977 and the United Flight 173 crash in 1978 were caused by captains overruling the concerns of their copilots. For this reason, the aviation industry created the Crew Resource Management (CRM) protocol. CRM emphasizes teamwork and communication and is based on the premise that everyone: captains, co-pilots and flight attendants alike, have equal responsibility in maintaining safety. CRM harkened a paradigm shift from the previous culture of ‘captain is god’. The incorporation of CRM into aviation training is a major reason why aviation is much safer today.  Hospitals should strive to create a similar safety culture in responding to COVID, where physicians, nurses and hospital leadership have equal responsibility in maintaining the safety of patients and providers.

How to maintain safety culture during a pandemic

Below, we have provided some principles and examples for a healthcare focused “CRM” to help hospitals maintain safety culture during the Covid-19 pandemic.

1. Adopt a team mindset: Hospitals are large hierarchical institutions often driven by top-down management. This can lead to operational disasters such as dwindling PPE supplies, prolonged turnaround times for COVID testing, inadequate safety protocols, and more. Hospital leaders should make themselves accessible to the input of frontline workers. They should view their staff as not just pairs of hands, but as employees with crucial knowledge to improve the hospital’s Covid-19 response.
2. Solicit feedback from frontline workers: Hospital leadership should actively solicit feedback from frontline workers to identify opportunities for improvement. At Stanford, we hold weekly virtual town hall meetings where all staff are invited to virtually send questions and concerns to Department of Medicine leadership. The answers are recorded and then made available on the internet for anyone to access. We also conducted focus groups with representatives from across Stanford Healthcare. Topics which consistently arose during discussion include access to appropriate PPE, being exposed to Covid-19 at work and taking the infection to family and not having rapid access to testing if needed, just to name a few.
3. Communicate with regularity and transparency: Covid-19 is a rapidly evolving situation. Timely and transparent communication regarding the state of the hospital and foreseeable challenges is critical to maintaining provider trust.  Many difficult truths will arise during these conversations, but if they are not shared, they will never be addressed.  At Stanford, we receive daily newsletters on how many SARS-cov2 tests were run in the last 24 hours, how many were positive, how many patients are on the floor/ICU and other metrics. We are also updated on how many days of PPE we have available, and what additional sourcing options the administration is pursuing in case we run out.
4. Maintain psychological safety: Psychological safety is defined as the belief that one will not be punished or humiliated for speaking up. Psychological safety occurs when there are safe spaces that expect individuals to speak up. The goal is to encourage honest feedback from every corner of the hospital, from the linens department to the intensive care nurses. Psychological safety is destroyed when management implements draconian consequences for speaking up, as was the case at St. John’s Health Center in Santa Monica, CA when nurses were suspended after they requested N95 masks while working in a coronavirus ward. In addition to the ongoing pandemic, another crisis emerges when healthcare workers are afraid to speak up.
5. Establish expectations of speaking up: Workers should not only feel empowered to express concerns regarding safety, they should know that there’s an expectation that they will do so. There are many widely available tools that healthcare workers can use to express and escalate safety concerns. At Stanford, one of the commonly used tools is ARCC (Ask a question, Request a change, voice a Concern, Chain of Command). For example, a resident physician could inform their chief residents, “Are you aware that we are only being limited to one N95 mask per week? I’d like to request more masks as I am concerned that reusing N95 masks will increase our risk for contracting Covid-19. This is a safety issue that endangers both residents and patients, I’d like to discuss this issue with the department of medicine leadership”. 

Healthcare workers should not be punished for speaking up, especially when their primary goal is to protect themselves and patients. We hope the above suggestions will help hospitals maintain a strong safety culture during their Covid-19 response. Maintaining a safety culture during the time of Covid-19 is not just important, it’s a matter of life or death.

Lisa Shieh, MD, PhD is a clinical professor in the division of hospital medicine and department of medicine at Stanford Healthcare, as well as an associate chief quality officer for Stanford Healthcare and medical director of quality improvement programs for graduate medical education.

Jingyi Liu, MD is a resident physician in internal medicine at Stanford Healthcare.  Twitter: @JingyiLiu8

Atul Gawande’s piece in the New Yorker elicited . In today’s post John Goodman writes .

This is Atul Gawande, writing about The Cheesecake Factory in The New Yorker:

You may know the chain: a hundred and sixty restaurants with a catalogue-like menu that, when I did a count, listed three hundred and eight dinner items (including the forty-nine on the “Skinnylicious” menu), plus a hundred and twenty-four choices of beverage.

How many different dinners — say with two food items and one beverage — can you draw from 308 food choices and 124 beverages? I used to know how to do this. It must be in the millions. So how do you make that work? Timing is everything:

Computer monitors positioned head-high every few feet flashed the orders for a given station. Luz showed me the touch-screen tabs for the recipe for each order and a photo showing the proper presentation. The recipe has the ingredients on the left part of the screen and the steps on the right. A timer counts down to a target time for completion. The background turns from green to yellow as the order nears the target time and to red when it has exceeded it.

The restaurant doesn’t just get plates on the table, however. It aims for perfection:

At every Cheesecake Factory restaurant, a kitchen manager is stationed at the counter where the food comes off the line, and he rates the food on a scale of one to ten. A nine is near-perfect. An eight requires one or two corrections before going out to a guest. A seven needs three. A six is unacceptable and has to be redone.

Gawande wants to know how we can make the health care system more like this restaurant. He envisions that this is the goal of health reform. I want to ask a more perverse, but perhaps more instructive question: What if we wanted to make the restaurant look like the health care system. What would we have to do?

Like Gawande, I too have been fascinated by the efficiency with which restaurant chains function. I’ve also talked to managers and owners about how they do it. My conclusion: a modern hospital might indeed function like The Cheesecake Factory. That it doesn’t is not the fault of doctors wedded to tradition or unimaginative hospital managers. It’s the fault of public policy.

Let’s return to my question: What public policy changes would we need to enact to make The Cheesecake Factory resemble a typical hospital?

To begin with, we need to eliminate out-of-pocket payment for restaurant food. To get that done, we need laws that encourage, or perhaps even require, restaurant food insurance. There can be many such insurers and each negotiates prices with the restaurants.

Third-party insurance immediately changes the incentives of the restaurant in radical ways. To begin with, we can dispense with the intricate and costly systems that achieve near perfect timing and quality control. So what if the hot meal doesn’t arrive at the table when it’s hot. Or if the ice cream arrives after it’s melted. Or if the two side dishes don’t arrive at the same time as the entree. Customers may be unhappy. But remember, food is now free at the point of consumption. We have patrons lined up outside, waiting to get in. The quality of the dining experience can decline quite a lot and not hurt the restaurant’s bottom line one whit.

To make sure of that fact, we can pass certificate of need laws for restaurants, just to make sure new entrants can’t steal customers away. We can suspend antitrust laws and allow the restaurant to buy up its competitors, just to make sure they don’t expand the market by increasing their capacity.

The elaborate systems designed to ensure proper timing and quality control will be replaced by equally elaborate systems designed for a new purpose: maximizing against the third-party payment formulas. Does toast and butter served separately command a higher fee than buttered toast? Then make sure we’re always out of buttered toast. Does BlueCross overpay for chicken added to Caesar salad? Then always rave about Caesar with chicken when a BlueCross diner comes in the door. Does Aetna underpay for the additional chicken? Then make sure you discourage that choice when an Aetna customer arrives.

I believe that skillful maximizing against third-party payment formulas is every bit as complicated, time-consuming and expensive as meeting the needs of cash paying customers. In fact it’s more so, for the following reasons.

On the demand side, the biggest problem with third-party payment is “moral hazard.” When food is free, people will select the most expensive items on the menu. They will order food they don’t need. They will order food they don’t even eat — and leave it on the plate! To deal with this problem, the insurers will have to invoke all kinds of rules and restrictions on what can happen in a restaurant. For starters, the insurer will greatly restrict the number of items it will pay for. Out of millions of possible food orders, it will pay for only a small subset. Instead of 30 different kinds of pasta, say, it might pay for only three. Then among the items it will pay for, the insurer will limit what any one customer can have. For example, you might be allowed ice cream or pie, but not both. The two together constitute “unnecessary” consumption. To enforce this rule, it might require servers to get pre-approval before placing a customer’s order. Or, it might just refuse to pay any bill that has the words “pie a la mode” written on it.

Then, of course, we will need a law prohibiting the corporate practice of food preparation. Food should be left to chefs, not to profit-seeking MBAs. But won’t the chefs’ decisions be tainted by the profit-making side of the restaurant? No problem. We’ll just pass a Stark law making it illegal for them to share in the profits or losses.

To put this in perspective, consider the problem of how much of each type of food the restaurant should order to be ready to meet the customers’ wants. Here is how Gawande describes the problem of wasted food:

Although the company buys in bulk from regional suppliers, groceries are the biggest expense after labor, and the most unpredictable. Everything — the chicken, the beef, the lettuce, the eggs, and all the rest — has a shelf life. If a restaurant were to stock too much, it could end up throwing away hundreds of thousands of dollars’ worth of food. If a restaurant stocks too little, it will have to tell customers that their favorite dish is not available, and they may never come back.

Remarkably, here is how The Cheesecake Factory handles this problem:

The company’s target last year was at least 97.5 percent efficiency: the managers aimed at throwing away no more than 2.5 percent of the groceries they bought, without running out. This seemed to me an absurd target. Achieving it would require knowing in advance almost exactly how many customers would be coming in and what they were going to want, then insuring that the cooks didn’t spill or toss or waste anything. Yet this is precisely what the organization has learned to do. The chain-restaurant industry has produced a field of computer analytics known as “guest forecasting.”

So if we want to end all this efficient ordering and make the restaurant resemble a typical hospital, how do we do that? Make the ordering of food the sole prerogative of the chefs and insulate them from the economic consequences of their decisions in the manner described above.

One more thing to consider: how often do the providers find it necessary to change whatever it is that they are doing?

Every six months, The Cheesecake Factory puts out a new menu. This means that everyone who works in its restaurants expects to learn something new twice a year. The March 2012, Cheesecake Factory menu included thirteen new items. The teaching process is now finely honed: from start to finish, rollout takes just seven weeks.

Contrast this with the experience in medicine:

One study examined how long it took several major discoveries, such as the finding that the use of beta-blockers after a heart attack improves survival, to reach even half of Americans. The answer was, on average, more than fifteen years.

So how do we get rid of the restaurant’s nimble response to market demand? Again, let the chefs make the decisions about what to prepare and how to prepare it, but completely insulate them from the economic consequences of their decisions. Remember, under rationing by waiting a restaurant doesn’t have to worry very much about whether it is responding to changes in demand. If patrons pay a price of zero for their food, the food has to be worth only a little bit more than zero to be a good buy.

There, I have shown you how to make The Cheesecake Factory function like a typical hospital. Any takers?

Ah, but if you want to move in the other direction — making the hospital look like The Cheesecake Factory — then you have to start repealing laws, not passing new ones.

John C. Goodman, PhD, is president and CEO of the National Center for Policy Analysis. He is also the Kellye Wright Fellow in health care. His Health Policy Blog is considered among the top conservative health care blogs where health care problems are discussed by top health policy experts from all sides of the political spectrum.

By PAUL KECKLEY

According to Gallup surveys, four of five Americans believe the quality of care they receive is good or excellent, and the majority think it is the best available in the world. Surveys by Roper, Harris Interactive, Kaiser Family Foundation, Harvard’s Chan School of Public Health, and others show similar findings. And the public’s view hasn’t changed in two decades despite an avalanche of report cards about its performance, a testy national debate about health reform and persistent media attention to its shortcomings and errors. But is the public’s confidence in the quality of the care we provide based on an informed view or something else? It’s an important distinction.

Two considerations are useful for context:

Measuring quality of care objectively in the U.S. system is a relatively new focus. And we’re learning we’re not as good as they think we are. Historically, the public’s view about “quality of care” has been anchored in two strong beliefs: 1-the U.S. system has the latest technologies and drugs, the world’s best trained clinicians and most modern facilities, so it must be the best and 2-the care “I receive” from my physicians and caregivers is excellent because they’re all well-trained and smart.

These beliefs are virtually unchanged since 2001 per Gallup. But since the turn of the Millennium, we’ve learned we’re probably not quite as good as they think we are. Three reports sparked the birth of the modern quality improvement era in our system almost 20 years ago:

• In 1999, the Institute of Medicine published To Err is Human: Building a Safer Health System concluding “as many as 98,000 people die in any given year from medical errors that occur in hospitals.” Patient safety and medication error were its central foci prompting every hospital to examine its medication management processes and related clinical operations.
• Shortly after, in 2001, it published a sequel, Crossing the Quality Chasm: A New Health System for the 21st Century expanded quality beyond safety to include care effectiveness, patient centeredness, efficiency, equitable access and timeliness. And it put an uncomfortable spotlight on unnecessary care and its pervasiveness in our system.
• And in 2003, a team of RAND researchers found gaps in quality pervasive: “Participants received 54.9% of recommended care. We found little difference among the proportion of recommended preventive care provided (54.9%), the proportion of recommended acute care provided (53.5%), and the proportion of recommended care provided for chronic conditions (56.1%). Among different medical functions, adherence to the processes involved in care ranged from 52.2% for screening to 58.5% for follow-up care. Quality varied substantially according to the particular medical condition, ranging from 78.7% of recommended care for senile cataract to 10.5% of recommended care for alcohol dependence…The deficits we have identified in adherence to recommended processes for basic care pose serious threats to the health of the American public. Strategies to reduce these deficits in care are warranted.”

And in tandem with these results, data from the Dartmouth Atlas showed widespread variation in Medicare spending and practice patterns across the country leading its iconic leader, Jack Wennberg, to offer that as much of one third of Medicare’s spending is wasted on unnecessary care and the zip code where a person lives a keen predictor of the quality of care the public gets.

Wow. Say it aint so. Can it be that quality is not uniform across the U.S. system? Can it be that some doctors get better results than others and some hospitals are safer for patients than others? Is it true that some approaches to care get better outcomes than others?

Inside the industry, these studies and hundreds since have revealed widespread variation in the quality of care we provide our patients. But the public remains largely unaware, and fewer than one in ten is predisposed to study our methods and results closely.

Quality of care in the U.S. cannot be readily compared to quality of care in other systems of the world. Data about the healthcare systems in 35 developed countries from the Organization for Economic Cooperation and Development allow comparisons of life expectancy, morbidity, access to providers and admission rates to hospitals among other metrics. For the most part, they’re accurate (though some are self-reported and dated). Supplemental analyses by academics like Robert Blendon and others also provide country comparisons. In these analyses, the U.S. system is always the most expensive, near the best in age-adjusted life expectancy, morbidity and mortality, on par with most for hospital admissions and access to specialized services, and lower than most for preventive health, public health and primary care services.

But these comparisons are misleading. Beyond the complexity of our pluralistic payment system, there are major differences between the U.S. system and other developed systems in the world:

• In the U.S., our “human services” programs like the TANF (Temporary Assistance for Needy Families), Supplemental Nutrition Assistance Program aka “food stamps” and others operate almost independently from our delivery system. De facto, the U.S. operates a “health system” that’s focused on hospitals, doctors, clinics, drugs and devices, and a set of “welfare” programs (TANF, Medicaid, Food Stamps, SSI, EITC and Housing Assistance) for 70 million lower income citizens and legal immigrants. We spend more on the health programs proportionately than other countries and less on the human services programs. That’s why private foundations, like Kresge, Robert Wood Johnson and many others supplement funding in the safety net. In other countries, safety net services for more directly integrated in their care delivery strategies; in the U.S., they’re not. So larger investments in safety net programs in other countries and their integration into their delivery systems are major differences between the U.S. system and others.
• The U.S. has unprecedented health challenges: the highest rates of suicide, gun violence and substance abuse in the world. The facts are startling: every day in the U.S., 123 commit suicide, 43 die from gun violence, and 175 will die from a drug overdose. It’s the health system that absorbs the responsibility for and expense associated with these deaths. No other country comes close.
• And in most developed systems, their federal/provincial government plays a larger role in paying for healthcare and thereby determining what’s appropriate and inappropriate care. Most use a strong primary care front door to their system, so preventive health and referrals for specialty care are appropriately maintained. Most have a mechanism whereby decisions about major interventions of guidelines for diagnosing and treatment are evidence-based and followed closely. Most negotiate with drug, device and technology suppliers directly and get significant discounts vs. what’s paid in the U.S. And most have a global budget for their health and human services investments, forcing regulators and providers to establish priorities and address tough decisions about end of life care, the usefulness of costly technologies and more.

Public opinion surveys in countries like France, Switzerland, the UK and others show higher levels of satisfaction with their systems that ratings of our system by Americans, so the U.S. system is NOT the world’s most popular as viewed by its constituents. It is our system’s complexity, uneven access and administrative red tape that push our ratings down while the majority believe our quality is “the best in the world”.

Here’s my take:

The quality improvement movement in the U.S. system has had profound impact. Clinicians and academicians have improved clinical processes for diagnosing and treating specific patient populations, addressing variability for virtually every diagnosis specific to signs, symptoms, risk factors, patient values and social determinants of their health. It has made household names of Deming, Juran, Crosby in healthcare C suites, recognition as Codman, Eisenberg, and Baldridge desirable and the roles of the National Quality Forum, National Association for Quality Assurance and others all the more relevant to our system’s future. The results of these efforts are clear and positive.

Health services researchers have correlated adherence to evidence-based clinical practices with better outcomes and lower costs. Accreditors and regulators have crafted rules and regulations based on process measures for which hospitals, physicians, and post-acute providers can he held accountable. Government agencies have become more aggressive in scrutinizing quality. And the sweeping change in incentives for providers from volume to value is premised on the assumption that achieving evidence-based thresholds of quality a basis for participating in savings. All these are derivatives of the quality improvement movement in the U.S. system about to begin its third decade.

The bottom line is this: there is no standard definition of quality in the U.S. health system but every sector is paying more attention. Hospitals have been the focus for most government-initiated efforts since they’re 32% of total spending, then physicians. Drug company scrutiny about the efficacy and effectiveness of their compounds and the underlying clinical research is getting more attention and how insurers manage their coverage and denial decisions is under the spotlight.  How regulators define and the public responds to “quality of care” appropriated through emergent disruptor-led systems sponsored by Amazon, CVS, Apple, and others is the next chapter.

We are improving, especially in high volume patient populations and care coordination but there’s much more to be done. Our society’s challenges around guns, drug abuse and income disparity render comparisons to other countries’ quality of care moot.

The U.S. public believes our quality of care is the world’s best. They think it’s unaffordable and complicated, but the world’s best based on what they’ve experienced for themselves. It’s a belief that’s strongly held, but not entirely based on an informed view of facts.

They’ve not been duped, but many might change their minds if they understood the gap between the quality they think we deliver and the facts. We’re making progress but we have a long way to go.

Data is not always the path to identifying good medicine. Quality and cost measures should not be perceived as “scores,” because the health care process is neither simplistic nor deterministic; it involves as much art and perception as science—and never is this more the case than in the first step of that process, making a diagnosis.

I share the following story to illustrate this lesson: we should stop behaving as if good quality can be delineated by data alone. Instead, we should be using that data to ask questions. We need to know more about exactly what we are measuring, how we can capture both the physician and patient inputs to care decisions, and how and why there are variations among different physicians.

A Tale of Two Doctors

“As soon as I start swimming, my chest feels heavy and I have trouble breathing. It is a dull pain. It is scary. I swim about a lap of the pool, and, thankfully, the pain goes away. This is happening every time I go to work out in the pool”.

Her primary physician listened intently. With more than 40 years of experience, the physician, a stalwart in the medical community, loved by all, who scored high on the “physician compare” web site listing, stopped the interview after the description and announced, with concern, that she needed to have a cardiac stress test. The stress test would require walking on a “treadmill” to monitor her heart and would include, additionally, an echocardiogram test to see if her heart was being compromised from a lack of blood flow.

“But, I have had three echocardiogram tests in the last year as part of my treatment for breast cancer and each was normal. Why would I need another”?

“Well, I understand your concern about more tests, but the echocardiograms were done without having your heart stressed by exercise. The echo tests may be normal under those circumstances, but be abnormal when you are on the treadmill. You still need the test, unfortunately. I want to order the test today and you should get it done in the next week”.

The patient had other ideas. She refused the stress test and, instead, sought a second opinion. The second physician (another experienced, well respected, multiple-board-certified physician) asked more questions than the first. The complaint was queried for the context of the pain. When did it occur? Did it occur only with swimming? How long did it last? What made it better? What made it worse? Any associated symptoms? What was its course—gradually better, or suddenly better? Were there any other medical issues in her life?

The pain abated as she exercised; it did not worsen as she swam and was gone in less than a minute. The pain never occurred in any other circumstance. The patient offered that she worked out daily and never had the problem when she ran or exercised robustly. She worked out faithfully for one hour, seven days per week and in a strenuous fashion. Her symptoms only occured during swimming. She had tried nothing to improve it, but since it had been occurring regularly, she sought help.

The physician asked about how she started her swim; she anxiously jumped in the pool, and, since she hated the sudden blast of cold water, she tensed and then swam vigorously to warm up. After listening to her lungs and her heart as part of a brief physical exam, the physician asked her to try a common sense solution, while not offering a cause. The physician advised her to enter the water slowly, perhaps, even, walk in the pool for a bit until she warmed up before swimming.

Approaches to Diagnoses Fall Outside of Data Analysis With Huge Impact
​What is going on with this person? Is she having trouble with her heart? Could this be her lung? Could she have asthma? The differential diagnosis for the cause of her symptoms is filled with worrisome conditions—or not. This is a diagnostic dilemma. What would you do? Which approach seems most reasonable to you?
This anecdote, like a TV program, is based on a true story, and I will tell you the outcome.

But, first, think about how varied were the responses of the physicians. One directed the patient to testing; the other performed no tests other than the history of the complaint and a focused physical exam. One physician’s care will be expensive, but will the stress test ordered by the first ease the worry of both the patient and the physician? Why did the first physician focus on the most serious cause and the second on the most likely cause?

The differences in approaches to this same patient are huge. Many quality improvement efforts aim to assess variations in physicians’ practice patterns. But these efforts of comparison are hampered on many fronts; foremost, that different patients, cared for by different physicians, vary in clinical context and disease burden. This blog’s story differs, pointedly. The patient variation is gone; the patient is the same, the variation is in the physicians’ responses to the patient’s complaint.

There is little information about the variations in how physicians approach a patient’s diagnostic dilemma beyond aggregate costs of care. This is a deficiency since physicians’ variation may surpass patients’ variation, as in this situation.

Variations in Care Are Not Fully Explained by Utilization in Care
A search for a diagnosis requires appropriate context. The complaint of “chest pain” has myriad causes; the complaint of “chest pain when jumping into a cold pool that abates as exercise progresses” has a much smaller list of causes. Too often, medical care improvement, safety and science truncate the discovery process regarding important aspects of a patient’s condition when a concerned physician fails, for whatever reason, to capture the nuance that is essential to medical care for an individual.

The second physician teased out the nuances of the complaint via questions and a physical exam appropriate to the contextual clinical situation. While we know, for example, that the physical exam done routinely will likely yield nothing of significance, a physical exam that aims to test a hypothesis based on the full “picture” of the patient’s complaint will bear more fruit.

More information about the value of a physician-patient encounter will always be found in the content of their communication than in what they ultimately do. The difference in these physicians’ behaviors will not be found in any database, electronic medical record, or machine-learning algorithm. I have yet to see data on the contextual information from a history of the present illness in any data set or quality improvement initiative.

Performance Improvement Must Include Both Physician and Patient Communications
We keep insisting on bland, poorly conceptualized ideas about how to improve care. We keep insisting on non-contextualized “data,” as if, somehow, that data includes something of value about what is going on between patient and physician. Medical care is a cottage industry of two and can be nothing else. It should be measured as such.

And now, the outcome: The patient followed the advice of the second physician. She changed her routine and now warms up before swimming; her complaint is gone and has not returned.

​She also changed physicians.

In an age where big data is king and doctors are urged to treat populations, the journey of one man still has much to tell us. This is a tale of a man named Joe.

Joseph Carrigan was a bear of a man – though his wife would say he was more teddy than bear.  He loved guitar playing,  and camp horror movies.  Those who knew him well said he had a kind heart, a quick wit and loved cats.

I knew none of these things when I met Joe in the Emergency Department on a Sunday afternoon.  I had been called because of an abnormal electrocardiogram – the ER team was worried he could be having a heart attack.  Not able to make sense of the story on the phone, I was in to try to sort it out.  Joe was gruff, short with his answers – but clearly something just wasn’t right.   He was only 54 but had more problems than the average 50 year old.   Progressive calcification of his aortic valve  some years ago  had caused intolerable shortness of breath resulting in  replacement with an artificial valve. Longstanding diabetes had resulted in kidney failure and dialysis,  and most recently abnormal liver tests had  revealed the presence of the early stages of cirrhosis from hepatitis C.  Yet Joe continued to live an active life – with only a tight circle of family and friends aware of the illnesses beneath the surface.

But on this particular day it was readily apparent Joe was not well.  It didn’t take long to figure out that Joe had an infection somewhere.  He had come in feeling hot at times, but having chills at others.  Review of his ECG, stored telemetry in the ER revealed a conduction disturbance not infrequently seen called a left bundle branch block, as well as a rhythm disturbance called atrial fibrillation , but no evidence of a heart attack  that had prompted the ER to call me.   I told Joe and is anxious wife that he needed to be admitted to the hospital to find where the infection may be coming from.

Joe had a number of places he could be harboring an infection.   While dialysis maybe one of the modern marvels of the world, the concept is fairly crude. In the absence of functioning kidneys an artificial machine does the work of the kidneys. Via large catheters the patient’s blood is removed, circulated through the machine and then returned to the patient.  The flow required to circulate enough blood over a course of three hours involves creation of large conduits between arteries and veins, or placement of artificial grafts.   Every access of these vessels with dialysis sessions  carries a risk of introducing bacteria into the bloodstream, making infections a common event in this population.  Complicating matters further, Joe had an artificial heart valve,  which lacking the natural immunity of native heart valves,  are especially prone to seeding by bacteria.

I told the admitting team that if no other source of infection became immediately obvious he would need an ultrasound to more closely examine his valve. That night I was called by the overnight resident in the hospital because Joe had a low heart rate. He felt fine,  and his blood pressure was OK .  I stopped a heart rate lowering medication that he had been on hoping this may be the cause.   The next morning a review of his electrocardiogram and recorded telemetry demonstrated an ominous finding – episodic heart block.   The heart’s four chambers are segregated into two small upper chambers that contain the pacemaker function of the heart and two lower chambers  that are the powerful mechanical pumps which circulate blood.   The upper chambers are electrically insulated from the lower chambers except for narrow specialized bundles of tissue that function as electrical cables.   The specialized conduction fibers form an intricate lattice that allows the normal heart to contract  in less than 100 ms.  The weak point is a narrow isthmus in close  proximity to the  aortic valve that the electrical bundles must navigate near their origin.   In Joe’s case, the electrical bundles were conducting in a stuttering fashion because bacteria was eating away at precious cardiac tissue surrounding the artificial valve.

There is only one treatment for this: surgery.  John needed a surgeon to open his chest, take out his infected  aortic valve, wash out the abscess and put another valve  in its place .

I called the surgery team immediately –  Joe was clearly a high risk case.   The  presence of kidney disease, liver disease, and prior open heart surgery were all independently high-risk markers.  Joe of course had all three.  Yet, he was an active functional 54-year-old. What was the other option? Let him die?  My initial conversation with the surgeon was a positive one, and surgery was scheduled for the following day.

The following day brought a decidedly different tone.   There was a brief discussion about a conversation with the hepatology team having changed the surgeon’s mind.  A prediction score for patients with liver disease suggested there was a 50% chance he would be dead in 6 months regardless of his current life threatening cardiac abscess.  That score seemed to be an overestimate, and hepatitis C is now a treatable illness but regardless, I pushed back and responded that the chances he would be dead in 2 weeks with his current condition was 100%.  After some more parrying back and forth – the real reason for the cold feet emerged.  Outcomes of cardiac surgery are publicly reported for all institutions and surgeons nationally, and in the state of Pennsylvania.  After discussing the case with the other surgeons at the institution, the decision at this hospital had been made: Joe had too high a chance of making the institution look bad.

How did we get here?

Public reporting

The desire for public reporting arises from an attempt to broadcast value to a public eager for information.  The Institute of Medicine’s seminal report in 1999 suggesting 98,000 patients die due to preventable medical errors every year added a sense of urgency and necessity to the push to grade providers.

Public reporting actually began more than thirty years ago – in 1984 – when the Health Care Financing Administration (HCFA), now known as the Centers for Medicare and Medicaid Services (CMS), began to publicly report the hospital mortality rates of Medicare patients.  The agency identified 269 hospitals that were outliers with regards to death rates. Although the analysis attempted to control for a variety of risk factors, it was heavily criticized and eventually HCFA stopped publishing the data.

There was no turning back though. New York state and Pennsylvania followed suit and started publicly reporting cardiac surgery outcomes.  The group in charge of reporting in Pennsylvania was the Pennsylvania Health Care Council (PHC4) -created by the Pennsylvania General Assembly with the charge of improving the quality of care and restraining health care costs.

As the HCFA learned, reporting outcomes is a complicated business.  Reporting outcomes  relies heavily on the ability to risk stratify patients.  The PHC4 struck on what is now a common model – calculate the expected mortality of each patient based on who they are and how they present and compare this to the actual reported mortality.  A number of factors are considered, and each factor is weighted with regards to its impact on mortality with surgery.

Figure 1 Clinical Predictor data table, Calculating expected mortality

The final data is then presented for the public’s perusal in an easy to understand table with actual and expected mortality for every hospital and surgeon in Pennsylvania.

The figure below demonstrates the results of three large hospital systems in the Philadelphia region.

Figure 2 . Observed/Expected Hospital mortality

Teaching to the test

Proponents of public reporting will point to improved outcomes in the era of public reporting.  In Pennsylvania, for instance, hospital mortality for Bypass (CABG) surgery dropped from 3.2% to 1.5%, and for Valve surgery dropped from 5.2% to 2.7%.

Figure 3. Cardiac surgery in-hospital mortality

The only thing more impressive than this is the drop in bypass procedures done statewide between 2000 and 2015 – which reflects an almost 60% drop from 20,029 procedures to 7,962 procedures.

Figure 4. Volume of cardiac surgery

An optimistic conclusion to draw from the data is that there are fewer unnecessary procedures being performed and those that are being done are of higher quality in a safer environment. The more troubling explanation is that surgeons are operating less, and improved mortality relates to avoiding high risk patients.

I spoke to a cardiac surgeon who had practiced in the public reporting era in New York State to ascertain his thoughts.  He told me that “without a doubt” this was a consideration that impacted who he and his colleagues would take to the operating room.  He spoke of how surgeon’s would “play the game” to improve their numbers by not necessarily denying the sickest patient, but rather waiting them out.  He gave me an example.

One of the stronger clinical predictors that impacted expected mortality was cardiogenic shock.  Cardiogenic shock refers to patients who suffer a heart attack so severe that their entire cardiac function is compromised to the point that the heart is unable to deliver an adequate amount of blood flow to vital organs such as the kidney, liver and brain.  A common approach by surgeons was to place these patients in the cardiac critical care unit or transfer the patient – the patients that survived a week were offered surgery.  The cold calculus makes sense – the expected mortality of the patient continued to be high as the patients were still labeled as cardiogenic shock – but clearly waiting allowed selection of a lower risk group of patients.

I despaired after hearing this about finding published data to support what I had heard, as I frequently find datasets simply aren’t granular enough to represent clinical practice. In this particular case I was rescued by the SHOCK registry that compared patients presenting in cardiogenic shock in NY patients and Non-NY patients.

Figure 5. Differences in treatment of Cardiogenic shock patients in NY and nonNY patients

The results are stunning.  As highlighted in the above table – the time to cardiac surgery in patients presenting with a heart attack and shock in New York State was different by almost 4 days!

The footprint of risk aversion is found even in the published data supportive of public reporting.  Peterson et. al, found declining CABG mortality in NY state with no apparent decrease in access, but troublingly, could not elucidate a mechanism for the decline in mortality.  If the point of profiling hospitals and surgeons as good or bad was to direct more patients to high performing centers, NY state officials found no evidence of migration of patients from high to low mortality hospitals.

Surveys of cardiac surgeons bear out that enmasse,  public reporting has markedly altered practice based on public reporting primarily by denying patients surgery.

Figure 6. Cardiac surgery survey results of public reporting

It used to be that a decision on surgery was between surgeon and patient.  Risks were outlined, and a decision arrived at.  The patient and doctor were in it together. Surgeons operated on high risk patients as long as they felt comfortable the patient and family understood the risks.  This construct exists no more.  Especially in an era where surgeons are employed by health care systems, the surgeon is now beholden to masters that never step into hospital rooms.  Surgeons with high mortality rates that make the institution look bad face serious repercussions, and even worse, put future employability at risk. Shared decision making is a joke – not because decision making isn’t shared but because the shared decision is between surgeon, risk score, and hospital system – the sickest patients don’t have a choice anymore.

Back to Joe

And so it was with Joe. Faced with certain death, he wanted to live, but I could not find a surgeon at my institution to operate on him.  The hyperefficient health care system moved quickly in this case.  The plan from the CCU team was now hospice/palliative care.  I couldn’t stomach it.  I told Joe that I was going to find a surgeon who would give him a shot.  So I called a center that did the most valve surgeries in the city.  This wasn’t the first time in my years in practice that a patient of mine had been turned down by a surgeon at a local institution.  Smaller volume centers are even more susceptible to risk aversion because their smaller volume amplifies the effects of bad outcomes.   The largest center in the city had bailed my patients out before from this predicament.

This time was different.  Unbeknownst to me, the academic nationally renowned busiest center in the city had been taken to task in the just published Pennsylvania public report  because they had been found to have worse than expected mortality.   The expected mortality was 1.2% – 4.1%, and the observed mortality was 4.3%.  Of 345 patients who underwent bypass surgery 15 died – one fewer death would have put the center in the expected range.  To rub salt in the wound, the local paper ran a story on the report and quoted a competing health system’s surgeon as noting the difference could relate to a minimally invasive surgery they did more of.  Not mentioned was the fact that the smaller competing health system routinely sent their sickest, most complex patients to the larger academic center.

So, the answer again was no.  Undeterred, I called Johns Hopkins next.  The cardiac surgeon patiently listened to my story – and inquired why no hospital in the city would operate on him.  I told him about public reporting, and he in turn told me about Maryland’s global payment system.  Maryland hospitals were not paid per admission anymore, they were paid a per capita amount related to the number of patients attributed to them.  The positives are that hospitals are incentivized to set up systems to keep patients out of hospitals.  The downside is that they have a powerful disincentive to take on a complicated high risk out of state patient like Joe.  I was told bluntly that Maryland was not paying for this.

I discussed the case with a prominent cardiothoracic surgeon who said explicitly –

“Politically impossible to do this case… Honestly, the big picture: he is a casualty of the public reporting system we have in the US and Pennsylvania where risk aversion is always with us.  It’s kind of sad.  There are consequences to all the decisions society makes”

Consequences.  Trade-offs.  This is decidedly not what the public hears.   The public hears things like the ‘triple aim’ – improve patient experiences of care, improve the health of populations, and reduce the per capita cost of health care.  You can have your cake.. and eat it too!  But there is a cost to all this, and it is a cost borne by our sickest, most vulnerable patients.

When CMS announced a policy in 2006 that attempted to restrict coverage of weight loss surgery to centers of excellence (COE), a study noted the unintended consequence of trying to make the surgery safer was a reduction in the proportion of minority patients receiving the surgery.  I imagine that had Joe been on the board of trustees of the hospital he would have a different set of options.

Figure 7. Proportion of minority patients offered bariatric surgery

There are so few of these patients relative to the total that they won’t move the needle when it comes to numbers population health devotees care about like life expectancy and overall cardiac mortality.  But the impact of this culture shift extends beyond the smattering of patients affected.

America has long been the envy of the world when it came to innovative surgical techniques.  The landscape for much of American history has consisted of brash physicians willing to push the envelope in dying patients with impossible odds.  Bennie Solis was one such 3 year old dying of an irreversible liver disease when he was operated on by Thomas Starzl in 1963.  Starzl thought he had perfected the technique of transplantation in dogs before attempting the feat – but he was wrong.  Bennie bled to death on the operating table, his damaged liver no longer able to make substances that would clot blood.  Starzl and the surgical team were devastated, but the lessons learned with Bennie’s death increased the chances of success for the patients that followed.  This is true of the history of every new surgical procedure attempted.  Innovation requires risk taking not risk aversion.  It is difficult to imagine a man like Starzl taking on the sickest of the sick patients at Johns Hopkins today.

The sad part is that the age of penury driving this behavior is one where overall health care spending continues to accelerate to the tune of 3 trillion dollars annually.  Our health care overlords may save some dollars on Joe, but won’t blink when spending billions of dollars on health care accountants, data entry clerks, hospital coding specialists, and any number of low yield primary care preventions geared to the worried well.

Some think the solution is simply better risk adjusting, or exclusions for centers performing at the frontiers.  Perhaps.  Or throw the whole system out and focus on the rotten apples among us.  I don’t know.  I do know that however well intentioned the practice of public reporting may be, the consequences may be severe.

I couldn’t find a surgeon who would operate on Joe.  So he died.

He never made it home to play his guitar.

The pictures and the story are presented with the permission of Joe’s wife, Debra.  Ever gracious, she hoped his story would teach us something.

Here’s hoping it will.

Anish Koka is a cardiologist in Philadelphia.  He can be found on twitter @anish_koka

HarvardX is offering a free online course, Improving Global Health: Focusing on Quality and Safety, starting on June 27 on edX.org. Participants in this 8-week course will engage with top experts in the field of public health as they grapple with the nature of high-quality healthcare: What is quality? How do we define it? How is it measured? And most importantly, how can we make it better? Whether you’re a healthcare provider; student of medicine, public health, or health policy; or a patient who simply cares about getting good care—this course is for you. The course is taught by Ashish K. Jha, MD, MPH, director for the Harvard Global Health Institute.

To learn more and register for free, visit: http://bit.ly/2oMMsch

How much does it matter which hospital you go to? Of course, it matters a lot – hospitals vary enormously on quality of care, and choosing the right hospital can mean the difference between life and death. The problem is that it’s hard for most people to know how to choose. Useful data on patient outcomes remain hard to find, and even though Medicare provides data on patient mortality for select conditions on their Hospital Compare website, those mortality rates are calculated and reported in ways that make nearly every hospital look average.

Some people select to receive their care at teaching hospitals. Studies in the 1990s and early 2000s found that teaching hospitals performed better, but there was also evidence that they were more expensive. As “quality” metrics exploded, teaching hospitals often found themselves on the wrong end of the performance stick with more hospital-acquired conditions and more readmissions. In nearly every national pay-for-performance scheme, they seemed to be doing worse than average, not better. In an era focused on high-value care, the narrative has increasingly become that teaching hospitals are not any better – just more expensive.

But is this true? On the one measure that matters most to patients when it comes to hospital care – whether you live or die – are teaching hospitals truly no better or possibly worse? About a year ago, that was the conversation I had with a brilliant junior colleague, Laura Burke. When we scoured the literature, we found that there had been no recent, broad-based examination of patient outcomes at teaching versus non-teaching hospitals. So we decided to take this on.

As we plotted how we might do this, we realized that to do it well, we would need funding. But who would fund a study examining outcomes at teaching versus non-teaching hospitals? We thought about NIH but knew that was not a realistic possibility – they are unlikely to fund such a study and even if they did, it would take years to get the funding. There are also some excellent foundations, but they are small and therefore, focus on specific areas. Next, we considered asking the American Association of Medical Colleges (AAMC). We know these colleagues well and knew they would be interested in the question.  But we also knew that for some people – those who see the world through the “conflict of interest” lens – any finding funded by AAMC would be quickly dismissed, especially if we found that teaching hospitals were better.

Setting up the rules of the road

As we discussed funding with AAMC, we set up some basic rules of the road.  Actually, Harvard requires these rules if we receive a grant from any agency. As with all our research, we would maintain complete editorial independence. We would decide on the analytic plan and make decisions about modeling, presentation, and writing of the manuscript. We offered to share our findings with AAMC (as we do with all funders), but we were clear that if we found that teaching hospitals were in fact no better (or worse), we would publish those results. AAMC took a leap of faith knowing that they might be funding a study that casts teaching hospitals in a bad light. The AAMC leadership told me that if teaching hospitals are not providing better care, they wanted to know – they wanted an independent assessment of their performance using meaningful metrics.

Our approach

Our approach was simple. We examined 30-day mortality (the most important measure of hospital quality) and extended our analysis to also examine 90 days (to see if differences between teaching and non-teaching hospitals persisted over time). We built our main models, but in the back of my mind, I knew that no matter which choices we made, some people would question them as biased. Thus, we ran a lot of sensitivity analyses, looking at shorter-term outcomes (7 days), models with and without transferred patients, within various hospital size categories, and with various specification of how one even defines teaching status. Finally, we included volume in our models to see if volume of patients seen was driving differences in outcomes.

The one result that we found consistently across every model and using nearly every approach was that teaching hospitals were doing better. They had lower mortality rates overall, across medical and surgical conditions, and across nearly every single individual condition. And the findings held true all the way out to 90 days.

What our findings mean

This is the first broad, post-ACA study examining outcomes at teaching hospitals, and for the fans of teaching hospitals, this is good news. The mortality differences between teaching and non-teaching hospitals is clinically substantial: for every 67 to 84 patients that go to a major teaching hospital (as opposed to a non-teaching hospital), you save one life. That is a big effect.

On Oct. 14 the Centers for Medicare and Medicaid Services (CMS) released detailed regulations for implementation of the Medicare Access and CHIP Reauthorization Act (MACRA). With so many changes to the Merit-Based Incentive Payment System (MIPS) and the Advanced Alternative Payment Model (APM) track, we at Health Catalyst have heard many questions and comments. This is understandable, as the substantial 962-page proposal has grown to the 2,398-page final rule. Also, since nearly all providers will be subject to the new Quality Payment Program (QPP), understanding MACRA and what it means for providers is imperative.

Earlier this year, Bobbi Brown, Health Catalyst Vice President of Financial Engagement, gave us a better understanding of the MACRA proposal. With the help of Dorian DiNardo and Dr. Bryan Oshiro, Bobbi is back to share her insights into the MACRA final rule and its implications for providers in a highly engaging question and answer format. Bobbi and the team will share the most frequently asked questions they have received since the announcement and their answers to them.

Some of the questions covered will be:

• Do I need to report individually or as a group?
• How should physicians prepare for MACRA?
• How do I qualify for an APM?
• What should be the implementation plan?

We look forward to you joining us. Click here to register.

One of the main goals of the Affordable Care Act (ACA), perhaps second only to improving access, was to improve the quality of care in our health system. Now several years out, we are at a point where we can ask some difficult questions as they relate to value and equity. Did the ACA improve quality of care in the ways it intended to? Did it do so for some people, or hospitals, more than others?

How did the ACA Attempt to Improve Quality?

Three particular programs created by the ACA are worthy to note in this regard. The Hospital Acquired Condition Reduction Program (HACRP) took effect on October 1, 2014 and was created to penalize hospitals scoring in the worst quartile for rates of hospital-acquired conditions outlined by the CMS. The Hospital Readmissions Reduction Program (HRRP), which began for patients discharged on October 1, 2012, required CMS to reduce payments to short-term, acute-care hospitals for readmissions within 30 days for specific conditions, including acute myocardial infarction, pneumonia, and heart failure. The Medicare Hospital Value-Based Purchasing Program (HVBP) started in FY2013, was built to improve quality of care for Medicare patients by rewarding acute-care hospitals with incentive payments for improvements on a number of established quality measures related to clinical processes and outcomes, efficiency, safety, and patient experience.

Did Quality Improve?

At first glance, it is tempting to expect that quality might improve through these initiatives. Who can argue with fewer re-admissions, fewer infections caused within hospital walls, and overall greater clinical outcomes, safety, efficiency, and patient satisfaction?

However, recent data suggests that not only is quality improvement hard—it is more confusing and unexpected than we might have realized.

With regards to the HACRP, we currently lack robust outcomes data on the reductions attributable to the program. A 17% decline prior to HACRP implementation in December 2014 further complicates the picture—if there are indeed future reductions in hospital-acquired conditions, are they because of HACRP, or because rates were already on the decline? However, concerning beyond the outcomes that we don’t know are the ones that we do. A recent study in JAMA found that the hospitals most penalized by the HACRP were teaching hospitals, safety-net hospitals, hospitals with the sickest patients based on case-mix indices, and ironically, hospitals with high quality scores on other instruments. Without adequate risk-adjustment, hospitals that are already burdened with the sickest and poorest are also bearing the brunt of financial penalties.

The HRRP has created a similar strain on hospitals with poorer and sicker patients. While readmissions under the HRRP have reduced from 21.5% in 2007 to 17.8% in 2015, Disproportionate Share Hospitals (DSH) are carrying a large share of the penalty burden without appropriate adjustment to their patient characteristics (household income, race, education level, Medicaid status), which are directly related to a higher risk of readmission. The environment a patient returns to is more than likely the cause of their readmission, especially for those in poorer communities that are primarily treated at DSHs.

Recent data unfortunately suggests that the HVBP program has also not shown promising results. HVBP hospitals have not reduced their mortality rates for particularly incentivized conditions relative to pre-HVBP era rates, even when stratifying for hospitals with higher financial incentives, poorer financial health, and greater market competitiveness. Similarly, the HVBP program may not appropriately account for the complexity of patient illness, and thus physicians that care for sicker patients may be subject to financial penalty. Ultimately, this begs the fundamental question that confronts all pay for performance programs: does a financial incentive significantly change the care physicians deliver to patients?

Recommendations and Conclusions

Overall, the ACA’s main quality improvement programs need to be adjusted, particularly to address the likely unintended consequences related to worsening inequity between hospitals treating the wealthy and the poor. This can in some parts be addressed through more thorough risk-adjustment that accounts for socioeconomic patient-level factors and case-mix indices, especially for the HACRP and HRRP. Additionally, the HACRP needs to re-examine its measurement protocols, given the cited incongruence with other quality measures. The HVBP program is a bit more complicated, particularly because we may still not be far enough out to properly assess effects on mortality. Furthermore, we need to more thoroughly examine impacts on morbidity as well. The HVBP might benefit from moving away from multiple specific measures of quality, and find a way to look more holistically at the overall picture of care. It must also account for the complexity of patient care delivered at the individual level so as to reward, rather than punish, physicians caring for our nation’s sickest and most complex.

Quality improvement is a complicated issue, largely because quality has a different meaning to every patient and every provider. However, in the effort to improve quality, the ACA, as related to the three specific programs highlighted, has at best shown that P4P initiatives have much to improve upon, and at worse has caused harm to hospitals caring for the sickest and most complicated patients. This, I am certain, is not an improvement, although it carries an important lesson: in creating measures to improve care, we must be careful to first do no harm, especially to the least well off in our healthcare system and those caring for them.

Abraar Karan MD is an MPH candidate in the Department of Health Policy and Management at the Harvard T.H. Chan School of Public Health.