by Thomas H Lee
For more than a century, the Mayo Clinic has promised that “the needs of the patient come first.” Like most great healthcare organizations, Mayo has a long, complicated mission and values statement, but that phrase about the needs of patients (uttered long ago by one of the founding Mayo brothers) is considered the primary value.
Mayo clinicians do more than mouth this phrase. They have differentiating patterns of care, such as the expectation that the first specialist to see patients will continue to follow them even when the focus of care has drifted into another area. For example, if a patient referred to a heart failure specialist turns out to have lung disease, that patient is connected to a pulmonary specialist, but the first doctor continues to follow the patient and makes sure all the care is coordinated. It’s obviously better for patients, but this kind of coordination occurs only by accident at most academic medical centers.
When I ask Mayo clinicians why they are willing to do what their counterparts elsewhere will not, I always get the same answer: “Well, here at Mayo, the needs of the patient come first, and. …” They don’t roll their eyes when they repeat what might seem a platitude in other settings. They say it with the same matter-of-factness with which they say, “It gets really cold here in the winter.”
Intermountain Healthcare in Utah adopted a new, short, and direct mission statement in 2014: “Helping people live the healthiest lives possible.” This replaced their prior mission statement, which was “Excellence in the provision of healthcare to the communities in the Intermountain West.” Note the shift in focus from reliability in what providers do to the more challenging, nebulous, but aspirational goal of meeting the needs of patients.
These and other effective statements of shared purpose have key features in common. First, they are unequivocally focused on patients. Second, they describe an aspirational goal—meeting patients’ needs—that will never be completely achieved. Improvement will always be possible and an imperative. Part of the subtext of these shared purposes is that hard work will not be enough; it will take greater group effectiveness to make progress.
Appreciative Inquiry
A valuable tool for fleshing out shared purpose is appreciative inquiry: focusing on the positive, identifying the cases that make caregivers proud, and trying to make the key elements of great care happen reliably instead of by chance. Cleveland Clinic makes every employee spend half a day in appreciative inquiry exercises with other staff members. Hennepin County Medical Center in Minneapolis has had “patient rallies” in which staff members sit with patients to discuss cases in which care has been excellent.
At the hospital where I practice, Brigham and Women’s Hospital in Boston, the emergency department began a multi-year journey that took its overall patient experience ratings from the sixth percentile to the ninety-ninth percentile by discarding its negative patient comments and focusing on the positive ones. The messages in the positive comments were the flip side of the negative ones. Patients appreciated when they were seen promptly and treated with respect and when coordination of care with their other clinicians was excellent. These positive comments shaped the Brigham emergency department’s overall goal: “VIP care for all.”
Of course, the emergency department needed much more than a slogan. It received more space from the hospital to expand its capacity. Department leaders created even more space by using lean management techniques, through which they concluded that the triage function was using space and personnel that could be better employed in direct patient care. And they deployed a range of financial and nonfinancial incentives to reward VIP care for all. The result was improvement in patient experience to the very top tiers of emergency medicine.2
The work began with the articulation of a shared purpose. It put the complex choices that had to follow in perspective and made clear which goal should be prioritized when there were difficult decisions to be made.
Financial Incentives
Financial incentives certainly catch people’s attention. After all, almost everyone is motivated in part by financial incentives and job security. Even if their organization’s shared purpose resonates deeply with them, people sit up and pay attention to whatever measures are being used to assess their performance. They want to know what data are available for those measures and how the data are analyzed. This interest is intensified tremendously if financial incentives are based on those measures.
The attention that results from financial incentives can do both good and harm. As has been noted, financial incentives draw attention to topics that might otherwise be ignored. However, when data are used for judgment—loss of a financial incentive qualifies as an adverse judgment outcome—the natural response is to try to reduce the risk of loss for that incentive, make the targets as easy to attain as possible, and raise concern about the quality of the data and the analyses.
There is no single right way to use financial incentives, but there are certainly some mistakes to be avoided. Some organizations make portions of clinicians’ compensation dependent on patient experience performance and are still struggling to find the best way to do it. Pennsylvania-based Geisinger Health System ties 20 percent of physicians’ compensation to their performance against certain goals, but improved patient experience ratings account for a very small portion of their performance-related income.3 Many other organizations are attaching 1 or 2 percent or more of clinicians’ income to improvement in patient experience scores. Some are basing incentive awards on reaching a minimum threshold, such as ranking in the seventy-fifth percentile of measures of physician performance for the group or for individual physicians.
This approach definitely gets patient experience on clinicians’ radar screens, but problems with financial incentives are intertwined with their strengths. People really hate to lose money. Prospect theory, one of the core concepts in behavioral economics, explains why. Prospect theory asserts two major themes, the first of which is that proportion matters: If you give people $100, they are happy. If you give them $200, they are happier, but they are not twice as happy. The second major theme is that losses affect people more than gains do. If someone gives you $100, you are happy, but you may forget to mention it when you get home for dinner that night. But if you lose $100, perhaps because you got a speeding ticket, you may be in a bad mood for the rest of the day.
The powerful message from prospect theory is that small packets of incentive dollars that are framed as potential losses provide the most return on the incentive dollar. If clinicians realize that failure to achieve a performance goal will lead to a loss of just 1 percent of their income, they will work hard to avoid that loss, especially if the performance goal is consistent with their notion of delivering excellent care. They hate the idea of losing a small incentive because of failure to do something they really ought to do. As a result, they may work long and hard to get it, even if means not doing things that might increase their income by 2 percent.
The dark side of financial incentives is intertwined with their strengths. When clinicians lose their financial incentive, they are deeply unhappy. They focus on the flaws in the measures, data, and analyses and demand an appeals process that can consume more resources than are actually at stake. Those dynamics are familiar to every leader who has used financial incentives to try to drive improvement in patient experience, and they should be recognized as the expected result of normal human behavior.
Financial incentives that are attached to getting to a specific percentile for individual or groups of physicians are particularly problematic. The intentions of such targets are noble. The leadership and governance of the organization may decide that they do not want to be just okay; they want to be outstanding. Therefore, they set a target such as the seventy-fifth or ninetieth percentile for overall physician performance.
Those targets may simply not be feasible for some groups or individuals with a high proportion of patients, such as Asians or New Yorkers, who tend to give lower ratings. Financial incentives for quali
ty also tend to create a floor but do not provide a motivation for pushing beyond what is needed to avoid the pain of missing the target. That dynamic is particularly problematic if the floor has been negotiated down to something that everyone can easily reach. Therefore, financial incentives for quality often are associated with one of two unattractive outcomes: they fail to motivate outstanding performance, or they infuriate large numbers of personnel who fall short of the target. Sometimes they do both.
Because of such issues, many organizations use financial incentives for improvement in patient experience as a transition tactic. Modest incentives draw clinicians’ attention to patient experience, and then the organization transitions to the nonfinancial tactics described below.
Nonfinancial Incentives: Increased Affection and Respect
Some organizations, such as Cleveland Clinic, Kaiser Permanente, and many academic medical centers, have avoided financial incentives and emphasized Max Weber’s affection model for social action, using peer pressure and the desire for respect to drive improvement. At Cleveland Clinic, for example, every physician is on a one-year contract and doctors undergo rigorous annual performance reviews. Intense feedback, including patient experience, is used to create accountability for performance and improvement in all areas. At other organizations, a mixed model is emerging in which financial incentives are being used for issues with very direct financial implications, such as the volume of services or the efficiency of care, and nonfinancial incentives are employed to improve the quality of care.
In fact, nonfinancial incentives are emerging as the preferred approach for issues that are predominantly related to the quality of care. No one wants to lose the respect of his or her colleagues, family, or friends. We all have egos to protect. This approach seems particularly appropriate to patient experience, in which no clinicians should be calculating whether they would be better off financially if they mustered the energy to demonstrate empathy to a patient. The decision should be based on the adoption of social norms. It should be a “can I look at myself in the mirror?” decision.
Transparency as Incentive
The fear of loss of affection and respect explains why many organizations have been successful in driving improvement without financial incentives. Some, such as Cleveland Clinic, use intense one-on-one annual performance reviews. An increasing number of institutions are ramping up the pressure through internal transparency, letting all of a clinician’s colleagues see his or her data, including patient comments.
Transparency, even when only one’s colleagues are seeing the data, is uncomfortable. That discomfort is essential for driving improvement, but clinical leaders have to find a middle ground in which they are pushing the organization ahead but not provoking open rebellion. That said, it is critical that provider leaders recognize when they are shying away from creating real pressure. Here are some potential warning signs that organizations may not be taking full advantage of the power of the affection model:
• Blinding. Some organizations share data but do not reveal the names of individual physicians.
• Chaos. Others may reveal the names but do not organize the comments by physician, making it hard for anyone to draw conclusions about any individual’s performance.
• No analysis. Many avoid rank ordering data, making it hard to discern who is doing well and who is not, thus softening the blow for those who may be well below average.
• Rose-colored glasses. Some share only positive comments.
The big breakthrough in the use of transparency to drive improvement in patient experience was created by the University of Utah Health Care, which was the first to make systematic full transparency of patient experience data part of its approach to improvement. The University of Utah’s leadership made this bold leap in the winter of 2012–2013 after a three-year journey that began with the decision to move from using mail or telephone surveys of a sample to employing electronic surveys of as many patients as possible. After they began feeding much more data and comments back to their physicians, they moved to internal transparency. After two years of viewing one another’s data, they decided to let the rest of the world see them as well.
The improvement achieved at University of Utah was startling. It went from having 1 percent of its physicians in the top 1 percent of Press Ganey’s overall physician ratings (exactly what would be expected) to 15 percent with internal transparency and to more than 25 percent of physicians in the exclusive group with public transparency. No organization that I know of has ever achieved anything like this kind of improvement with any financial incentive. University of Utah Health Care made this startling progress without putting a single dollar on the line for any of its physicians. It was all due to Weber’s affection lever.
Although transparency is a bit unnerving to clinicians, movement in this direction is accelerating in part because of the proliferation of online consumer review websites. A 2013 Pew Research Center study determined that 72 percent of adult Internet users had searched online for information about health issues in the previous 12 months.4 A study in the Journal of the American Medical Association reported that 59 percent of American adults strongly consider online ratings when choosing a physician. Over one-third of individuals using online reviews chose their physicians on the basis of positive reviews, and 37 percent avoided doctors with negative reviews.5
Clinicians are not happy with the information available on many of these sites, which generally require that patients take the initiative to log on and enter a review, as opposed to reviews stimulated by a patient’s response to a prompt delivered by e-mail or another type of communication. The numbers of surveys or comments used to profile physicians on many sites are often extremely small, and the sites assign “star ratings” that are based on as few as two or three reviews. In addition, the proportion of negative reviews tends to be higher when patients have to take the initiative to provide information, presumably because unhappy patients are more likely to act than are patients who have had the good experiences they expected.
In addition, there is no guarantee that the information actually comes from patients who saw that physician or visited that facility. Researchers from MIT and Northwestern University examined 325,869 online consumer reviews and found that almost 5 percent were submitted by customers with no confirmed transaction. Those questionable reviews also included a significantly higher proportion of negative comments.6
When an organization works hard to get e-mail addresses and surveys are sent electronically to all patients by e-mail, the number of responses per clinician goes up dramatically. That has the happy effect of moving provider find-a-doctor websites to the top of the page on search engines. The reason is that search engines prioritize sites with more and fresher data. Thus, the constant influx of new comments to provider-driven transparency sites is rewarded with more attention, as it should be.
Before Utah began publishing online reviews, its physician profile page views totaled 32,144 per month. Three months after implementation, monthly page views increased nearly fourfold to 122,072. Piedmont Health had similar results with a 200 to 300 percent increase in web traffic. Both organizations’ find-a-doctor websites display significantly more reviews than any third-party review site, and there has been a marked increase in their visibility in search results.
It has not taken long for Utah’s example to spread. Within two years, Wake Forest and Piedmont had rolled out their own programs, and Stanford followed shortly thereafter. When you talk to colleagues at University of Utah and other places that have gone down this path, you learn that transparency changes the way physicians look at each patient encounter. They recognize that each interaction is high stakes for both the patient and themselves. The patient has a high probability of being surveyed, and he or she may write a negative comment, but in fact, the patient is about 10 times as likely to write a positive one.
The focus of transparency is not on any individual metric. It is on trying to take better care of
patients and feeling real pressure to do so for every single one of them. The pressure is relentless, but it is pressure to be the kind of clinicians that we want to think of ourselves as being. I haven’t met a physician or nurse yet who thinks this kind of pressure is perverse.
Using the Tradition Lever
People working in healthcare ought to be proud of what they do and where they work. They should be motivated by that pride, and they should worry about the implications of deviating so markedly from the norms of the organization that they are asked to leave. No one wants to lose his or her income, of course, but there are important nonfinancial losses as well: the prestige, security, and sense of self-worth that come with being part of a respected organization.
One of my favorite examples of tradition in action is the Mayo Way, which characterizes the approach to a wide range of activities of the Mayo Clinic not just at its mother ship in Rochester, Minnesota, but throughout its system. For example, as was described in Chapter 5, there is a clear expectation that clinicians will answer their pagers immediately. They don’t finish their conversations. They don’t finish their e-mails. They don’t finish their television shows or movies. They answer at once.
I learned about this social norm from a leader of radiology at Mayo’s facility in Arizona. She told me that her radiologists tend not to recommend additional testing at the end of their interpretations of tests. Instead, they simply page the physicians who are directly involved in the patients’ care and discuss what to do next. The radiologists are willing to do this because they know they can get a response from the other clinicians right away. It makes the radiologists feel more like they are part of a team taking care of patients as opposed to being technologists sitting in a dark room working their way down a long list of images. I can’t prove that this pattern of care is better or more efficient, but I’d be shocked if that was not the case.