by Thomas H Lee
Data show that there is a strong association between an athlete’s happiness and the happiness of his or her teammates, and when a player’s teammates are happier, performance improves. These findings naturally raise the question of whether happier healthcare providers might do a better job taking care of their patients. I’m not sure this is a question that requires a randomized controlled trial. There are plenty of data from employee engagement surveys that support the hypothesis that the mood and values of the social networks in a clinical workforce affect quality of care.
Emotions such as happiness are not quite the same thing as values, such as the sense that if I am not delivering empathic care to this patient, I will have trouble looking in the mirror later today. However, research on the way happiness spreads provides insight into how values can spread in a network. Indeed, one would expect values such as empathy to spread more quickly and reliably than emotions do, since they can be enforced through peer pressure.
Christakis and his colleagues have studied the question of whether happiness spreads just through dyads (i.e., between pairs of people) or whether there is also hyperdyadic spread: between you and your friends’ friends, and their friends, and so on. They had previously found that the impact of changes such as obesity or changes in cigarette smoking habits fell off dramatically once one got beyond three degrees of separation.
When they mapped happiness among 1,020 people in the town of Framingham for whom they had detailed information on friends and other social relationships, there were two major findings. First, unhappy people tended to cluster with other unhappy people and happy people tended to cluster with other happy people. Second, unhappy people were more peripheral in social networks. They were much more likely to appear at the end of a chain of social relationships than in the middle.12
This kind of clustering could be the result of birds of a feather flocking together, but Christakis’s research was able to show that there was also a causal effect of one person’s happiness on another. Their analyses indicated that someone was about 15 percent more likely to be happy if she was directly connected to another person who was happy. At two degrees of separation, there was a 10 percent increased probability of being happy, and there was a 6 percent increase at three degrees. Beyond that, there was no measurable impact.
There is now a wealth of data examining various issues within the question of whether and how happiness spreads, whether distance in the relationships matters, and whether the transitivity of the social network matters. In short, these data indicate that emotions spread from person to person to person; in effect, they are contagious. However, emotions such as happiness do not sweep through a community and affect every person the way a plague can wipe out a town; the ripple effect peters out. For an epidemic of empathy to keep going, it must be driven from many different foci in an organization and constantly reinforced.
Social Isolation in Healthcare
The opposite of being connected is being unconnected, a lonely state that is all too common in modern medicine. As was detailed earlier, medical progress has led to both a massive increase in the number of people required to provide state-of-the-science care and a narrowing of their focus. Although there are many more people around, the irony is that the delivery of care is a lonelier experience than it was a generation ago.
For example, when I was an intern and resident at Peter Bent Brigham Hospital from 1979 to 1982, I felt that I knew everyone. It would take 20 minutes to walk 50 yards down the long hallway known as the Pike because I would run into friends and colleagues every few steps. During those interactions, we would discuss issues related to patients we shared and also plan to get together for dinner or the movies. We knew everything about everyone else’s lives.
Once I was trying to read the newspaper before morning rounds, and another resident, a young woman who was trying to get pregnant, asked if I would cover her responsibilities in the emergency department for an hour during lunchtime. I had learned from a senior resident when I was an intern, “Never say no to a colleague, and there is a good chance that they will never say no to you.” I did not even look up from the newspaper as I said, “Sure.”
She continued standing there as I read the sports section. “Don’t you want to know why?” she asked.
I put down the paper, smiled, and said, “I think I know.” She told me anyway.
I bring this story up not to show what a generous colleague I am but to show what happens naturally in settings in which clinicians are thrown together, know that they are stuck together, and adopt a social norm such as “Never say no to a colleague.” There were certainly some residents among us who did say no. Three decades later, my former cotrainees still talk about them, and we are a little cool to them when we see them at various events. When a critical mass of clinical personnel are connected as my coresidents were back then, good things happen, possibly including the births of the children of the resident whom I covered for that day.
I am sure there are still groups of friends in my hospital in which clinicians know one another as well as our group of residents did, but the fact is that healthcare has become a lonely life for many. Many outpatient physicians do not go to the hospital, and when they do, they do not recognize the clinicians who are taking care of their hospitalized patients. They may walk onto the floor of the hospital and no nurse or other doctor knows them by name.
The inpatient specialists have offices next to those of other clinicians with the same expertise, and they eat lunch in those offices—if there is time to eat. There is little time and opportunity for sitting in the cafeteria talking about cases, medical articles, or life itself. No one goes down to radiology to talk to the radiologists anymore. Lots of activity happens virtually and occurs much faster as a result, but the side effect is loneliness and social isolation.
For leaders hoping to drive an epidemic of empathy, this loneliness and social isolation should be seen for what it is: a major problem. Aside from the obvious challenge it poses for the spread of a value such as empathy, the fact is that isolated people are less happy and unhappy people are less able to feel and convey empathy.
Loneliness is a problem that feeds on itself. Research shows that people who feel lonely all the time lose friends faster than others do, are less likely to attract new friends, and are less able to name social connections. Loneliness can be epidemic in some organizations. Data suggest that being directly connected to someone who is lonely increases your risk of loneliness by 52 percent. At two degrees of separation, the risk increases 25 percent, and at three degrees of separation, the increase is 15 percent.13
Experienced researchers will have plenty of questions about what is cause and what is effect in these data, but experienced managers will recognize that there is a problem to be solved. That solution should begin by targeting people at the edges, to repair their social networks. It should also begin by spreading the connectedness of people who understand the meaning of “The Opposite of Loneliness,” the 2012 graduation speech given by the Yale senior Marina Keegan, who died in a car accident five days later, before she could begin her postgraduation job at The New Yorker.14 She called it “this feeling that there are people, an abundance of people, who are in this together. Who are on your team.” Her essay captured the sense of security and possibility that came with being a part of the social network of her graduating Yale class.
The Implications of Networks
The implications of the work of Burt, Christakis, and others in their field go well beyond the simple message that we all need to work more closely together, know one another better, and be nicer to patients. It would be naive to believe that the clock can be turned back to a simpler time when everyone knew everyone at a hospital. The to-do list for building social capital and driving an epidemic of empathy in medicine’s modern era includes the following:
• Make the accumulation of social capital through brokerage and closure an important focus for leadership. Idea brokering should be as impo
rtant an activity as any other type of research. Organizations should recognize that they need to build trust so that they can achieve closure on the best practices that create value for patients.
• Segment patients and providers into teams in which patients’ needs can be better met through coordinated multi-disciplinary groups. Give those groups data on outcomes, including patient experience and costs. Concentrate patients on those teams so that there is critical mass. Colocate those teams and have them meet formally and informally.
• Identify the lonely and disconnected people in your network and try to bring them into the fold.
• Identify many of the positive and central influences in your network and put them together to create a persuasive powerful influence.
• Use an array of incentives—financial and others—to drive your epidemic of empathy, as is described in Chapter 6.
CHAPTER
6 Changing Behavior and Creating the Epidemic
WE KNOW THE problem: the chaos of modern medicine, which is a side effect of the marvelous progress in our times. We can do so much more for patients than we could a generation ago, but there has been an exponential increase in the number of people involved in delivering care even for routine issues. Those clinicians often have narrow foci and overlook the big-picture issues of greatest importance to patients. Frequently, they are not working together as well as patients hope and expect.
We have the imperative to improve. Healthcare providers are entering a marketplace driven by real competition on value. We have to deliver good outcomes and meet patients’ needs, including their psychological need for peace of mind, and do this as efficiently as possible. Providing higher-value care is good for patients and good for society and ultimately leads to professional pride among healthcare personnel. That pride makes coming to work more pleasant and leads to lower turnover of personnel, which leads to better business performance.
We know what we need to do. We understand the nature of empathy, and thanks to researchers such as Nicholas Christakis, we have the knowledge from social network science of how values such as empathetic care can spread.
The question is, How do we bring it all together? How do healthcare organizations and their leaders seize the competitive opportunity? How do we help individual clinicians live up to their aspirations to reliably deliver care that is safe, up to date, evidence-based, and efficient but also coordinated and compassionate?
Achieving reliable excellence is the essence of delivering healthcare—nothing more and nothing less. How do we do it for every patient every time? It’s not really a change in behavior of healthcare providers that we seek. It is the achievement of consistency in being at our best.
Changing the Focus
Being reliably at our best in healthcare means doing something wonderful, remarkable, and uncontroversial, but the pursuit of consistency and the attempt to create the necessary teamwork can pose cultural issues. All too often, it seems that the strengths and weaknesses of healthcare are inextricably intertwined, creating a Gordian knot that defies unraveling. To visualize the resistance to redesigning care around the needs of patients, many healthcare leaders have to look in the mirror. That means people like me—doctors, mostly age 50 and older—who came into medicine in a different climate and different marketplace but are now in charge. Many of us still hold the belief that medicine is an art and a science but isn’t a business, and most of us are resistant to change.
Nuanced Approach
One size will not fit all when it comes to driving improvement in healthcare. A nuanced, multidimensional approach is needed, because there are so many different people in healthcare and they are motivated on a day-to-day basis by different types of incentives. The truth is that we ourselves vary in the care we provide from hour to hour and patient to patient. This variability in healthcare personnel suggests that an epidemic of empathy cannot be driven solely through the use of the two major approaches employed by most healthcare organizations: carpet bombing and identifying bad apples.
In the carpet bombing approach, everyone is rallied to the organization’s missions, often by being asked to watch videos, listen to speeches, take tests, and sometimes even take short courses on skills such as communication and listening. In the bad apple approach, the physicians and other personnel with the worst performance are taken aside, told they have a problem and must improve, and sometimes connected to coaches or courses to help them do so. If the bad apples cannot improve, they may be asked to leave the organization.
There are definitely important roles for carpet bombing and finding the bad apples, but driving an epidemic of empathy requires a more sophisticated approach than giving pep talks and playing Whac-A-Mole. In light of the forces that work against empathic care described in Chapter 1, these approaches are only going to slow deterioration, not accelerate improvement.
To drive improvement and make empathic care the norm, we need to use a range of tools, including financial and nonfinancial incentives, followed closely by giving clinicians help with the methods most likely to improve the care they provide. It doesn’t take much to nudge clinicians in the right direction. After all, everyone in healthcare wants to deliver care that is compassionate and coordinated. However, although we are all humans—good humans—we are humans all the same. Left to our own devices, we are not as consistent as we would like to be both individually and collectively. We need incentives and nudges to be at our best and to try to improve.
Four Models for Social Action
The endlessly interesting question in patient experience pertains to how incentives for better healthcare performance should be designed. Should they be financial or nonfinancial? How can they be applied ethically and effectively?
In my 2014 article with Toby Cosgrove and an earlier essay with the ethicist Nikola Biller-Andorno,1 I discussed an adaptation of the four models for social action described a century ago by the German economist and sociologist Max Weber:
1. Tradition. For example, a standard of behavior such as a dress code that if violated threatens one with expulsion from an institution
2. Self-interest. For example, the use of financial bonuses or penalties for achieving performance targets
3. Affection. For example, the use of peer pressure from one-on-one performance reviews or transparency with performance data
4. Shared purpose. For example, achievement of consensus that the overarching goal of a healthcare organization is to reduce the suffering of its patients
I will explore each of these four in detail, but the overall conclusion is that we need to use all four models in healthcare. Let’s start with the creation of a shared purpose.
The vision that underlies a shared purpose—for example, Cleveland Clinic’s Patients First or Mayo Clinic’s constant reminders to its personnel for more than a century that “the needs of the patient come first”—comes from stories that capture what makes us proud or ashamed. Trying to deliver care with the elements that consistently bring pride to providers is what performance is all about, and no one can perform at a high level without first having a clear idea of what he or she is trying to achieve.
A shared purpose is necessary but not sufficient to drive a real epidemic of empathy. After all, tears dry and goose bumps subside. We need data and incentives to drive constant improvement and need to use the other three of Weber’s levers to push clinicians to be at their best for every single patient. The reason we need to start with the creation of a shared purpose is the harsh reality that no measure is perfect and the data used to bring a measure to life will never be perfect either. Even if data are collected from every single patient, one can never be sure that all the risk factors that might bias the analysis in one direction or another have been taken into account.
If every measure has inherent weaknesses, if the data are unlikely to ever be complete, and if analyses cannot adjust for every risk factor, what can be done? Not measuring anything is obviously a bad idea. The implication is that we
have to think clearly about what we are trying to do with the data and how we do it.
There are two very different questions one can ask when looking at performance data:
1. “Who is good enough?” or its variants, such as “Are we the best?” and “Am I good enough to have some credential, such as board certification, or to earn some financial incentive?”
2. “How can we get better than we currently are?” Here the question is how we can improve in the pursuit of a shared purpose. The assumption is that no matter how good we currently are, we should try to get better, and the data are a tool—albeit an imperfect one—to aid that progress.
Brent James of Intermountain Health has described this dichotomy, pointing out, “There are data for judgment, and there are data for improvement.” When data are being used for improvement, they are intended for use for the second goal: trying to be better than we currently are.
The second question sounds nobler, of course, and one thing I have learned as life goes on is that when something sounds nobler, it usually is nobler. But the reality is that we need to address both questions. After all, organizations do have to decide whether to hire and fire people, and society does want to know whether a physician has achieved a certain level of mastery of his or her area of expertise as suggested by board certification.
Nevertheless, to drive an epidemic of empathy, healthcare organizations need to make a distinction between these two questions. The reason is that the two questions elicit very different responses from the personnel who are being evaluated. When data are being used for judgment, the issues have high stakes for the individuals involved, and the normal human response is to try to minimize the risk of failure. The best way to minimize that risk is to work hard to be as good as possible. But even for those who take that high road, an attractive backup plan is to attack the measures and the data that put one at risk for failure.