Editor's note: Gerald B. Hickson, MD, is one of the world's leading experts on physician behavior and its connection to clinical outcomes and medical malpractice. He is a Professor at the Vanderbilt University School of Medicine, where he is also the Joseph C. Ross Chair in Medical Education and Administration, Associate Dean for Clinical Affairs, Director of the Vanderbilt Center for Patient and Professional Advocacy, and Director of Clinical Risk and Loss Prevention. We asked him to speak with us about high-risk physicians and malpractice.

Dr. Robert Wachter, Editor, AHRQ WebM&M: Take us through the key findings in your research over the last decade or two.

Dr. Gerald Hickson: Our early studies simply affirmed what all good risk managers already knew: patients often contact plaintiff attorneys because they believe that they never could get answers to their questions, and sometimes concluded that professionals engaged in cover-ups. Other studies that followed examined why certain physicians have a dark cloud for malpractice risk. From 2% to 8% of physicians by discipline are associated with far more than their fair share of claims experience. We and other investigators identified that these high-risk physicians do not treat sicker or more suit-prone patients. Nor are there measurable differences in technical competence. These findings led to a series of studies that demonstrated that individuals at high risk are more challenged in establishing and maintaining rapport with their patients and that such high-risk physicians can be reliably identified by families' unsolicited complaints about their care experiences.

RW: Do you think these kinds of physicians actually do have more adverse outcomes, or the adverse outcomes that they have are more likely to generate a lawsuit?

GH: Our initial hypothesis was that everyone has adverse outcomes; therefore, we are simply superimposing adverse outcomes on either reasonable or poor preexisting relationships. Now that more research has been done examining factors contributing to adverse outcomes, the link between poor communication skills and poor outcomes is clear. When medical team members don't play well together—sometimes because a physician engages in disruptive behavior—we are more likely to fail to achieve our intended outcomes and therefore experience more errors.

RW: Take us through the pathophysiology of the development of a problem physician. Is this someone who had personal issues at the time that he or she applied to medical school? Or are these behaviors learned over time?

GH: I can only offer observations based on our patient complaint data and feedback we get from peer messengers who conduct interventions with high-risk physicians. These observations and experiences suggest that some at-risk physicians (perhaps 15% to 20%) are individuals who demonstrated behaviors in medical school that suggest that they have always been challenging. Studies show a relationship between behavioral and other performance measures in medical school and reporting to state licensing boards. Other individuals join medical families in which they simply never received early feedback that certain behaviors are not appropriate and pose a threat to care quality. We find that many such individuals who are identified and receive peer-delivered interventions are stunned; they often assert that they had no idea. Without feedback, disruptive behavior can become accepted as the norm.

There also are individuals with personality disorders. Certainly, we see individuals who model narcissism. Then there are others whose disruptive behaviors may be triggered by life-changing events. Sometimes problems at home are taken to work and vice-versa. There are also a few who are impaired due to drug and alcohol use.

RW: You were talking about a cultural norm in certain fields in which this behavior didn't receive feedback. Did it go even further than that?

GH: One of the things we often hear when we provide feedback to high-malpractice-risk physicians is "everybody in my field gets sued and everyone gets complaints." So we have examined, as an example, whether all trauma surgeons are equally at risk. When using unsolicited patient complaint reports to identify risk, we find that in any 4-year period, 40% of trauma surgeons generate no unsolicited complaint reports, while a small group of such surgeons have an extraordinarily high number of complaints, mirroring the distribution of lawsuits within the discipline. So, the notion that "everyone in my field gets sued" is just not true.

Now the second issue you raised was about culture and feedback. As I said, high-risk physicians often respond, "I never had any notion that my risk was higher than others in my field." This is why we feel so strongly that early sharing of risk status is one element needed to promote a culture of safety.

RW: What is your threshold to identify a physician as being at risk? And once a physician is identified, walk us through what happens in your system.

GH: We don't believe that we ever will create a culture without any nonprofessional conduct. Humans are human; what we want to do is create a culture that discourages development of patterns of disruptive behavior and that has no tolerance for episodes of the most egregious behaviors.

To address nonprofessional conduct, we use a "disruptive behavior pyramid" to help us match circumstances with the appropriate level of intervention. The pyramid is based on a foundational concept that the vast majority of team members never demonstrate any disruptive behaviors. That's important because when leaders are beginning to contemplate addressing the challenge of disruptive behavior, they need to understand that the vast majority of people who walk in the door are outstanding. In fact, we honor outstanding colleagues by addressing colleagues who need a little help. Therefore, whenever an "event" is reported, we recognize that there may be two sides to the story. Sometimes patients have encounters when they believe providers responded in nonprofessional ways. Whenever possible, physicians need to have those events brought to their attention through an informal "cup of coffee" process. It's not a control contest. It's simply a way for a medical group to say, "We want you to know what the patient shared." The individual physician involved needs to understand that the complaint is based on the patient's perception, which may or may not have merit, and we want the provider to know.

The problem is that when an event occurs, does it represent an anomaly or just one more event that identifies a pattern of disruptive behavior? We believe that patients and other members of the health care team serve an important surveillance role. Their eyes and ears are incredibly effective in identifying problems if organizations are committed to listening, recording observations, and looking for patterns.

Once a pattern is suspected, there needs to be action. So as an example, at Vanderbilt we routinely code all unsolicited complaint reports and assign the complaints to 34 categories. The complaints are aggregated, and those physicians with more than their fair share receive a letter requesting a visit with a trained peer messenger. "Bob, for whatever reason, you seem to be associated with more complaints than the vast majority of your colleagues. I'm not here to ascertain why. My goal is not to tell you what to do, but suggest that you review the material I am sharing with you and reflect on what families are saying about your practice. Furthermore, follow-up will occur in a defined number of months." The process needs to be fair and apply to all members of the team. In addition, we believe that messages about patterns should be delivered by a trained physician peer messenger. One mark of a profession is its self-regulating entity. One way we demonstrate our commitment to our profession is to sit down and share with our high-risk colleagues. The peer messenger is able to say, "Good colleague, 50% of our physicians don't get any complaints. Some get an occasional complaint. But, Bob, you get more complaints than 95% of all our colleagues, and I thought you would want to know."

RW: All right. And I hear you. But in my heart of hearts, let's say I really don't want to know. And this behavior continues. What does monitoring look like? And what does an escalation strategy look like over time?

GH: The process begins by making the peer aware, letting them know there will be follow-up. But when the data indicate there is no improvement, an institutional policy directs that the individual's authority figure be notified. The leader needs to be held accountable for developing a plan. This is the point at which I revert to my experiences as a pediatrician. The pattern of complaint generation indicates there's a fever. But most of the time, with an antipyretic and fluids (personal physician reflections), the fever goes away. But if it persists, it is time to bring the leader in to conduct a much more detailed evaluation (authority intervention). That detailed evaluation may include an assessment of the practice. It may include coaching or a comprehensive mental health evaluation for the professional. But we're looking at a circumstance where the high-risk colleague has received regular feedback from a peer and has not responded. When the high-risk physician is either unable or unwilling to respond, it's time for the organization to take a detailed look and hold the individual and his or her designated leader accountable for change.

RW: And what does that mean?

GH: It means that sometimes leaders need to be encouraged, sometimes by executive coaching, about why there might be a problem and what the resources are that the group or organization can use to help. There needs to be a written plan to define expectations for the high-risk physician, what the identified deficiencies are, what interventions will occur, how success will be measured, and what the consequences are for failure to respond.

RW: If 100 physicians are identified at the earliest stage of an unusual pattern of problematic behaviors, how many of them will be remediated simply through being made aware of their behaviors? How many end up having this escalated? How many cannot be fixed? And what happens to those people?

GH: Based on 10 years of experience with almost 40 medical groups, we find that 60% of high-risk physicians respond to the Level 1 "awareness" intervention—when a peer sits down, shares confidentially, and reminds the high-risk individual that they will return with follow-up data. Of those who respond to the awareness intervention, the recidivism rate appears to be about 4%. Of the 40% that do not respond to the awareness intervention, we find—which surprised us—that about half (20% of the total) leave the group. Some individuals simply choose to take what we call the "geographic" solution. They recognize the tiered intervention process. They may think, "I'm not going to put up with scrutiny any longer. I'm leaving and going someplace where my technical skills will be appreciated." Unfortunately, we now have many circumstances in which we've identified individuals at one institution. They leave and go to a second site where we are working and, lo and behold, those same practice challenges follow the high-risk physician to the second site.

RW: Shocking.

GH: It is a tragedy, and at this point I do not have a solution. One thing that we in medicine need to do is collectively think about what our responsibility is in addressing individuals with high malpractice claims who are highly mobile. Of the half (20% of the total) who continue to collect complaints and who do not move, they receive a Level 2 "authority" intervention. About half of this group respond to the more directive interventions. Some reorganize their practices, add capacity, obtain coaching, or receive comprehensive mental health evaluations and therapy. Unfortunately, about half of this group do not respond, and in those circumstances, the organization must consider the possibility that the individual simply doesn't practice in a way that is consistent with the group's commitment to quality and professionalism.

RW: How often does that happen in the life of a medium-sized hospital over a year?

GH: We've learned that these processes must be run fairly. Fairness says that you need a system that is applied to all. The purpose of surveillance and peer intervention is to provide an opportunity for the individual to develop insight. Those who develop insight can improve. For those who cannot develop insight, you need to try again through an authority intervention. But at the end of the day, the entire process is about choice, and some high-risk physicians will not respond. Over the past 10 years, we have parted company with a colleague every other year.

RW: You've made a major point of fairness. I assume that by fairness you mean that the same behaviors in a neurosurgeon bringing in $20 million a year and a pediatrician will be handled the same way.

GH: When medical groups contact our center at Vanderbilt requesting help in establishing a tiered intervention process built around an analysis of patient complaint data, we always warn leaders that they must be prepared to deal with the problems identified. That means addressing the high-risk "rainmaker" neurosurgeon in the same way as a generalist who brings in little revenue and is more easily replaced. Success depends on real leadership. You can create policies about professional conduct, you can create surveillance systems, and you can even train messengers. However, the question at the end of the day is, when a high-risk physician is also highly valued by whatever currency, and he or she will not respond to the intervention, will the leader refuse to blink?

RW: That's hard to do.

GH: It is exceedingly hard. I am not unsympathetic to the challenges that medical leaders face in circumstances where the high-risk physician is a high producer and the only specialist in the area.

RW: How do you do this across disciplines? Because obviously doctors are not the only ones who have personality issues and behavioral issues. How do you have fairness when often there are different sets of rules and laws or union issues?

GH: We started our efforts to address high-risk physicians because it was the logical next step in our medical malpractice research. It was fortuitous because so often when we attempt to bring about culture change in medicine, we focus on everyone but the physician. So our notion was that we would deal with each other as colleagues first, and then extend the work to other members of the medical team. The phenomenon of disruptive conduct, however, affects all medical team members and what I refer to as "medical families." For example, we find a 10- to 20-fold variation in unsolicited complaints per 1000 hospital bed days across the various inpatient units at our hospital. Sometimes units have been inadequately supported by administration. Sometimes the unit is being poorly run, creating a toxic environment for staff and patients. The same differences that we see among our physician colleagues are also observed among our advance practice nurses. We believe that whenever variation is identified by whatever surveillance system, you apply the same tiered intervention strategy, beginning first with a "cup of coffee," advancing to "awareness" and then authority as needed. We don't think this is a physician-only phenomenon.

You identify another challenge with application of any of our work in environments with different rules or bylaws for different members of the team. Consequently, all efforts to address high-risk colleagues, regardless of their roles, must begin with a careful review of the rules of engagement. At the end of the day, I go back to the principles of fairness and providing an opportunity for individuals, regardless of their role, to gain insight and improve.

RW: I'm going to focus on a few external drivers of all this. One is the malpractice system and what you've come to understand about the cost-effectiveness of this kind of intervention. The second is the involvement of The Joint Commission and the National Patient Safety Goal around disruptive behavior. For example, do malpractice carriers believe that this kind of a vigorous program ultimately saves them money?

GH: Malpractice carriers have recognized for years that it doesn't rain on everybody equally and that some physicians have a malpractice dark cloud. In addition, when we are invited to partner with an organization, someone always says, "I already know who you're going to identify." I routinely respond, "You might be right." What I want to ask, but don't, is, "So what have you been doing to address the known challenge?" The good news is that a pilot study examining 12 years of claims experience of a defined cohort of surgeons suggests that a peer-based intervention program will reduce complaints and perhaps malpractice claims.

But it is far more than just reducing malpractice expenses. In my view, it is about promoting professionalism and teamwork. When individuals think about disruptive behavior, they think in terms of aggressive or bullying behaviors. But disruptive behavior represents any behavior that impacts the ability of the team to achieve the intended outcome. When we look at the management literature, experts have long written about the "bad apple" effect. While I don't like the term, because it implies no hope for improvement, the term helps us understand that when certain individuals walk into an ICU, other team members begin having to monitor the high-risk colleague, not just their assigned task. I believe that such tracking promotes slips and lapses and contributes to poor patient outcomes. So addressing disruptive behavior is not just about medical malpractice prevention—it's about delivering quality care, it's about retaining qualified staff, and it's about maintaining morale. All those things make the return on investment of addressing nonprofessional conduct strong. But there needs to be more research in the area.

Now the other thing you ask about is imperatives. The Joint Commission sentinel event alert, Behaviors That Undermine a Culture of Safety, helped focus attention on the impact of nonprofessional conduct and the need for hospitals to have a plan. But at the end of the day, it is still about local leadership. This is what we refer to as the question of implicit versus explicit culture. It is also about developing a process, establishing a surveillance system, using a tiered intervention strategy, training individuals to deliver the message, and providing the appropriate resources to help those who are identified. The Joint Commission alert reminds us that we need to address a challenge, not simply write a new conduct policy so we can pass a review.

RW: How do apologies and disclosure fit into all of this?

GH: Full disclosure about adverse outcomes and medical errors is a measure of professionalism. Failure to disclose is an example of nonprofessional conduct. All of us need to recognize that disclosure is neither an event nor a moment in time but a process. And that process is built best upon a relationship of respect. I worry when I hear the sound bite "sorry works"—an apology is powerful when superimposed on a patient–professional relationship of trust. But as the medical team member, if I have not demonstrated respect for you or if I have sent you the message that you're not valued, when I walk in and apologize, it may not be so well-received. In circumstances where an error has occurred, we need to be forthright, apologize, and provide an explanation of what has occurred. We need to commit to the patient that not only are we going to try and make it right, we're going to try to understand why and prevent future errors.

I also worry about what I refer to as premature disclosure and apology in circumstances where there is, at least initially, uncertainty. We promote the notion that in a circumstance of uncertainty, we want professionals to be deliberate. We want to let patients know that we're committed to understanding what happened, and as soon as we get more information we're going to sit down, share, and answer questions. Apology is important, but in our view even more important is recognizing the value of respect and effective communication from the first encounter.

RW: I want to see how you respond to push back. If you have your way, you'll create this environment where everybody is being so careful that you have an environment that's robotic, and the kind of individualistic person who pushes the envelope and helps us innovate will be drummed off the medical staff. Did you ever hear that before?

GH: I have heard many high-risk physicians assert: "I am the only person in this place who really cares about patient outcomes." Advocating for patients is important, but advocacy does not provide license to behave in a disruptive manner. When a colleague behaves in a disruptive manner about a real quality concern, my notion is that we have two issues to address—the quality issue and the behavior. In the end, it is about promoting professional conduct and quality, and recognizing that a profession should be a self-regulating entity.