Logo of annsurgLink to Publisher's site
Ann Surg. 2006 Nov; 244(5): 649–652.
PMCID: PMC1856582

When Good Doctors Go Bad

A Systems Problem

It is a real pleasure to be back and see so many old friends from my former life. In the 20 years since I left pediatric surgery, I seem to specialize in getting people to talk about things they don't want to talk about. And I want to do exactly that again today. This time it's problem doctors.

A few examples:

The chief resident pulls the senior resident aside and says, “I want you to go scrub with Dr. McAllister on that thyroid and keep him out of trouble.”

The chairman of surgery has a meeting with the vice-president of nursing and says, “Yes, I know Dr. Smith is a problem, and yes, I know he doesn't really treat the nurses very well, but you have to understand, he's one of our very best surgeons. And when we have a stable of thoroughbreds, some of them are going to be a little high strung from time to time.”

Two surgeons are walking along and one says, “You know, I really can't believe that Jim is an alcoholic. I've never seen him drunk.”

The CEO is talking with a member of the department of pediatrics about the chairman of pediatrics and says, “Look, he's dying of cancer. I'm not going to tell him he has to step down at a time like this.”

I have had all of those experiences.

How many of you had in your class in medical school someone who was psychopathic and should not have been allowed to become a doctor?

Look around! Look how many hands are up! (Laughter)

How many of you know at least one physician right now who is practicing who you think is a risk to his patients? (Again, many hands are raised.)

So we are not talking about some ethereal abstract, we are talking about something that is very real. I would submit to you that our failure to ensure that all of our colleagues are competent and safe is ethically indefensible.1 It is also unnecessary. It doesn't have to be this way. We can change it. And we must. And only we can change it. Because nobody can be responsible for the quality of practice except those of us who know what surgical practice is because we do it. This is not news to you. We know, we clearly know, we have a responsibility. It is in the Code of Conduct of the American College of Surgeons as well as in the code of ethics of the American Medical Association and others that we are responsible for maintaining and enforcing standards of practice.

What I want to talk to you about today are some thoughts about how we might do a better job of doing that.

First, let's start with the facts. What are we talking about?

Well, there is clearly the psychopathic physician, but they are, fortunately, rare. Much more common are the impaired physicians who have a problem with substance abuse, mental or physical illness. There is the really serious problem of declining competency, or maybe lack of ever having it. Finally, there are behavioral problems, the disruptive physician who is abusive to the people he works with or abusive to patients. Well, we have been interested in this, and we tried to see what kind of evidence there was in terms of frequency. Is this a big problem or is it a small problem? I don't know of any measure of the number of psychopathic physicians. I would estimate maybe one in a thousand. But judging from the number of hands I just saw raised, maybe it is one in a hundred! In any case, it is a serious problem. But it is one that ought to be taken care of in medical school. It ought to be taken care of before medical school, but certainly it should be taken care of then. And we need to think about how to do that. But I don't want to spend our time today on that particular problem.

There are some very good data about impaired physicians. We know, for example, that, just like civilians, about 10% of doctors have a problem with alcohol dependence and about 5% have a problem with drugs. With civilians, of course, it is illegal drugs. With doctors, it is legal drugs used illegally. But the numbers are pretty firm. There have been many studies of this. So 15%, I think, is a number that is realistic. In terms of mental illness, we have good information on that also. A recent very excellent study on a large population base showed that 16% of people had a disabling episode of depression at least one time in their life. That is certainly true of physicians. The suicide rate, as you may know, is higher in physicians than in civilians, so it may well be that our mental illness problem is even greater, but there is certainly no reason to think it is less.

In terms of physical illness, I have never seen a good study of the incidence of physical illness in physicians. But we are, like everybody else, subject to cancer, heart disease, stroke, fractures, and whatever, and it doesn't seem to me that it is unreasonable to say that at one time or another at least 10% of us are unable to practice for a period of time because of a physical illness.

In terms of declining competency, it is really tough to get a measure on this. Nobody has any really good measures overall. But there is one little bit of data that I found that is interesting and relevant, and that is the recertification examination failure rates. Pediatric surgery was the first of the surgical specialties to require recertification. Now all specialties require it. We have data from recertification exams. And it turns out that, with the exception of pediatrics, which is different because it is a nonproctored open book examination, the failure rate on recertification exams runs around 5% or 10%. I am sure that is but the tip of the iceberg of competency, but at least it is a hard number that we can have some confidence in.

With regard to behavioral problems, this is where we really lack good data because there haven't been any good studies, no respectable surveys. But some surveys have been done. One survey that I had access to was of a large number of nurses but with a very incomplete response rate. So I think these numbers are an indication, but don't take them as gospel. Almost all respondents had witnessed an episode of disruptive behavior by physicians. Two thirds of those reporting said they had had a problem within the last 2 months. Respondents estimate that it is not a very large number of physicians, probably in the range of about 5%. And that fits with what hospital executives will say, that about 1% to 5% of doctors will exhibit disruptive behavior. So let's say it is about 5%. Probably more, but at least that is a reasonable number.

Some of you are familiar with Gerry Hickson's work at Vanderbilt looking at patient complaints. From those data, he estimates that 6% of doctors (I rounded off to 5%) are abusive to patients at one time or another. So 5% here, 10% there, 15% there, you add them up, it gets to be pretty impressive. Obviously, there is overlap. Clearly, a physician might have mental illness, an alcohol problem, and declining competency all at the same time. But allowing for all of that, and most of the time there really isn't a lot of overlap, it looks like probably 30% to 40% of physicians at one time or another have a problem that poses a threat to safe patient care.

So it is not them, it is us: 30% to 40%. If you look at a hospital with 100 on the staff, that means at any one time there are probably on average 1 or 2 physicians who need help. Again, I ask you to look at your own experience in your hospital. Are there 1 or 2 physicians that need help? Yes. And most of them don't get it. And that is what I want to talk to you about.

One reason, of course, is we are very reluctant to do anything about it. These are our friends. These are our family. These are the people we work with. None of us is perfect. Who am I to point a finger? I make mistakes, too. And as many of you know from sad personal experience, when you try to do something about really major problem doctors, they fight back and you may find yourself at the receiving end of a lawsuit. So it can be a very messy business. I would suggest that there is another reason for us not to want to do it: we don't have a good mechanism for assessing physician performance. And that is exactly what I want to get to. We don't really have a system for assessing physician performance. We have a nonsystem. It is more implicit than explicit. It is more based on personal evaluation than on data. Now, we are using more and more data. Most credentialing processes now look at physicians' records. But many hospitals don't have very good data on that.

We tend to think of it mostly in terms of our assessment of someone when they have slipped; let's have somebody talk to them about it. We deal with it on a very personal level. In addition, to a large extent, we feel it is an all-or-nothing game: we either have to get rid of the person or tolerate it. And that puts us in quite a bind. Who wants to take away someone's livelihood? Most of us, when we think about this, are hung up on punishing. We think that the answer is to get rid of them. When you have a bad physician, let's get them out of here, let's make them somebody else's problem. That, of course, is not the approach to anything in patient safety.

It is exactly what we have been trying to get away from. Quit punishing for errors; instead, try to fix the causes and prevent them. I suggest to you this same approach is needed here because the safety challenge shouldn't be what to do about a bad doctor. The safety challenge is how do you prevent doctors from going bad, and how do you prevent anyone from hurting people?

So what we need is a system. We would like to have a system that is objective, that would identify doctors early, that would do something about it, and that would do something about it in a prompt fashion. I would like to suggest 3 characteristics for such a system. First, it needs to be objective. That means it needs to be based on data not on opinion, on objective information not on personal feelings and the biases that go with them. Second, it needs to be scrupulously fair. If we are going to have a system to assess physician performance, we need to assess everyone's performance, not just that of the doctors we are concerned about. So an assessment must not be a stigmatizing event, it should be a routine event, such as pilots go through and people in many other walks of life. Third, it needs to be responsive. If we are going to assess people, if we are going to set up standards and then assess physicians, then we need to do something about it when we find a problem. Otherwise, it is a waste of time. It has to be responsive both in terms of getting information back to the physician and in terms of taking action when it is needed.

The goal, of course, is not to weed them out but to enable physicians to continue to practice, but safely. So instead of weeding them out, if I may continue the agricultural metaphor, we need to cultivate them in. Here is one idea of how it might look. Other people may have other ideas. My hypothesis is very simple: first, that it is possible to identify physician performance problems before doctors get in trouble and hurt people. Second, if we do that, we will not only reduce harm, we will also make it much easier to help those doctors. Time and again, I hear from programs that work with problem doctors, “If we could only get them earlier.” I believe that is true, although there are not much data, so it is open to debate.

The kind of system I would like to see is first, objective, and therefore it should be based on standards. So, therefore, in my hospital, I would want to adopt performance standards, to have some standards that we all agree on. This is what counts; this is what we are going to be measured on. Then I would want to communicate to my staff that we are serious about this. As a condition of appointment here, you sign a statement indicating that you understand those standards and agree to follow them. You understand your compliance will be monitored, and you understand that if you don't follow them you will be asked to leave. So we are serious. We are not just playing games. The next thing I would do would be to monitor everyone using measures that are related to the standards. Everyone, so that it is fair, across the board, probably on an annual basis. Then, when we find a problem, feed back the information—the individual certainly has a right to know everything anyone else knows—but also take appropriate action to help the individual get back on track. Now, to provide meaningful help, we need a much broader repertoire than we now have of methods for assessment and remediation. Again, it doesn't do any good to identify problems if we are not going to take the next step. So this is a big agenda and not a simple one. But it is based on the concepts that we can define performance that is subpar in an objective way, that routine monitoring is a necessary requirement, and finally, that our response has to be not only prompt but also constructive, oriented toward getting the person back into practice safely and competently.

To do all this, we have to have standards, we have to have measures, and we have to have programs. We already have some of each. What we need is much more. In terms of standards, many of the people in this room have been involved in what I think is one of the most significant activities of the past 10 years in terms of safety of our patients, the development of competency standards in every specialty. These are the 6 areas with which the standards have been developed. I think it is an exciting development, and it is clearly the bedrock upon which any kind of an assessment and remediation program would be based. So we are off to a very good start in terms of competency. There is more work to be done, clearly, but we have made a very good start.

In terms of behavioral standards, we don't have them; we need a comparable national effort. I think it should probably be done by groups other than the American Board of Medical Specialties, but it is certainly an area that needs work. However, many of you work in hospitals that do have behavioral standards. Here is, for example, one from a hospital that I know of: “Treat coworkers with respect.” It says very clearly we do not tolerate hostile behavior, we do not tolerate demeaning behavior or humiliation of residents and nurses, nor derogatory comments about colleagues, and we expect people to gracefully accept challenges to their authority. So when the nurse or the resident questions what you are doing, you should thank them for it, thank you very much, not put them down. This kind of thing is already in place in some hospitals. It clearly needs to be in place in all hospitals.

In terms of measures, again, the American Board of Medical Specialties competency measures are now—and again, many of you in the room are working on this—being turned into measures for testing. Many of you know of the excellent work being done by the American Board of Internal Medicine in terms of competency testing. I think we have moved along quite a bit. And these methods clearly can all be used in the hospital or practice environment.

I mentioned Gerry Hickson's work, very useful, using patients' written letters of complaint as an indication of a physician's performance problems. There is another approach that many of you may not have heard of which I find particularly exciting, and that is the PAR program in Alberta, Canada. PAR stands for Physician Achievement Review, a “360” evaluation. For more than 5 years, every physician in Alberta has undergone this evaluation. The PAR review consists of a questionnaire of about 30 or 35 questions relating to everything to do with practice: your competency as well as your interrelations with people, how you keep your records and respond to requests, and so forth. And this questionnaire is filled out by the individual physician on himself or herself, by 7 or 8 nurses, 7 or 8 colleagues, and 20 patients. And the individual gets back the anonymized information. The result is that almost everybody has a surprise. Almost everybody finds out something about their own behavior and performance that they had no idea about. So it is very positive for them. It is done confidentially. It is not done as part of the credentialing or disciplinary process, it is done purely for personal improvement. We now have a study going to try to see how this it work out in a surgical department, and I hope to have some exciting news about that at another time.

We need to get serious about getting information from all sources. As all of you know, if you want to find out who the good or bad surgeon is, you ask the residents. It is the same sort of thing here, to get information from the people who really know in a safe and useful way, in a constructive, positive environment that is oriented purely around how we are going to help you be all you want to be—if I can steal that from the Army.

But what about some other things that have been suggested? Why don't we have annual physical exams? What about random drug testing? A hot issue. But, wait a minute, which is more important, the safety of our patients or our personal privacy? And what about cognitive testing? Don Trunkey called for this a year or 2 ago. Many of you may remember. We have the methods. We have the tools. Sadly, there is no question about it: everyone's cognitive ability declines with age. Now, you and I are in the top echelon, so if we go down 20% from where we were, we are still about 80% better than everybody else, right? Well, maybe not! The question is whether you are good enough. And we can measure that. The only other bit of evidence I have on this is from the Colorado C-PAP program. They find that 25% of the doctors that are referred there for assessment have serious problems in cognition. We have to take this one on, folks. So it isn't easy. And the hardest challenge of all, of course, is that we don't have the assessment and remediation programs we need. We need 5 times as many programs. We really need to be able to provide every doctor who needs it with a program and the remediation training they need. Who is going to do it? Clearly, you and I can't do it as individuals. Hospitals can't do it. It has got to be a national effort. We have called on the organizations that have the fiduciary responsibility for quality of physician performance to do this: the American Board of Medical Specialties, the Federation of State Medical Boards, and, of course, JACHO. There is no reason they couldn't do this, and it is time for people like you and me to tell them that.

A lot of other problems, too: Who is going to pay for it? What is a doctor going to do for a living while he is out for 3 or 6 months getting remediated? Are we willing in all of our residency programs to make place for a retrained doctor? We'd better. Are we willing to mentor them? Are we willing to take the responsibility for them? Are we willing to have them with us while they are getting up to speed and have them take care of our patients? Tough questions. But questions we as a profession need to answer. This is a problem we can handle. The means are there; all we need is the will.


Reprints: Lucian L. Leape, MD, Harvard School of Public Health, 677 Huntington Avenue, Boston, MA 02115. E-mail: ude.dravrah.hpsh@epael.


1. Leape LL, Fromson J. Problem doctors: is there a system-level solution? Ann Intern Med. 2006;144:1–8.

Articles from Annals of Surgery are provided here courtesy of Lippincott, Williams, and Wilkins
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


  • PubMed
    PubMed citations for these articles

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...