From 3M Health Information Systems
Podcast episode transcript: What makes a good doctor? The role of board certification
Dr. Gordon Moore: Welcome to 3M’s Inside Angle podcast. This is your host Gordon Moore and with me today is Dr. Richard Baron, board-certified in internal medicine and geriatric medicine. He’s president and chief executive officer of the American Board of Internal Medicine and the ABIM Foundation.
He was a former chair of the American Board of Internal Medicine’s board of directors and served on the ABIM Found board of trustees.
Welcome, Dr. Baron.
Dr. Richard Baron: Great to be here, Gordon.
Dr. Moore: Tell me, for those who may not be familiar, what is ABIM? What do you guys do?
Dr. Baron: When you hear the expression board-certified doctor, ABIM is the largest board that does the certifying. We certify one in four doctors in the United States. Internal medicine includes not just the primary care discipline of general internal medicine, but it also includes hospital medicine and a whole variety of internal medicine subspecialties like cardiology, oncology, pulmonary.
So internists focus on nonsurgical diseases of adults and the work of the American Board of Internal Medicine is in a publicly recognizable way to certify doctors at the beginning of their career who’s successfully completed their training and are ready for independent, unsupervised practice in their discipline. And, over the course of career, to speak to whether doctors have stayed current in their field.
That’s the core work of the American Board of Internal Medicine.
Dr. Moore: So again, for those who may be unfamiliar with all the different aspects of this, why is this necessary on top of I have a med school diploma and I’m licensed to practice?
Dr. Baron: We have, in this country, a trajectory of lots of specialization. If you go back to the 19th century, then doctors were—everybody was a GP and people might deliver a baby. They might take care of an infant. They might take out a gallbladder or do an appendectomy all in the course of one day. And what we realized over time was that doctors focusing their knowledge and expertise in more specialized areas was something that could really benefit patients.
So the idea—there’s a wonderful historical story, Gordon. In 1865 the New York delegation to the AMA convention came with a complaint. They said there were doctors in New York who are advertising that they were specialists in heart disease and the New York delegation said they’re not specialists in heart disease. This is a naked attempt to grab patients. And they asked the AMA to declare claims of specialization unprofessional.
What the AMA wound up doing was saying no, specialization is a good thing. It can help patients. But self-declared specialization, that’s a bad thing. We shouldn’t have people just saying hey, I’m an expert whatever. That if people were going to make those claims, it was important that they be validated by an independent third party and that’s the role that the boards play.
So these days, medical training will include four years of medical school and the people will do at least one-year postgraduate training referred to as internship and after that, they start differentiating. They may become a pediatrician, or an OB/GYN, or a surgeon, and the years that they do in postgraduate training qualified them. Trained them to practice in a specialty and our board certifies some of those specialties.
I misspoke a minute ago in talking about it’s after internship the differentiation occurs. Back in the ‘30s and ‘40s internship was generic and people did lots of different things in internships. These days people differentiate when they graduate medical school. So a pediatric internship looks different than a surgical internship, looks different than an internship in internal medicine.
Dr. Moore: And there’s another aspect. You had mentioned sort of ongoing certification. Used to be I could come out of med school and residency, I could become board certified by taking a complicated exam in my specialty and that’s it. I’m board-certified. I’m done. Nowadays that’s not the case.
Dr. Baron: That’s right. The board that represents the discipline in which you trained, family medicine, was the first board in 1969 to say we’re not going to issue certificates for life because medicine is a high-stakes profession. Knowledge changes a lot. The skills and ability somebody had 20, 30 years ago when they trained, that’s not especially relevant today.
So what your board realized and other boards including ours followed suit was we started issuing time-limited certificates where periodically people needed to demonstrate that they had stayed current in the field. And the overwhelming number of internists and internal medicine subspecialists today have these time-limited certificates and if they aren’t maintaining their skills they will lose those certifications and that is an important part of our program today.
Dr. Moore: I certainly hear that as I have to go online and take courses on the different aspects of family medicine to keep up with what’s happening. But you know, as I think about all of that work, I also think about insurance plans saying I have to go through some certification process to be one of the doctors who can accept their insurance. That’s yet another layer of certification.
Dr. Baron: Well, I think it’s kind of different. The continuing certification that the boards oversee is about changes in knowledge in the discipline, changes in the literature in the discipline and patients really have an important interest in seeing a doctor who has stayed current. You don’t want to be getting advice that’s ten years old.
The leading drugs for treating high blood pressure and diabetes didn’t even exist when I was in training. So knowledge changes a lot and the board’s commitment to be focused on demonstrating current knowledge is a critical part of certification.
Now, insurance companies may have a variety of requirements for their doctors that might be related to their willingness to accept reimbursement from the insurance company or their willingness to follow certain kinds of insurance company policies. One of the challenges with insurance company-issued credentials is they’re issued by each insurance company in a very fragmented delivery system. So Aetna might have one set and Blue Cross might have another and United might have a third.
One of the values and virtues of board certification is it’s a truly national credential. In fact, it’s an international credential. We give exams all over the world.
Dr. Moore: Well, then if it exists why do we have this fragmented system with the insurance companies doing a different certification process?
Dr. Baron: I can’t speak with any intimate knowledge of an insurance company’s certification process. I’m not entirely sure I even am familiar with what you’re describing. Insurance companies create networks and they may have qualifications to be in a network but those might relate to, for example, the hours that a doctor’s willing to offer, or the after-hours coverage that a doctor has in place.
My guess is that insurance company expectations that physicians focus more on the service aspects of their practice whereas the board certification focuses exclusively on medical knowledge and skills.
Dr. Moore: Yeah. And I appreciate your pointing out. I was using certification loosely there. Yeah. I think they don’t label it that. It’s something else. To be on panel with an insurance company has some other name.
So I think about the other requirements. I see health insurance plans and other groups that are asking physicians who are affiliated with them to demonstrate their capacity to deliver good care by having them bring up a lot of process indicators and quality indicators of care and I hear that that’s a huge burden.
So why do we have that on top of a certification process?
Dr. Baron: Well, that’s a great question and I think that the answer is a combination of history and information technology. The original creation of quality measures as we know it was something that actually happened on the employer side and the insurance side. It didn’t happen on the physician side and I think, frankly, as physicians, we can and should be a little embarrassed about that. That we weren’t the ones who developed the original quality measures.
A quality measure by its nature is trying to get at how often do you do the thing you were trying to do. How often do you get a good outcome that you were trying to get? How often do you get a bad outcome that you were hoping not to get? By its nature, those kinds of data are always numerator/denominator data. The question is the rate at which you’re achieving an outcome.
So early on in my career when I spent 30 years in community practice and I had a leadership role in a Medicaid managed care organization in the ‘80s and ‘90s when managed care was just getting started, I would meet with groups of doctors. And I could ask a group of doctors how many of you can tell me what the American Cancer Society recommendations for mammography and every hand would go up.
In those years, the recommendation was every woman between age 50 and 65 should get a mammogram every year. That recommendation’s changed but that was the recommendation then. All the doctors could tell me that. But if I asked them okay, doctors, how many of you can tell me the rate at which the women that you saw last year who were between age 50 and 65 when you saw them, how many of you can tell me the rate at which they had a mammogram last year? And the doctors would all look quizzically at me and say how would I know that? I’ve got a bunch of paper charts on a wall. I have no way to know the rate at which the women got mammograms, or the children got immunizations or any of those things. I can’t give you rate data because I’m on paper.
The first quality measures were developed by the insurance companies because they had computers. If a woman had a mammogram they paid a claim for it and they knew looking at their computer the gender of the patient and the age of the patient. And so they could generate a report that I in practice with paper charts could not generate that would tell me the rate at which certain things happened. And I think those are very important measures and I think doctors are still wrestling with how those are helpful, how those help us practice better because there’s no question that they can. But for that to happen, the doctors and the data need to enter into conversations and dialogue and relationship that doesn’t reliably happen.
Dr. Moore: It makes me think there, as you describe it that way, the insurance companies have the claim data. They can look to see what’s coming through their information flow and come to some conclusions about what of those elements demonstrate what looks like or could be quality indicators. But the lived experience I think of many of our colleagues is that it’s not just coming out of a claims data flow. It’s coming out of a lot of requirements for bringing supplemental data to bear and that comes at a very high cost of documentation. The big burden on reporting which I’m hearing is diverting attention and time away from patient care.
Dr. Baron: I think that is a factor on the ground today. I think the world that we imagined was once doctors got to a world where we were on electronic platforms, we should be able to understand numerator/denominator data about our own patients.
For a whole variety of reasons, most doctors in practice don’t understand that about their own patients and they wind up either being asked to produce data like that for an insurance company—but again, most doctors’ practices in this country—when I was in community practice, a certain percent of my practice was Aetna patients and a certain percent was Blue Cross patients and a certain percent was United patients and a certain percent was Medicaid, and certain percent was Medicare.
Each of those insurers was only interested in the performance indicators of the patients of theirs that I took care of. And I was taking care of patients from six or eight or ten different insurers. A much better way to do this is to try to do it in the medical office space and for it to happen on the provider’s side, but we’re still in a period where the doctors are working through their ability to generate these kinds of data and the hospitals and health systems they work with are generating that. And the insurers who are often working with employers directly are trying to produce data for their employers who actually aren’t that interested in all of the insured’s population. They’re interested in what’s happening to their employees.
So you have different groups looking at different segments of the population. As somebody who was in practice, the systems that I built served all of the patients I took care of and I, therefore, thought it was an advantage to be doing that stuff at the practice level because it was less fragmented and gave me more good data to get better.
The insurance company data when I got it was always just a portion of my practice and of course, on the employer level it may just be some number of employees as opposed to something that can really be interpreted because those employees are being cared for across a variety of different systems.
Dr. Moore: It’s an interesting problem. I mean that is where the idea of how you frame the population to come to some understanding of level of care delivery in your practice versus what’s happening to my employees, that’s just how I sort different datasets and make that work. But there’s another piece and you somewhat touched on this early on.
I’m wondering when we have this “measure what’s available to us” strategy, there’s the risk that we’re not measuring what really matters and that what’s really important. And I wonder if you have a sense at ABIM that we’re measuring the essence of what a general internist is as a physician, as a professional, when we’re measuring the rate at which they’re appropriately capturing A1Cs and getting them to a good level.
Dr. Baron: So I know we are not measuring the essence of what makes a good internist. I know we’re not. And I think it is a sad reality that one confronts both the limits and the value of what one can achieve. From a patient’s point of view which is I think the one that really matters, the excellent internist is going to have spectacular communication skills. We don’t have very good ways to get at that.
The excellent internist is going to be accessible. You’re going to get an appointment when you need one. From the board’s perspective, we don’t have any insight into appointment systems.
The excellent internist is going to work with a wonderful team of courteous, friendly, compassionate, empathetic staff. We don’t have much visibility into that.
The excellent internist is going to be affordable. Is going to be in a network that the patient has access to. Again, as a national-level credential of doctor expertise, we don’t have insight into that.
So there are lots of ways in which one can devalue what a board credential might mean because it doesn’t capture some of the essence of what makes an excellent doctor. That said, we do a very good job identifying which doctors have stayed current in the field and which doctors have not.
And for example, we just published a study in BMJ Open where we looked at how doctors perform on the exam that we give. The doctors compared to—half the questions on the exam are diagnostic questions. Questions about diagnostic ability and diagnostic reasoning. Comparing the doctors in the top third performance on that to the bottom third of performance for some presenting complaints that are frequent sources of diagnostic error, things that could be something or could be nothing. A headache. Maybe it’s just a headache, maybe it’s a warning sign of meningitis or a stroke.
Well, looking at the way doctors performed in real life on complaints like that, we found the top third was 30 percent better in avoiding unexpected hospitalization, emergency room visits, or death than the bottom third.
So diagnostic ability matters to patients and doctors who struggle to pass our exam are doctors who may not have the knowledge they need to take care of patients. That’s not the only thing that’s the essence of being a good doctor, but it’s an important thing to being a good doctor and one that these days we can’t just presume that people have.
Dr. Moore: That is very interesting as I think one of the challenges for me is I was thinking about how I was being defined as a family medicine physician around my appropriate antibiotic prescribing for strep throat. Of course, that’s important but what’s the relative importance compared to am I doing well for all the people who come so that they’re less likely to end up in emergency room, in the hospital, readmitted after a hospitalization. And so you were able to study that against the terciles in your competency exam.
Dr. Baron: And demonstrate that performance on the exam correlated importantly with clinical performance.
I mean look, one of my favorite quality measures ever is the rate at which when a surgeon takes a patient to the OR for appendicitis what’s the percentage of the cases that they take to the OR for appendicitis that actually turn out to have appendicitis. And this was one of the first quality measures, Gordon, because the surgeon is the one who makes the decision to go to the OR but it’s a pathologist in the laboratory downstairs who makes the decision on whether that thing they took out actually was an inflamed appendix. And you know and I know that appendicitis can be a very difficult diagnosis to make and you really can’t be 100 percent sure as a surgeon when what’s going on is something that’s going to be appendicitis or not.
So in the ‘50s and ‘60s, there were creative leaders who started saying you know what? For the surgeons that we have, we’re going to have all the path specimens. We’re going to create numerator/denominator data. What’s the rate at which when Dr. Baron takes a patient to the OR for appendicitis how often is it appendicitis?
And you might think—if you were not knowledgeable about how clinical measures work and how uncertainty works, you might think I want to see the 100 percent surgeon. I want to see the surgeon who always gets it right. But it turns out that when you look at this in practice, rates of appendicitis will range from 25 percent to almost 100 percent. But the best surgeons are around 85 percent. And why do I say the best surgeons? Because one of the challenges in making the diagnosis of appendicitis is if you wait too long, if you sit on it, you’re letting an infection fester inside the abdomen and the patients have much worse outcomes if it’s left in there a long time.
So 100 percent’s the wrong number, but 20 percent’s the wrong number too. A surgeon who’s only getting it right 20, 25, 30 percent of the time is a surgeon who could be doing better.
I’d love to see a world where all of us in clinical practice are asking ourselves questions about how effectively we are achieving the goals we go to work to achieve using the tools we have in practice to measure that, and thinking about ways we can make that better. I think that’s core to being a good doctor.
I think that’s very difficult to measure. It’s very difficult for boards to measure. And quality, of course, is a team sport. But I think that we would all be in a better place if doctors had the core skills of knowing how to ask whether they were achieving the goals they were trying to achieve.
Dr. Moore: One of the problems that I think we sense is that we—how do I put this? It’s like we study to the test in a sense. We know we are being measured and we’re very aware of that. And if we’re being measured in a way that’s weird, it may somewhat engender weird behavior.
So for instance, I read a study and that guideline-driven care that’s insensitive to patient variation and comorbidity may drive some clinicians to push on a medication which would be inappropriate and therefore actually put somebody at risk. For instance, aggressive pursuit of hemoglobin A1C in somebody over 75 and now you’re putting them at higher risk of mortality.
And so I wonder how well we’ve defined what it is to be a good physician and how much we’re falling into the trap of studying for the test and invoking weird behavior.
Dr. Baron: There’s no question that it’s a risk when you apply measurement that you will generate weird behavior. And what I love about the appendectomy example is it’s very clear that 100 percent is patient damaging bad behavior.
So I think that physicians have a responsibility to understand that even when there’s a measure out there the goal is not to get 100 percent. That mammography rates—if a woman is dying of a glioblastoma and she’s 57 years old, sending her for a mammogram is not good care.
But I think we have to take seriously that unmeasured care is not good quality care either. The idea that you think oh, I’m just going to work with a good intention and I’m sure I’m doing well. Every time I ever measured the outcomes I was achieving in care, they were lower than I wish they had been. And I think that’s our continuing reality.
We’re not doing as well as we think we are. We often mistake our intention to do the right thing for achieving doing the right thing. And I think objective measurement helps us focus.
Now, we have to collaborate and partner with people who are measuring in a way that is constructive. That we have to understand that—people who are measuring shouldn’t create systems expecting 100 percent because then they will get really bad behavior. But people who are practicing should be able to understand the difference between 20 percent, 85 percent, and 100 percent as being very meaningful differences to patients and we should be striving to be in the right place.
Dr. Moore: So now let me posit that I think—I have the sense that we have a lot of indicators that talk about processes of care that are important because they can lead to better health outcomes as you’ve described. And that if we’re not attending to these where we fall into the heuristic where we believe we’re doing the right thing but it’s not always true. So it’s good to know and look.
But you touched on in the BMJ Open paper the idea that you can track things like hospitalization rate, ED utilization rate, and those are attractive to me because in a sense they—to get a terrific risk-adjusted rate it requires that you’ve done a lot of things right for people.
Obviously, there are confounding factors that have to do with society and social determinants and other things that are difficult for physicians to control. But presuming that some elements of those outcomes are health care delivery driven, what about using those as performance indicators?
Dr. Baron: So I think we’re on a kind of national and international journey to find the right performance indicators for sure. When I talked earlier about doctors on paper not knowing the rate at which things happen, I think that is true. And I think even with doctors on electronic platforms it’s often still true that they don’t have a reliable sense they trust about whether the measure that they’re being given is accurately and fairly a measure of them.
I think that continues to be an open question. But what’s interesting to me at the time was when a doctor makes a recommendation to a patient to take a blood pressure drug or to try to control their diabetes, what the doctor’s relying on is numerator/denominator data that somebody else gathered.
Somebody designed a study and said if I take patients with high blood pressure and I give half of them this drug and half of them I don’t, which ones wind up with fewer strokes. So the methodology that I think is now available in practice to be thinking about numerator/denominator performance data, that’s the very same methodology that we use in medicine to construct our standards of care.
We know when we recommend a blood pressure medicine we are recommending it because there was a population-level well-designed study that said people who took it did better than people who didn’t take it. That said, we always have to customize that for the patients in front of us because maybe the patient had a bad reaction to the drug. Maybe that drug isn’t covered by the patient’s insurance plan and they’re going to have to pay some outrageous amount of money for it and then they’re not going to be able to buy healthy food. Or they’re not going to be able to live in a safe place. Or they’re not going to be able to feed their children.
So we have to understand that the numerator/denominator data that we’re talking about as performance measures in practice is itself the basis of the standards we have for how to practice and that those standards themselves don’t have 100 percent applicability and never will. That doesn’t mean they’re not worthwhile and they’re not helpful. Just like it doesn’t mean the performance measures aren’t worthwhile and aren’t helpful. It just means that you have to understand the limitations of them and you want to use them to make care better.
Dr. Moore: The National Academies of Sciences, Engineering, and Medicine just came out with a report for implementing high-performing primary care. And I would love your thoughts on how we could define a high-performing primary care physician. What would they look like and how would those signals be recognized by others?
Dr. Baron: Well, I think the challenge in primary care right now and has been a challenge for a while is I don’t think we have either inside the profession or outside the profession a very well-articulated understanding of what we really want primary care doctors to do and what the resources they need to have to do that turn out to be.
So to be more concrete, we may expect a gastroenterologist to do colonoscopies and we understand that if they’re going to do a colonoscopy, well we need their time. So there’s some amount of money being paid for their time doing the procedure. But we also understand that they need a colonoscope and a monitor and a table and a nurse to assist them and that that colonoscope has to be cleaned. And if we’re going to have them do colonoscopies, we’ve created a whole revenue stream for the colonoscopy and a different revenue stream for the doctor.
Same thing with a surgeon and an operation. Surgeon goes into the OR and operates. They get a fee for doing the surgery but there’s a bunch of money being paid to create the environment in which they do that work. And I think we’re still stuck with a primary care model of “Dr. Wellby” in an office with a lady who answers the phone and another lady who takes the blood pressure and does the EKG and “Dr. Wellby” doing everything else. And I think that’s an undertheorized understanding of what we want primary care to be. And the National Academies report calls out a lot of other aspects of primary care that we want to see.
One of the reasons why we don’t see them is the payment system doesn’t have the resources for the primary care doctor to have the team of people who could be managing the information technology. And the primary care doctor shouldn’t be the one calling people up and saying gee, the last time we measured your A1C it was poorly controlled and you—how are you doing? Or we started this new medicine last week. How’s it going?
Those are things that can be structural properties of a team but you have to have a team. And you have to have a system that will put forward who are all the patients we started on new medicines two weeks ago. What are all their phone numbers? Who could call all of them and say did you fill the prescription and how’s it going? It’s pretty basic stuff but if the only person who’s funded to be supportive in the model is the doctor you’re not going to get all that stuff.
So I think part of the problem is the way that the insurance system doesn’t provide the resources that people need to practice primary care effectively. But the other part of it is that I don’t think we’re training primary care doctors effectively to know what to do if they had the resources. They’re not reliably trained to create the team. They’re not reliably trained to think about the patients who aren’t in front of them. They’re not reliably trained to use the information technology to help identify the patients who may be in special need of closer follow-up.
That’s where primary care can contribute value. In the emergency visit that doesn’t happen, in the hospitalization that doesn’t happen. But for primary care doctors to do that, they need to move from a model where it’s visit based which is what they get paid for and then they’re waiting for the patients to come in. They don’t have the capacity to do the proactive outreach that would head this stuff off faster but that, from a business point of view, requires new overhead for a primary care office that is not supported by the revenue that flows to a primary care office when it’s all visit based.
Dr. Moore: Yeah. I think about some of the chronic models that talk about working within a multidisciplinary team creating opportunistic care opportunities. So for instance, having signals in the electronic medical record that this person has a care gap for prevention or chronic disease management and that person shows up and needs a parking voucher or a refill and a secretary can say oh, wait a second, I see this flag. Let me grab a nurse because there’s something we need to be doing for you.
So that IT, that technology is necessary but comes with a cost. And then, as you said, the people who aren’t showing up. And so having the technology to create the lists that then go to people who do the outbound calling. Again, that’s a cost. I heard from some children’s hospitals during COVID times that they use some sophisticated algorithms to identify kids with high burden of illness so that they could call out when they weren’t coming in for necessary follow-up to say we’re worried about you. How are you doing? Even though we’re not going to make you come in during COVID times because that’s scary too. All of that comes with cost.
Dr. Baron: Exactly. And I think there’s a return on investment for that. And I know many of the listeners to your podcasts are employers and purchasers of health insurance on behalf of their employees and they have a role here. There are now models of advanced primary care, the patients that are in medical home, various kinds of federally sponsored comprehensive primary care initiatives that are designed to expect a higher level of collaboration teams in exchange for more resources and I think employers are critical in supporting that.
I think increasingly also larger health systems are owning and purchasing primary care practices. So the resources that go to primary care practices are not just determined anymore by insurance companies. They’re determined by the large health systems.
So the large health systems are in effect businesses that receive some amount of revenue and then decide how to allocate that revenue internally and they’re making decisions that don’t support the creation of teams in primary care.
So I think employers are entitled to ask about what the health system investment in primary care is, as well as what the insurance company investment is in primary care. And there are a variety now of primary care models that can be recognized, can be supported that purchasers should be demanding that their employees, the beneficiaries that they are creating with the insurance they are purchasing, that those folks have access to that kind of primary care.
Dr. Moore: There are so many interesting avenues that we could go down but I also want to be cognizant of the time. This has already gone on a little bit longer because the conversation’s just fascinating. But Dr. Richard Baron, I want to thank you so much for your time and your insights.
Dr. Baron: Thank you, Gordon. It’s really a pleasure to talk to you. I appreciate it and thanks for helping getting out the message of how important primary care is in the high-functioning health system.