Managing Physician Behavior and Diagnostic Errors to Improve Patient Outcomes

Key Highlights

• Clinicians should distinguish diagnostic errors caused by physician misjudgment or delay from systems failures and behavior-related practice patterns, such as excessive or fear-driven testing.

• Cognitive biases, particularly premature closure and overreliance on initial impressions, remain major contributors to missed or delayed diagnoses.

• Psychologically safe, peer-based review processes are more effective than top-down audits in identifying root causes and improving diagnostic performance.

• Despite advances in technology, accurate diagnosis continues to depend on thorough history-taking, targeted testing, careful data interpretation, and ongoing reassessment of the patient’s clinical response.

In this episode of Podcast360, Norman Weinberg, MD, a retired primary care internist with over 40 years of experience, discusses the intricacies of managing physician behavior and diagnostic errors.

Diagnostic errors occur when a physician fails to arrive at the correct diagnosis initially or is delayed, while systems problems are external factors affecting diagnosis, and behavior problems relate to the physician’s habits in data collection. Despite technological advances, the physician remains central to diagnosis, requiring careful questioning and data interpretation.

Transcript:

Kate Young: Hello and welcome to another installment of Podcast360, your go-to resource for medical news and clinical updates. I'm your moderator, Kate Young, with Consultant360, a multidisciplinary medical information network.

Today, Dr. Norman Weinberg is here to talk about managing physician behavior and diagnostic error. Dr. Weinberg is a retired primary care internist from Lexington, Massachusetts with over 40 years of experience working in healthcare. Before his retirement, he served as the medical director of Emerson Hospital Physician Hospital Organization, an expert reviewer for the Board of Registration and Medicine for the state of Massachusetts, and has reviewed for several medical journals. He has also lectured at Harvard and several national conferences about physician errors and has served as the chair of the Patient Care Improvement Committee for Partners Community Healthcare. Thank you for joining us today, Dr. Weinberg.

Let's start by defining what we mean when we're talking about diagnostic errors and physician behaviors.

Norman Weinberg, MD: The definition of a physician error is when the physician is presented with a patient and begins a history, takes a physical, interprets the data and does not arrive at the appropriate diagnosis initially or is delayed in arriving at the diagnosis. So this involves those aspects of the diagnosis that are totally the responsibility of the physician. A systems problem are those issues that supplement a physician's ability to care for the patient and make a diagnosis. For example, if a patient is sent to a hospital from a nursing home and all the lab data and clinical information about that patient is not sent to the admitting physician, that's a systems problem. That is not a responsibility of the admitting physician. A behavior problem does not involve the diagnosis, but the habits and the way that a physician goes about making the diagnosis such as the number of tests that they order, how long they keep a patient in the hospital, it is in my view, independent of making the diagnosis. It is how he goes about accumulating the data to make the diagnosis such as is ordering or she an excessive number of laboratory tests or x-rays. So those are the three types of issues that I have dealt with in my career.

Young: How can clinicians help create a psychologically safe environment where diagnostic errors can be openly discussed?

Weinberg: What literature has shown is what is the safest environment for the physicians to feel comfortable being open about discussing their errors? And it is clear from a number of studies that the safest environment is among colleagues with whom they practice and whom they trust, whom they have respect for, who offer a diversity of opinion. And in addition to that, knowing that let's say it's a group practice, that each of the members when presented with their errors and having to discuss it, each of the members has their turn in the hot seat. And so they know that their colleagues are looking out for their best interest and looking out for helping the physician get better, which in turn helps that covering doctor take better care of their patients.

Young: Why do you think missed or delayed diagnoses remain so prevalent despite advances in technology and access to information?

Weinberg: Why is there an incidence of up to and perhaps over 25% of patients admitted to the hospital subject to adverse events? Given all of the advanced technology, it is still the physician who is the architect of making the diagnosis. So this is basic medicine. The physician has to ask the relevant questions depending on the chief complaint, has to do a careful examination according to what he feels may be the diagnoses, order the appropriate tests and interpret information into a coherent diagnosis. The advanced technology allows gathering of information in a much more sophisticated way than it used to. So you can have advanced MRI or nuclear medicine, anything that you could consider state-of-the-art these days in the most advanced hospitals. All of that still depends on the physician asking the right questions, formulating a diagnosis and interpreting the data that he requests. He is the architect no matter what the technology is that allows him to gather information. The physician is the architect of making the diagnosis and that has not changed no matter how sophisticated the technology is.

Young: Why do audit and feedback processes often fail to change diagnostic behavior from the clinician's perspective?

Weinberg: The main reason is that doctors are not very good at unveiling the root cause of why a problem occurred. How do you discover a root cause? And what I'm going to start with is explaining to you the theory of five whys, which actually was a business concept, and it was actually a professor at Harvard Business School that explained to me, and I'll give you an example. In a factory in Japan, there were an inordinate number of fans that the factory assembled that were coming back with problems of the fan blades loosening. So the first question is, why were the fans… Why is the product being returned? It's because the fan blades were loosening. Why were the fan blades loosening? Because there was no information or instruction or guidance of how tight the factory workers should turn the screws. So they had no way of knowing what was too much. Why was that? There was no guidance, no information. They were left on their own to determine what is the appropriate tension to apply to the screws that attach the fan assembly. Understanding that, the solution to the problem was they suspended the wrenches from the ceiling so that the wrenches could not be turned past a certain point and overtighten the blades. And that's the five whys arriving at a solution that eliminated over-tightening the screws.

The same is true for medical errors as well as systems and behavior examples. So a physician example of the five whys is, and this is a clinical example, it's an actual one. A patient comes in to the emergency room, he's in his mid-eighties. He's a bit confused, no fever. Family comes in and says he hasn't had a bowel movement. He's really constipated. I think that's his problem. The clinician does an examination, orders an x-ray. The patient had a Foley catheter because of incontinence, and the x-ray of his abdomen shows a bowel that had a large amount of stool in it. And so the diagnosis of the emergency room physician was, this is constipation, and they admitted the patient for cleansing enemas.

The other lab data was that he had no fever, that his white blood cell count was normal with a little bit of a left shift and that the urine analysis showed a number of bacteria and white blood cells. So he was admitted with the diagnosis of constipation. A day later, he developed a very high fever. He becomes hypotensive and cultures are ordered, both urine and blood, and the next day, the urine and blood cultures all came out positive for a gram-negative bacteria. And the diagnosis was actually urinary tract infection with sepsis and a paralytic ileus. I should say that the bowel sounds on admission were absent. So the admitting diagnosis from the emergency room physician was wrong because he ignored the bacteria in the urine. He looked at the x-ray, which showed a lot of stool and listened to the family about constipation and came to the conclusion it was constipation. And he ignored that there was a left shift even though the white blood cell count was normal. And why was that? Is because he reached a premature diagnosis based on what the family told him, which influenced how he interpreted the x-ray. So that was the root cause was being accepting the family's diagnosis to the extent that he ignored the clues that were there for bacterial urinary tract infection, sepsis. That was the root cause. And the emphasis was to listen to what the family says, but do not jump to conclusions without addressing all of these little clues such as a left shift and a little bit of confusion. That was that.

A systems example, and this is a real example, a systems example. I was reviewing one day all of the admissions to our hospital of patients with chronic lung disease. And every admission said for the past number of days, the patient has been increasingly short of breath, having a cough, no fever, and eventually reached a point where he could not be maintained at home and was admitted to the hospital. And I said to myself, where was anybody during the five days? And the first why was, why did this go on for five days without any input from physicians or nurses? The second why is because the family and the patient had no way of understanding what are the signs that the patient is clinically deteriorating, and then why is that? Because there was no system in place to educate the patient about what are the signs to watch for when to call for help before the patient has to be admitted. So my answer was I gathered the home care nurses and social service and we developed a specialty team of nurses who are specifically trained in understanding when patients with chronic lung disease are deteriorating, who would go out, instruct the patient and let the patient know at the beginning of when they started to deteriorate to contact the nurses, have the nurses assess and make an intervention before they became too sick to monitor at home. That was a systems problem. There was no diagnostic error there. The patient was admitted because they weren't treated. The patient didn't know that how sick they were and that was the reason why.

A behavior example is regardless of how accurate diagnoses are, a physician is ordering an excessive number of tests to make a diagnosis. So when the data came back from the hospital and from insurance, he was way out of line with other physicians in terms of his costs per patient in laboratory tests. So in a case like that, why was this doctor ordering so many tests? And further analysis showed that it tended to be newer doctors, first year residents, doctors just out of training. And it's clear from studies that experienced clinicians can make a diagnosis with less data than new physicians. So they needed more data to make the diagnosis. And then when you talk to the doctors, they would say, I was afraid of not knowing certain tests because of what the attending doctor would criticize me for not knowing that. So I ordered lots of tests. Educating this doctor on ordering of tests is not helpful because the doctor was afraid of being found to be incompetent in the eyes of the attending physician. It wasn't that he didn't know to order the test, they were just number one, he was new and needed more data. And number two, he was afraid of what the attendings would say if he didn't know the answer to some of the questions regarding tests. So educating this doctor on tests is irrelevant to addressing the real problem, which was his fear of being castigated by the attending physician. So that's an example of behavior had nothing to do with making a diagnosis.

Young: How should clinicians measure whether a behavior change has truly improved diagnostic performance?

Weinberg: The failure to improve clinically is fundamental basic medicine. You look at the patient, are the symptoms getting better? Is he or she more alert, eating better, less pain, less cough? Is the x-ray improving? Are the x-rays improving? Are the blood counts, the white blood cell counts improving? So this is basic medicine to assess is the patient getting better or not? And that's just a clinical assessment by the doctor of reinterpreting data on the second, third, fourth day compared to what it was when the patient was admitted. And that includes not only the clinical status such as the mental status of the patient, how the appetite is, what the fever is, what the white count is, what the chest x-ray. These are basic things that allow the physician to determine whether or not the patient is improving. And it really is not so much a question of how sophisticated the data is that was used to collect the information.

Young: What role does professional culture play in either reinforcing or reducing diagnostic error?

Weinberg: The professional culture is what I had referred to before, which is in order to improve diagnostic performance and reduce the rates of physician errors, one needs to understand, first of all identify physician errors and then identify the root causes and the culture that allows one to expose root causes is what I referred to before, which is the physician's confidence in the person or the people that they're exposing themselves to. So if these are people that they have relationships with, that they trust, that they go beyond, sometimes it's a relationship that goes beyond medicine, that each one of them knows that the other doctors that they're exposing themselves to have only in mind the betterment of their treatment and that these other doctors also face the same treatment when it gets their turn. This is different than having someone, an expert or an expert clinician or a trained nurse or someone from the insurance company or an administrator come in and talk to them about their errors. They are not going to feel open about that. They will be concerned about where is this going to go when I expose myself. So the culture is one of a small group of physicians with whom the doctor feels comfortable, preferably one, a group with whom they practice with meet with regularly, and each of them is comfortable exposing their fallibility.

Young: Given everything that we've talked about today, walk me through what a successful diagnostic process looks like in 2026.

Weinberg: A successful diagnostic process is really unchanged. It's old fashioned. A successful diagnostic process is has the appropriate history been taken that is relevant to the chief complaint that the patient presents with? Are the appropriate laboratory and x-ray tests ordered? Are they interpreted correctly? And then initiate a therapy and then monitor things that I'd mentioned before. The clinical status of the patient. Are they more alert? Is the fever down? Are they eating better? Are the blood counts getting better? Is the x-ray getting better? So that's old fashioned. How do you assess whether the diagnostic process is successful, is no different now than it was 30, 40 years ago. It's looking at the patient and seeing whether they're getting better by the various modalities in 2026. There certainly are more modalities, and we have more information and more sophisticated collection of data than we did before an MRI was invented or any of the other nuclear medicine tests that we now have that they did not have available then. But it still comes down to looking at that data and interpreting it that fits into what the doctor has made as an initial diagnosis that's relevant to the chief complaint.

Young: Thank you so much for joining us today, Dr. Weinberg.

©2026 HMP Global. All Rights Reserved.
Any views and opinions expressed are those of the author(s) and/or participants and do not necessarily reflect the views, policy, or position of Consultant360 or HMP Global, their employees, and affiliates.