The most critical of a physician’s skills… “It is every doctor’s measure of his abilities; it is the most important ingredient in his professional self-image” . This quote resonated with me and started my journey about how physicians make diagnoses and how we are vulnerable to making errors. This following story of mine keeps me humble as I care for someone’s child. I have made errors that have almost led to harm or have inadvertently harmed a child. There was a little boy admitted to the pediatric hospitalist service after evaluation in the emergency room. He was given a presumptive diagnosis of pneumonia with effusion. This assessment traveled from physician to physician until a wise radiologist who called himself an “old country doctor” looked back at a chest X-ray and noted a mildly enlarged heart. This child who was being treated for pneumonia and effusion actually had myocarditis and was in early congestive heart failure. I remember this child well because I was rushing, had some personal issues at home and was overwhelmed with pages, teaching and the pressure to discharge patients in a timely manner. When I realized my error, I was aghast. Because of initially missing the diagnosis, the child almost underwent an unnecessary surgery. My “physician thinking” had been compromised because of numerous external pressures and the lack of realization that I must reflect upon my own thinking process so as not to harm children.
We need to make accurate diagnoses as inaccurate diagnoses may lead to patient harm. My inspiration was initially drawn from Dr. Jerome Groopman’s book, How Doctors Think , in which he brings to light errors that physicians may make in patient assessment and diagnosis. Dr. Groopman, an internist from Harvard Medical School, begins with a story about a woman suffering from chronic weight loss, intermittent vomiting and other gastrointestinal symptoms. Over many years, multiple practitioners told her that she had diagnoses ranging from anorexia nervosa to irritable bowel syndrome. This woman, not surprisingly, became depressed because of her chronic illness and, as a last hope, went to see a famous gastroenterologist. The patient then recounts that this physician simply listened to her story, despite an enormous medical file full of consultant reports, laboratory results and radiology studies. This gastroenterologist was not influenced by the past workup of this patient or by others’ opinions. He told the woman that he thought she had celiac disease which was ultimately confirmed after appropriate testing. This woman regained her weight, spirit and life.
My personal journey: After reading this book, I decided to learn more about how doctors think and make clinical decisions. This has turned into my life work. While conducting an informal internet search, I stumbled upon an entire literature of physician cognition and learned that there are over fifty types of diagnostic errors that physicians can make. I read about what I have struggled with for years: as a pediatric hospitalist, I have made errors in diagnosis, judgment and communication because I was in a hurry or because I listened to someone senior to me and not to my inner voice. In fact, numerous studies, especially in the patient safety internal medicine literature, show that diagnostic errors comprise the largest category of errors made in medicine. I read all that Pat Croskerry has published, a well-known emergency physician who writes extensively about the impact of cognitive biases on diagnostic errors. He describes cognitive errors such as “anchoring” in which the physician may decide too early in the presentation of a patient his or her diagnosis and then unconsciously looks for information to confirm that diagnosis . Another error that I have seen in the hospital setting, which the woman in the clinical vignette may have experienced, is “diagnostic momentum” in which “once diagnostic labels are attached to patients they tend to become stickier and stickier.” All other possibilities are excluded in the treating physician’s mind.
Caring for children: Unique characteristics in caring for children while being cognizant of diagnostic errors brings forth challenges. As I tell my residents and medical students, if a child is hospitalized, consider the family members to be your patients as well. Parental emotions that arise when their child is ill are often deep and may unwittingly bias a physician when he or she is attempting to make an accurate diagnosis. When parents are tearful and angry, the history can become extremely convoluted and trying for the physician to take accurately. I recall interviewing many scared and exhausted mothers who are choking back tears while trying to explain the symptoms that their child is experiencing. In the instance of non-verbal children, pediatricians have to rely upon the caregiver to explain the pain that their child is experiencing. Many a time I have asked mothers if the babies are having stomach pain, and I can then ask the mother if she picks up her baby, when the baby’s stomach is against her own, does the baby cry or is the baby consoled? There is a well described phenomenon of “paradoxical irritability” when a baby may lie quietly in the crib, but when held, the baby cries more. It suggests meningitis as the meninges are under stress when the baby is held. These time honored “tips” for diagnosis, however, can be tough to teach and to transmit to a younger generation of doctors. Younger doctors, while facile in gathering and locating up-to-date information using electronic resources, may not be as not experienced in relating to the parents’ emotions and identifying subtle cues in non-verbal children. For example, babies with mildly increasing abdominal distension or a slight pallor to the skin can be a foreshadowing of an evolving ominous diagnosis. The challenge is for the pediatrician to recognize the subtle cues in ill neonates and children.
Pediatricians also care for older children who are non-verbal. We care for special needs children and teenagers, who may present with a chief complaint of, “something has changed in my child. He is not the same.” At these times, I rely heavily on the caregiver for assistance in making a diagnosis, as the physical examination may not be revealing or serve as a guidepost as much as we would like. When I tell a parent/caregiver my diagnosis, I look for agreement. I often tell the parents that, “you are the experts in your child and we will work together to make the right assessment and plan of care for the child.” I often tell my learners the old adage that, “If mama ain’t happy, no one is happy.” I find that if the parents trust me and agree with my diagnosis, I am often more right than wrong. If the parent looks like they do not agree with my diagnosis, I force myself to reevaluate my assessment. I try to role model and teach the learners that “I do not know” is a safer answer than telling your thoughts to a family that are based on inaccurate or limited data. As I often tell learners, the art of medicine is to make an accurate diagnosis while also showing caring towards the patient and the family. One day, during inpatient family-centered rounds, our team saw a 6-year-old with increasing abdominal distension that was worsening over the weekend. When I walked in, the parents were tearful and frightened. I myself felt nervous upon seeing the patient and we asked the critical care team to transfer the patient to the intensive care unit. To facilitate this, I was walking in and out of the patient’s room, making phone calls to both the charge nurse and intensive care unit attending. I then walked the child to the unit. The parents followed behind us, holding hands and crying softly. I then turned to them and said, “I am so sorry I was walking in and out of the room. I just wanted to make sure that we were taking care of your child as quickly as we could. I do not know what is making your son sick, but he will be watched closely and be safe.” They actually then hugged me and thanked me for caring for their son, while I was feeling badly that we had no clear diagnosis.
As pediatricians, our greatest ally in caring for the children are the parents/caregivers. In most cases, if the parents are smiling on rounds, I can tell, almost without examining the child, that my patient is well. When I round and the parents are hovering anxiously, look scared, angry and/or sad, I know that their child is ill and I must address their concerns and rethink my initial impressions about the patient. Pediatricians are also trained to be very thoughtful and questioning of all laboratory/radiology tests ordered. For example, if a child presents with abdominal right lower quadrant pain and vomiting, I would hesitate to order an abdominal CT scan because of concerns about radiation exposure to a growing child. A limited right upper quadrant ultrasound may be a better choice. Similarly, there are many drugs that are not approved for use in pediatrics and we often use drugs “off label” as they have not been studied in these populations. Of no surprise, parents can be hesitant to enroll their children in investigational drug studies. Pediatricians should pause before ordering multiple laboratory tests: children, especially neonates, have much lower blood volumes and literally I have had to transfuse patients because so much blood was drawn to establish a diagnosis. While this is not always wrong, I encourage pediatricians to think cautiously about all tests that are ordered. Overall, encourage your colleagues to make honest and thoughtful assessments of their patients and never hesitate to rethink their initial impressions. This is someone’s loved one, after all.
Teaching clinical reasoning
- Make it fun! I have had the opportunity to teach in many venues, to many different learners, from medical students. People enjoy hearing about examples from popular culture so I often begin with Malcolm Gladwell’s book entitled Blink! In which he writes about art historians are able to tell (in a “blink of an eye”) whether the artwork is a fraud or not and how these historians are using their intuition, or system 1 thinking, and the pitfalls associated with this process. Marketing companies are aware of system 1 processes and their effect on what we buy. For example, a candy company prints the calories for their product in green because the consumer associates “green” with a good feeling, although the candy is calorie-laden. These examples help learners understand the concepts, keep the information interesting and inspire discussion.
- Make it real. Relate the didactics back to your patients. Use cases to teach clinical reasoning and ask your learners many questions that show their reasoning skills. As an educator, I have decided to speak less and guide more. I ask questions during rounds to assess what my learners are thinking. When a learner asks me if a consultant should be called, I rephrase the question back to them. “What do you think? What do you think is best for your patient?” for example. I have decided that I am not the “thinker” on rounds and that the learners are the “doers” although many times it would be more efficient to do all the thinking myself and then the learners carry out the orders. My value system as an educator is that we are teaching our future doctors that will be caring for me and my own children some day and they need to be taught to reason and interpret data in a thoughtful healing, caring and not in a harmful manner. Share your stories. Learners remember concepts when illustrated by stories about patients. Even more powerful is sharing your mistakes.
- Share cases and then use templates to assist learners in organizing their thinking (Table 1). For example, I have used the following case in my teaching.
A 13-year-old male with no past medical history is admitted with fever, bilateral thigh swelling, and abnormal liver transaminases. Physical exam reveals an alert and febrile adolescent whose thighs are warm, edematous, and tender to palpation. The symmetry of the findings is striking and unexpected. He is unable to walk secondary to thigh swelling and pain. You are perplexed by his clinical presentation. The patient’s mother is Spanish-speaking and after speaking to her with an interpreter, you are still unsure of the diagnosis.
Osteomyelitis crosses your mind but does not make sense given the symmetry of the findings and the elevated transaminases. While serum sickness is considered, you are reassured that the right consultants, i.e., rheumatology, are on board and can assist with the management of that illness. Later, the rheumatology attending pages you and relates his suspicion that the patient has scleredema. “What is that?” you ask, and he explains that this illness can be associated with streptococcal infections and is treated with penicillin. Of course, while ASO titers and other laboratory findings are helpful, ultimately the diagnosis is a clinical one. You review Pubmed, Up-to-Date and also conduct an internet search, but cannot find any published case reports in support of the rheumatologist’s diagnosis. However, you cannot think of any other diagnostic possibilities.
The next day, the residents suggest ordering a MRI. However, the patient has been started on empiric intravenous penicillin and appears improved, so you decide to wait on further testing. That evening, you receive a call from the rheumatologist who recommends starting high-dose steroids. You agree – you are already late for home and the recommendation seems reasonable in light of the possible diagnosis of scleredema, even though you do not fully understand it. Subsequently, a new rheumatology attending evaluates the patient and recommends a MRI of the thighs. The MRI revealed mutifocal osteomyelitis with moth-eaten bone and the patient is diagnosed with chronic multifocal methicillin-resistant Staphylococcus aureus osteomyelitis and myositis. He remained hospitalized for six weeks because of persistent fevers. The patient underwent three separate drainage procedures by orthopedics during his hospital course. The patient is now well and will remain on chronic suppressive therapy for at least 2 years.
This case included several examples of cognitive errors that ultimately led to the wrong initial diagnosis and treatment, as well as a delay in the institution of effective therapy for the correct diagnosis. Challenge yourself by reading the case again and identifying the cognitive errors that occurred. Answers are listed below.
- Framing effects: The emergency room inadvertently created a framing effect by consulting rheumatology and packaging the patient as an “unknown.” The ward team subsequently viewed the patient as having an array of symptoms that only rheumatology could address.
- Premature closure: The hospitalist considers osteomyelitis as a diagnostic possibility, but ultimately settles on serum sickness as a more likely etiology for the child’s condition. The hospitalist hoped that rheumatology would confirm the diagnosis and guide the treatment.
- Blind obedience: Despite evidence to the contrary, the hospitalist did not challenge the diagnosis of scleredema made by the specialist and did not pursue other diagnostic possibilities.
- Anchoring heuristic: The hospitalist and the rheumatologist did not entertain other diagnoses despite questions from other team members. Furthermore, the hospitalist was influenced both by time pressures and the need to “tuck in” the patient before the weekend. Both things, in effect, “anchored” the diagnosis to the patient.
- 4.Admit your mistakes to yourself and others. Share your stories of your mistakes. A physican who cannot admit any mistakes needs to gain more insight and needs feedback, in my perpsective. Get over the shame and learn so you can teach others to be better thinkers which then leads to better doctors. I have shared my mistakes through these endeavors to teach clinical reasoning, and frankly, while sometimes embarrassing, the process is powerful in starting a dialogue about why errors are made.
- 5.Reflect upon your own process of clinical reasoning.
Framework to analyze cases as a teaching tool about diagnostic error.
Used with permission of the author, Thammasitboon, S .
Specific strategies for reducing diagnostic errors 
- Emphasize the importance of the clinical exam.
- Promote a systematic approach to common problems.
- Expand your knowledge base using the best available evidence in the literature for easy access when needed.
- Promote the use of time-outs or pauses. Ask the team to step back and rethink the diagnosis or ask a colleague for a second opinion.
- Consider using the “worst case scenario” strategy to generate a differential diagnosis. But take care not to order unnecessary laboratory or radiographic studies, as this too might lead to an error.
- Keep asking questions. Why does this lab value not make sense? Why does the family seem skeptical of the diagnosis? What can I not explain?
- Acknowledge your feelings about a patient or family. Positive or negative feelings may bias your approach.
- Slow down. When individuals are rushed, more mistakes may occur.
- Learn, understand, and recognize the features of the cognitive process that can lead to a diagnostic error.
- Admit your mistakes. This can lead to reflection and change in behavior.
- When you make a diagnostic error, conduct a “cognitive autopsy.”
- Tell your history well and take good notes.
- Be the link between your physicians.
- Accept the fact that medicine is more of an art than a science.
- Research your diagnosis on the internet.
- Be aware of the red flags of misdiagnosis.
- Trust your gut instincts.
- Ask “Doctor, what else could this be?”
- Ask your doctor why she chose this diagnosis.
- Ask if there are any tests that would be helpful.
- Seek a second opinion.
We must teach medical trainees how to “think” like the expert to enhance diagnostic performance, and reduce errors. The content can be taught via a case-based approach with an open and honest discussion about individual diagnostic errors and experiences. A case-based didactic session can address theories about cognitive bias, clinical reasoning and their impact on medical decision making. The concepts of expert thinking using a blend of intuition (rapid unconscious thinking) and metacognition (deliberate, conscious thinking) to solve clinical problems can be discussed . Encourage yourself and others to reflect upon the “practice” of being a physician and to find a way in which to role model the importance of thinking about my own thinking to medical students and residents. In my journey to become a better physician and teacher, I have realized that the more that I am aware of my failings, the more that I am able to step back and take a look at the whole picture, to think, listen, and feel for patients and not just “do”, the stronger my healing has become.
- I still make mistakes in diagnosis even though I study and teach about this field. That continues to boggle my mind.
- Sharing your stories with others encourages open dialogue diagnostic error.
- Reflect upon your own mistakes. Is there a pattern of errors that you make?
- Work within your system to raise awareness about these types of errors.Define initiatives to combat them. Be persistent and continue to speak with people about this important area of patient safety. Ask families to share their misadventures with misdiagnosis. Ask administrators and other leaders to listen. Be a change agent.
- Support on-going and high-quality research in this field. You have a voice as I realized that I do. I knew nothing about diagnostic errors 5 years ago, and only because I cared, I have been able to share my thoughts with others through my commitment to raising awareness about the occurrence of diagnostic errors with others.
My human imperfections as a physician are on the forefront of my mind every day. As Atul Gawande writes in his book, Complications: A Surgeon’s Note on An Imperfect Science , “medicine is …a strange…and in many ways a disturbing business…The thing that still startles me is how fundamentally human an endeavor it is.” Dr. Gawande then notes, “You have a cough that won’t go away – and then? It’s not science that you call upon but a doctor. A doctor with good days and bad days…. A doctor with three patients to see and…gaps in what he knows and skills he’s trying to learn.” . That sounds like me!
In summary, my philosophy and devotion to the science of diagnostic error is typified by this statement: I will do no harm or injustice to them. (patients)-Hippocratic Oath.
Croskerry P. A universal model of diagnostic reasoning. Acad Med 2009;84:1022–8.
Groopman J. How doctors think. Houghton Mifflin Company, 2007.
Croskerry P. The importance of cognitive errors in diagnosis and strategies to minimize them. Acad Med 2003;78: 775–80.
Dr. Satid Thammasitboon MM. 2013. Personal Communication.
Graber ML, Franklin N, Gordon R. Diagnostic error in internal medicine. Arch Intern Med 2005;165:1493–9.
Quirk M. Intuition and metacognition in medical education: keys to developing expertise. New York: Springer Publishing Company, 2006.
Gawande A. Complications: a surgeon’s note on an imperfect science. New York: Metropolitan Books, Henry Holt and Company, 2003.