Dr. Seo
It is better to be feared than loved.
The associate dean of student affairs at my medical school embraced this motto. Although the dean of the medical school was titularly in charge, it was the associate dean who kept the school running. And we all feared her, just a little bit. Without ever raising her voice, she managed to communicate that she was not to be trifled with. When she stood to speak, the idle chatter stopped, and we all sat in the civilian equivalent of parade rest, awaiting whatever important information she had come to communicate.
In the 1940s, the Association of American Medical Colleges (AAMC) tried to rectify this problem by establishing the Cooperative Plan, in which medical schools agreed to withhold information about their students from the hiring hospitals until a pre-selected date, so offers of employment would come to students around the same time.
Human nature being what it is, training programs continued to game the system by giving medical students progressively shorter periods of time to decide whether to accept an offer. When the Cooperative Plan started, students typically had a few weeks to think things through; toward the end, students were given less than a day.
In 1952, the AAMC decided that enough was enough. It established a new system in which candidates and programs would list each other in order of preference. An algorithm was developed that used cooperative game theory to make the matches that worked out best for both students and programs. Thus, the residency match was born.1
The residency match is the last leg of a long race. Prior to the establishment of the Electronic Residency Application Service (ERAS) by the AAMC in 1995, that race was particularly arduous because each program was allowed to set its own idiosyncratic requirements for residency applications.2 The one enduring requirement has been the personal statement, in which candidates are invited to write about what drew them to a given specialty.
When one of my classmates asked the dean for specific advice on how to approach the essay, she said, “Well, it depends on what you’re applying for. If you’re going into ortho, you’re going to write, ‘When I was in high school, I broke my leg playing football, and I met an orthopedic surgeon, who was the coolest guy ever, and I knew I had to become an orthopedic surgeon, too.’”
After the laughter subsided, the dean smirked and said, “You laugh, but it’s true.”
Sherlock Holmes, Rheumatologist?
As a program director, I can see the truth to those words. When you read a large number of applications for rheumatology fellowship, some common themes emerge. The most frequent is an anecdote regarding a patient with an autoimmune disease who inspired an early interest in rheumatology. A common variation on that theme is when the patient is a family member. And then, of course, are the legions of applicants who want to become a medical Sherlock Holmes.
In Chapter 6 of The Sign of the Four, Sherlock Holmes and Dr. James Watson are trying to divine how a perpetrator entered a locked room. When Dr. Watson fails to come up with a plausible explanation, Mr. Holmes famously replies: “How often have I said to you that when you have eliminated the impossible, what remains, however improbable, must be the truth?”3
Our profession seems to have become inextricably linked to this style of investigation. A Google search for the keywords “rheumatology” and “Sherlock Holmes” yields 194,000 hits, the majority of which likely lead to case reports of improbable diagnoses made by exclusion of all the other possible diagnoses.
It’s our own fault, really. I’ve written countless consult notes in which I made the obligatory statement that many rheumatic diseases are diagnoses of exclusion and should be considered seriously only after the other consult services have done their part to rule out more common possibilities. Where did the concept of diagnosis of exclusion come from, and is it time for it to go away?
Excluding Diagnoses
Apparently, it started in Vienna.
In 1365, Pope Clement VI gave Duke Rudolf IV of Austria permission to establish the University of Vienna, which included a medical school.4
For a few centuries, the medical school wasn’t particularly good.
This wasn’t their fault. The pinnacle of medical understanding at that time was based on the works of Galen, whose conception of medicine was based on humorism (i.e., the belief that illness was the result of an imbalance in one of the four humors: black bile, yellow bile, blood and phlegm). Throughout the English language, the persistence of such words as bilious, sanguine and phlegmatic serve as reminders of this school of thought.5 Anatomy, on the other hand, was a purely descriptive science. Through vivisection, Galen gained an understanding of the circulatory, respiratory and nervous systems, but never connected anatomy to the clinical manifestations of human disease.6
This changed in the 19th century, when Carl von Rokitansky, a pathologist, and Joseph Škoda, an internist, founded what we now call the Second Viennese Medical School, which became the preeminent institution for medical learning throughout Europe. This was largely thanks to Dr. Rokitansky, who realized that clinical syndromes were often an incomplete expression of abnormal organ function and that organ pathology was directly linked to human disease.7 Rokitansky also realized that, contrariwise, clinical signs and symptoms could be used to infer the degree of organ dysfunction. Together, Drs. Rokitansky and Škoda developed the concept of clinicopathologic correlation, as in the radiologist’s catchphrase: clinicopathologic correlation is suggested.
Dr. Škoda was responsible for bringing the scientific method to the bedside. Previously, physical exam findings were described idiosyncratically, by comparing the sounds of auscultation and percussion to animal noises, sawing wood or creaking leather. Dr. Škoda systematically examined countless patients until he could classify these sounds by pitch and tone. He was the first, for example, to differentiate normal heart sounds from murmurs and pericardial friction rubs. Our tendency to use such words as bronchial and tympanic is a direct descendant of his work.
Later in his career, Dr. Škoda earned renown for his lectures on the application of the scientific method to clinical medicine. He taught physicians to rely only on their own physical examination, rather than hearsay. He also taught his method of diagnosis: “Recognizing morbid anatomy as the sole basis for an accurate diagnosis, he endeavored to unravel the complex phenomena of disease, to reduce them to their simple elements, and per viam exclusionis to arrive at an exact diagnosis.”8
One of the reasons we know about Dr. Škoda’s method of exclusion is because of the impact it had on Ignaz Semmelweis, who applied Dr. Škoda’s method to his analysis of puerperal fever.
The Viennese hospital operated two maternity clinics, which admitted patients on alternating days. The First Clinic, which was a teaching service for medical students, had a maternal mortality rate due to puerperal fever of 10%. The Second Clinic, which was a teaching service for midwives, had a mortality rate of only 4%. The difference was so stark that women begged to be admitted to the Second Clinic—or chose to deliver in the streets.9
Dr. Semmelweis excluded overcrowding as a cause for the difference because the Second Clinic was almost always more crowded than the First Clinic. He excluded other variables because both clinics were housed in the same facility and, therefore, had the same environment. Eventually, he decided the difference must be the medical students. In particular, he realized that, unlike midwives, medical students also dissected cadavers.
Dr. Semmelweis concluded the medical students must be carrying cadaverous particles on their hands, which they subsequently brought with them to the maternity wards. These cadaverous particles induced puerperal fever in their patients. He came to this conclusion over a decade before Louis Pasteur proposed his germ theory of disease, in which microorganisms were found to be the cause of puerperal fever.10 Dr. Semmelweis had no proof these microorganisms existed. He concluded they must exist through his application of Škoda’s method of exclusion, which left no other possibility.
Beyond Exclusion
Michael Putman, MD, and Anisha Dua, MD, MPH, argue that enough is enough.
In their recent editorial, titled “There Is No Diagnosis of Exclusion in Rheumatology,” they aver:11
Like other ‘pseudo-probabilistic aphorisms,’ the phrase ‘diagnosis of exclusion’ suggests a preceding thoughtful diagnostic evaluation … . But should it? In our experience, a ‘diagnosis of exclusion’ is subjective and conditional, which misaligns it with principles of diagnostic reasoning, resulting in unnecessary testing and premature closure … . Such lists are subjective because a list of exclusions depends upon a combination of prior experience, cognitive dispositions, and physician biases, which vary greatly from rheumatologist to rheumatologist. They are conditional because they depend upon the patient’s presentation … . Even in cases where the ‘diagnosis of exclusion’ framework has resulted in an adequate workup, it closes the imagination to reasonable future possibilities.
They have a point. I think the main problem with diagnoses of exclusion, however, lies less with the concept and more with its sloppy use.
When I was a medical student, I owned a pocket book called Differential Diagnosis. For every chief complaint, the book had a list of diagnoses that needed to be considered, which I dutifully copied into my admission note, along with a brief statement as to how each diagnosis was excluded.
Dr. Škoda would have been horrified.
For him, his method of exclusion was not a Gedankenexperiment; diagnoses needed to be made at the bedside, supported by a scrupulous examination of the patient, rather than by reviewing a checklist of possibilities. Also, because making a clinical diagnosis is fraught, a diagnosis of exclusion must always be treated as a hypothesis, rather than the last word. In its ideal form, the method of exclusion opens the imagination to less common possibilities, like the existence of a hypothetical particle that medical students carried from the morgue to the obstetrical wards.
The story of hepatitis C is a classic example of a diagnosis of exclusion. A serologic test to identify hepatitis B from blood donors was developed in 1974. A serologic test to identify hepatitis A was available a decade later. Screening the blood supply for these viruses, however, eliminated only about half of the post-transfusion cases of hepatitis. Because the evidence indicated the remaining cases were caused by a third virus, it was dubbed non-A, non-B hepatitis. It wasn’t until 1989 that scientists at the biotechnology company Chiron identified the culprit, which is now known as hepatitis C.12 For decades, this disease was a diagnosis of exclusion, until the diagnostic tests caught up.
Anyone can read a coronary calcium score off of a printout to diagnose a patient with atherosclerosis. Only a true clinician can diagnose a patient with … a rheumatic disease for which the perfect diagnostic tests do not yet exist.
U.S. Secretary of Defense Donald Rumsfeld stated, “As we know, there are known knowns; there are things we know we know. We also know there are known unknowns; that is to say, we know there are some things we do not know. But there are also unknown unknowns—the ones we don’t know we don’t know. And if one looks throughout history …, it is the latter category that tends to be the difficult one.”13
He might as well have been talking about rheumatology. The unknown unknowns represent the outer edge of our knowledge. We have a good sense of what remains unknown in rheumatology. The unknown unknowns represent the aspects of medicine we only think we understand—the chapters in the textbooks that are destined to be heavily rewritten. It’s the unknown unknowns that drive us to drag our puzzling patients back to the clinic so we can re-examine them, even after they have been given a putative diagnosis. These are precisely the patients who are routinely failed by diagnostic tests.
The diagnosis of exclusion has become in extricably linked with rheumatology because our diagnostic tests are so lacking. Only in rheumatology are we allowed to ignore antibody tests that don’t seem to speak to the patient’s diagnosis or to make diagnoses even when the supportive tests have refused to cooperate. Anyone can read a coronary calcium score off of a printout to diagnose a patient with atherosclerosis. Only a true clinician can diagnose a patient with adult-onset Still’s disease, seronegative rheumatoid arthritis or any one of a number of rheumatic diagnoses for which the perfect diagnostic tests do not yet exist.
This will not be true forever. But because our patients can’t wait for perfect diagnostic tests, the diagnosis of exclusion remains a useful concept. It reminds us all to think nimbly about our patients, to have faith in our clinical acumen and—most importantly—to remain open to evidence of things not seen.
Philip Seo, MD, MHS, is an associate professor of medicine at the Johns Hopkins University School of Medicine, Baltimore. He is director of both the Johns Hopkins Vasculitis Center and the Johns Hopkins Rheumatology Fellowship Program.
References
- Roth AE. The origins, history, and design of the resident match. JAMA. 2003 Feb 19;289(7):909–912.
- AAMC History. https://www.aamc.org/who-we-are/aamc-history.
- Doyle AC. The Sign of the Four. The Project Gutenberg eBook [ebook 2097]. Release date: March 2000.
- Editors. University of Vienna. Encyclopaedia Britannica. https://www.britannica.com/topic/University-of-Vienna.
- ‘And there’s the humor of it.’ Shakespeare and the four humors. National Library of Medicine. https://www.nlm.nih.gov/exhibition/shakespeare-and-the-four-humors/index.html.
- Rothemich K. Chapter 4. Galen’s Anatomical anomalies and discoveries. In Science, Technology, & Society: A Student-Led Exploration. Clemson University. (CC BY-NC 4.0) https://opentextbooks.clemson.edu/anne1/chapter/ancient-anatomy-galen-2.
- Glorious peaks and painful depths—Viennese medicine. Medical University of Vienna. https://tinyurl.com/f9xw8tuy.
- Obituary. Professor Josef Skoda. Edinburgh Medical Journal. 1881;27(part1):286–287. https://tinyurl.com/rsfuuemy.
- Semmelweis IF. The Etiology, Concept, and Prophylaxis of Childbed Fever. 1859. Translated by Carter KC. Madison, Wis.: University of Wisconsin Press; 1983. https://archive.org/details/etiologyconcepta0000unse.
- Pasteur L, Ernst HC. [On the extension of the germ theory to the etiology of certain common diseases]. Comptes Rendus de l’Académie des Sciences. 1880 May;XC:1033–1044.
- Putman M, Patel JJ, Dua A. There is no diagnosis of exclusion in rheumatology. Rheumatology. 2022 May 12.
- Houghton M. Discovery of the hepatitis C virus. Liver Int. 2009 Jan;29(Suppl 1):82–88.
- DoD news briefing. Secretary Rumsfeld and Gen. Myers. U.S. Department of Defense archive. 2002 Feb 12. https://archive.ph/20180320091111/http://archive.defense.gov/Transcripts/Transcript.aspx?TranscriptID=2636.
Author’s note: I would like to thank Anisha B. Dua, MD, MPH, for graciously allowing me to comment on her editorial, which you should definitely read for yourself.11
