When I was younger, I could remember anything, whether it had happened or not; but my faculties are decaying now and soon I shall be so I cannot remember any but the things that never happened. It is sad to go to pieces like this but we all have to do it.
There was a recent study that raised a mild brouhaha. Titled “Diagnostic Errors in the Emergency Department: A Systematic Review“, it found that there was often a misdiagnosis made in the ER. And there was gambling, gambling! in Ricks.
The study found 15 medical problems that accounted for the bulk of the misdiagnosis:
The top 15 clinical conditions associated with serious misdiagnosis-related harms (accounting for 68% [95% CI 66 to 71] of serious harms) were (1) stroke, (2) myocardial infarction, (3) aortic aneurysm and dissection, (4) spinal cord compression and injury, (5) venous thromboembolism, (6/7 – tie) meningitis and encephalitis, (6/7 – tie) sepsis, (8) lung cancer, (9) traumatic brain injury and traumatic intracranial hemorrhage, (10) arterial thromboembolism, (11) spinal and intracranial abscess, (12) cardiac arrhythmia, (13) pneumonia, (14) gastrointestinal perforation and rupture, and (15) intestinal obstruction.
With 130 million U.S. ED visits, estimated rates for diagnostic error (5.7%), misdiagnosis-related harms (2.0%), and serious misdiagnosis-related harms (0.3%) could translate to more than 7 million errors, 2.5 million harms, and 350,000 patients suffering potentially preventable permanent disability or death.
So, not an insignificant problem.
My first thought was to contrast that finding with the practice of fantasy-based medicine, those who follow the nonsensical ravings of a lunatic mind (NRLM). You know, alternative medicine.
Practitioners of Traditional Chinese-Pseudo Medicine are, after examining the tongue and pulse, 100% wrong in their diagnosis. Chiropractors? As long as they are finding and treating subluxations, 100% wrong. Naturopaths? There I am less certain, as that is a form of NRLM that excels more in goofy treatment than in goofy diagnosis. Most of the other forms of NRLM hover close to 100%. The alterative medicine diagnosis is definitional for misdiagnosis.
As this post goes up, I have one working day left as an Infectious Disease doctor. If you include the second two years of medical school, I have spent 41 years working mostly in the hospital taking care of acutely ill people. I have easily seen 50,000 patients in consultation. Really. 4-5 consults a day, 48 weeks a year, 1 in three to 1 in 5 weekends on call. It adds up to a small city worth of patients.
As I head into my inevitable decline, like all old geezers, I want to reflect on the past and pontificate to all you youngsters who will likely ignore my sage opinions.
Getting the correct diagnosis in the fog of medicine can be damn difficult and, like Grampa Simpson, I will take this opportunity to give a mostly SBM free rambling on medical diagnosis and my experiences, both failures and triumphs.
For I flatter myself that I am a diagnostician. Making the correct diagnosis is key to all that follows in patient care and, as I mentioned, tough. Anyone can look up the correct treatment once a diagnosis is made. Although I remain shocked at often that simple google is not done. Hence, a career.
As another aside, Google Scholar can be invaluable in making the right diagnosis. Has influenza been associated with sterile cardiac vegetations? Yep. Can vomiting lead to esophageal tear and empyema? You betcha. Two examples from this month where the diagnosis of atypical diseases presentations was explained by a quick google search. I am so glad the Index Medicus has gone the way of the dial up modem. Now if we could just get rid of medical references behind paywalls.
Making the diagnosis is what makes the real doctor. Petty as it is, I derive enormous satisfaction from reaching a diagnosis missed by others, especially if the diagnosis is not an infectious disease. Brings out the gloat like nothing else.
For example, I had a patient who came home to Portland to die of AIDS. This was the bad old days before HAART. He had been at Brigham and Women’s Hospital in Boston for two months slowly fading away. When I saw him? Tanned like he had been in Jamaica in August instead of a Boston winter. Low blood pressure and sodium, high potassium. Classic Addison’s, induced by his rifampin. Hard to miss. A bit of steroids and he perked up, living another year. Gloat central to this day.
As Gag Halfrunt noted, “Vell, Brigham and Women’s is just zis hospital, you know?” But I learned as a resident in Minnesota that the so called great hospitals? Meh. Patients often went to Rochester for a second opinion only to come back with no new insights but significantly poorer. Hold the Mayo, we used to say.
When I am asked about a second opinion, I always think of the Henny Youngman joke: My doctor told me I was fat. I said I wanted a second opinion. He said, “OK, you’re ugly, too”.
Not all doctors are diagnosticians. What is a diagnostician? Someone who, after doing a history and physical, comes up with the diagnosis for what ever ails the patient. I remember once I had an AIDS patient, a home brewer, who had Saccharomyces cerevisiae in his blood. So I knew what aled him. Yep. I went there.
Pathologists and radiologists do not qualify as diagnosticians. Nor do many other physicians. Some are fine within their organ/specialty. A cardiologist is fine for a cardiac diagnosis. A gastroenterologist is fine for a GI diagnosis. Larry is fine for a Three Stooges diagnosis.
ER docs? Most assuredly not diagnosticians. Sorry guys. The usual attitude of physicians at the receiving end of an admission from the ER is that the ER is wrong about the diagnosis. But that is how it should be. Part of the ER docs job is to determine whose illness needs admission. Every doctor has their specific training and expertise. Just as you would not want me to remove your appendix or deliver a baby, you would not expect all doctors to be equal diagnosticians.
If you want a doctor who has both breadth and depth of medical knowledge, you want either an internal medicine/hospitalist or an ID doctor. I, like Zaphod Beeblebrox, would survive the total perspective vortex.
As an aside. Kids today. I remain shocked – shocked! – at how few house staff or medical students have any knowledge of the classics. Monty Python and The Hitchhiker’s Guide to the Galaxy are unknown to them. Doctors are not the dorky nerds we used to be.
Infectious diseases involve all organ systems and many processes can mimic an infection. As my old boss used to say, the ID doc has to be the second-best cardiologist, the second-best pulmonologist, etc. in the hospital.
So what does it take to be a diagnostician? Besides overweening self-esteem?
Here is where the old geezer wanders down memory lane with an onion on my belt, as was the style of the time.
Experience helps. While I have long maintained the three most dangerous words in medicine are in my experience, that has more to do with choosing a treatment than making a diagnosis. The practice of medicine, with emphasis on practice, is as much a trade as a profession, and like building cabinets, the more patients you have seen, the better diagnostician you become. Unlike, say, golf, where no amount of practice will allow you to escape mediocracy. Sigh. That is life’s motto: Practice makes mediocre.
A couple of years ago I introduced myself to a patient who apparently was admitted from central casting as a stereotypical Russian. A bear of a man, he got out of bed and gave to a crushing hug.
“Finally,” he said. “An old doctor. Someone who knows what they are doing.”
While I have not seen it all, I have seen more than most. It is seeing diseases that are important. You can read about a disease, but it never really gives the flavor of a patient presentation.
The textbooks will say 89% of patients will have this symptom and 67% that symptom, and 5% will have the other thing. But patients do not read the textbooks and so often present with signs and symptoms kind of sort of resembling the textbook description, but not quite. Like Nutrimatic tea.
Years ago I saw a patient in clinic on Friday afternoon. A bad time to be sick. And she was sick. High fevers, severe polyarticular arthralgias, rash, toxic appearing. It had been going on for a week. I had no clue what she had, but she was sick. I was betting on endocarditis with immune complex arthritis and admitted her.
Monday, I came back to work and the patient was gone. I asked the hospitalist what happened, fearing the worst.
“Oh. It was Still’s disease. Got better on steroids and we sent her home. Classic.”
Classic? I had vague memories of Still’s and read up on it. Yep. Classic case. Totally missed it. But I learned about Still’s and vowed not to miss a case again.
Wrong. Three months later I was consulted on another case, not classic, and missed the diagnosis again. Rheumatology made the diagnosis on the basis of a sky-high ferritin. Since that time? I have not missed a case and have made the diagnosis several times. If you have never seen a case, you will not likely recognize it the first time. Or the second time. There is no better experience for becoming good at diagnosing a disease than missing a few cases. From the patient’s perspective, you do not want to be the first for some doctors learning experience. There is a saying in medicine, you don’t want to be a great case.
When you see a patient, we are trained to generate a differential diagnosis: a list of diseases that can cause the patients symptoms. The 10 reasons for chest pain. The 12 reasons for shortness of breath. The 4 causes of fever of unknown origin. I actually didn’t bother to count them up, so do not quibble in the comments.
The weird thing about that process of generating lists is how, over time, it has become internalized. It used to be that I would consciously go through a list for whatever the problem was, considering the pros and cons of each option. Now? More often than not, I know the diagnosis and with remarkably little thought and scant information. The answer just burbles up like a bubble in a tar pit: pop. It’s leptospirosis. It is really weird how much processing goes on beneath my geriatric consciousness. And it impresses other docs no end when you give the diagnosis after just a few words.
I like to swing for the fence that way. Progressive headache for two months and an abnormal LP? Cryptococcus. Pancytopenia and fevers for a month in a Hispanic? Brucella. And Bingo was his name-o. It is important to remember that while Babe Ruth led the league in home runs, he also led in strikeouts. People always remember the former and forget the latter. You can make quite a name for yourself swinging for the fences and being right only occasionally.
Pulling information seemingly out of thin air impresses patients as well. Years ago I saw a young female with, well, a relapsing fever. So rather than the usual ‘any travel lately’ question, I asked, “How was your vacation in Black Butte?” The look of shock when she asked how I knew she had been to Black Butte was priceless. I understand how mentalists must feel. But Black Butte is the only part of Oregon where Borrelia, the spirochete that causes relapsing fever, is found. I went to the lab and there was the wee beastie on a smear.
I still mentally go through the list. Gotta be safe. And I often make sure I consider four lists: the list of what the diagnosis might be. The list of what diagnosis would be potentially catastrophic to miss. The list of cool diagnoses. Of course, I am unnaturally drawn to the cool diagnosis, no matter how remote the possibility, and it is painful to not go looking for diseases that I know have almost zero prior plausibility of being the cause of the patient’s symptoms. And lastly, depending on the case, I might mentally run through organism class: could it be viral, rickettsia, spirochete, bacteria, fungal, parasite, etc.? I once counted about 1,400 bugs I need to know, or at least know of, to be an ID doc. More than Pokémon (1,008), but I still gotta catch ’em all.
There is another way to categorize and consider disease presentation for diagnosis:
- The common presentation of a common disease.
- The uncommon presentation of a common disease.
- The common presentation of an uncommon disease.
- And the worst: an uncommon presentation of an uncommon disease.
- And even worser: an uncommon disease presenting like a common disease.
I was not surprised to see that epidural abscesses were on the miss list in the ER. It is an uncommon disease that often presents like a common disease. Many of the cases I see have had a delay in diagnosis because it starts as persisting back pain after minor trauma. No signs of infection. It is not until the disease progresses to some sort of paralysis that the MRI is ordered. We can’t do every test on every patient every time. There is not the time, resources, or money.
The fog of war has nothing on the fog of diagnosis.
Medicine is the realm of uncertainty; three quarters of the factors on which action in medicine is based are wrapped in a fog of greater or lesser uncertainty. A sensitive and discriminating judgment is called for; a skilled intelligence to scent out the truth.
You know. An ID doctor. But it is not unusual when initially seeing a patient to not have a clue as to what is going on. Lymphoma? Endocarditis? Still’s again? Often early in a disease the presentation is too vague to point to a specific disease. I am looking for a pattern and there is no pattern I can see. What then? That is a common problem in the ER.
Early in my career it was don’t just stand there, do something. Now it’s don’t just do something, stand there. So often time will clarify what the diagnosis might be and what needs to be done. As Laozi noted, and I often quote, “Let muddy water stand and it will become clear.” It is important to explain to the patient why you are doing nothing, as they often want something, anything, done. I often compare the problem to unripe fruit; it takes time for the disease to mature to the point it can be recognized. The ER is filled with green bananas that are yellow by the time I am called to see the patient.
ID docs do have a reputation for doing excessively detailed histories to aid in making a diagnosis and for writing extremely long notes.
Another aside, if you have not discovered Dr. Glaucomfleken, you are in for a real treat. I have never met the man, but he has evidently followed me on a hike.
Producing War and Peace for each consult is a real fetish that I do not understand. Many cases are simple, especially if you know what you are doing. By the way, if you really want an experienced diagnostician, avoid a University Hospital. Think about it. Often the attending is on service 3 months a year. So they might have 25% of the experience of someone who works every day in the trenches. And it is often indirect experience: patient to medical student to resident to fellow to attending and back, the perfect opportunity for medical telephone.
My time in skepticism and SBM has made me a far better physician and diagnostician. Skeptics are, I wager, more aware of how we think and the errors in thought that lead to faulty diagnosis:
The most common cognitive problems involved faulty synthesis. Premature closure, i.e. the failure to continue considering reasonable alternatives after an initial diagnosis was reached, was the single most common cause. Other common causes included faulty context generation, misjudging the salience of findings, faulty perception, and errors arising from the use of heuristics. Faulty or inadequate knowledge was uncommon.
Because of my time at SBM, I have spent an inordinate amount of time thinking about thinking. As a result I think I think better. One of those thought processes is to consciously consider the outliers. Often all the patients presenting signs and symptoms can be accounted for with one diagnosis, or at least we try to. Occam, not Hickam, rules. Newbies and the unexperienced focus on the what is consistent with the presumptive diagnosis. I make a point of considering the inconsistencies, occasionally resulting in a different diagnosis. Like the aforementioned empyema. Huge pleural infection with an oral Streptococcus but almost no lung infection. There was no pneumonia to cause the empyema. It was thought the empyema was due to aspiration, but careful questioning found the aspiration event occurred during an episode of prolonged and violent vomiting. Boerhaave’s syndrome was a better explanation. Focusing on the outliers is a learned habit that does not come naturally.
There are two, and likely more, patterns in medicine that I think will adversely affect the future development of good diagnosticians.
One is a lack of continuity. Everyone is a shift worker. A week on and a week off. In my hospitals I am the only doc that is there day in and day out, except for vacation and some weekends. I suspect physicians are not getting the understanding of disease that only comes with following a patient from admit to discharge and then outpatient follow-up. Continuity of care is important to gain understanding of disease and is the best opportunity to see your, well, screwups. I always tell residents it is OK to make a mistake once, it is how you learn. Nothing imprints a disease like a mistake. Just don’t do it twice. Except in diagnosing Still’s disease.
The other is EPIC aka the electronic medical record. The way rounds works now is everyone sees their patients then vanishes to the computer room to chart. I rarely see other docs, attendings or residents, as I wander the halls. The constant interaction with other providers has vanished and with it the cross-fertilization that occurs with discussing the odd cases. I used to get multiple questions a day from docs I would run into. Now? Maybe one a day. On a busy day.
I do not know what the solution is to improving diagnostic acumen. The comment section of the NYT article suggests doctors should listen to patients better. I doubt that. Patients don’t give histories, they tell stories. Digression-filled stories from which we need to extract the pattern of the patient’s disease. It is amazing how much of what I need to know (symptoms, onset, progression, current status) seems at times to be almost deliberately avoided by patients in favor of details I do not need or information I have ready access to in the chart. Time, concentration, stamina, empathy are all limited resources in a busy day. I understand why doctors interrupt the patient around 11 seconds in. Often they are not telling you anything you need to know to figure out the problem at hand.
For a given disease, nonspecific or atypical symptoms increase the likelihood of error
These issues are not unique to the ED—they are seen across clinical settings, regardless of study method.
I am not certain if there is any way to make physicians better diagnosticians. Most of the solutions I find online seem a bit, well, trite. They come down to thinking better. Which is hard to do. Or teach. Just look at the cognitive bias codex. It is a wonder that humans can force themselves to think clearly. Rational thought, I wonder, is perhaps at its core an unnatural act. And there are those who just seem unable to process the information to come up with the right diagnosis. 1+1= cheese. Always odd to see. But how to counter it?
One of the sobering experiences of my career is coming across the charts of patients I took care of long ago. Reading my notes I wrote as an intern when I was a third-year resident or coming across my notes from early in my practice.
Man. As Mark Twain famously sort of noted:
When I was an intern, I was so ignorant I could hardly be called a doctor. But when I got to be an attending, I was astonished at how much I had learned in seven years.
My current skill set is a result of 40 years of hospital clinical medicine and all that comes with it. I do not think there is a shortcut to all the lessons learned. Unfortunately. I expect some misdiagnosis will always be part of the human condition.