I often get called on to be a diagnostician. The referring doctor is uncertain what is going on in the patient, often a fever of unknown origin, and they call me to help figure it out. Sometimes I do, sometimes I don’t.
Making the correct diagnosis is not easy, even after 35 years. The classic phrase is the fog of war, but the fog of medicine is equally confusing. In retrospect sometimes a diagnosis becomes clearer, but in real time? It is so easy to be wrong and so difficult to be correct.
I remember the first case of my career. A patient developed neutropenia (low white count) from a sulfa antibiotic and over the next two weeks went into multi-organ system failure and died. At autopsy it was discovered that he had miliary tuberculosis. I totally missed the diagnosis, despite all the tests, including liver and bone marrow biopsies before he died.
A decade later a case was presented at conference about a patient with a fever who went into multi-organ system failure and died. At the time of the conference I had zero recollection of the case from a decade earlier, but knew this was a case of miliary TB based on the data presented, although it was far from a textbook case. After the conference the presenter let me know that it had been my case, the one I had missed years ago.
I was ten years wiser in my medical career and more aware of the subtleties and variations of disease presentation. Experience has made me a somewhat better diagnostician. Or so I hope. Making the diagnosis of a common presentation of a common disease should be simple, but patients rarely read the textbooks and we are more likely to see an uncommon presentation of a common disease, a common presentation of an uncommon disease, and most dreaded, an uncommon presentation of an uncommon disease.
I still have difficulty making the correct/accurate diagnosis on patients and so often I am left with best probability based on the often incomplete and contradictory data. I could make an accurate diagnosis on every patient, but for some reason they are hesitant to allow an autopsy while still alive. Go figure.
I kill me, hopefully not patients
Being in a teaching hospital not only do I need to teach the residents about the diseases I see, but how I get the diagnosis. I think my time in the skeptical world has made me a better physician as I try to be aware the various cognitive biases that make us poor thinkers. I like to use “Observations on spiraling empiricism: its causes, allure, and perils, with particular reference to antibiotic therapy” as a good Infectious Disease example of cognitive biases and try and apply its lessons to myself and my teaching.
There has been a recent focus on diagnostic error by physicians as a source of patient harm and the numbers are remarkable.
About 5 percent of adults who seek outpatient care annually suffer a delayed or wrong diagnosis.
Postmortem research suggests that diagnostic errors are implicated in one of every 10 patient deaths. Not every death is scrutinized, however, so the findings can’t be generalized to all hospital patients.
Chart reviews indicate that diagnostic errors account for up to 17 percent of hospital adverse events.
Diagnostic errors are the principle cause of paid malpractice claims and are almost twice as likely to end in a patient’s death than claims for other medical mishaps. They also represent the biggest share of total payments.
Those numbers do not surprise me. As Barbie said, medicine is hard. There are systems reasons for making the wrong diagnosis that boil down to too much data and too little time to process it. The use of Electronic Medical Records was supposed to help solve some of these issues. Hahahahahahahahahahahahahahahahahahahaahahahahahaha. Wipes eyes. I kill me sometimes.
There are knowledge deficits; we tend to diagnose what we know and miss the diseases of which we are unaware. Knowledge deficits are correctable with education. Hahahahahahahahahahahahahahahahahahahaahahahahahaha. Wipes eyes. I kill me again.
I once had a case of a young female with fevers, severe joint and muscle pains, high white count and a sed rate of 120 who I saw in clinic as an add on at the end of a Friday. She was sick, but I had no idea with what. I thought maybe endocarditis, so I admitted her and left for the weekend. When I returned on Monday she had been discharged. What did she have? I asked the hospitalist. Oh, she said nonchalantly, as if it were oh so obvious, it was Still’s. And upon review it was a textbook case. If I had ever read the textbook, I didn’t remember it. Totally missed the diagnosis. And even more annoying is I missed another case about three months later. Needless to say now I always consider Still’s. But one of the odd medical phenomena is how often you see your first case of a disease shortly after you learn about it. I so hate the Baader-Meinhof Phenomenon. But it makes me wonder how many similar cases I had missed in the past.
But the why and how of misdiagnosis are, I suspect, much deeper than lack of knowledge and systems problems. As skeptics know all too well, humans are fundamentally poor thinkers. We have evolved to survive reality, not to understand it, and the biggest problem with making the correct diagnosis is probably being human. I think the IOM is going to have their work cut out for them.
Last week I attended a Grand Rounds that was an excellent review of the cognitive biases as applied to medical diagnoses. No new cognitive biases were mentioned, but it was interesting to hear them in the context of medical diagnosis.
One nice review has 32 biases as applied to medical decision making. All of them are medical examples of cognitive errors and shortcuts we all use every day. The issue, of course, is how to get physicians to not only recognize the biases, but recognize that the biases apply to them. And there is the impenetrable reality distortion field generated by the Dunning–Kruger effect to contend with.
Perhaps more AI computer support. I Google/Pubmed everything I see and I often get ideas and directions I would not otherwise have considered. Perhaps crowdsourcing diagnoses will help. Perhaps edumatcating physicians on the cognitive biases. The best bet is to use CRISPR technology to build humans with improved cognitive abilities, the first step to the Zombie Super Soldier Army I so want to control. Or is that just me?
The problem is that cognitive error is high-hanging fruit and difficult to get at, and there will be a tendency to pursue more readily attainable goals…Real solutions to cognitive diagnostic errors lie in the shadows, and they will be difficult to find.
As I say medicine is hard and solutions to some problems may lie at the heart of being human.
Complementary and alternative medicine, and poor decision-making
That’s medicine. How about SCAM? How does one reduce diagnostic errors when the whole field of practice is fundamentally based on the unknowing application of every cognitive bias? In medicine we see the 32 biases of clinical decision making as a problem. SCAM must deny them to exist. To recognize them as applicable means the end.
The alternative medicine world is notorious for how multiple practitioners are unable to reach the same diagnosis when presented with the same patient. TCM practitioners can’t agree. Chiropractors can’t agree. Acupuncturists can’t agree. And naturopaths, proudly having no standard of care, can’t agree.
The entire field of Integrative/Alternative/Complementary Medicine is based on making the wrong, often imaginary, diagnosis then running with it. No need to worry about cognitive biases in world that is the medical equivalent of Lake Wobegon:
where all the diagnoses are correct, all the therapies are effective, and all the providers are above average.
Medicine may be plagued by misdiagnosis and its consequences, but we have some hope that these issues can be corrected or ameliorated.
SCAM is defined by misdiagnosis and its consequences.
I wonder if the IOM will ever tackle that problem.