番茄社区

In deciding to go to medical school, I had visions of heroic moments like I'd seen on TV. You know the type. A patient has some mysterious problem that nobody can figure out until the super genius doctor somehow cracks the case just in time. The answer is often some rare disease that no ordinary doctor would have even known about, let alone be able to recognize outside of a text book or medical journal.

In these shows -- of which "House, MD," is the most prominent modern example -- the "great doctor," thinks of some test that had not been done. Once the diagnosis is made, the right treatment is given, and everyone lives happily ever after.

What I find in practice is actually the opposite of this, in two distinct ways.

Mistake #1: Disease Hunting

First, when a patient has symptoms that don't quite make sense, the temptation is to think of rare diseases. That is often a mistake. The reality is that mystery cases are usually not rare diseases. Rather, they are unusual presentations of common entities. Another possibility is that typical symptoms of a disease are present, but there are other unusual ones that throw the clinicians off the scent.

Here's a nice example from a case I had years ago. A woman came to the emergency department with classic symptoms of a stroke affecting the left middle cerebral artery. But she was young and fit. She was wearing sneakers. She'd come from the gym, we were told. All of that was rare enough, but not impossible to believe. People have genetic risks that can mean a stroke at age 35. But something else was strange, in hindsight. She was extremely sleepy. She could barely stay awake. Now, patients with this kind of stroke can have lethargy, but it looks less like someone struggling to stay awake. Something just looked different about this situation.

Well, part of the protocol of a rapid stroke evaluation is to do a few quick tests before we send them over to get a CT scan of the brain. One of the tests we always do is a "fingerstick" test to check the glucose level of the patient's blood. This patient had a level that was close to zero!

We quickly administered a form of sugar (dextrose) into her IV, and within seconds, she was awake, and the stroke symptoms were gone. Life-threatening low glucose (in a patient who, unbeknownst to us, had diabetes) had faithfully mimicked a stroke. She easily could have died within minutes. (Codes in the radiology unit are never a good scene.)

Another case I had recently really threw me off. This guy came in with a constellation of symptoms that really made no sense at all. He'd been seen by a bunch of doctors and he'd come to us looking for help.

I discussed his case with the resident physician who had already seen him. We spent many minutes talking it through. Now, usually it takes me about 10-30 seconds to narrow down what is going on with a patient when I'm talking with an experienced resident. In this case, I was just confused and nothing was adding up.

I went to see the patient and his son. I told him I'd been thinking about his case a lot. But I was also honest. I told him, "I have no idea what you have."

What he said in reply kind of surprised me. "Oh, I like you! You're the first doctor I've seen through this whole process who has admitted that."

He'd seen a bunch of specialists, each of whom had developed a pet theory that had not worked out. I defended them, saying, "Well, when you're a hammer, it's easy to see nails everywhere."

I laid out two broad possibilities. One was that he had a typical presentation of some rare condition that they'd have on a TV show like "House, MD." If that was the case, I had no chance of diagnosing it, but maybe one of the specialists in the hospital would eventually figure it out. The other -- the more likely one -- was that he either had an unusual presentation of a common condition or he had two somewhat common conditions at the same time, which was muddying the clinical picture. Either way, we'd admit him to the hospital and it would get sorted out.

The solution ended up being the latter. He had one problem that had basically been asymptomatic for months or years, and a second problem that both added new symptoms and uncorked the other condition's symptoms. Now, he had symptoms that were all over the place due to the coincidence of two reasonably common conditions that just don't tend to travel together.

Mistake #2: Getting Too Creative With Testing

The second big mistake clinicians make when facing a strange case is to try to think of new tests to do. In a sense, we are chasing the temptation to try to diagnose a rare disease and be a hero. The potentially big mistake is not repeating tests that have already been done and, instead, sending off some new (and often unusual) ones that had not yet been considered.

This brings to mind a quote (often attributed to Mark Twain): "It ain't what you don't know that gets you into trouble. It's what you know for sure that just ain't so."

What often happens is that we are told things about prior test results, and that information misguides us. Now I specifically chose the word "misguide" for a reason. That's because the information may have been true or not. Think about two situations:

1. It's possible that prior test results have been inaccurately reported or conveyed. I remember a case in which someone told me, "His MRI from 2 days ago was normal." But it wasn't normal! Had I not double-checked, I'd have been in big trouble, and the patient would've been too.
2. The prior test results have been accurately reported but they are either out-of-date or they were themselves inaccurate. In a pretty recent case, I was told that a patient had tested negative for a certain infection just 1 day prior. That information was correct! But I also knew that the test may have been incorrectly performed (or performed too soon to detect the pathogen). For example, if you have COVID-19 and you barely touch your nostril, you'll get a "false negative." Alternatively, if you take a test too soon after infection, your body may not yet have enough virus for the test to turn positive.

So, the big error clinicians make is putting too much faith in previous test results. Those results could have been right at the time they were reported, or the results could have been misunderstood or misreported. Games of telephone in medicine do happen -- and that can be dangerous.

A Face Off Between Rare Diseases and Inaccurate Tests

Let's do some "fun" math on this to see how likely this actually might be in practice.

Consider a patient who has influenza. A good nasal swab should pick up 90% of the cases. But let's say this patient has had fever for 2 days, a cough, and the nurse got a really good swab. In this case, our test might be able to diagnose 99% of the cases. Even so, there's that one in 100 chance that the test could be falsely negative.

If I saw such a patient, I could easily be misled into thinking "She tested negative for flu yesterday, and she tested negative on a handful of other common tests as well. This must be some rare condition. Let's dig in!"

I might then proceed to order a handful of tests that each carry a one in 5,000 chance of being the real cause of a person's illness at that point. The math says that I'd have to check for 50 rare diseases before the odds would exceed the false negative rate (1%) of the flu test taken yesterday. If each of my "Hail Mary" tests costs $50 to $200, I could easily spend $2,500 to $10,000 chasing rare conditions, when the smarter move would have been simply to retest for flu (at a cost of $50 to $100).

I've had some cases recently where the hero move was not to think of some rare condition. Instead, the right move was to take nothing for granted and start from the beginning. We can't do this on every case. But when someone is getting worse, it's important not just to zoom out, but to back up.

The biggest mistake clinicians make when trying to tackle a medical mystery is not a failure to be clever enough to think of something nobody else considered. It's a failure to remember that every test has its limitations in accuracy and that they must be administered at the right time. That is, there are false negatives and there are "window period" errors (i.e., tests done too soon in the disease course to be reliable).

In my experience, remembering these points leads to the right diagnosis in tricky cases -- often of a relatively common condition -- much more often than the earnest search for rare diseases that we heard about in medical school, read of in a medical journal case report, or saw on a TV drama.

This piece was originally published in .

Comment