After my post yesterday, "Differential Diagnosis, Defensive Medicine and Medical Malpractice: Coumadin Edition," the original physician responded on WhiteCoat’s Call Room at length:

Max Kennerly is another lawyer that posted a response on his blog “Litigation & Trial.” He accused me of being afraid to use the “basic principle of clinical medicine known as differential diagnosis” – which he defines as “a process of elimination by which physicians reach a diagnosis by eliminating the most serious and unlikely diagnoses first before continuing their basic evaluation.”

What Mr. Kennerly is apparently suggesting is that, rather than use medical education and heuristics, physicians “shoot the moon” and order “million dollar workups” on every patient complaint. Forget that a runny nose and cough in a child are highly likely to be a viral upper respiratory infection. According to Mr. Kennerly, physicians have to “eliminat[e] the most serious and unlikely diagnoses first … before continuing their basic evaluation.” Because runny nose and cough could also be signs of serious and unlikely diagnoses like bronchopulmonary dysplasia, pandemic bird flu, and inhaled foreign bodies, Mr. Kennerly is apparently asserting that every child with a runny nose and a cough requires a NICU admission, full isolation precautions, viral cultures for H5N1 influenza virus, a call to the CDC (just to be sure), and bronchoscopy before physicians can breathe a sigh of relief and recommend nasal suction and honey (cold syrup is much too dangerous – just ask all the pediatricians). Did I miss anything in my “differential,” sir?

Mr. Kennerly then takes issue that I would consider discharging a woman with a mild head injury who developed a headache 5 days later and who was also taking coumadin. Bleeding in the brain must be ruled out “even after minor accidents,” according to an article he cited from the NIH. But Mr. Kennerly does not stick to his own script. Many “serious and unlikely diagnoses” can cause a headache. Using Mr. Kennerly’s logic, it is likely that “differential diagnosis” algorithm he proposes would require me to get an MRI and MRA to rule out vascular causes of headache and to perform a lumbar puncture to rule out pseudotumor cerebrii. While he may have some success getting a jury to believe that “his” is the way medicine should be practiced, it just isn’t so.

I removed the second half; I’ll have to answer that later.

I responded in his comments section:

Thanks for the link! It’s great to get a dialogue going.

Just to be clear, I didn’t use the phrases “shoot the moon” or “million dollar workups,” but I did suggest that physicians should rule out severe and life-threatening conditions first.

I’m surprised you’d disagree. Truth is, you don’t. Think back to all of the examples you provided in your prior post — why did you order all those x-rays and CT scans? To avoid a lawsuit?

Nope — no physician has ever been held liable for not performing an x-ray or a CT scan. There’s no harm from simply not performing a test.

Physicians are liable for not ordering tests when they should have and when harm was caused by that failure. You ordered all those tests because, in your judgment, there was a reasonable chance that the ‘unlikely’ scenario was the actual diagnosis.

Let’s take the 60yo woman on coumadin with the head injury. You tried to dodge those initial facts by recasting it as “a headache” and then listing all the potential but unlikely causes.

Well, she didn’t have “a headache.” She was on coumadin, had a fall, and then had a headache serious enough to bring her to the hostpial, which is why you ordered a CT scan looking for brain bleeding, and not a lumbar puncture looking for pseudotumor cerebrii.

You applied your judgment, saw an unlikely by possible serious complication, and ruled it out. That wasn’t “defensive,” it was “appropriate” — if you didn’t think there was a reasonable chance of her having a brain bleed, then you’d have absolutely no reason to fear a lawsuit.

Moving on to your child with the runny nose and the cough, it’s hard to take your example seriously when you first propose the “child” go to a unit reserved for neonates. If a neonate has an obvious infection, that is a very serious issue that will be treated accordingly, likely with multiple antibiotics and multiple x-rays to repeatedly check pulmonary function.

If by “child” you mean the typical toddler going to a pediatrician, then, yes, I submit to you that if the pediatrician has reason to suspect something more serious than a typical cold then they should rule out that serious possibility. You gave no other facts than “every child with a runny nose and a cough.” I have kids. They’ve had runny noses and coughs. My pediatrician ordered no tests. That’s fine; it was a typical kid with a cold.

But let’s mix it up, the way it happens in real life: my child has had a severe cough for over a week now, has shown trace blood in her mouth, can’t sleep, and won’t eat.

Now what? Go home?

Or should you look for something more?

“Defensive medicine” doesn’t exist — the concept requires a doctor somehow see enough of a risk to fear litigation but not enough of a risk to warrant testing. What sense does that make? Either the doctor fears the serious outcome or they don’t.

But the ball is in your court — what would you have us do different? Set up a, say, 5% rule? As in, if something has a less than 5% likelihood, physicians as a matter of law need not look for it?

You tell me. I hold doctors to the standard of keeping people safe by making sure patients don’t have any serious or life-threatening complications that are reasonably foreseeable. You want something less than that.