[Note: I wrote this article back in 2009. Since then, the FDA approved Pradaxa as a substitute for Coumadin. Pradaxa has an even greater risk of intracranial hemorrhage than Coumadin, and there’s no reversal agent, unlike how Vitamin K reverses Coumadin. The concerns raised below thus apply even more to patients on Pradaxa. On the opposite of the spectrum, since this article has come out a ton of studies have showed that common hormonal contraceptives like the NuvaRing dramatically increasing the risks of blood clots, stroke, and death.]

The magazine Emergency Physicians Monthly hosts a blog called WhiteCoat’s Call Room, which recently posted a complaint about “defensive medicine:”

Why was I ordering all of these things when my clinical judgment led me to believe that they would “probably” not lead to any changes in the patient’s management?

The answer is because in our culture, “probably” doesn’t cut the mustard any more. Clinical medical judgment has been supplanted by the demand that physicians disprove the improbable. Society has made it so that physicians are more concerned with proving that unlikely diagnoses with the possibility of a “bad outcome” don’t exist and with maintaining good Press Ganey scores. Many physicians are afraid to practice rational medicine based upon clinical judgment and physical examination skills. No one wants to face the liability.

The author is complaining about a basic principle of clinical medicine known as “differential diagnosis,” a process of elimination by which physicians reach a diagnosis by eliminating the most serious and unlikely diagnoses first before continuing their basic evaluation.

Using a “differential diagnosis” compels physicians to evaluate patients in a systematic, rational and logical fashion, free of any distractions or other biases that might cloud their judgment. One example is the well-known psychological effect of “confirmation bias,” through which people who hold a particular belief tend to review the available evidence in a way that confirms that belief.

The failure to diagnose — the failure to use the differential diagnosis — may be the most common form of medical malpractice.

The author lists a number of situations where they thought they were practicing “defensive medicine” instead of being “rational” (the bulk of which were x-rays that confirmed the absence of bone fractures), including this situation:

A patient in her 60s fell and hit her head 5 days ago. She was having a headache. I couldn’t find a mark on her and was inclined to send her home with pain medications. But she was on Coumadin which put her at risk of bleeding. So I did a CT scan of her head to “make sure” that she didn’t have a bleed. She didn’t.

I’ve lost count of the number of people I’ve seen killed because their physician or hospital ignored the dangers of Coumadin. Here’s what the NIH says about Coumadin (PDF):

Because Coumadin reduces the ability of the body to form blood clots, a patient on Coumadin will bleed longer after an injury than a patient not on Coumadin. …

Bleeding inside the brain, even after minor accidents, can also be deadly.

If a patient in her 60s on Coumadin falls and then has a headache of sufficient severity to bring her to the hospital, then warning bells should be going off.

The fact that this physician would not have ordered the appropriate test — a CT scan — but for fears of medical malpractice liability suggests to me we need more liability, not less, since the physician obviously didn’t recognize the standard of care dictated that a brain bleed be ruled out before sending her home.

[UPDATE: WhiteCoat replied to me, so I replied to him.]