Multiple choice questions are a terrible way to test doctors


Most people are surprised to hear that the way doctors are recertified every ten years is through a multiple choice test. “Really?” they’ll say. “You take a multiple choice test? As a doctor?”

Unfortunately, after all these years that’s the most efficient way we can think of to evaluate professionals. But the method is so flawed. And your ability to answer questions on the multiple choice test isn’t necessarily linked with your ability to be a good clinician. Take a look at this question, for example:

Which of the following statements regarding deep venous thrombosis is correct?

  1. Most calf vein thrombi extend into the proximal deep veins, usually within a week after presentation.
  2. Most patients have classical physical examination findings.
  3. Pregnancy is a predictor according to the Wells criteria.
  4. Up to 10 to 15 percent of calf vein thrombi result in pulmonary emboli

Every good doctor knows that a patient with a swollen leg raises the concern for a deep venous thrombosis (DVT): a blood clot in the veins of the leg. This question is trying to ascertain whether the clinician understands the approach to the diagnosis of DVT. But look at the answers.

The important concept being tested with answer choice “A” is whether or not calf vein clots (“thrombi”) can extend into the thigh veins, and indeed they do. Does it matter if most of them do or some of them do? Probably not. Many aspects of this answer are just tricks. As a doctor, I know that it’s possible for a calf DVT to extend to a thigh DVT. But answer choice “A” is actually wrong because only some calf clots extend. Not most.

Choice “B” says, “Most patients have classic physical exam findings.” Studies have shown that the sensitivity for leg swelling (classic physical exam finding) is anywhere from 35 percent to 95 percent. So would you say that most have classical physical exam findings? I guess it depends whether you go with the study that showed 35 percent or the one that showed 95 percent. The examiner is expecting you’ll go with 35 percent, as he means for this to be an incorrect answer.

The next answer choice “C” says, “Pregnancy is a predictor according to the Wells criteria.” This statement is in here to throw you off, because pregnancy is not technically part of the Wells criteria. The Wells criteria help determine what patients are at high risk for DVT. Despite this, you would be hard pressed to find a clinician with a pregnant patient with a swollen leg who would not order an ultrasound to rule out DVT. We’ve all seen it enough times. The pregnant lady with the clot in her leg. So this answer, while technically false, is practically useless.

The last answer choice “D” is the correct one according to the examiner. It says up to 10 to 15 percent of calf thrombi result in PE (a blood clot in the lungs). But how important is this actual percentage number? The important concept here is, “Do calf vein thrombi ever result in pulmonary emboli?” and the answer is, “Yes, some of them do.”

But is it 10 to 15 percent? Is it is it 30 percent? Is it 45 percent? Who cares. The truth is, I’m a practicing emergency physician for the past 13 years, and I don’t know the exact percentage of calf thrombi that result in pulmonary embolism. I do know that it can happen, but not very often. Isn’t that good enough? No. According to the writers of the standardized tests, I should know the exact percentage

I get that test questions are difficult to write. But often the veracity of one answer choice rests on some statistic that most physicians are not going to know on the tip of the tongue. Sure, I can study for a few months and try to cram as many of these facts in my head as I can. But the point is: When I’m at your bedside, and you have been pregnant, and your calf hurts, I know that DVT is a possible diagnosis. I will search for it, and exclude it, or find it. Isn’t that more important than whether I can tell you the exact statistical numbers?

There’s got to be a better way. Recently the American Board of Anesthesiology decided to stop administering its 10-year multiple choice “recertification exam” to experienced anesthesiologists. This just seems like common sense. The state bar does not administer multiple choice question choice questions to seasoned lawyers after decades of practice. And clinicians who are doing a good job of keeping up with the literature should not be asked to do the same. Continuing education requirements already help ensure that doctors are keeping on top of things. Why continue this archaic practice of giving multiple choice tests over obscure facts every ten years? It’s just not necessary.

Jeremy Gabrysch is an emergency physician.

Image credit:


View 4 Comments >

Most Popular

✓ Join 150,000+ subscribers
✓ Get KevinMD's most popular stories