USMLE Step 2: Summarizing Stuff That’s Made More Complicated Than Necessary

1.  Antiarrhythmic Drugs:  Atropine speeds shit up. Adenosine slows it down. Amiodarone fixes anything, but kills your lungs. Everything else is 6th line.

2.  Antipsychotic Drugs:  All that’s worth knowing about the 2nd gens are their major side-effects, which luckily mirror their names: Olanzapine causes weight gain cause “o” is the fattest letter. Aripiprazole and Ziprasidone sound zippy and thin, and are both weight-neutral.  (Ziprasidone is, in fact, so zippy that it can cause arrhythmias).  And finally, Quetiapine sounds like “quiet”, which is appropriate because it has the quietest side-effect profile (the least risk of extrapyradimal S/Es or increased prolactin).

3:  Sensitivity of a test vs its Positive Predictive Value:  Tricky because they seem so similar, so the temptation is to think of them as the same exact thing.  Textbooks do a crappy job of explaining the difference, I think, so here’s my attempt at it:

sensitivity vs PPV

Say your friend makes an instant swab test for The Bubonic Plague.  The test correctly catches all but 1% of her PI’s Yersinia Pestis Orientalis samples at the expense of also catching 5% of the perfectly healthy controls- so her screening test is 99% sensitive and 95% specific.  You are impressed.

She swabs your cheek with it to while she’s explaining the test to you.  If the test then gives a positive result, how worried about imminent death would you really be?

Answer: you shouldn’t be worried at all, because The Bubonic Plague is so rare that the odds of you being one of the 5% false positives vastly eclipses the chance you are the index case of a modern pandemic.

So despite being an extremely sensitive test, it has a low PPV… in the modern population.  The key point is that if you were a medieval European peasant getting that same swab and testing positive, you’d correctly be far, far more worried.  The Positive Predictive Value of the exact same test would be higher in the 14th century.

And that’s the point: sensitivity & specificity depend only on the test, whereas PPV and NPV also depend on the prevalence of the disease of the population being screened.

10 thoughts on “USMLE Step 2: Summarizing Stuff That’s Made More Complicated Than Necessary

  1. You’re a good learner, which could make you a good teacher. Any thoughts of pairing teaching duties with your practicing of medicine? I think you’d do quite well, educating both patients/family members and future medical professionals.

  2. I always think of sensitivity and specificity as something that is useful to the researcher / person developing the test, and the NPV and PPV being more useful to the patient.

    I.e. the patient wants to know: Given I got a positive result, what are the chances I actually have the disease.

    The researcher wants to know: Given this population of people with the disease, how good is this test at identifying them.

    I like the part you added about prevalence.

  3. Hey there Action Potential,

    Great sensitivity vs PPV explanation! I think you might have made an error in your example though: you state “The test correctly catches all of herr PI’s Yersinia Pestis Orientalis samples at the expense of also *catching 5% of the perfectly healthy controls* – so her screening test is *95% sensitive.*” (asterisks added to draw attention). However, in order to have a sensitivity of 95%, it ought to be that the test misses 5% of the disease cases (not picks up 5% of healthy controls). This would allow the example to be consistent with the table shown.

  4. The best and finest thing I’ve ever read since forever 😍
    Thank you so much
    My exam is in a couple of days and I’m just going over things that usually gets me mixed up and thank God that I’ve stumbled into your post!

    Life is beautiful now.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s