How do you correct PPV/NPV for prevalence rate?

[12floz]

Hi everyone,

As I am sure you all know, PPV and NPV are affected by prevalence. In all the UWorld examples, they always somehow specify that the disease incidence is 50% so that you don't have to do any extra correction, but I was wondering on how to do that if I ended up needing to? Does anyone know? Thanks!

---

Members don't see this ad.

 

[lymphocyte]

Hi everyone,

As I am sure you all know, PPV and NPV are affected by prevalence. In all the UWorld examples, they always somehow specify that the disease incidence is 50% so that you don't have to do any extra correction, but I was wondering on how to do that if I ended up needing to? Does anyone know? Thanks!

PPV isn't "corrected" for prevalence. You simply cannot calculate PPV without knowing some of the same numbers you need to calculate prevalence. PPV and prevalence are therefore interrelated. (Another way to think about it: how can you know the number of true positives without knowing how much disease was in the population to begin with?)

If they give you prevalence and sensitivity/specificity, and ask about PPV, you can do two things: 1) regurgitate a heinously long formula or 2) fill in the simple table from FA and do a little algebra, knowing that TP + FN = the numerator of the prevalence and N is the denominator of the prevalence.

Also, incidence is not the same as prevalence.

I found this link very helpful. Don't just memorise the formulas, but plug and play random numbers till the problems become second nature.

http://www.medcalc.com/bayes.html