Ideas

I'm full of ideas! Here's one.

Medicine is full of tests. These tests can be useful, or not useful depending on what their "sensitivity" and "specificity" are. What does this mean? Let's use an example, say, complete blood counts (CBCs) and bleeding.

Sensitivity tells us how good positive CBCs are at picking up bleeding. Specificity tells us how good negative CBCs are at telling us there's no bleeding. When a test has lots of false positives, it means it's got great sensitivity, but crappy specificity. When a test has lots of false negatives, it means it's got crappy sensitivity but great specificity.

All of medicine loves these sensitivity/specificity/positive predictive value/negative predictive value garbage.

I'm here to talk about false false positives.

Say you use a CBC as a test for bleeding in a patient (all you doctors shut up; it's the best example I could come up with). The CBC shows anemia, and the patient is indeed bleeding. Yay, true positive. OR IS IT?????

What if the patient IS bleeding... what if the patient DOES have anemia... but what if the patient's bleeding isn't severe enough to cause anemia? What if the cause of the anemia is something else? This is what I mean by false false positive. A false false positive is a false positive in someone who just happens to have the illness you're testing for.

Someone should really study this. I mean, it won't really make a BIG change to the statistics out there already, but it's something that would impact the reliability of EVERY medical test. False negative rates would be adjusted upwards to reflect false negatives and false false positives.

Anyway. That's my brilliant idea. If you need a thesis, go right ahead. I didn't articulate it well, but I'm postcall so up yours, freeloading moochfaces.