Step I High sensitivity test to rule out dz?

[MudPhud20XX]

Advertisement - Members don't see this ad

So why isn't this the other way around? Intuitively, it makes more sense that you would want to use high sensitivity test to rule in some one has a dz, since the definition of sensitivity is the probability of that a person with a dz has a positive test. What am I missing here?

Let's say that a sensitivity is 100% and you test some one and get a positive result, then you are 100% confident to rule in the dz, right? Why should we think the other way around? Here is what FC says:

• Use high-sensitivity tests to rule out disease: Tests with sensitivity close to 1 have a low false-negative rate

 

[thehundredthone]

The attribute of utility of the test is not that of it's true results, but of its false results. A highly sensitive test has a very low false negative rate, and that's why it is useful in ruling out disease. If you were to use it to rule in disease, you would also rule in all the false positives, which the sensitivity of the test does not account for. A highly specific test with a very low false positive rate will help you rule in disease the same way.

 

[dyspareunia]

You can also think of it as what you would tell a patient.

High sensitivity, low specificity for HIV: Patient has positive result. "Hi patient, you might have (we ruled in) HIV". <-- Probably wouldn't do that as a doc.
Low sensitivity, but high specificity for HIV. Patient has positive result. "Hi patient, you probably (we ruled in) HIV". <--- That's more like it.

 

[MudPhud20XX]

Thanks guys! That makes much more sense now.

 

[bulldogmed]

So why isn't this the other way around? Intuitively, it makes more sense that you would want to use high sensitivity test to rule in some one has a dz, since the definition of sensitivity is the probability of that a person with a dz has a positive test. What am I missing here?

Let's say that a sensitivity is 100% and you test some one and get a positive result, then you are 100% confident to rule in the dz, right? Why should we think the other way around? Here is what FC says:

• Use high-sensitivity tests to rule out disease: Tests with sensitivity close to 1 have a low false-negative rate

Also, don't forget this easy mnemonic: SPIN, SNOUT --> specific, rule in; sensitive, rule out.