<<Disclaimer: Verify this information before applying it to your situation.>>

On Thu, 19 Sep 1996 [log in to unmask] wrote:

> I am not sure I am going to get an answer to my question about the
> conditional probability, but the way I arrived at the figures I quoted
> was as follows:
>
> For EmA:
> Sensitivity 98%  Specificity 98%
>
> For the ocurrence of CD in the general population I was using .4% (1:250).
>
> So, out of 10,000 people there would be 40 celiacs.
>
> If those 40 were tested for celiac supposedly 39.2 (98%) would show
> positive for EmA.

Yes, that makes sense to me.

> Of the 9,960 in the general population who did not have CD were tested
> for CD then 98% would show negative on the CD test.
>
> That would mean 9760.8 would show negative on the test, leaving 199.2 of
> the people who did not have CD showing as positive on the EmA test.

I think that the level of specificity refers to those who test positive,
not all those tested. I think that would be referred to as incidence.
One in 250 healthy blood donors show elevated levels of EmA.

> My problem was with the predictive values given for the various blood
> tests.  They were fairly high, but as seen above, in a population of
> 10,000, 39.2 who had the disease would test positive and 199.2 who did
> not have the disease would test positive.  Thus a positive test result
> would not be a good predictor of having the disease.
>
> One would need to go through the above for all the blood tests and figure
> out the probability for testing positive on all the blood tests.  I was
> just wondering if it had been done before I tried to do it (especially
> since I am by no means proficient in this area).

Since the European Society of Pediatric Gastroenterology and Nutrition
(ESPGAN) accepts elevated levels of ARA and EMA as diagnostic of celiac
disease, your interpretation of the predictive value of the tests is
unlikely.