One of the ways tests can be shown to be accurate or not is to see if their results can be replicated. I have had four tests over the years -- two RAST tests and two ELISA tests. I decided to do a bit of a comparison between each set of tests.
Below are my findings in chart form. What the results show are exclusive to me.
Checking the Accuracy of Food Allergy Tests Performed on Me
ELISA RAST
Number of foods tested that were the same foods in each test* 190 14
Percentage of severe** allergens overall that stayed the same in both tests (positive in one and positive in the next) -- this number approximately indicates the accuracy of the test in my case 30% 71%


What do the above percentages mean? Here's where it gets confusing (actually, I'm already confused). Let's say you're not pregnant and you have a specific pregnancy test twice, and both times it gives you a negative reading. That means that you got the same results both times (100%). But you would also get 100% if it gave you a positive result both times -- even though you're not pregnant! The percentage is thus not an accuracy percentage, but a percentage based on how closely the first test results resembled the second test results. To find out the accuracy of one test (ELISA or RAST) over the other, I would have to challenge the results of each test by trying each food, and comparing my reaction/lack of reaction to the test's findings. But then, I'd be accused of being subjective, so unfortunately that type of chart wouldn't be very helpful.
If nothing else, I hope this chart shows that:
  • Test results themselves cannot be taken at face value; they change over time and are not 100% accurate.
  • Allergy tests are still helpful in determining what foods should be "challenged."
  • A very accurate food allergy test does not yet exist, and still needs to be developed (currently RAST and skin testing are considered the most accurate).
    • * The RAST tests that were performed on me tested for different foods -- though one test tested me on 15 foods, the other was a test for 20, and many of them were not the same foods. Thus, I factored in only the foods that were tested in both tests (i.e., 14). The ELISA test tested for the exact same foods in both tests, so I needed to do this in order to ensure accuracy.

    ** Only severe allergens are factored in. Test numbers naturally go down over time, so borderline and moderate allergies were excluded.

    The tests were performed on one person over the course of a few years, from 1994 to 1997. This chart was created in January, 2002.

    Back to FAST