There are various types of food-allergy tests, but beware: They're not all equally accurate. Some, in fact, can be very misleading.
By far the best are food challenge tests, in which a person is fed carefully regulated amounts of food, under close supervision, and the amounts are gradually increased.
Other tests examine blood or skin reactions. These are more common -- and also less reliable.
The most common blood test looks for blood proteins called IgE antibodies that respond specifically to particular foods. (Doctors also often order a test that looks for a different antibody, IgG, though that test has no scientific validity, allergy experts say.)
IgE tests are very good at confirming the lack of an allergy -- but only when the antibody is just plain missing. In many cases, for reasons scientists don't understand, just about everybody has antibodies to foods they don't react to.
And even when people are legitimately allergic to a food, levels of antibodies don't necessarily predict to what degree.
"No two patients are the same," says David Fleischer, a pediatric allergist at National Jewish Health in Denver.
"Diagnosing food allergies is in some ways more art than science."
The same is true for skin tests, which involve dipping a needle into a solution of the suspected allergen, then pricking the skin and watching for mosquito bite-like bumps to appear.
These skin tests are fairly reliable when they're negative. Positive reactions to foods, on the other hand, are wrong up to 50% of the time.
Someone with a milk allergy might develop bumps from beef and pork pricks, and someone with an egg allergy might react to chicken and turkey pricks, simply because the proteins in those food groups are similar enough, or because a person has sensitive skin.
Those bumps don't necessarily mean that the patient will have any sign of allergy after eating the suspected food. Follow-up testing is always recommended.