Abstract
The accuracy of the modified prick skin test (Pepys) for the diagnosis of food sensitivity reactions was evaluated, noting both immediate and delayed responses. Single‐blind tests were performed on 19 patients who had stable patterns of responsiveness to elimination and repeated open food challenges. Foods to which the patients were sensitized were tested, together with an equal number of foods to which (by challenge) the patient had been shown not to be sensitive. Skin test responses were compared with those to the open food challenges, the latter of which are assumed to be accurate. Using these criteria, the prick skin test had an accuracy of only 28%, with the best two extracts (cow's milk and cheese) scoring respectively 7 correct and 4 incorrect results and 6 correct and 2 incorrect results. The least accurate extracts were pork and sugar, scoring respectively 2 correct and 8 incorrect results, and 1 correct and 5 incorrect results. The author concludes that the modifed prick skin test is not a valid technique for diagnosing food sensitivity responses in humans.