| Abstract: | A study was made to determine how well the results of skin tests for sensitivity to various foods agreed with observation of clinical reactions to those foods. Test reactions were divided into several categories-negative, and 1, 2, 3 or 4 plus. Then the strong reactions, that is the 3 and 4 plus reactions, the milder reactions and the negative results were studied separately to determine the agreement of results, in each category, with the clinical response. Wide variations were noted. For some foods the agreement was high, for others low. For some foods, the agreement was high in some categories of reaction, low in others. For example, negative results of skin test might match with nonreaction to the food clinically in a high proportion of cases, and 3 or 4 plus reaction to skin test might be in close agreement with the incidence of distress upon ingestion of the food, yet for the same food there might be very poor correlation between mild reaction to skin test and clinical response. This being the case, accuracy of skin tests cannot be determined simply by combining all data on reactions, of whatever degree, and taking the aggregate of agreement in all categories as an index of the validity of the test. Each category of reaction must be considered separately.Combined data and categorized data on accuracy of skin tests for sensitivity to 26 foods were tabulated in the present study. |
