The erythema and induration measurements in the tuberculin test: decomposition of the frequency distribution into normal components by a graphic method

Abstract

The tuberculin test by the Mantoux technique is used in order to classify people according to the size of the reaction caused by a single standard shot of tuberculin. Empirical data seem to demonstrate that such reactions obtained in general populations really represent a mixture of at least two separate groups. Aiming at the separation of these groups, in order to solve the classification problem later on, the Bhattacharya method was used. A decomposition of the frequency distribution of the erythema and induration measurements into normal components was thus obtained. The results of this decomposition showed that the number of components belonging to the distribution of the erythema variable was not the same as that for the induration one.