[Multidimensional discriminative analysis for predicting the outcome of Graves' disease (author's transl)]

Attempts to identify at an early stage of Graves' disease patients who will respond favorably to the various treatments available have been largely unsuccessful. However, all such studies have been based on the measurement of a single discriminant at a time, and in an attempt to improve predictive accuracy we have made use of multidimensional techniques which analyse several factors simultaneously. The overall accuracy in predicting the outcome in drug-treated patients, before treatment was started, was 79% ; in RAI treated patients 88 %. There remains a residuum of patients whose fluctuations in disease activity are such as to defy all prediction of outcome.