[Improved method for the diagnosis of diabetes mellitus (author's transl)]

Based on a sample of 953 subjects without known metabolic abnormalities and a sample of 186 diabetics a classification rule was developed by a quadratic discriminant analysis. The measurements of blood glucose 1 and 2 h after the oral administration of 100 g oligosaccharids are used to classify the patients into the categories of "healthy", "suspect" and "diabetic". The rates of false positive and false negative classification were estimated from a second sample of 128 non-diabetic and 73 diabetics. By administration of 100 g oligosaccharids these rates were found to be 3.1% and 0%, respectively. An analogous study for the administration of 100 g glucose was based on a different sample of N = 472, 99 diabetics included. Misclassification rates were found to be 1.6% and 2.7%, respectively. For practical application of the rules two-dimensional netplanes are given.