Multivariate analysis techniques applied for the performance measurement of Federal University Hospitals of Brazil

The presence of teaching hospitals represents a strategic role in improving the quality of life of the Brazil Unified Health System patients, based on the qualified training of health, medical and complementary professionals, which means efficiency in hospital production. Therefore, the purpose of this paper was to apply the multivariate analysis techniques of principal component (PCA) and clusters analysis in the performance measurement of Federal University Hospitals, aiming to contribute for the improving of Brazilian health care services. The research is descriptive, with exploratory purpose, being organized according to three stages of development, that is, the use of principal component analysis; in the second stage, the selection of inputs and outputs for the management of hospital organizations and; in the third, the use of clusters analysis technique, for the inputs and outputs defined. Results show that the groups' formation represented divergences between both techniques applied. Thus, it was verified that two hospitals, of twenty, remained alone in both clusters and PCA, demonstrating that these organizations differentiate from the others, regarding the similarity of their characteristics described by the original variables used in the analysis.