Abstract

ALVAREZ ALIAGA, Alexis et al. Prediction tree for hypertensive heart disease. Rev cubana med [online]. 2014, vol.53, n.3, pp.266-281. ISSN 0034-7523.

Introduction: hypertensive heart disease is associated with high morbidity and mortality. The complexity of its prevention lies in the number of risk factors involved in its development. Objective: design and validate a prediction tree for the development of hypertensive heart disease on the basis of hemodynamic and non-hemodynamic factors. Methods: a prediction tree for the development of hypertensive heart disease was designed and validated following the procedure of knowledge discovery in databases and data mining. The patients surveyed were being cared for at the hypertension specialized consultation of the secondary care polyclinic in Carlos Manuel de Céspedes General University Hospital, municipality of Bayamo, Granma province, Cuba, from 1 January 2004 to 31 December 2009. Results: the tree predicted risk of developing hypertensive heart disease in 82.598% of the patients, with an area of 0.861 under the ROC curve and a true positives rate of 0.733 and 0.921 for classes 1 and 2, respectively. The most important factor was C-reactive protein, followed by glycemia, uric acid, cholesterol and microalbuminuria, in that order. Conclusions: the decision making tree and the set of rules deriving from it allowed prediction of the risk to develop hypertensive heart disease in high blood pressure patients. Their incorporation into clinical practice will improve the evaluation of these patients.

Keywords : hypertensive heart disease; arterial hypertension; risk factors; prediction tree.

        · abstract in Spanish     · text in Spanish     · Spanish ( pdf )