Hypertension risk assessment based on a trend prediction methodology

Abstract

This work presents a new strategy for the prediction of biosignals' future evolution trend, based on the Haar "à-trous" wavelet transform. The proposed scheme is based on the hypothesis that the future evolution of a given biosignal (template) can be estimated from similar patterns existent in a historic dataset. The proposed approach, which does not use an explicit model, considers the wavelet decomposition of the signals (template and similar patterns) to determine the most representative trend at each of the several decomposition levels. Then, a set of distance-based measures, able to assess the likelihood of the representative trends in contributing to a consistent prediction, is introduced. From these measures and through an optimization process, a subset of these trends, called optimal trends, is selected and aggregated to derive the required biosignal future estimation. The effectiveness of the methodology was tested in the assessment of hypertension risk using blood pressure signals collected in the context of two tele-monitoring studies: TEN-HMS and MyHeart. The obtained results, in terms of sensitivity and specificity, showed the potential of the approach.