Validation of a waveform delineator device for cardiac studies: Repeatability and data mining analysis

Abstract

This paper envisages showing the potential of innovative non-invasive techniques based on affordable and easily operated instrumentation as well as user-friendly computer aided algorithms in the screening of cardiovascular (CV) diseases. These techniques are based on the assumption that arterial stiffness is currently an important predicator of the CV diseases development and can be assessed by analyzing the arterial pressure waveform (APW). A previously developed PZ based device for non-invasive APW capture is currently under test in clinical environment, using a heterogeneous population constituted by healthy and unhealthy subjects. A dedicated Matlab analysis tool was designed and developed to extract relevant information and further APW analysis. Several recordings of the APW in the same day and in consecutive months are being performed by trained observers, to evaluate its reproducibility. Data mining analysis is subsequently the last task where the Weka 3-6-5 package software is used. The usefulness of developing data mining algorithms for cardiovascular applications can benefit the CV screenings contributing for the early identification of arterial stiffness related patterns.