Linear Regression in Haemodynamics

Abstract

We hypothesized that the dynamic relation between diastolic, systolic blood pressure and pulse pressure is a novel measure of individual homeostatic features of the cardiovascular system, which can be easily determined from ambulatory blood pressure recordings and which can predict acute hypotensive episodes of the patients. Using the clinical blood pressure data of the patients of intensive care units, we plotted diastolic and systolic blood pressure against pulse pressure and used the linear regression coefficients for distinguishing between two groups of patients who experience an acute hypotension episode, and patients who do not. We exploited Support Vector Machines-based classifier for separating these disjoint populations and finding the indices that are prognostic of acute hypotensive episodes in these individuals.