An Adaptive Complex Algorithm for Shockable Heart Rhythm Detection and Its Use in Automated External Defibrillators

Abstract

Cardiovascular mortality in Russia is very high [1]. Eighty percent of deaths caused by cardiac arrest are observed outside hospitals. Automatic external defibrilla tors (AED) are intended to eliminate life threatening arrhythmia caused by cardiac arrest. The AED is designed to provide cardiopulmonary resuscitation under inpatient and outpatient settings. AED induced cardiac defibrilla tion is mediated by powerful 1 10 msec electric pulses. Each minute of retarded successful defibrillation causes a 7 10% decrease in the probability of successful outcome. The AED processor is based on the shockable heart rhythm (SHR) algorithm. The sensitivity and specificity of SHR detection algorithms determine efficiency of the defibrillation procedure. There are a large number of SHR algorithms [2,3]. Progress in computer technology has increased the overall efficiency of SHR algorithms [4,5]. The resulting SHR algorithm combines efficiency of several algorithms included in the software, thereby increasing sensitivity and specificity of the method [6].The goal of this work was to construct an adaptive complex algorithm for shockable heart rhythm detection and its use in automated external defibrillators. This approach is based on ECG mediated defibrillation effi ciency. MethodsThe sensitivity and specificity of an SHR algorithm determine its efficiency. The SHR algorithm sensitivity is the ratio ТР/АР, where TP is the number of correctly identified SHR events; AP is total number of SHR events. The SHR algorithm specificity is ratio TN/AN, where TN is the number of non SHR (NSHR) events; AN is total number of NSHR events.Individual sensitivity and specificity characterize each method of SHR detection. This should be taken into account in construction of the resulting algorithm.The weighting coefficientwas suggested in [5]. This coefficient is calculated from the following equation:where Se is algorithm sensitivity; Sp is algorithm speci ficity; a is the boundary parameter derived in [5]. When the boundary parameter is passed, SHR is changed to NSHR. The contribution of specificity to Eq. (1) is more significant than the contribution of sensitivity. This is due to SHR international standard [7]. This approach allows parameter a to be optimized and contribution of each algorithm component to be determined based on weighting coefficients.Parameter of efficiency Eff(a) provides information about not only weighting coefficients but also about the group of cardiac rhythm. Efficiency coefficient provides adaptation to parameter modification. This coefficient is calculated fromThe sensitivity and specificity of shockable heart rhythm detection algorithms should be nearly 100% in auto matic external defibrillators. To attain these values, a method combining several algorithms and based on the weighting coefficients and coefficients of efficiency is reported in this work. The algorithm was tested on anno tated databases of the ECG signals (AHA, MIT, and CU).