Telemedicine assisted secondary prevention with individual forecasting based on ECG monitoring

The aim of our study was to apply two HRV methods (PD2i and wavelet-CART) for the individual forecasting of ventricular tachycardia (VT) and death in telemedicine setting during the cardiac rehabilitation of postinfarction patients. Eighty-eight postinfarction patients were ECG monitored monthly during 36 months. The predictive accuracy of time-frequency HRV variables for ventricular tachycardia (VT), wavelet (W;Daub-4) decomposition parameters from level 2 (W2) to level 256 (W256)) analyzed. The best variable was W32 (W 0.799, p<0.0001). Th eCART's decision sensitivity and specificity reaching 84.5% and 91.9% respectively. The role of telemedicine management (TM+), and 94 age-matched control group without it (TM-) were also studied. Nine deaths in the TM+, and 21 death in the TM-group were observed (p<0.01). The sensitivity, specificity, positive and negative predictive accuracy of the PD2i values (cutoff 1.9) was 66.4, 79.5, 64. and 83.7% respectively.