Improved prediction of defibrillation success for out-of-hospital VF cardiac arrest using wavelet transform methods.

We report an improved method for the estimation of shock outcome prediction based on novel wavelet transform-based time-frequency methods. Wavelet-based peak frequency, energy, mean frequency, spectral flatness and a new entropy measure were studied to predict shock outcome. Of these, the entropy measure provided optimal results with 60 +/- 6% specificity at 91 +/- 2% sensitivity achieved for the prediction of return of spontaneous circulation (ROSC). These results represent a major improvement in shock prediction in human ventricular fibrillation.