Background: Dynamic systems theory and entropy proportions of the cardiac attractors allow developing a predictive diagnostic methodology for Holter. This method differentiates normality, chronic and acute disease, as well as evolution between states in a quantitative way, being useful as a clinical tool for diagnosis and prediction of dynamics that evolved to more acute conditions.
Objective: In this work the clinical applicability of this method is confirmed by means of a blind study applied to 600 holter, including normal and with different pathologies, which conventional diagnosis, taken as Gold Standard, was masked.
Results: For each holter, it was built a computational simulation of the cardiac dynamic, starting from the heart rate and the total beats per hour values, taken in 18 hours. This information was taken to build the chaotic attractor of each dynamic. Next, it was evaluated the probability, non-equiprobable entropy, and entropy proportions of the occupied spaces of the obtained attractors, and based on this information the physical diagnosis was established. Finally, the conventional diagnosis was unmasked and compared with the obtained results.
Conclusions: Normal cases had entropy proportions inside the limits previously established for normality in the physical diagnosis, while abnormal cases showed always two or more of this values outside this limits. The blind study showed values of sensibility and specificity of 100% and a Kappa coefficient value of 1, confirming the clinical applicability of the method.
The quantitative measures of the new method diagnose each particular case in an objective and reproducible way, and they allow to predict the impact of surgical or pharmacological interventions, improving the conventional clinical method with physical predictions.
