In the electrocardiogram (ECG), the QT interval is clinically one of the most important signals. The challenge in the interpretation of the QT measurement is its dependence on the heart rate, i.e., the RR intervals. To normalize out this dependency, a multitude of approximative correction methods have been developed, e.g., the power-law formulas of Bazett and Fridericia that are still in clinical use. We have developed a new QT correction method, which is superior to the existing QT correction tools.
Our method originates from transfer entropy (TE) in information theory. Previously, we have applied TE to solve the probabilistic dependence of the QT intervals on the RR intervals. Based on this dependence, we have developed a numerical software that takes the set of RR and QT intervals as the input, and computes the QT correction as an output without approximations or external models.
The difference between our method and the conventional QT correction methods is drastic: our method dynamically adapts to a multitude of previous RR intervals and gives the true QT correction as an output. The detection rate for abnormal QT times is very high compared to the conventional methods. On the other hand, in standard measurement conditions our method does not yield false positives in contrast with the conventional QT correction.
In summary, we have developed a QT correction method based on the dynamical RR-QT dependence. The results are superior compared to the conventional methods. Our method is readily applicable to ECG monitoring devices for clinical use, as well as to all the phases of drug development including QT measurements.