Methods of frequency-correction for systolic time intervals.

To judge the influence of drugs and diseases on systolic time intervals (STI), the connection to the covariate belonging to heart frequency has to be eliminated. For this purpose, we adopt linear dependency of the interval between the R-waves in the ECG and STI. In this paper, according to different medical questions, well-fitting correction formulae are derived, based on new biometric analysis of published data. The methods used include L1-norm and L2-norm regression, rank-test connected regression Theil and Sen, and rank-scores analysis of covariance. It is concluded that the biometric methods must be well adapted to medical methods.