An ECG compression algorithm for full disclosure in a solid-state real-time Holter monitor

In a real-time Holter ECG analyzer, a limited amount of solid-state memory is available for storing the 24 hours of data required for full disclosure capability. In the analyzer discussed here, 25 hours of a single channel of ECG data had to be stored in 1.5 Mbyte of RAM. Meeting the necessary data-compression goals while maintaining adequate waveform fidelity was not possible using standard data-compression techniques. The authors developed an ECG compression method for full disclosure data that meets the design requirements without imposing much of a computation burden on the real-time device. The method makes full use of the high degree of redundancy in the ECG signal. In addition, it uses filtering and subsampling techniques that, while inherently distorting, do not distort the waveform in such a way as to alter the clinical interpretation that can be made from a minaturized full disclosure report format. A quantitative analysis of the distortion is included