Evaluation of monitoring cardiac output by long time interval analysis of a radial arterial blood pressure waveform using the MIMIC II database

Abstract

We recently proposed a technique to estimate relative cardiac output (CO) change by unique long time interval analysis (LTIA) of a radial arterial blood pressure waveform. Here, we evaluated the technique in 169 critically ill patients, while comparing it to previous "pulse contour analysis" techniques, using the public MIMIC II database. The LTIA technique achieved an overall calibrated CO error of 18.8% against reference (single determination) thermodilution measurements. This level of accuracy was not better than the previous techniques. However, the average absolute thermodilution CO change in each patient was only 12.3%. As the absolute CO change increased, the LTIA technique became increasingly more accurate than the previous techniques.