Inter-rater reliability for noninvasive measurement of cardiac function in children

Abstract

Introduction: A transcutaneous ultrasound monitor has recently been developed which noninvasively and quickly measures cardiac output. Validity and reliability testing has been reported in adults. No reliability testing has been undertaken in the pediatric population. Objective: Our objective was to evaluate the inter-rater reliability of a transcutaneous Doppler ultrasound technique to measure cardiac index (CI) and stroke volume index (SVI) in pediatric emergency department patients. Methods: An 8-month prospective observational study was conducted on a convenience sample of emergency department patients younger than 18 years old. Five raters were trained to use an ultrasound cardiac output monitoring device. Two raters, blinded to each other's results, obtained independent measurements from the same patient within 15 minutes of each other. Inter-rater agreement was measured with the Pearson product correlation coefficient. Bland-Altman analysis demonstrated the extent of deviation from a line of agreement between raters. Results: Ninety-seven patients were enrolled. Major diagnostic categories included infection, trauma, and gastrointestinal disorders. There was significant inter-rater correlation for CI (r = 0.76; 95% confidence interval, 0.66-0.83; P < 0.0001) and SVI (r = 0.79; 95% confidence interval, 0.70-0.86; P < 0.0001). Bland-Altman analysis of CI measurements between 2 raters showed bias of 0.06, SD of bias 1.00, and 95% limits of agreement -1.91 to 2.02 L/min/m. Stroke volume index showed bias of -0.5, SD of bias 11.01, and 95% limits of agreement -22.08 to 21.08 mL/m. Conclusions: Transcutaneous Doppler ultrasound technique demonstrates acceptable inter-rater agreement for measuring CI and SVI in children. Copyright © 2008 by Lippincott Williams & Wilkins.