Inter-rater reliability of manual QT measurement and prediction of abnormal QT,HR pairs

OBJECTIVES: To assess interobserver agreement when trained healthcare staff measure the QT interval using a standardized approach across a range of QT lengths. METHODS: A sample of 110 electrocardiograms (ECGs) was taken from general and psychotropic overdose admissions to the emergency department including drugs known to cause QT prolongation and Torsades de Pointes. Four of the authors measured the QT interval in all ECGs using a previously developed approach. Each rater was blinded to the ECG admission details and the measurements of the other raters. The primary outcome was the inter-rater agreement for the median QT and interobserver classification as to whether the QT interval was abnormal or not, based on the QT nomogram. RESULTS: There was good agreement between raters (intraclass correlation coefficient, 0.61; 95% CI = 0.53 to 0.69). When classifying the QT as abnormal there was good agreement between the raters with a Fleiss' kappa of 0.61. There was perfect agreement between all four raters on 86 of 110 ECGs (78.2%), agreement between three raters on 18 of 110 (16.3%), and a split between the four on 6 of 110 (5.5%). Disagreement between the automated QT measurement and the majority of the raters was slightly more than within raters. CONCLUSIONS: The study demonstrates that there is good agreement between trained observers when manually measuring the QT interval using a standardized approach. This provides an inexpensive method for the measurement of QT in clinical studies of drug overdose and a potentially useful approach in the clinical assessment of patients with possible QT prolongation.