THE ACCURACY AND INTEROBSERVER AGREEMENT IN DETECTING THE GALLOP SOUNDS BY CARDIAC AUSCULTATION-ASTERISK

Study objectives: To determine the observer accuracy and interobserver agreement in identifying S-4 and S-3 by cardiac auscultation and whet her they improve with increasing observer experience. Design: Prospect ive, blinded study. Setting: Cardiology and general internal medicine wards in a university-affiliated teaching hospital. Patients: Forty pa tients with a cardiac diagnosis and 6 patients without were studied. M easurements and results: Two cardiologists, one general internist, thr ee senior and two junior postgraduate internal medicine trainees, blin ded to the patients' characteristics, examined the patients and docume nted their findings on a questionnaire. Computerized phonocardiogram w as obtained in all patients as a gold standard and was interpreted by a blinded, independent cardiologist. The mean positive predictive valu es for S-4 and S-3 were 51% (range, 24 to 100%) and 71% (range, 50 to 88%), respectively. The mean negative predictive values for S-4 and S- 3 were 82% (range, 67 to 94%) and 64% (range, 56 to 85%), respectively . The overall interobserver agreements for detecting S-4 was K = 0.05 (95% confidence interval [CI], 0.01 to 0.09) and S-3 was K = 0.18 (95% CI, 0.13 to 0.24). There was no apparent trend in the accuracy or int erobserver agreement with regard to the level of observer experience. Conclusion: The agreement between observers and the phonocardiographic gold standard in the correct identification of S-4 and S-3 was poor a nd the lack of agreement did not appear to be a function of the experi ence of the observers. The overall interobserver agreement for the det ection of either S-4 or S-3 was little better than chance alone.