Madness and method in stress echo reading

Abstract

Aim: To assess whether 'eye education' through short-term, high-intensity joint reading sessions may improve diagnostic accuracy and inter-observer agreement among beginners.

Methods and results: Seventeen cardiologists with absent to minimal (<100 studies performed) previous stress echo experience independently and blindly read 18 stress echo studies, nine at the beginning ('pre-training' set) and nine at the end ('post-training' set) of a 2 day stress echo school which included a joint reading session of 50 tapes. The two sets were balanced as far as type of stress and image quality. The 17 observers had an average accuracy score of 51+/-16.4 before and 64.3+/-8.7% after the training (P<0.005). Concordant (i.e. >14 readers giving the same response) interpretation occurred in three out of nine studies before and in eight out of nine studies after the training (33% vs 88%, P<0.01). Kappa values went from 0.14 (poor) before to 0.39 (fair, close to moderate) after the training.

Conclusion: Short-term, high-intensity dedicated training in stress echo, with joint reading sessions and consensus development of reading criteria significantly increased accuracy and markedly reduced the inter-observer variability in the reading of stress echoes by beginners. If there is a Shakespearean madness in stress echo reading, 'yet there is a method in't' (Hamlet, II, II, 205-206).