The accuracy of ECG screening by GPs and by machine interpretation in selecting suspected heart failure patients for echocardiography

National Institute for Health and Clinical Excellence (NICE) guidelines in the UK state that suspected heart failure patients should have an ECG in order to select patients for echocardiography. The research underpinning this recommendation comes from studies in which cardiologists interpreted the ECGs. In practice, however, it would be general practitioners (GPs) interpreting ECGs. The aims of this study were to assess both GPs and ECG machine interpretation in their ability to use ECGs to select suspected heart failure patients for echocardiography. Six GPs were asked to classify 90 ECGs taken from suspected heart failure patients either as normal or as having an abnormality present. The ECG machine report was also used to classify the ECG in the same way. These results were compared to a gold standard interpretation. The GPs and the ECG machine report would have not referred 17.8% and 8.3%, respectively, of the appropriate patients for echocardiography. In doing so, the GPs would have missed 5.6% of patients with left ventricular systolic dysfunction (LVSD) whereas the ECG machine report would have missed none. We conclude from our findings that there is heterogeneity between GPs in their interpretational skills. Some GPs could successfully use ECGs to select patients for echocardiography. The difference in cost between performing echocardiography on all patients and GPs screening with ECGs is £74 more per case. Screening with ECG machine interpretation costs virtually the same as performing echocardiography on all patients.