The diagnostic accuracy of pocket-size cardiac ultrasound performed by unselected residents with minimal training

Abstract

Pocket-size imaging devices may represent a tool for fast initial cardiac screening in the emergency setting. Pocket-size cardiac ultrasound (PCU) examinations performed by experienced echocardiographers yield acceptable diagnostic accuracy compared to standard echocardiogram (SE). However, the success of this method when used by unselected non-cardiologists remains unexplored. The current study studies the diagnostic accuracy of PCU when used by unselected internal medicine residents with minimal training. All residents were given a 2-hour introductory course in PCU (Vscan) and reported PCU results for up to 15 predefined cardiac landmarks. These were arbitrarily divided into 3 priority groups, such that left ventricle (LV) and pericardium were of first priority. Diagnostic accuracy [sensitivity/specificity and negative/positive predictive values (PPV/NPV)] and agreement were evaluated using a subsequent SE as reference. During a 9.2 months period a total of 303 patients were included in the study, the majority on the basis of presenting with chest pain or suspected heart failure. In the pooled LV and pericardial (1st priority) data, sensitivity/specificity/PPV/NPV were 61/92/70/89% respectively. Similar specificities and NPVs were observed for the 11 remaining indices, as were lower sensitivities and PPVs. The best PCU sensitivity (76%) was attained for the assessment of LV wall motion abnormalities. Overall agreement was k = 0.50. PCU examination performed by internal medicine residents with minimal training could provide a suitable means of ruling out cardiac pathology, as reflected in the high specificities and NPVs. It is not, however, a satisfactory tool for identifying patients with various cardiac disorders.