Kristyan B Guppy-Coles1, Sandhir B Prasad1, Kym C Smith2, Ada Lo2, Patrick Beard2, Arnold Ng3, John J Atherton4. 1. Department of Cardiology, Royal Brisbane and Women's Hospital, Brisbane, Qld, Australia; Faculty of Medicine, University of Queensland, Brisbane, Qld, Australia. 2. Department of Cardiology, Royal Brisbane and Women's Hospital, Brisbane, Qld, Australia. 3. Faculty of Medicine, University of Queensland, Brisbane, Qld, Australia; Department of Cardiology, Princess Alexandra Hospital, Brisbane, Qld, Australia. 4. Department of Cardiology, Royal Brisbane and Women's Hospital, Brisbane, Qld, Australia; Faculty of Medicine, University of Queensland, Brisbane, Qld, Australia. Electronic address: John.Atherton@health.qld.gov.au.
Abstract
BACKGROUND: Three-dimensional echocardiography (3D-Echo) performed by novice health care staff to measure left ventricular ejection fraction (LVEF) could allow cost-effective screening and monitoring for left ventricular systolic dysfunction (LVSD) prior to the development of heart failure. The aim of this study was to determine feasibility and accuracy of cardiac nurses (after completing focussed training) independently acquiring 3D-Echo images, and measuring LVEF using semi-automated software when compared to an echosonographer. METHODS: One echosonographer and three cardiac nurses acquired 3D-Echo images on 73 patients (62 ± 16 years, 62% male) with good image quality, and subsequently measured LVEF using a semi-automated algorithm. RESULTS: Overall feasibility was 89% with the three nurses successfully acquiring 3D-Echo images suitable for LVEF assessment in 65 of the 73 patients. High accuracy (r = 0.82; p < 0.0001) with minimal bias (+0.1, -10.6 to +10.8 limits of agreement; p = 0.91) was observed comparing the nurses to the echosonographer for measuring LVEF. Individual nurses demonstrated high feasibility (86%-92%), accuracy (r = 0.83-0.87; all p < 0.0001) and intra-observer reproducibility (r = 0.96-0.97; all p < 0.0001), with good inter-observer consistency in accuracy compared to the echosonographer (one-way analysis of variance p = 0.559). CONCLUSIONS: We have demonstrated that, following a focussed training protocol, it was feasible for cardiac nurses to acquire 3D-Echo images of sufficient image quality to allow measurement of LVEF using a semi-automated algorithm, with comparable accuracy and intra-observer variability to an expert echosonographer. This could potentially allow the broader application of echocardiography to screen for LVSD in high-risk cohorts.
BACKGROUND: Three-dimensional echocardiography (3D-Echo) performed by novice health care staff to measure left ventricular ejection fraction (LVEF) could allow cost-effective screening and monitoring for left ventricular systolic dysfunction (LVSD) prior to the development of heart failure. The aim of this study was to determine feasibility and accuracy of cardiac nurses (after completing focussed training) independently acquiring 3D-Echo images, and measuring LVEF using semi-automated software when compared to an echosonographer. METHODS: One echosonographer and three cardiac nurses acquired 3D-Echo images on 73 patients (62 ± 16 years, 62% male) with good image quality, and subsequently measured LVEF using a semi-automated algorithm. RESULTS: Overall feasibility was 89% with the three nurses successfully acquiring 3D-Echo images suitable for LVEF assessment in 65 of the 73 patients. High accuracy (r = 0.82; p < 0.0001) with minimal bias (+0.1, -10.6 to +10.8 limits of agreement; p = 0.91) was observed comparing the nurses to the echosonographer for measuring LVEF. Individual nurses demonstrated high feasibility (86%-92%), accuracy (r = 0.83-0.87; all p < 0.0001) and intra-observer reproducibility (r = 0.96-0.97; all p < 0.0001), with good inter-observer consistency in accuracy compared to the echosonographer (one-way analysis of variance p = 0.559). CONCLUSIONS: We have demonstrated that, following a focussed training protocol, it was feasible for cardiac nurses to acquire 3D-Echo images of sufficient image quality to allow measurement of LVEF using a semi-automated algorithm, with comparable accuracy and intra-observer variability to an expert echosonographer. This could potentially allow the broader application of echocardiography to screen for LVSD in high-risk cohorts.