Accuracy and cost-effectiveness of two handheld electrocardiogram recorders to screen for atrial fibrillation in a hospital setting

Abstract

Introduction: Early detection of asymptomatic atrial fibrillation (AF) allows earlier institution of measures to prevent morbidity and mortality, mainly related to stroke and heart failure. Handheld electrocardiogram (ECG) devices could facilitate the detection of patients with silent AF as they are easy to use, portable, low-cost and have built-in algorithms allowing immediate interpretation of the ECG. Yet, their usability in hospital populations at high risk has never been evaluated. Purpose: The aim of this study was to evaluate the accuracy and cost-effectiveness of two single-lead handheld ECG devices to detect AF in a hospital setting. Methods: In total, 445 patients hospitalised at the departments of cardiology and geriatrics who were able to consecutively hold the MyDiagnostick (Applied Biomedical Systems, The Netherlands) for one minute and the AliveCor (AliveCor Inc., USA) for 30 seconds were included. A simultaneous 12-lead or 6-lead ECG recording was used as a ‘gold standard’ to calculate the sensitivity and specificity of the automated algorithm of each device. Time investments by nurses and physicians were tracked and used to estimate the hospital costs per newly identified AF patient and the costs per preventable stroke (based on the mean CHA2DS2-VASc score and the assumption that 64% of strokes can be prevented using proper oral anticoagulation therapy). Results: Even after exclusion of patients with an implanted device (n=67), the sensitivity and specificity of the automated algorithms were suboptimal (Cardiology: 81.8% respectively 94.2% for MyDiagnostick; 54.5% respectively 97.5% for AliveCor; Geriatrics: 89.5% respectively 95.7% for MyDiagnostick; 78.9% respectively 97.9% for AliveCor). One patient was newly detected with AF at the cardiology ward. At geriatrics, two new AF patients were discovered on the 6-lead ECG recording. However, the automated algorithm of both devices detected only one of these two patients. In patients without an AF history, the AliveCor was the most cost-effective method to identify one new AF patient, with a direct hospital cost of €193 at the cardiology ward and €82 at the geriatric ward. For the MyDiagnostick these values were €293 respectively €134. Translating these costs into hospital screening costs to prevent one stroke per year, €1916 (AliveCor) or €3115 (MyDiagnostick) would be needed at the geriatric ward (average CHA2DS2-VASc of 5.0±1.5). For the cardiology ward (average CHA2DS2-VASc of 3.9±2.0), costs were €7535 using AliveCor and €11439 using the MyDiagnostick. Conclusion: Using AliveCor or MyDiagnostick handheld recorders for AF screening in a hospital setting seems challenging as incorrect automatic analyses are still common. However, when screening is performed in a hospital population without AF history and without an implanted device, the expenses per diagnosed new AF patient and preventable stroke are reasonable.