Three different algorithms for identifying patients suffering from atrial fibrillation during atrial fibrillation free phases of the ECG

The paper presents and compares three different methods to detect atrial fibrillation using algorithms. Each of these algorithms can identify patients suffering from atrial fibrillation even if there is no atrial fibrillation visible on the ECG. All methods are based on RR-intervals; thus, only a single channel ECG is required. These algorithms have been tested using the MIT-BIH atrial fibrillation database and the MIT-BIH normal sinus rhythm database. The sensitivity and specificity of method 1 is 91.5% and 96.9% respectively. Method 2 results in a sensitivity of 93.3% and a specificity of 92.8% and method 3 in a sensitivity of 94.1% and a specificity of 93.4%. Even if an atrial fibrillation burden of 0% is assumed, the sensitivity still proves satisfactory (sensitivity of method 1, 2 and 3 is 82.9%, 96.3% and 94.1%, respectively).