1-year follow-up in patients identified with possible atrial fibrillation: sub-analysis of the omron heart study

Abstract Background As mobile health (mHealth) devices become more widespread, the opportunities for early detection of atrial fibrillation (AF) have increased. However, it is known that even when a device suggests the possibility of AF, few patients actually seek medical care. Purpose We conducted a sub-analysis of the Omron Heart Study, tracking the healthcare facility attendance of patients in whom "possible AF" was detected. We also explored strategies to encourage clinic visits and developed a simple scoring system. Methods We recruited participants by sending study invitations to users registered with the OMRON Connect application and volunteers via website. Patients aged 60 years or older with a history of hypertension and currently on antihypertensive medication were eligible for the study. After receiving a blood pressure (BP) monitor equipped with an electrocardiogram (ECG) measurement function, subjects recorded both ECG and BP measurements for three months. A follow-up questionnaire was administered at 3 months and 12 months after enrollment. Results A total of 4,078 participants were enrolled. Due to insufficient data and other reasons, 258 were excluded, leaving 3,820 subjects for analysis. Among these, 1,700 were detected with "possible AF." Of the 1,700 participants with "possible AF," 1,682 (98.9%) responded at the 3-month follow-up, and 1,323 (77.8%) responded at the 12-month follow-up. Only 399 (23.5%) visited a medical facility within 3 months, and 289 (17%) underwent additional examinations. Thirty-five patients (2.1%) were diagnosed with AF at 3 months and a total of 115 patients (6.8%) were diagnosed with AF over 12-month period. After the trial ended, the "possible AF" ECGs were interpreted by three cardiologists. As a result, 220 patients were interpreted to have AF. Logistic regression analysis was performed on the 1,700 participants with "possible AF" to identify factors associated with AF diagnosis. In the multivariate analysis, Sex, Metabolic syndrome, Age, Daily Palpitation, Burden, and Palpitation on "possible AF" detection were significantly associated with AF diagnosis. Using the initial letters of these variables, we developed a scoring system named the SMAP-BuP score, which demonstrated moderate diagnostic performance (AUC = 0.70). If the cut-off point was set to 3 points, sensitivity was 85% and specificity was 39%. Conclusion A sub-analysis of the Omron Heart Study highlighted a significant data gap, revealing that despite possible AF detection, few patients sought medical consultation or received AF diagnosis even after long-term follow-up. The SMAP-BuP score is a moderate predictive model for AF diagnosis in possible AF patients and holds potential for use in daily clinical practice.Logistic Regression Results SMAP-BuP Score