Arrhythmia has a high incidence, and is a serious threat to human health. However, due to concealed symptoms and unpredictability of onset, it is difficult for traditional ECG equipment to capture the ECG data at the onset of the arrhythmic events, so it could be misdiagnosed and under-treated. Fortunately, the diagnosis rate of arrhythmia could be greatly enhanced by analyzing the uploaded real-time ECG data of individuals measured at home using the wearable single-lead ECG monitoring system under the guidance of community doctors, but there is a lack of relevant evidence from real-world studies.

To assess the diagnostic accuracies of three schemes for screening arrhythmia in the community using a wearable single-lead ECG monitoring system.

A real-world, community-based study design was used for comparing three schemes for screening arrhythmia using a wearable single-lead ECG monitoring system: scheme 1 was used for collecting 24-hour ECG data on any three nonconsecutive days in two weeks, scheme 2 was used for collecting 24-hour ECG data on any day and at least three 1-hour ECG data in two weeks, and scheme 3 was adopted for collecting 72-hour ECG data on any three consecutive days and at least one 1-hour ECG data on any one day in two weeks. Subjects were 435 community-living residents, who were recruited from Yinchuan from September 2020 to September 2021. They were divided into male group (177 cases) , female group (258 cases) ; young group (135 cases) , middle-aged group (200 cases) and elderly group (100 cases) by age; primary school group (77 cases) , middle school group (165 cases) and university group (193 cases) by educational level; arrhythmia group (233 cases) and non-arrhythmia group (202 cases) by the history of arrhythmia. Subjects measured the ECG either at the onset time of perceived arrhythmia or not using any one of the three screening schemes chosen voluntarily with the wearable single-lead ECG monitoring system, then uploaded the measurement results to the cloud platform. The number of participants using each of the three screening schemes was counted. The correlation of age, education level or history of arrhythmia with scheme selection was analyzed. And detection rates of the three screening schemes were compared.

The number of subjects who selected the three screening schemes was 321, 40 and 74, respectively. The average age of the subjects was significantly different (P=0.047) . There was no correlation between gender, education level, medical history and protocol selection (χ²=0.670, P=0.715; χ²=2.994, P=0.559; χ²=2.225, P=0.893) . There was a significant correlation between different age groups and protocol selection (χ²=9.939, P=0.041) . The arrhythmia detection rates of the three screening protocols were 85.67%, 82.50% and 85.14%, respectively, and the difference was not statistically significant (χ²=0.286, P=0.867) . There was no significant difference in the arrhythmia detection rate between the male group and the female group (χ²=0.966, P=0.707; χ²=0.917, P=0.678) . There was no significant difference in the detection rate of arrhythmia among young group, middle-aged group and elderly group (χ²=2.102, P=0.350; χ²=0.871, P=0.706; χ²=1.063, P=0.622) . There was no significant difference in the detection rate of arrhythmia among the three screening schemes in primary school group, middle school group and university group (χ²=2.421, P=0.271; χ²=1.115, P=0.633; χ²=2.181, P=0.353) . There was no significant difference in the arrhythmia detection rate between the three screening protocols in the history group and the no history group (χ²=1.442, P=0.507; χ²=0.548, P=0.818) . The frequency of 1-hour ECG data collection in protocol 2 was positively correlated with arrhythmia detection rate (r_s=0.912, P=0.011) . The frequency of 1-hour ECG data collection in protocol 3 was positively correlated with arrhythmia detection rate (r_s=0.852, P=0.026) . In protocol 2, the detection rate of arrhythmia in 24-hour ECG data was 72.5%, and that in 1-hour ECG data was 77.5%. There was a strong consistency between the two kinds of long-term ECG data (Kappa=0.601, P=0.001) . In protocol 3, the arrhythmia detection rate of 72-hour ECG data was 82.4%, and the arrhythmia detection rate of 1-hour ECG data was 63.5%. There was a medium consistency between the two kinds of long-term ECG data (Kappa=0.410, P<0.001) . In protocol 2, there was a strong consistency between the diagnosis results of 1-hour ECG data and the total protocol (Kappa=0.844, P<0.001) . There was a strong consistency between 24-hour ECG data diagnosis and total protocol diagnosis (Kappa=0.717, P<0.001) . In protocol 3, there was a moderate consistency between the 1-hour ECG data diagnosis and the total protocol diagnosis (Kappa=0.466, P<0.001) , and the consistency strength was general. The results of 72-hour diagnosis were strongly consistent with those of the total protocol (Kappa=0.901, P<0.001) .

There is no significant difference in the arrhythmia detection rate among the three arrhythmia screening schemes based on community mobile health care, which can be used regardless of whether there are symptoms or not. Subjects of different ages have different tendencies to choose the three screening schemes, and the frequency of 1-hour ECG data collection is positively correlated with the arrhythmia detection rate, which suggests that the community doctors should select the optimal compliance screening scheme according to patients' age, occupational characteristics, economic income and other factors, so as to truly enable the screening and management of arrhythmia in the community using mobile technologies.

心律失常发生率高且严重威胁人类健康，但由于其症状的隐匿性和发作的不可预测性，传统的心电设备很难捕捉到发作时心电图，无法得以确诊和对因治疗。在社区医生指导下，居民居家自行应用单导联可穿戴心电设备采集数据并实时上传可以明显提升心律失常的诊断率，但目前国内外相关研究大多缺乏真实世界数据的支撑。

探讨基于社区移动医疗的3种心律失常筛查方案的真实世界研究。

选取2020年9月至2021年9月银川市所属社区纳入的435例居民作为受试者，根据性别分为男性组（177例）、女性组（258例）；根据年龄分为青年组（135例）、中年组（200例）、老年组（100例）；根据受教育程度分为小学组（77例）、中学组（165例）、大学组（193例）；根据既往有无明确心血管病病史分为有病史组（233例）和无病史组（202例）。应用单导联可穿戴远程心电设备，分别按以下方案采集心电数据：间断采集3次24 h心电数据（方案1）；采集1次24 h及至少3次1 h心电数据（方案2）；连续采集1次72 h及至少1次1 h心电数据（方案3）。由受试者自主自愿选择筛查方案中的任意1种，不论有无症状均自主选择时间，自行佩戴单导联可穿戴远程心电设备采集心电数据并上传至云平台。统计选择3种筛查方案的受试者数量、不同分组与方案选择的相关性，及3种筛查方案的心律失常检出率。

选择3种筛查方案的受试者分别为321、40、74例。其中3种筛查方案受试者平均年龄比较，差异有统计学意义（P=0.047）；受试者性别、受教育程度、有/无病史与方案选择无相关性（χ²=0.670，P=0.715；χ²=2.994，P=0.559；χ²=2.225，P=0.893）；不同年龄分组与方案选择有相关性（χ²=9.939，P=0.041）。3种筛查方案心律失常的检出率分别为85.67%、82.50%、85.14%，差异无统计学意义（χ²=0.286，P=0.867）；男性组、女性组受试者的3种筛查方案心律失常检出率比较，差异无统计学意义（χ²=0.966，P=0.707；χ²=0.917，P=0.678）；青年组、中年组、老年组受试者的3种筛查方案心律失常检出率比较，差异无统计学意义（χ²=2.102，P=0.350；χ²=0.871，P=0.706；χ²=1.063，P=0.622）；小学组、中学组、大学组受试者的3种筛查方案心律失常检出率比较，差异无统计学意义（χ²=2.421，P=0.271；χ²=1.115，P=0.633；χ²=2.181，P=0.353）；有病史组、无病史组受试者的3种筛查方案心律失常检出率比较，差异无统计学意义（χ²=1.442，P=0.507；χ²=0.548，P=0.818）。方案2的1 h心电数据采集次数与心律失常检出率呈正相关（r_s=0.912，P=0.011）；方案3的1 h心电数据采集次数与心律失常检出率呈正相关（r_s=0.852，P=0.026）。方案2中24 h心电数据心律失常检出率为72.5%，1 h心电数据心律失常检出率为77.5%，两种时长心电数据诊断结果间具有较强一致性（Kappa值=0.601，P=0.001）；方案3中72 h心电数据心律失常检出率为82.4%，1 h心电数据心律失常检出率为63.5%，两种时长心电数据诊断结果间具有中等一致性（Kappa值=0.410，P<0.001）；方案2中1 h心电数据诊断与总方案诊断结果间具有强一致性（Kappa值=0.844，P<0.001）；24 h心电数据诊断与总方案诊断结果间具有较强一致性（Kappa值=0.717，P<0.001）。方案3中1 h心电数据诊断与总方案诊断结果具有中等一致性（Kappa值=0.466，P<0.001），一致性强度一般；72 h诊断与总方案诊断结果具有强一致性（Kappa值=0.901，P<0.001）。

基于社区移动医疗的3种心律失常筛查方案心律失常检出率无显著差异，不论有无症状均可使用，不同年龄受试者选择3种筛查方案倾向性不同，1 h心电数据采集次数与心律失常检出率呈正相关，提示社区医生应结合患者的年龄、职业特点、经济收入等因素进行多维度评估后，选择能达到最优依从性筛查方案，才能真正使移动医疗服务助力于社区心律失常患者的筛查及管理工作。