Katapadi Aashish, Chelikam Nikhila, Rosemas Sarah, Higuera Lucas, Colombowala Ilyas, Bansal Shanti, Darden Douglas, Pothenini Naga Venkata K, Koerber Scott, Tummala Rangarao, Park Peter, Gopinathannair Rakesh, Lakkireddy Dhanunjaya, Kabra Rajesh
Kansas City Heart Rhythm Institute, Overland Park, Kansas, USA.
Medtronic Cardiac Rhythm Management, Mounds View, Minnesota, USA.
JACC Adv. 2025 Apr;4(4):101656. doi: 10.1016/j.jacadv.2025.101656. Epub 2025 Mar 19.
Insertable cardiac monitors (ICMs) are essential for managing arrhythmias but often generate large numbers of transmissions and false alerts. Integrating artificial intelligence (AI) as part of the ICM workflow can reduce this burden. However, its impact on clinic workflow and resource utilization must be better understood.
The aim of the study was to assess the impact of AI-enhanced ICMs on clinic workflow and resource utilization.
A cross-sectional analysis was conducted using real-world, deidentified ICM remote monitoring data from Octagos Health, which included 140 U.S. device clinics between July 2022 and April 2024. Nonactionable alerts (NAAs) were defined as false or repetitive alerts transmitted on the remote monitoring platforms but dismissed by device technicians and not forwarded to clinicians for review. We compared NAAs generated by AI-enhanced vs non-AI-enhanced ICMs and estimated associated staffing hours, resources, and costs extrapolated for a clinic managing 600 ICM patients.
Among 19,320 patients (mean age: 69 ± 13.5 years; 47.3% male), 68% had non-AI-enhanced ICMs, and 32% had AI-enhanced ICMs. The mean annual NAA volume per 600-ICM clinic was 5,078 for non-AI-enhanced ICMs and 2,110 for AI-enhanced ICMs, resulting in 559 fewer staffing hours (956 vs 397 hours; 95% CI: 513-605 hours; P value < 0.001) and $29,470 in annual savings ($20,929 vs $50,399; 95% CI: $27,035-$31,904; P value < 0.001).
Compared to non-AI-enhanced ICMs, AI-enhanced ICMs significantly reduce NAAs, leading to a projected decrease in clinic workload and associated costs, potentially improving workflow and health care efficiency.
植入式心脏监测器（ICM）对于心律失常的管理至关重要，但常常会产生大量传输信息和误报。将人工智能（AI）整合到ICM工作流程中可以减轻这种负担。然而，其对临床工作流程和资源利用的影响仍需更好地了解。
本研究旨在评估人工智能增强型ICM对临床工作流程和资源利用的影响。
采用横断面分析，使用来自Octagos Health的真实、匿名的ICM远程监测数据，该数据涵盖了2022年7月至2024年4月期间美国的140家设备诊所。不可操作警报（NAA）被定义为在远程监测平台上传输的虚假或重复警报，但被设备技术人员驳回，未转发给临床医生进行审查。我们比较了人工智能增强型ICM和非人工智能增强型ICM产生的NAA，并估计了管理600名ICM患者的诊所的相关人员工时、资源和成本。
在19320名患者中（平均年龄：69±13.5岁；47.3%为男性），68%使用非人工智能增强型ICM，32%使用人工智能增强型ICM。每600例ICM诊所中，非人工智能增强型ICM的年平均NAA数量为5078条，人工智能增强型ICM为2110条，从而减少了559个人员工时（956小时对397小时；95%置信区间：513-605小时；P值<0.001），每年节省29470美元（20929美元对50399美元；95%置信区间：27035-31904美元；P值<0.001）。
与非人工智能增强型ICM相比，人工智能增强型ICM显著减少了NAA，预计可降低临床工作量和相关成本，可能改善工作流程和医疗保健效率。
