Use of an Artificial Intelligence Algorithm to Increase Productivity in Implantable Loop Recorder Monitoring: A Multicentre Observational Study.

BACKGROUND

Insertable loop recorder (ILR) services are increasingly constrained by high volumes of transmitted episodes, many of which are false, clinically irrelevant, or non-actionable. Cloud-based artificial intelligence (AI) algorithms have the potential to suppress false atrial fibrillation (AF) and pause alerts while preserving clinically relevant events. We present the first real-world impact of an AI algorithm activated across a fixed multicentre ILR cohort.

METHODS

We performed a retrospective, multicentre before-and-after cohort analysis of consecutive patients. Eligible patients had continuous ILR monitoring for 12 months before and 12 months after service-wide activation ("switch-on") of the Medtronic AccuRhythm AI platform (Medtronic plc, Galway, Ireland), enabling within-patient comparison. All clinician-facing transmitted alerts were counted in each period. Secondary analyses examined the concentration of alert burden across patients and estimated workflow impact using published time-and-motion data for remote transmission review. Differences in alert counts were tested using paired t-tests, reporting mean paired differences with 95% confidence intervals (CI).

Results

The cohort included 445 patients (Reveal LINQ n=438; LINQ II n=7); 440 (99%) were implanted for syncope (mean age 67±15 years; 49.6% male). Total transmitted alert volume fell from 4,261 pre-AI to 2,509 post-AI (29% reduction). The mean paired within-patient change was -3.94 alerts (95% CI -7.5 to -0.4; p<0.05). Alerts were highly concentrated: 7% of patients generated >90% of alerts, and 104 patients had intermittent disconnection from remote monitoring. Applying established workflow timings (11-13 minutes per remote transmission review) translated the reduction into 185-218 hours of physiologist review time released annually: approximately four hours per week of "virtual physiologist" capacity.

CONCLUSIONS

In routine practice, AI activation in a multicentre observational study was associated with a statistically significant reduction in ILR alert burden and a clinically meaningful release of staff capacity. Parallel management of high-alert patients and connectivity optimisation may further amplify the operational benefit.