Can AI-led remote ECG monitoring result in service improvement for cardiac investigations

Type of funding sources: Private company. Main funding source(s): iRhytm

Ambulatory ECG monitoring is needed to investigate intermittent heart rhythm symptoms. Traditional monitoring approaches are resource intensive, requiring multiple patient hospital visits and time-consuming analysis, using fixed capacity monitoring equipment.

We carried out a 4000-patient pilot with an AI-led remote ECG monitoring service (Zio XT) to determine its impact on patients and the service. Single-use monitoring patches were sent directly to the patient to fit at home or fitted in outpatients. Patients wore the patch for 14 days of continuous monitoring and then returned it by Freepost for outsourced analysis and reporting.

We undertook a retrospective analysis of 200 consecutive patients attending for standard Holter monitoring, and compared this cohort with 200 consecutive patients who underwent monitoring with Zio. The Holter sample was taken from a time period prior to the introduction of the new remote service. Patients had been referred to cardiology for investigation of heart rhythm symptoms.

The age ranges in the two groups were broadly comparable. In the Holter group, 100% of patients required a separate outpatient appointment to fit the monitor compared with 0% in the Zio group. Median time from monitor fitting to result was 18 days in the Zio group compared with 106 days in the Holter group (p<0.001). Test results were unavailable at follow-up in 16% of the Holter group compared with 2% of the Zio group (p<0.001). Arrhythmias were detected in more patients in the Zio group vs the Holter group (p<0.05), (SVT ≥ 30s 7.4% vs 0.5%; VT ≥ 8bts 10.5% vs 2.0%; Pause ≥ 3s 6.3% vs 2.5%; PAF 7.4% vs 2.5%). In the Holter group, 24 patients (12%) had no follow-up in cardiology outpatients. Of these, 9 were lost to follow-up and 3 deceased before completing the pathway. By comparison, 3 patients (1.6%) in the Zio group had no follow-up. In patients who had a follow-up appointment, 75% of patients in the Zio group had a non-face-to-face follow-up vs 48% in the Holter group (p<0.05). Repeat test was required in 11.5% of the Holter group compared with 4.7% of the Zio group (p<0.05). In the Holter group, 6 (3%) of the repeat tests were ILRs, compared with 2 (1.1%) in the Zio group.

Patients in the Zio group received their test results more quickly. Use of Zio XT resulted in fewer visits to hospital and a reduced need for repeat testing. Earlier detection of significant clinical arrhythmia, especially AF, facilitates rapid clinical decision making, and is expected to result in more patients receiving appropriate therapy.