Real-World Diagnostic Performance for Heart Failure Events of TriageHF^® Algorithm

Area of Science:

• Cardiology
• Medical Technology
• Health Informatics

Background:

• Remote monitoring programs for heart failure (HF) leverage data from cardiac implantable electronic devices for early intervention.
• These programs aim to reduce healthcare burdens by proactively managing HF patients.
• Characterizing HF patient profiles, alert burden, and alert associations with decompensation is crucial.

Purpose of the Study:

• To characterize heart failure patients within a remote monitoring program.
• To analyze the burden and characteristics of alerts generated by the TriageHF® algorithm.
• To determine the association between TriageHF® alerts and heart failure decompensation events.

Main Methods:

• A retrospective observational study involving 134 patients in an outpatient HF remote monitoring program.
• Data collected included alert frequency, alert-driving data, and alert duration from February 2022 to July 2023.
• Clinical outcomes comprised unplanned ER visits, HF hospitalizations, all-cause death, and a composite endpoint.

Main Results:

• Over 1.7 years, 77 high-risk alerts occurred (0.44 alerts/patient-year).
• Patients with alerts showed higher rates of chronic kidney disease, stroke history, and furosemide use.
• High-risk days had a significantly higher event rate (7.9%/month) compared to standard-risk days (0.8%/month).
• TriageHF® alerts showed 35% sensitivity and 71% specificity for predicting adverse events.

Conclusions:

• The TriageHF® algorithm effectively identifies periods of increased HF-related adverse events.
• This capability supports efficient resource allocation in HF management.
• The low alert burden indicates the algorithm's feasibility, even in resource-limited settings.