Abstract 10571: Triage HF Plus: Predicting, Preventing and Managing Heart Failure Episodes in Ambulatory Patients With a Cardiac Implantable Electronic Device Using Device Based Diagnostics and Automated Carelink Alerts

IntroductionChanges in physiological parameters, such as thoracic impedance, arrhythmias, pacing percentage, diurnal heart rate variability and patient activity may precede clinical decompensation in HF. In retrospective analyses, the Medtronic ‘Heart Failure Risk Score (HFRS), which uses the input from the integrated device diagnostics to detect changes in key physiological parameters, has been shown to identify individuals increased risk of heart failure (HF) event in the next 30 days. In the current analysis we evaluated the utility and predictive accuracy of the HFRS for the a priori identification of patients at increased risk of a HF episode in a prospective real world evaluation.Hypothesis:Carelink HFRS alerts can be used to identify the ambulatory HF patient who may benefit from specialist review.Methods640 patients with HFRS enabled cardiac implantable electronic devices had Carelink alerts enabled. HFRS results stratify patients as low-, medium- or high-risk of a HF event in the next 30 days. High HFRS results automatically alert clinical staff (via Medtronic Carelink) who telephone the patient with screening questions for clinical indicators of HF decompensation. If positive, the need for medication changes or additional clinical review was recorded.Results87 (14%) patients had high-risk HFRS results. At telephone triage, 55 (63%) were confirmed to have HF decompensation requiring medication changes or additional clinical review; 6 (7%) were determined to have an alternative acute medical problem requiring medical attention; and 26 (30%) did not report worsening HF symptoms. The positive predictive value (PPV) of HFRS to predict HF decompensation was 63% as a standalone test; and 90% when combined with telephone triage. The sensitivity, specificity and overall diagnostic accuracy of the combined tool were 100%, 81% and 93% respectively.ConclusionsRecent advances in CIED artificial intelligence systems offer potential to monitor the ambulatory HF population for changes in physiological parameters, which may signal a decompensation.