Interim analysis of the HF-TRACK multicenter crossover RCT in heart failure
T Castiello, P SalahshouriAbstract
Background
Heart failure (HF) remains a leading cause of hospitalisation and mortality, with poor adherence to self-monitoring contributing to adverse outcomes.(1) The HF-TRACK trial is a randomised controlled trial designed to assess the effectiveness of an artificial intelligence (AI)-driven remote patient monitoring device in tracking peripheral oedema(2) to reduce HF-related hospitalisations. The primary objective is to evaluate the device’s impact on unscheduled hospitalisations, with secondary objectives including data availability and acceptability.
Methods
Seventy-eight patients (median age 77.3 years, 45% female) were recruited from general practice (72%), pharmacies (26%), and hospitals (2%). Patients were randomly assigned to alternating periods of standard care and AI-based monitoring in a crossover design. The primary endpoints were HF-related hospitalisation events, mortality and device-related complications. Secondary outcomes included all-cause hospitalisations and comparisons of data availability between AI monitoring and conventional weighing scales. This preliminary analysis presents safety and efficacy data collected during the first six months of the study.
Results
Three HF-related hospitalisations were recorded in the control arm (0.29 per patient-year) none in the AI-driven device arm. The total number of all-cause hospitalisation events was seven in the control arm (0.48 per patient-year) and six in the AI-device arm (0.39 per patient-year). Two deaths occurred in the control group, with no HF-related deaths recorded in either arm. No device-related complications were recorded in either arm. The AI device demonstrated substantial superiority in data availability, with a median of 5.7 monitoring days per week compared to 0.3 days for conventional weighing scales.
Conclusion
Preliminary findings suggest that the AI-driven monitoring device is safe and does not raise efficacy concerns. The device significantly improved data availability, addressing a key limitation of traditional self-monitoring methods.(3) Further analysis with larger cohorts and extended follow-up periods may provide clearer insights into the potential of AI-based monitoring to enhance clinical outcomes and healthcare efficiency in HF management.For image description, please refer to the figure legend and surrounding text.For image description, please refer to the figure legend and surrounding text.