The Foot Study, Observational Retrospective Multi-center Clinical Trial Evaluating Alerts From The Heartfelt Device

• Published in: Journal of cardiac failure 2024-01, Vol.30 (1), p.182-183
• Main Authors: Dewhurst, Matthew, Hann, Karen, Matthews, Iain, Gallagher, Lisa, McCafferty, Gemma, McKie, Hayley, Pick, Sara, Hughes, Debbie, Clayton, Louise, Nicolson, William, Gent, Susan, Kumar, Ganesan, Keeling, Philip
• Format: Article
• Language: English
• ISSN: 1071-9164; 1532-8414
• DOI: 10.1016/j.cardfail.2023.10.160

The Heartfelt device is a recently developed in-home remote patient monitoring device for heart failure patients, utilizing 3D imaging and AI processing to measure foot and lower leg volume, and thus quantitative changes in peripheral edema. The device autonomously recognizes and measures patients as they go about their daily life, without the patient having to do anything, wear anything, or remember anything (Figure 1). Alerts are automatically raised with the patient, relatives, or care team. The Foot Oedema Observed Over Time (FOOT) Study (NCT04072744) was designed to assess the effectiveness of these alerts in predicting clinically important heart failure events. Our hypothesis is that the Heartfelt device can directly detect the increase in peripheral edema associated with heart failure decompensation and that, on average, the number of days with missing data collected by the Heartfelt device will be lower than that of the weighing scales. The FOOT study recruited 31 patients from 5 hospital trusts in the UK from February 2020 to June 2022. Of these 31 patients, 26 had the Heartfelt HF3 device installed in their home. The devices were operated normally, with data and alerts logged rather than being raised to patients, families or the care teams, so as not to influence outcomes. After a 6 month follow-up period, clinical notes were used to identify deaths, non-routine outpatient or inpatient appointments related to heart failure. Patients were also provided Bluetooth-connected weighing scales to collect daily weight measurements as usual care. Of 26 patients with the Heartfelt device installed, there were 9 HF events recorded across 8 patients. These were 2 outpatient events, 6 hospitalizations, and 1 death. One patient was hospitalized twice during the period. The Heartfelt device provided alerts within the two weeks prior to HF events for 5 hospitalizations, and 1 death. The device had raised alerts in the two weeks prior to both outpatient events, however in both cases the devices had been persistently alerting approximately weekly for several weeks prior, potentially due to excessively sensitive thresholds, so the outpatient events were counted as missed. Therefore 6 of 9 HF events are judged to have been usefully predicted in advance. Weighing scale data of sufficient regularity to interpret was unavailable for most patients due to adherence issues, however, 1 hospitalization and potentially the death were predictable based on weighing scale data.