Study Objective: Evaluating the wearable Biobeat sensor which measure hemodynamic and respiratory main variables and temperature.

Design: Prospective multicenter observational study without human intervention.

Setting: Private non-profit hospitals.

Patients: Patients having had abdominal surgery.

Interventions: Carrying the precordial sensor for 72hr whose information was transmitted to a cloud-based repository through a Wi-Fi router.

Measurements: Incidences of total or partial lack of measurement, artifacts, hemodynamic (particularly hypotension), respiratory and temperature abnormalities. Sensor measurements were compared with nurses‘ ones using Bland and Altman analysis and Clarke-Error grid analysis including five zones, from A (excellent concordance) to E (opposite treatment).

Results: 114 patients were included, 90 were analyzed. The median duration of recording per patient was 63.8 hrs [25-75^thpercentiles: 42.6-72.3]. Overall complete data loss was important (45% of time), and, when data was available, pulse oximetry was missing the most often (16.1% of the remaining time). Artefacts were infrequent (0.1%). Biobeat patch detected a severe hypotension in 10.0% of the patients vs 2.9% according to nurse follow up (p 0.034). The bias between Biobeat and nurses measurements was small but with large limits of agreement. However, most Biobeat’s measurements were in zone A except for temperature (53.4% in zone B) and respiratory rate (55.6% in zone B and 4.4% in zone D).

Conclusions: Our findings suggest that wireless monitoring using Biobeat sensor is reliable, provided that the problems associated with data transmission are resolved.

Biobeat measurements were stored automatically on a cloud web platform (Instamed) and retrieved at the end of the study. 

Patients' demographics and nurses’ measurements were retrieved from the hospital database.