A mixed-methods assessment of a multi-sensor wearable band during normoxia and hypoxia

Wearable devices in the sport and healthcare industries have grown exponentially in recent years as we move towards a personal measurement approach. Whilst this technology is not new, hardware that incorporates multiple sensor measures are limited. The PROTOTYPE used here aims to monitor up to seven “vital physiological signs” including heart rate (HR) and oxygen saturation (SpO2). Despite this, the validity of the measurements, and the usability of the device by practitioners, are currently unknown. The purpose of this study was to assess the validity of the PROTOTYPE to measure two of the key variables (HR & SpO2) at rest in normoxia and hypoxia. The usability of the PROTOTYPE by practitioners was also evaluated. Following ethical approval, 44 healthy participants (31 ± 9 years) completed one experimental trial (VAL) and 11 healthcare professionals completed one trial in a simulated ward (USE). During VAL, the PROTOTYPE was secured laterally on the upper-left arm and output from device sensors were recorded continuously using bespoke software. Reference gold standard (GS) outputs were recorded simultaneously (HR = Polar H7, SpO2 = Masimo Pulse Oximeter). Participants then completed two experimental conditions (T1, T2). During T1, participants lay supine for 10 min in normoxia. During T2 participants were exposed to 5 min Hypoxia (Altium i10). Data were filtered to exclude data affected by movement and random artifacts. Concordance correlation coefficient (CCC), Bland and Altman plots and Wilcoxon rank test assessed the levels of agreement between the PROTOTYPE and the GS. During USE, participants were provided with video and written user guides before initiating the device on a participant. Sessions were video-recorded and participants completed a “user questionnaire” and focus group (n = 5). VAL highlighted substantial agreement between GS and PROTOTYPE variables (HR: CCC = 0.88, SPO2: CCC = 0.68). PROTOTYPE slightly overestimated HR by 2 beats·min−1 (95% CI = –8 – 12 bpm). The mean bias for SpO2 approached zero (1%, 95% CI = –4 – 6 %). Evaluation of the questionnaire and focus group suggests a positive level of agreement between the experience and usability of the device, and it was concluded that the device is functional, usable and could be of major benefit to practice. These findings indicate that the PROTOTYPE device is capable of measuring HR and SpO2 within reasonable limits of GS measures at rest and appears functional in practice. Despite this, algorithms should be enhanced to improve the power for individual agreement.