ABSTRACT
RATIONALE: Reliable remote respiratory monitoring that is acceptable to patients is crucial for healthcare systems and clinical research in the post-COVID-19 era. Existing methods (such as peak-flow, diaries, and pulse oximetry) are limited by adherence and technique or confounded by subjectivity and recall bias. Nocturnal periods provide important signs in respiratory disease activity, yet accurate and unobtrusive methods for home monitoring are lacking. Current gold-standard tools - wearable polysomnography (PSG) devices - capture objective signs such as respiratory rate (RR) but require uncomfortable sensors, which are unsuitable for use beyond a few nights. Emerging monitoring solutions must minimize patient-burden to facilitate long-term engagement in clinical care and research. METHODS: In healthy adults, we evaluated accuracy of a passive, non-contact bedside device (Albus Home RD), that uses wireless motion sensors to capture RR, compared to gold-standard, wearable PSG (SOMNOtouch™ RESPIRATORY, Somnomedics). The table-top Albus Home RD was positioned on the bedside adjacent to the participant in their normal home bedroom environment. Participants slept with usual clothing and bedding;sleeping arrangements ranged from single- to king-size beds with single- and co-sleepers. Gold-standard PSG RR data were recorded using manual count of the raw respiratory traces derived from thoracoabdominal respiratory-effort belts. 10-minute periods from each hour of monitoring were chosen, where sufficient data were available and free from confounding movement and artefacts. Data from Albus Home RD were then analyzed using proprietary signal processing algorithms to output corresponding 30-second RR segments (as breaths/minute). RR results for each device for the selected periods were time-synchronized and compared for each 30-second segment. As per previous validation literature, accuracy was reported as proportion of RR measurements within +/-10% or +/-2 breaths/minute of the PSG RR. RESULTS: 16 healthy adults (9 males, 7 females) participated in overnight monitoring;ages and BMI ranged 20-74 years and 19-38 respectively. Albus Home RD RR measurements for 1540 thirty-second segments were compared against the gold-standard with overall accuracy of 92.4%. Mean Absolute Percentage Error was 0.06 (SD=0.07). CONCLUSIONS: Albus Home RD passively measured RR with 92% accuracy in adults compared to gold-standard in 770 minutes of analysis. Using wireless sensors and proprietary signal processing algorithms, the Albus Home RD is a valid bedside, non-contact monitor of RR in real-world environments. The non-touch, passive nature of this monitor can enable low-burden, long-term home nocturnal monitoring. This system provides new possibilities for remote clinical care and objective data gathering in longitudinal research studies. .