TL;DR
This paper validates Albus Home, a novel contactless, automated bedside system for nocturnal respiratory rate (RR) monitoring. The system demonstrates high accuracy (over 98%) compared to gold-standard polygraphy in real-world home environments, offering a less burdensome alternative for continuous patient monitoring.
What Problem Does This Paper Address?
Accurate and continuous monitoring of respiratory rate (RR) is crucial for detecting cardio-respiratory deterioration, yet current methods are problematic. Manual RR counting is time-consuming, prone to variability between clinicians, and only provides sporadic readings. Existing automated solutions often involve uncomfortable wearable sensors with poor adherence, or are validated only in controlled laboratory settings, not reflecting the complexities of real-world home environments. This leaves a gap for a non-intrusive, automated, and validated system for long-term nocturnal RR monitoring at home.
Methods
The study validates the Albus Home™ system, a bedside device that uses a radiofrequency radar-based motion sensor to detect chest and abdominal movements indicative of respiration. This contactless approach eliminates the need for patients to wear any devices. The system's RR readings are processed by proprietary software.
For validation, participants simultaneously underwent overnight monitoring with the Albus Home™ and a gold-standard polysomnography device equipped with thoraco-abdominal respiratory effort belts. Reference RR readings were obtained by manual clinician counting of the polysomnography data. Accuracy was defined as the percentage of Albus RR readings within ±2 breaths/minute of the reference counts. Correlation and agreement were assessed using Pearson's correlation coefficient and Bland-Altman analysis.
Datasets
The study utilized data from 32 participants (18 healthy volunteers and 14 with chronic respiratory conditions, including adults and children) who underwent simultaneous overnight monitoring at home. This dataset represents a real-world validation scenario, distinct from controlled laboratory settings.
Key Insights and Ideas
- Contactless Monitoring is Feasible and Accurate: The Albus Home™ system proves that accurate RR monitoring can be achieved without any physical contact with the patient, addressing a major barrier to adherence with wearable sensors.
- Real-World Validation is Crucial: Evaluating the system in participants' actual home environments, with their usual sleeping conditions, bedding, and clothing, provides a more robust assessment of performance than laboratory studies.
- Automated Analysis Reduces Burden: The system's automated analysis of data eliminates the time-intensive and error-prone manual scoring of polygraphy data, making long-term monitoring practical.
- Nocturnal RR as a Key Indicator: The study reinforces the importance of nocturnal RR as a sensitive indicator of health status and potential deterioration, especially in chronic respiratory conditions.
Key Results
- High Accuracy: The Albus system achieved a mean accuracy exceeding 98% across all participants, defined as readings within ±2 breaths/minute of the reference.
- Strong Agreement: Bland-Altman analysis showed excellent agreement between the Albus system and reference, with a mean difference of -0.4 breaths/minute and 95% limits of agreement of ±1.2 breaths/minute.
- Robust Performance: Accuracy was maintained across diverse participant characteristics, including age, body mass index, sex, and bedroom occupancy (single sleepers vs. bed-sharers).
- High Data Availability: The Albus system reported an RR result 97.7% of the time when a reference RR was available.
- Hourly Mean RR: For a subset of participants, hourly mean RR readings also showed strong correlation and close agreement with reference data.
Limitations and Future Work
- Clinical Significance Not Assessed: The study focused on the accuracy of RR measurement, not the clinical significance of the measured rates or patterns.
- Participant Cohort: Participants were required to be stable enough for home monitoring, meaning the study did not capture extreme RR variations seen in acutely unwell patients in hospital.
- Single Night Monitoring: The validation was conducted over a single night per participant, limiting the assessment of repeatability and reliability over longer periods. Future work is planned for longitudinal studies.
- Absence of Respiration Detection: While the system captures RR, the detection of apnoeas (cessation of breathing) was not the focus of this study but is considered feasible with the deployed sensors and a subject for future work.
- Thoracic/Abdominal Movement vs. Airflow: The system infers RR from chest/abdomen movement, which may not perfectly align with direct airflow in rare cases of obstruction.
- Future Work: The authors suggest future research to understand the variability of nocturnal RR, define minimum clinically important differences (MCIDs), and explore the system's potential for detecting apnoeas and other nocturnal metrics. They also plan to deploy the system in longitudinal studies to assess performance and reliability over extended periods.
Summary updated