Abstract
Continuous monitoring of heart rate (HR) and respiratory rate (RR) has shown potential to reduce hospital mortality and morbidity. Devices based on ballistocardiography (BCG) can continuously monitor vital signs in a continuous and contactless manner. The BCG-based Dozee VS (part of the Dozee family of devices) from Turtle Shell Technologies Pvt. Ltd. is designed for continuous and contactless measurement of HR, RR, movement, and bed occupancy. This paper focuses on the final field validation of the Dozee VS system. Three clinical investigations, (India 2 & USA 1), were conducted on 120 patients to validate Dozee VS's claims and safety. These studies compared HR and RR from Dozee VS to a US FDA-cleared BCG system, EarlySense 2.0, and a US FDA-cleared patient monitor that employs capnography (EtC02) waveforms to determine RR and electrocardiography (ECG) waveforms to determine HR.
In addition, we compared Dozee and EarlySense's movement and bed-exit detection accuracy to manual observations. Over 14000 HR and RR data points for both Dozee VS and EarlySense were analyzed. Dozee's HR mean absolute error (MAE) was 2.48 beats/minute and RR MAE was 1.67 breaths/minute against the gold standard. For Dozee against EarlySense, the HR MAE was 2.31 beats/minute and RR MAE was 1.42 breaths/minute. Dozee VS had a 94.5% movement detection accuracy and an 88.5% bed-exit detection accuracy, whereas EarlySense had a 98% movement detection accuracy and a 96.5% bed-exit detection accuracy. Thus, in terms of accuracy as well as a comfortable contactless patient monitoring solution, BCG, specifically Dozee, can be considered a clinically viable system.
Introduction
Vital signs are indicators of the body's most fundamental processes and almost all patient-caregiver interactions begin with a check of the patients' vital signs. The ability to continually monitor patients improves patient care beyond typical spot checks and allows hospital staff to provide better care and timely interventions. Trends and patterns in vital signs data give caregivers a better understanding of the patient's condition and aid in clinical decision making [1]. While signs of decline appear quickly, they can start up to 24 hours before a crucial phase. Regular vital indicators are not sensitive enough to pick up these modest alterations. According to The American Journal of Critical Care, changes in heart rate and respiratory rate are the most important indicators of health deterioration.
Traditionally, outside ICU settings, infrequent manual spot checks are carried out to measure vital signs such as HR and RR. These only present a snapshot view of the patient and are not enough to detect early signs of patient deterioration or change in clinical conditions over time. In ICUS, where continuous monitoring needed, Electrocardiography (ECG) has been used for HR measurements [2]. RR, on the other hand, is obtained by measuring airflow with a nasal cannula (Capnography). These techniques are skill and resource intensive as well as require strong patient cooperation, thereby making them unfit for use outside ICU settings [3]. This necessitates the development of monitoring systems that have the ability to remotely and continuously check vital signs. These techniques allow for patient mobility, which may lessen discomfort for the patient and ease nurse workload [4]. Ballistocardiography (BCG) is one such promising method which could help bridge the gap in continuous vitals monitoring. These publications [5],[6],[7] go into considerable length about the factors that have contributed to the current resurgence of BCG, most notably the creation of cutting-edge equipment and signal processing techniques for measurement and analysis. In brief BCG can detect any physiological parameter that causes a motion, including heart beats, breathing, snoring, and limb movements. These devices have the potential to monitor data such as HR, RR, blood pressure, heart rate variability characteristics, sleep quality, and cardiac beat-to-beat intervals.
Dozee VS device developed by Turtle Shell Technologies Pvt. Ltd. uses the BCG technology to provide contactless remote monitoring for vital signs, movement detection, and bed-exit detection. Dozee VS captures micro-vibrations caused by heart contractions, upper respiratory tract activity, and body motions using piezoelectric sensors. HR and RR, are then derived by analyzing these micro-vibrations using a proprietary unsupervised clustering algorithm that is capable of effectively isolating movements from cardiac contractions and respiratory events from raw BCG signals. The vitals can be viewed on the dashboard, exported and even be directly transferred to the in-house hospital management software. The Dozee VS consists of the following sub - units: The sensor sheet, pod, and a power adapter. The sensor sheet is placed below the mattress (Fig. 1.) and captures data while the user is resting or sleeping. The sensor sheet's contactless nature mitigates any dangers connected with direct touch, particularly long-term contact from continuous monitoring. Through a mobile application or a web-based platform, the patient, doctor, caregiver, or user can access the computed data. The range of HR is 30-170 beats/minute (bpm) with an accuracy of ± 3 bpm. The range of RR is 6-45 breaths/minute (brpm) with an accuracy of ± 2 brpm. The movement and bed-exit accuracy are both over 80%.
Before such devices are used in clinical settings, they must undergo several verification and validation stages, as well as final clinical testing to satisfy the claims of the product. The Dozee VS device was subjected to verification and validation activities as well as bench testing, which included the following: Risk Analysis (ISO 14971:2019), Electrical Safety and Electromagnetic Compatibility (IEC 60601-1 and IEC 60601-1-2), Environmental Conditions Testing and Performance Bench Testing. To ensure data and information security, the organization is also ISO/IEC 27001 compliant. The final step before a medical device is used commercially is to validate its performance in a clinical setting. Hence, clinical studies have been conducted [6][7][8] for HR,RR, movement detection and bed-exit detection by contrasting these with those procured from applicable FDA-cleared devices. The FDA-cleared EarlySense 2.0 System (K131379) developed by EarlySense Ltd utilizes a similar BCG technique and serves as a predicate for Dozee. The EarlySense 2.0 System is designed to automatically and contactlessly measure HR, RR, movement and bed-exit in a home, hospital, or clinic setting. The system is indicated for use in children, adolescents and adults. The operation of the EarlySense has been studied in children (weight 2 10 Kg) and adults (weight <111 Kg) during sleep and resting conditions. The range for RR is 6-45 brpm and the range for HR is 30-170 bpm The accuracy for RR is 1.5 brpm and the accuracy of HR 5 bpm.
This paper focuses on the clinical performance validation of the Dozee VS device in comparison to its FDA-cleared predicate device as well the gold standard techniques for continuous HR and RR monitoring, movement detection accuracy and bed-exit detection accuracy. The goal of these investigations was to validate the Dozee's claims for HR, RR, movement detection accuracy, and bed-exit detection accuracy. These studies were carried out on healthy volunteers as well as patients at three sites: Eastside Research Associates in Seattle, USA, Sparsh Hospital in Bangalore, India, and Apollo Hospital in Jubilee Hills, Hyderabad.