Clinical Validation of Non-Contact Vital Signs

Overview

Vital signs are critically important to modern emergency medical care. These metrics (heart rate (HR), respiratory rate (RR), blood pressure, temperature, and pulse oximetry) have been shown to predict a patient's morbidity, mortality, need for ICU-level admission, and suitability for discharge. As such, vital signs are used at multiple points of emergency department (ED) care, from triage through treatment through disposition. Unfortunately, the acquisition of vital signs can be a time-intensive process. Time-and-motion studies have shown the measurement and recording of vital signs may take over five minutes per patient and longer with interruptions.5 Increased crowding, lack of available staff, and length-of-stay have all been shown to delay or decrease the frequency of vital sign measurements. Some studies have shown a majority of both children and adults may be missing vital signs during their ED stays. Other studies have shown repeat vital signs to be taken less frequently every two hours. With increasing boarding and crowding in ED settings, the frequency of vital sign acquisition may be significantly worse than this reported data. Many emergency departments have sought to improve vital sign acquisition through various quality improvement initiatives. Some researchers have also evaluated alternative means of vital sign acquisition. Telemedical providers, for instance, have evaluated the accuracy of patients' self-reported vital signs. Device studies have also looked at non-contact vital signs via radar and thermal imaging. More recently, studies have evaluated the feasibility of acquiring vital signs using cameras. These studies of camera-acquired vital signs have shown promise, though they are limited by small sample sizes. Here, we present the clinical validation of camera-acquired vital signs obtained from commercially available products over a larger sample size.

Vitals signs are a key component of modern emergency medical assessment. They are acquired at the initiation of most emergency department ("ED") visits and inform triage, diagnostic workup, and therapeutic care. Unfortunately, vital signs are often resource-intensive. Acquisition of vital signs may take a provider several minutes. Accuracy of vital signs may be affected by the noise and other environmental factors within the ED. Decreased frequency of vital sign acquisition affects ED care in two separate ways. First, patient care is directly affected. Increased frequency of vital signs would allow for better re-evaluation of a patient's condition and a given therapy's effect. Conversely, failure to acknowledge abnormal discharge vital signs has been correlated to worse outcomes in both adult and pediatric patient populations. Secondly, ED operations are also affected. Vital signs are often a key component of disposition. Patients with abnormal vital signs are often held in the ED until their vital signs improve. Patients with more abnormal vital signs may similarly be held for ICU evaluation instead of floor-level admission. Delayed vital signs thus impede patient disposition and increase ED crowding. Of note, ED crowding has also been shown to delay vital sign acquisition, itself. Presage Technologies has developed software which can acquire heart rate and respiratory rate non-invasively using commonly available cameras. This technology can theoretically allow for decreased time to vital sign acquisition on patient arrival, increased frequency of vital sign reassessment during patient stays, increased patient safety as the device is non-contact, and decreased time to vital sign acquisition at time of discharge. It may also allow for increased vital sign acquisition for agitated and other potentially dangerous patient encounters. These benefits could theoretically allow for better single-patient care and ED operations, ultimately improving patient outcomes. Presage's SmartSpectra Vital Signs Monitor is a software only medical device ("SaMD") which uses optical camera-based measurements for estimation of pulse rate and breathing rate. The SmartSpectra device includes a preprocessing client that is a software development kit ("SDK") installed into a host software application and includes algorithms to analyze video signals and output pulse rate and breathing rate measurement results. The SmartSpectra Vital Signs Monitor is intended for non-invasive spot measurement of pulse rate and breathing rate for humans aged 18 and older of all skin types, who do not require critical care or continuous vital signs monitoring. The device assesses imaging of an individual's face and chest wall movements after installation on an approved, statically mounted mobile platform - either the Apple iPhone 16 Pro or a Samsung Galaxy S24e. The SmartSpectra device is intended for prescription only use ("POU") in hospitals, general care, and secured environments, from a statically mounted position installed by approved Presage Technologies personnel, where a single human face is visible in the frame of the camera and where a framework exists for periodic checks by a trained professional to ensure subject safety. It is not intended for use in an at-home environment. The SmartSpectra device is intended for use by appropriately trained individuals and should not be used by untrained users. The SmartSpectra device is not intended for use in transportation or domestic (at-home) environments. The device is not intended to be the sole method of checking the physical health of a subject. Additionally, it is not meant for continuous monitoring or for over-the-counter use. CLINICAL STUDY The objective of the study was to evaluate the safety and effectiveness of the SmartSpectra Vital Signs Monitor version 1.0.0 vis-a-vis standard-of-care, FDA-cleared Philips Intellivue MX450 patient monitor device (FDA Clearance K130849)("MX450"). The MX450 measures end-tidal CO2 ("EtCO2") and three-lead electrocardiogram ("ECG") waveforms. This study compared the SmartSpectra Vital Signs Monitor pulse rate and breath rate acquired from two mobile devices (iOS and Android systems), as well as from one laptop + USB camera, against the MX450 pulse rate and breath rate data. The FDA approved MX450 measures heart rate and respiratory rate (clinician scored electrocardiogram and capnography, respectively) within a 30 second spot check period. A concurrent 30 second spot check period was used for each of the iOS mobile device, Android mobile device and MX450 device. A third party validation of rate measurements will provide data for FDA to use in evaluation of SmartSpectra for clearance as a class II SaMD. The secondary objective of this study was to ensure that the safety, effectiveness, and performance of the SmartSpectra device was consistent across demographic subgroups of skin color, age, sex, body mass index ("BMI"), and device types. Study Design A study design was chosen to directly compare algorithm-estimated vital signs from SmartSpectra with concurrently collected clinician-scored reference data under standardized conditions, reducing inter-subject variability. By using each subject as their own control, the design maximizes statistical power and minimizes confounding variables without requiring external control groups. The ECG and etCO2 systems serve as clinical gold standards for heart rate and respiration rate, respectively. No placebo or inactive comparator is included, as the goal is performance validation against these accepted references rather than clinical outcome measurement. The stratified subgroup analysis evaluated any performance degradation or demographic bias in a real-world, intended use environment. Selection of Study Subjects Baylor College of Medicine ("BCM") conducted a non-randomized enrollment to ensure capture of statistically significant sample sizes of target subgroup population. All study related activities took place in a clinical room at Harris County Hospital District ("HCHD") Ben Taub Emergency Department. Initial enrollment criteria was based on convenience sampling in order to target the observable demographic subgroups. Therefore, relatively healthy patients, visitors, or employees located in the Ben Taub Emergency Department of HCHD were requested to participate voluntarily. Potential subjects were screened based on exclusion/inclusion criteria. Subject decision making capacity was determined by the research staff approaching the subject. No one was approached for participation while undergoing medical treatment, and a subject must be oriented to person, place, time to participate. Other than decision making capacity, inclusion criteria was patients, visitors or employees greater than 18 years of age and less than 75 years old. Exclusion criteria, other than lack of decision making capacity, were patients requiring immediate medical care, persons with facial scarring, age less than 18 years old, age greater than 75 years old, employees of Presage Technologies, self-reported pregnancy and anyone who does not pass the screening questionnaire. Removal of Subjects from Study It was intended that subjects requiring immediate medical attention or those withdrawing consent would be removed from the study. Zero persons who were assigned a unique identifier number were removed from the study due to requiring immediate medical attention or upon withdrawing consent. No subjects required medical treatment in the course of the study. Data from subject 56 was excluded after review. Data appears to have been recorded for this patient twice, with differing baseline vital signs between the measurements. The data recorded in this row is a mixture of the two measurements and appears to be a clerical error. The cause of this is unknown. Study Methodology This study is a prospective, non-randomized, within-subject observational evaluation of a contactless vital signs monitoring software (Presage Technologies SmartSpectra Vital Sign Monitor). The software under investigation was used in an iPhone, an Android as well as a laptop with a USB webcam (described interchangeably as C++) to estimate pulse rate and respiratory rate. These estimates are compared to reference standards: electrocardiogram (ECG) for pulse and capnography (etCO2) for respiration; the reference standards measured by a MX450 device. Each subject underwent a protocol-defined measurement session lasting approximately 10 minutes in a clinical setting. The study assessed accuracy, precision, and failure modes across a demographically diverse sample, stratified by age, sex, skin tone phenotype (Fitzpatrick scale), and Body Mass Index ("BMI"). Data collection took place in a controlled clinical setting and involves no investigational drug or invasive procedure. Subjects serve as their own control in all comparisons.