prospective validation of smartphone based heart rate and respiratory rate measurement CORD-Papers-2021-10-25 (Version 1)

Title: Prospective validation of smartphone-based heart rate and respiratory rate measurement algorithms
Abstract: Measuring vital signs plays a key role in both patient care and wellness, but can be challenging outside of medical settings due to the lack of specialized equipment. In this study, we prospectively evaluated smartphone camera-based techniques for measuring heart rate (HR) and respiratory rate (RR) for consumer wellness use. HR was measured by placing the finger over the rear-facing camera, while RR was measured via a video of the participants sitting still in front of the front-facing camera. In the HR study of 95 participants (with a protocol that included both measurements at rest and post exercise), the mean absolute percent error (MAPE) {+/-} standard deviation of the measurement was 1.6% {+/-} 4.3%, which was significantly lower than the pre-specified goal of 5%. No significant differences in the MAPE were present across colorimeter-measured skin-tone subgroups: 1.8% {+/-} 4.5% for very light to intermediate, 1.3% {+/-} 3.3% for tan and brown, and 1.8% {+/-} 4.9% for dark. In the RR study of 50 participants, the mean absolute error (MAE) was 0.78 {+/-} 0.61 breaths/min, which was significantly lower than the pre-specified goal of 3 breath/min. The MAE was low in both healthy participants (0.70 {+/-} 0.67 breaths/min), and participants with chronic respiratory conditions (0.80 {+/-} 0.60 breaths/min). Our results validate that smartphone camera-based techniques can accurately measure HR and RR across a range of pre-defined subgroups.
Published: 3/12/2021
DOI: 10.1101/2021.03.08.21252408
DOI_URL: http://doi.org/10.1101/2021.03.08.21252408
Author Name: Bae, S
Author link: https://covid19-data.nist.gov/pid/rest/local/author/bae_s
Author Name: Borac, S
Author link: https://covid19-data.nist.gov/pid/rest/local/author/borac_s
Author Name: Emre, Y
Author link: https://covid19-data.nist.gov/pid/rest/local/author/emre_y
Author Name: Wang, J
Author link: https://covid19-data.nist.gov/pid/rest/local/author/wang_j
Author Name: Wu, J
Author link: https://covid19-data.nist.gov/pid/rest/local/author/wu_j
Author Name: Kashyap, M
Author link: https://covid19-data.nist.gov/pid/rest/local/author/kashyap_m
Author Name: Kang, S H
Author link: https://covid19-data.nist.gov/pid/rest/local/author/kang_s_h
Author Name: Chen, L
Author link: https://covid19-data.nist.gov/pid/rest/local/author/chen_l
Author Name: Moran, M
Author link: https://covid19-data.nist.gov/pid/rest/local/author/moran_m
Author Name: Cannon, J
Author link: https://covid19-data.nist.gov/pid/rest/local/author/cannon_j
Author Name: Teasley, E S
Author link: https://covid19-data.nist.gov/pid/rest/local/author/teasley_e_s
Author Name: Chai, A
Author link: https://covid19-data.nist.gov/pid/rest/local/author/chai_a
Author Name: Liu, Y
Author link: https://covid19-data.nist.gov/pid/rest/local/author/liu_y
Author Name: Wadhwa, N
Author link: https://covid19-data.nist.gov/pid/rest/local/author/wadhwa_n
Author Name: Krainin, M
Author link: https://covid19-data.nist.gov/pid/rest/local/author/krainin_m
Author Name: Rubinstein, M
Author link: https://covid19-data.nist.gov/pid/rest/local/author/rubinstein_m
Author Name: Maciel, A
Author link: https://covid19-data.nist.gov/pid/rest/local/author/maciel_a
Author Name: McConnell, M V
Author link: https://covid19-data.nist.gov/pid/rest/local/author/mcconnell_m_v
Author Name: Patel, S
Author link: https://covid19-data.nist.gov/pid/rest/local/author/patel_s
Author Name: Corrado, G S
Author link: https://covid19-data.nist.gov/pid/rest/local/author/corrado_g_s
Author Name: Taylor, J A
Author link: https://covid19-data.nist.gov/pid/rest/local/author/taylor_j_a
Author Name: Zhan, J
Author link: https://covid19-data.nist.gov/pid/rest/local/author/zhan_j
Author Name: Po, J
Author link: https://covid19-data.nist.gov/pid/rest/local/author/po_j
sha: 79a5d77a33aa904f900f1a12ddd262c0e5dbbdd4
license: medrxiv
source_x: MedRxiv
url: http://medrxiv.org/cgi/content/short/2021.03.08.21252408v1?rss=1
has_full_text: TRUE
Keywords Extracted from Text Content: tan brown participants patient heart colorimeter-measured skin-tone ± 2-3 B. Left neck Funder/Sponsor medRxiv preprint Figure 1 medRxiv pulmonary Self-identified LLC https://doi.org/10.1101/2021.03.08.21252408 doi upper body Supplementary Table 5 Supplementary Table 1B melanin's San Diego skin-tone medRxiv preprint Pantone Capsure color matcher colorimeter Shcherbina medRxiv preprint Figure 3 wrist shoulder line tan https://doi.org/10.1101/2021.03.08.21252408 brown atrial medRxiv preprint Supplementary Figure 2 participants body CA red lines 39,41 left Skin ±2 extract heart medRxiv preprint Supplementary Table 4 p=0.70 medRxiv preprint Figure 2 COVID-19 ± blood PPG-based HR medRxiv preprint A B Figure 3 GA RGB in-app Supplementary Figure 1 blood pressure-is Heart A. ROI PPGs Artemis patients people Nashua, NH sinusoids cardiovascular health ventricular Food upper torso per-channel Irvine B Bandpass-filtering quadrants blue lines patient skin extract heart signs-HR direct-to- fingertip cheek skin
Extracted Text Content in Record: First 5000 Characters:Measuring vital signs plays a key role in both patient care and wellness, but can be challenging outside of medical settings due to the lack of specialized equipment. In this study, we prospectively evaluated smartphone camera-based techniques for measuring heart rate (HR) and respiratory rate (RR) for consumer wellness use. HR was measured by placing the finger over the rear-facing camera, while RR was measured via a video of the participants sitting still in front of the front-facing camera. In the HR study of 95 participants (with a protocol that included both measurements at rest and post exercise), the mean absolute percent error (MAPE) ± standard deviation of the measurement was 1.6% ± 4.3%, which was significantly lower than the pre-specified goal of 5%. No significant differences in the MAPE were present across colorimeter-measured skin-tone subgroups: 1.8% ± 4.5% for very light to intermediate, 1.3% ± 3.3% for tan and brown, and 1.8% ± 4.9% for dark. In the RR study of 50 participants, the mean absolute error (MAE) was 0.78 ± 0.61 breaths/min, which was significantly lower than the pre-specified goal of 3 breath/min. The MAE was low in both healthy participants (0.70 ± 0.67 breaths/min), and participants with chronic respiratory conditions (0.80 ± 0.60 breaths/min). Our results validate that smartphone camera-based techniques can accurately measure HR and RR across a range of pre-defined subgroups. Measurement of heart rate (HR) and respiratory rate (RR), two of the four cardinal vital signs-HR, RR, body temperature, and blood pressure-is the starting point of physical assessment for both health and wellness. However, taking these standard measurements via a physical examination becomes challenging in telehealth, remote care, and consumer wellness settings. [1] [2] [3] In particular, the recent COVID-19 pandemic has accelerated trends towards telehealth and remote triage, diagnosis, and monitoring. 4, 5 Although specialized devices are commercially available for consumers and have the potential to motivate healthy behaviors, 6 their cost and relatively low adoption limit general usage. On the other hand, with smartphone penetration exceeding 40% globally and 80% in the US, 7 up to 3.8 billion individuals already have access to a myriad of sensors and hardware (video cameras with flash, accelerometers, gyroscope, etc) that are changing the way people interact with each other, and their environments. A combination of these same sensors together with novel computer algorithms can be used to measure vital signs via consumer-grade smartphones. [8] [9] [10] [11] [12] Indeed, several such mobile applications ("apps") are available, some with hundreds of thousands of installs. 13 However, these apps seldom undergo rigorous clinical validation for accuracy and generalizability. Only a limited number of apps have undergone clinical evaluation for HR measurement (and/or atrial fibrillation detection), 14 and the authors are unaware of any RR measurement apps that underwent clinical validation. In this work, we present and validate two algorithms that make use of smartphone cameras for vital sign measurements. The first algorithm leverages photoplethysmography (PPG) acquired using smartphone cameras for HR measurement. PPG signals are recorded by placing a finger over the camera lens, and the color changes captured in the video are used to determine the oscillation of blood volume after each heart beat. 15 In the second algorithm, we leverage upper-torso videos obtained via the front-facing smartphone camera to track the physical motion of breathing to measure RR. Herein, we describe both details of the algorithms themselves, in addition to reporting the performance of these two algorithms in prospective clinical validation studies. The studies sought to demonstrate reliable and consistent accuracy on diverse populations (in terms of objectively-measured skin tones, ranging from very light to dark skin) for HR and health status (with and without chronic pulmonary conditions) for RR. We conducted two separate prospective studies to validate the performance of smartphone-based HR and RR measurements ( Figure 1 ). The user interfaces of the two custom research apps are shown in Supplementary Figure 1 . The HR algorithm used PPG signals measured from the study participants placing their finger over the rear camera, and the enrollment for the corresponding validation study was stratified to ensure diversity across skin tones. The RR algorithm used video captures of the face and upper torso, and the enrollment for the corresponding validation study was stratified to capture participants with and without chronic respiratory conditions. The following sections detail each of the two studies. Prior work in computer vision to extract heart rate from RGB (red-green-blue) video signals has leveraged manually extracted features in PPG signals from the finger for arrhythmia detection, 16 ballistocardiographic movements from
PDF JSON Files: document_parses/pdf_json/79a5d77a33aa904f900f1a12ddd262c0e5dbbdd4.json
G_ID: prospective_validation_of_smartphone_based_heart_rate_and_respiratory_rate_measurement
S2 ID: 232200205