C72-28 Simultaneous Heart and Respiratory Rate Extraction Reduces Triage Time in Remote Auscultation

Abstract Rationale Remote auscultation is an augmented method of performing physical examination in virtual settings and should be a cornerstone of any respiratory-focused teleconsultation, analogous to traditional auscultation. Current teleconsultation workflows require separate devices and sequential measurements to obtain heart and respiratory rates, prolonging assessment time and potentially delaying triage decisions. Beyond qualitative lung sound assessment, these remote auscultations contain quantifiable physiologic information useful for clinical assessment. This study evaluated the accuracy of heart rate (HR) and respiratory rate (RR) values extracted from remotely recorded chest sounds using an acoustic signal processing algorithm. The objective was to determine whether dual physiologic parameters derived from standard remote auscultation recordings could streamline clinical triage during teleconsultation, reducing time to obtain vital signs. Methods Twenty-six adults (aged 21-80 years) had two five-minute auscultation recordings made using a wearable digital stethoscope placed on the upper chest: one during rest and another post-movement following two minutes of self-paced jumping jacks to elevate both HR and RR. HR and RR were extracted from the same acoustic signal using an algorithm that isolates pseudo-periodic cardiac and respiratory components. Concurrent reference measurements were obtained using a Masimo MightySat® Rx pulse oximeter (K181956) and a Philips NM3 capnograph. Agreement was assessed using mean bias, limits of agreement, and mean absolute error (MAE). Acceptable accuracy thresholds were predefined as MAE ≤ 6 bpm for HR and ≤ 6 brpm for RR. Results Algorithm-derived HR and RR values showed close agreement with reference measurements. Mean HR biases were 2.8 bpm (rest) and 3.0 bpm (post-movement); mean RR biases were 0.7 brpm (rest) and 1.1 brpm (post-movement). All MAE values were within predefined limits. Bland-Altman analysis demonstrated narrow limits of agreement and no systematic bias across the measured physiologic range, with HR from 47 to 132 bpm, and RR from 8 to 26 brpm. Conclusion Simultaneous heart rate and respiratory rate extraction from a single remote chest sound recording achieved clinically acceptable accuracy under both resting and active conditions. By consolidating two vital sign measurements into one five-minute acquisition, this approach eliminates the need for separate pulse oximetry and respiratory monitoring devices during teleconsultation. Integrating this streamlined workflow into remote auscultation provides healthcare professionals with rapid, objective physiologic assessment, enabling early identification of cardiorespiratory instability and more efficient prioritization of patients requiring urgent evaluation. Acoustic signal analysis may augment respiratory teleconsultation with time-efficient capability and contribute to improved virtual care delivery. This abstract is funded by: None