A Novel Infant Wearable Device Intended for Interstage Cardiorespiratory Monitoring: First Proof-of-Concept

S. Duffy¹, K. Calligy¹, W. Golden¹, J. Nogee¹, D. Gottlieb Sen^2
1Johns Hopkins University, Baltimore, Maryland ²Johns Hopkins Medicine, Baltimore, Maryland

Purpose: Interstage monitoring for single ventricle congenital heart disease (CHD) has been shown to significantly reduce mortality (Ghanayem, 2003). However, the current standard of care is limited by reliance on untrained parents to collect and interpret vital sign data, as well as data transfer to a specialized team (Taylor, 2016).

Methods: We designed and built a wearable vital sign monitor intended to be integrated into the clinical care of infants with CHD. This monitor is worn under a diaper, uses primarily ECG and bioimpedance (Bio-Z) waveforms to derive heart rate (HR), respiratory rate (RR) and tidal volume, and collects data related to temperature, infant cries, and motion. We then studied healthy newborns to demonstrate proof-of-concept with an early prototype. We used MATLAB code to derive HR and RR from the abdomen and conventional chest positions. Next, we calculated the average signal-to-noise ratio (SNR) of the ECG and Bio-Z waveforms from the chest and abdomen to validate acceptable signal quality. Finally, we calculated average standard deviations of HR and RR over time at one location and between locations to determine whether variability seen at the abdomen and between the chest and abdomen locations is acceptable.

Results: MATLAB-derived HR and RR values from the chest and abdomen demonstrated no significant difference in average measurement (p,HR=0.30; p,RR=0.15) and showed a significant correlation between the two locations (r,HR=0.94; r,RR=0.83). ECG waveforms had stronger signal quality at the chest (SNR=19.8) with acceptable signal quality at the abdomen (SNR=14.3). Conversely, Bio-Z waveforms had a stronger signal at the abdomen location (SNR=24.2) compared to the chest location (SNR=19.0). The standard deviation of ECG ‘biological’ variation (calculated over time at one location) and ‘locational’ variation (calculated between the abdomen and chest location), also were determined (Table 1). We identified no significant difference in the standard deviation observed in the abdomen compared to the chest for both HR (p=0.63) and RR (p=0.87). Additionally, we observed a higher standard deviation in RR values, which may be explained by the high variation and irregular breathing patterns of newborns (Njeru, 2022).

Conclusion: We demonstrated that HR and RR can be acquired from the abdomen of infants using our wearable monitor. From the abdomen, ECG signals are weaker, and respiratory signals are stronger. Unobtrusive monitoring of vital signs in infants has important potential application to interstage monitoring.

Identify the source of the funding for this research project: Brett Boyer Foundation