Evaluation of a contactless neonatal physiological monitor in Nairobi, Kenya

Background Globally, 2.5 million neonates died in 2018, accounting for 46% of under-5 deaths. Multiparameter continuous physiological monitoring (MCPM) of neonates allows for early detection and treatment of life-threatening health problems. However, neonatal monitoring technology is largely unavailable in low-resource settings. Methods In four evaluation rounds, we prospectively compared the accuracy of the EarlySense under-mattress device to the Masimo Rad-97 pulse CO-oximeter with capnography reference device for heart rate (HR) and respiratory rate (RR) measurements in neonates in Kenya. EarlySense algorithm optimisations were made between evaluation rounds. In each evaluation round, we compared 200 randomly selected epochs of data using Bland-Altman plots and generated Clarke error grids with zones of 20% to aid in clinical interpretation. Results Between 9 July 2019 and 8 January 2020, we collected 280 hours of MCPM data from 76 enrolled neonates. At the final evaluation round, the EarlySense MCPM device demonstrated a bias of −0.8 beats/minute for HR and 1.6 breaths/minute for RR, and normalised spread between the 95% upper and lower limits of agreement of 6.2% for HR and 27.3% for RR. Agreement between the two MCPM devices met the a priori–defined threshold of 30%. The Clarke error grids showed that all observations for HR and 197/200 for RR were within a 20% difference. Conclusion Our research indicates that there is acceptable agreement between the EarlySense and Masimo MCPM devices in the context of large within-subject variability; however, further studies establishing cost-effectiveness and clinical effectiveness are needed before large-scale implementation of the EarlySense MCPM device in neonates. Trial registration number NCT03920761.


Appendix B
Figure shows time synchronization of signals from EarlySense and Masimo Rad-97 devices over a 1-hour period. a EarlySense raw data signal. b EarlySense (in red) and Masimo Rad-97 (in blue) heart rate (HR) detections synchronized. c EarlySense and Masimo Rad-97 respiratory rate (RR) detections synchronized.
BMJ Publishing Group Limited (BMJ) disclaims all liability and responsibility arising from any reliance Supplemental material placed on this supplemental material which has been supplied by the author(s)

Appendix C
Figure shows time synchronization of signals from EarlySense and Masimo Rad-97 over a 2h:15min period. a EarlySense raw data signal. b EarlySense (in red) and Masimo Rad-97 (in blue) heart rate (HR) detections synchronized. c. EarlySense HR detections using 10 second median (in pink) and 1 minute filtered median (in black). Real-time measurement of HR is dependent on a 6 second signal and may suffer from artifacts. Anomaly detections are filtered out by only taking HR processed data between the 10 th and 90 th percentiles. A median value is calculated over the filtered data. If this median value and the median value from the last 10 seconds (non-filtered) crosses a threshold, an alert is raised.

Data processing and selection
We retrieved raw data collected in real-time from the reference device with a custom Android application. Data was parsed in C (Dennis Ritchie & Bell Labs, USA) to obtain plethysmograph waveform and plethysmograph quality index (PO-SQI) data at 62.5 hertz (Hz), and capnography (carbon dioxide (CO2)) waveform data at approximately 20 Hz. We analyzed CO2 waveform data using a breath detection algorithm developed in MATLAB (Math Works, USA) based on adaptive pulse segmentation. We developed a custom algorithm based on capnography features to determine the capnography quality index (CO2-SQI).
The HR and RR from the EarlySense device were calculated every second from the output signal which was sampled at 240.5 Hz, using neonatal algorithms adapted from previously validated algorithms for adults. 21 Signal quality was determined for each measurement to assist with data analysis.
A sample size of 200 randomly selected epochs of data was used to ensure a narrow 95% confidence interval of +/-0.24 standard deviation of between-device-measurement differences around the limits of agreement (LOA).

Additional results
Of the recordings that were included for analysis, there was on average 21% of data missing from the EarlySense device, and 25% of HR and 30% of RR data missing from the Masimo Rad-97 device. Reasons for missing EarlySense data included the neonate being out of bed or sensor disconnections due to power outages. Reasons for missing Masimo data included the capnography and pulse oximetry probes being removed when the neonate was out of bed and sensor disconnections due to power outages. The Masimo Rad-97 capnography nasal cannula at times needed to be removed due to the fragility of these neonates, some of whom did not tolerate the nasal cannula. Power outages also contributed to missing data.
In a retrospective review of all of the data, we were unable to detect any significant episodes of apnea (> 20 seconds of no respiration associated with bradycardia).
BMJ Publishing Group Limited (BMJ) disclaims all liability and responsibility arising from any reliance Supplemental material placed on this supplemental material which has been supplied by the author(s)