Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
As neonatal heart rate (HR) is a vital sign used to assess the need for and response to resuscitation,1 measuring it rapidly, accurately and affordably is important to clinicians around the world. We aimed to assess the accuracy, speed and reliability of a novel low-cost digital stethoscope (DS) attached to a smartphone running real-time newborn HR detection software in determining the HR of clinically stable infants.
We studied infants >26 weeks’ corrected gestation excluding those receiving high-frequency oscillation ventilation. We applied the Stethocloud V.0.2beta DS head to their exposed precordium. The DS was connected to an Apple iPhone 5s running Neorate 0.1a software (CliniCloud, Melbourne, Australia) which displayed a real-time HR (DS HR). Infants were simultaneously monitored with three-lead electrocardiography (Philips Intellivue MP70, Philips, Andover, USA). Video recordings captured both HR displays and data were extracted at 1 s intervals. We excluded seven of 57 recordings based on our a priori definition of reliability, data available within 30 s of activating the DS software and at least 10 s of output during the 90 s recording period.
Participants were a convenience sample obtained in the neonatal intensive care unit when a parent was available to provide informed consent. Corrected gestational age and weight at time of study ranged from 26.7 to 54.7 weeks and 656–3690 g, respectively. Thirty-four per cent had respiratory support via an endotracheal tube, continuous positive airway pressure or heated humidified high-flow nasal cannulae; 24% had cardiac lesions (table 1).
A Bland–Altman plot constructed to assess agreement between ECG and DS HR compared 3972 paired data points (figure 1). The mean difference (SD) between ECG and DS HR was 7.4 (24) beats per minute (bpm). Only six recordings had a mean difference exceeding 20 bpm (data not shown). The median (IQR) time to first DS HR display was 2 (1–7) s.
To our knowledge, this is the first time DS technology for real-time HR measurement has been studied in newborns. Accuracy of DS HR is greater than chest auscultation and umbilical cord palpation which underestimate ECG HR by 14 and 21 bpm, respectively.2 Pulse oximetry in comparison correlates more strongly with ECG but takes >1 min to establish a reliable HR,3 whereas DS HR was displayed in <5 s on average. Figure 1 shows that for 8.6% of displayed values, DS HR is >20 bpm different to ECG, mostly due to underestimation of ECG HR. Such imprecision, despite occurring infrequently, represents a software algorithm flaw which renders the current device unacceptable for clinical use. Interestingly, we found external noises and murmurs were not associated with inaccurate DS HR. The lower frequency of neonatal heartbeats allows these high-frequency sounds to be filtered.4
A key limitation of the device is shown by the proportion of recordings excluded due to poor performance. Accidental loss of stethoscope contact with the precordium and loose wiring connections were major contributors. The DS and smartphone technology represent important developments in HR monitoring but further refinement is necessary before they are useful in clinical practice.
Contributors All authors were involved in the planning, conduct and reporting of the work.
Competing interests None declared.
Ethics approval The Human Research and Ethics Committees of The Royal Women’s Hospital, Melbourne, Australia.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement Unpublished data from the study can be made available. Please contact the corresponding author.