Diagnostic accuracy of an over-the-counter infant pulse oximeter for cardiorespiratory events.

Abstract

Objective: To determine the diagnostic accuracy of an over-the-counter infant pulse oximeter for cardiorespiratory events.

Design: Single-centre prospective diagnostic accuracy study.

Setting: University of Alabama at Birmingham.

Patients: Infants weighing ≥1500 g, <44 weeks' postmenstrual age (PMA) and off ventilator/continuous positive airway pressure support.

Interventions: Test device for 48 hours in addition to standard hospital monitors, ECG and pulse oximetry.

Main outcome measures: Data were time aligned and analysed using MATLAB. The coprimary outcomes were the diagnostic accuracy of the test device for the detection of events with heart rate (HR) <50 beats per minute (bpm) and events with oxygen saturations (SpO₂) <80% for ≥3 s.

Results: 66 infants with a median gestational age of 31 weeks (range 23-40) were studied at a median 35 weeks' PMA (range 32-42) weighing 1930 g (range 1500-3605 g) from April to July 2023. The sensitivity for detection of HR <50 bpm ≥3 s was 6% and 39% for smoothed and raw data, respectively, while the specificity was >99% for both smoothed and raw data. The sensitivity for SpO₂ <80% ≥3 s was 14% and 74%, while the specificity was >99% and 96% for smoothed and raw data, respectively. Sensitivity for bradycardia events was higher for events with longer durations and/or when using higher thresholds. Sensitivity was higher for hypoxaemia events with longer durations and/or when using higher thresholds.

Conclusion: An over-the-counter infant pulse oximeter had high specificity for bradycardia and hypoxaemia events consistent with a low false alarm rate. Sensitivity improved with longer events and higher event thresholds.

Trial registration number: NCT05774470.