The study was done in the postnatal ward in a tertiary- care hospital in India during June-August 2017. Institutional Ethics Committee approval was obtained and informed written consent was taken from one of the parents.

We used 98.6 Fever Watch (Helyxon Health Care Private Limited) for continuous monitoring. It consists of an insulated thermistor with Bluetooth connectivity [2]. The thermistor was connected to an iPod (Apple Inc) kept within 30m via Bluetooth, which in turn was connected to physician’s smartphone via Internet.

BET thermistor was attached to baby’s skin in right hypochondrium using a transparent film dressing. BET measured baby’s temperature every minute, which was displayed in the iPod. When the temperature was abnormal (<36.5 or >37.5ºC) the iPod gave an alarm. The nurse would check baby’s temperature using DT and take appropriate measures. If hypo/hyperthermia persisted for 15 minutes, an alert was escalated in the physician’s mobile, who would examine the baby and plan further management.

We recruited 30 term infants with mean (SD) birthweight of 2838 (418) grams and median (IQR) postnatal age of 3 (0.5, 4) days; 15 late preterm infants with birthweight of 2087 (464) grams and postnatal age of 4 (4, 5) days; and 15 preterm infants with mean (SD) gestational age of 31+4 (1+6) weeks, birthweight of 1685 (511) grams and postnatal age of 15 (9, 22) days. Duration of recording was 24 hours for term and 48 hours for preterm infants.

There were 1537 temperature readings with concurrent measurements from both BET and DT. Intra-class correlation co-efficient (ICC) and Bland Altman plot were used to measure the agreement between BET and DT. The ICC between DT and BET was 0.80 (0.78-0.82) (P<0.001). The mean (SD) bias was 0.024 (0.29). The limits of agreements were (- 0.55, +0.60). None of the babies developed skin rash or reaction due to the thermistor probe.

Our study shows that continuous temperature monitoring identifies hypo/hyperthermia episodes that would have been missed or detected late during intermittent monitoring. The agreement between DT and BET was good. The false alarms were due to reasons such as improper attachment of the probe to the skin or handling of the baby. The stress and manpower needed to deal with false alarms should also be considered.

The limitations of the study were small sample size and lack of controls. The clinical outcomes of infants were not analyzed in the study.

Continuous temperature monitoring using BET will be useful in postnatal wards, where the thermistors of many infants can be connected to a single iPod or smartphone. It can also be used for temperature monitoring of LBW infants at home, using the parents’ smartphone.

Contributors: SMS,TA: Concept, design, data collection, data analysis, manuscript writing and review; AC,UB,BN: Concept, design, data collection and manuscript review.

Funding and Competing interests: The equipment supply and payment for the nurses were done by Helyxon Health care Solutions Private limited. The device manufacturers had no input in study design, data accrual and analysis or manuscript preparation and had no access to the study data. The study investigators had no financial gain from the study.

1. Lunze K, Bloom DE, Jamison DT, Hamer DH. The global burden of neonatal hypothermia: Systematic review of a major challenge for newborn survival. BMC Med. 2013;11:24.