Non-contact thermometers: a method comparison study assessing agreement with electronic axillary and infrared tympanic thermometers.

Talk Code: 
Gail Hayward
Jan Y Verbakel, Fatene Abakar Ismail, George Edwards, Kay Wang, Susannah Fleming, Gea A Holtman, Margaret Glogowska, Elizabeth Morris, Kathryn Curtis, Ann van den Bruel
Author institutions: 
University of Oxford, KU Leuven, University of Groningen, University of Bristol


Temperature measurement is commonly performed in children aged 5 and under in primary care and is an important compnent of guideliens for management of acute illness in this age group. Currently either axillary or tympanic thermometers are used. Non-contact infra-red thermometers (NCITs) have the potential to reduce both the distress of the child and the risk of cross-infection. However our systematic review found limited evidence of comparative accuracy to standard methods. We aimed to compare two different NCITs to axillary and tympanic thermometers in children aged 5 and under presenting to primary care.


Wr performed a methods comparison study comparing 2 different NCITs (Thermofocus and Firhealth) to axillary and tympanic thermometers in children aged 5 or under attending the GP with an acute illness. Measurements were made in a pre-determined order using a random number generator. We evaluated reproducibility and sensitivity for fever of 38 °C or more using the axillary thermometer. Parents and children used scales to rate their discomfort.


401 children were recruited with median age 1.6 years (IQR 0.79-3.38). Our primary outcome was the comparison between the Thermofocus NCIT and axillary thermometer. The mean difference between the two methods was -0.14°C (95% CI -0.21 – (-0.06)), with the lower limit of agreement being -1.57°C (95% CI -1.69 – (-1.44)) and the upper limit 1.29°C (95% CI 1.16 - 1.42). Similar agreement was seen for the Firhealth NCIT. The Thermofocus had a sensitivity for fever (38°C using axillary measurement) of 29.3% (95% CI 16.1 - 45.5). The Firhealth had a sensitivity of 4.9% (95% CI 0.6 - 16.5). Reproduciblity was acceptable for both NCITs. The majority of parents found all methods acceptable, although discomfort ratings were highest for the axillary thermometer.


Temperature measured using an NCIT can vary by more than 1°C compared with measurements from currently recommended thermomter types, and sensitivity for fever was poor for both non-contact thermometers. However the agreement between aciallary and tympanic thermometers and core body temperature has also been shown to be poor. There is a clear opportunity for new technology to improve the accuracy of non-invasive temperature measurement.

Submitted by: 
Gail Hayward
Funding acknowledgement: 
This project was funded by the NIHR HTA Programme 16/45/01