Measurement accuracy of fever by tympanic and axillary thermometry

Abstract

As the basic sciences develop, temperature measurement methods and devices were improved. For hundreds of years both in clinics and home, mercury-in-glass thermometer was the standard of human temperature measurements. In this study, we aimed to compare tympanic infrared thermometers with the conventional temperature option, mercury-in-glass thermometer, which is historical standard in the clinical conditions.

Methods: A total of 102 randomly selected pediatric patients who admitted to our hospital were enrolled, and simultaneous temperature measurements were performed via axilla and external auditory canal with 3 different techniques. For external auditory recordings, infrared tympanic First Temp Genius for clinical use and Microlife IR 1DA1 for home usage were used. Classic mercury-in-glass thermometers were used for axillary recording. For each method, 886 measurements were performed.

Results: The mean results of the axillary mercury-in-glass thermometers, infrared tympanic First Temp Genius, and Microlife IR 1DA1 were 36.8 +/- 0.7, 37.5 +/- 0.9, 36.9 +/- 0.8, respectively. The Bland-Altman plot of differences suggests that 95% of the infrared tympanic clinical use thermometer readings were within the limits of agreement, which is +0.27 and -1.75 degrees C range of mercury-in-glass thermometer. The Bland-Altman plot of differences suggests that 95% of the tympanic home-use thermometer readings were within the limits of agreement, which is +0.98 and -1.27 degrees C range of mercury-in-glass thermometer. In our group, 15% of the patients were misdiagnosed as febrile with home-use tympanic thermometer, whereas this percentage was 4% with clinical tympanic thermometer. Also, 5% and 31% of febrile patients were misdiagnosed as afebrile with clinical tympanic and home-use tympanic thermometer, if axillary mercury-in-glass thermometer recording defines fever.

Discussion: Our results showed that there is a significant difference in each recording with different thermometers, and this variance was present in both higher and lower readings. We recommend thathome-use infrared tympanic thermometer could be used for screening but must not be considered as a tool to decide patients follow-up.