Accuracy and Precision of Tympanic Temperature in the Reflection of Core Temperature

Monitoring of patient's temperature is an important nursing clinical procedure. Body temperature is usually taken at different sites including pulmonary artery, rectal, oral, and axillary areas. Currently, infrared tympanic temperature measurement method is increasingly used as a convenient, noninvasive and rapid method. However, accuracy and precision of this method has been questioned. The aims of this study were (a) to determine accuracy and comparing tympanic temperature with rectal temperature as a gold standard and (b) to determire Precision (Repeatability) of tympanic temperature with single measurement. This was a descriptive-comparative study with the sample size of 336 subjects, executively selected from the patients who admitted to Rasool Akram Hospital in Tehran. The patients were 16-85 years old. The tympanic temperatures were measured using the Braun ThermoScan type 6020 and the rectal temperatures were measured using the type M9380226 that both thermometers had 0.1 ºC internal error. The tympanic temperatures were measured twice per side with 1 minute of cooling off in each subject. Paired t-test showed that there was statistical difference between right and left tympanic temperature with rectal temperature (P<0.0001), however, no differences between the second measurement of right tympanic temperature (P=0.172), twine measurements of left tympanic temperature (P=0.100) and also between right and left tympanic temperatures (P=0.500) were found. The Bland-Altman analysis showed that the mean rectal temperatures were 0.23ºC more than the mean of tympanic temperatures, with ranges of agreement between 0.98ºC to 0.51ºC for right tympanic temperatures and the limits of agreement between -0.88ºC to 0.41ºC for left tympanic temperatures. Rectal temperatures showed that there was good correlation between right and left tympanic temperatures (r=0.78, P<0.0001 and r= 0.84, P<0.0001) respectively. Despite high correlation between the variables (rectal and tympanic temperature), there were clinically and statstically significant differences between them. Also the sensitivity of tympanic temperature was unacceptably low with the number of patients with fever. Therefore, it is concluded that tympanic temperature can’t exactly reflect rectal temperature. Due to acceptable precision and ignorable bias, the tympanic method for temperature measurement can be used, though it is not perfect alternative for rectal route.