Purpose: Infrared tympanic thermometry (ITT) is increasingly used as a convenient, noninvasive assessment method for febrile children. However, the accuracy of ITT for children has been questioned, particularly in relation to specificity and sensitivity. This study was designed to (a) determine the correlation and extent of agreement between rectal temperature (RT) readings obtained by electronic thermometer and ear-based temperature readings obtained by ITT, and (b) determine the accuracy of detecting fever in children under 6 years of age.
Methods: This correlational study used a sample of 241 paired ear and rectal temperatures obtained in the emergency department (ED) of a 920-bed regional hospital. All children under the age of 6 years who routinely received a rectal temperature measurement were eligible to participate. According to the ED protocol, rectal temperatures were obtained on all patients less than 3 years or patients 3-6 years that presented with a complaint of fever. For the study, tympanic measurements were also taken.
Results: Correlation between rectal and tympanic temperature readings was statistically significant (r = 0.84, p < .001). The mean difference between rectal and tympanic temperatures was -0.60 degrees C. Threshold-adjusted accuracy in screening for fever was determined by sensitivity (80%), specificity (85%), positive predictive value (87%), and negative predictive value (85%).
Conclusions: Sensitivity, specificity, positive predictive value, and negative predictive value are unacceptably low and the number of children with fever who would be missed by screening with a tympanic thermometer is unacceptable. Findings of this study do not support the use of tympanic thermometers to detect fever in children under 6 years of age.