Background Rectal measurement is considered a gold standard in many healthcare systems for body temperature measurement in children. Although this method has several disadvantages, an ideal alternative thermometer has not yet been introduced. However tympanic and infrared skin thermometers are potential alternatives.

Methods A prospective cohort study was performed including 100 children between 0 and 18 years of age admitted to the general paediatric ward of Spaarne Hospital in The Netherlands between January and March 2009. The objectives of this study are to evaluate the accuracy of tympanic and two types of infrared skin thermometers (Beurer and Thermofocus) compared to rectal measurement and furthermore to evaluate the influence of different variables on temperature measurements.

Results Compared to rectal measurement (37.56°C), the mean temperatures of the tympanic (37.29°C), Beurer (36.79°C) and Thermofocus (37.30°C) thermometers differed significantly (p<0.001). Mean and SD of differences between rectal temperature and temperature measured with these alternative devices varied significantly (p<0.001). Sensitivity, specificity, positive and negative predictive values for detecting rectal fever measured with the tympanic, Beurer and Thermofocus thermometers are unacceptable, especially for the Beurer thermometer. This difference in temperature between rectal and the alternative thermometers remained after stratification on gender, age, skin colour and otoscopic abnormalities.

Conclusions In this study the authors demonstrated that the tympanic, Beurer and Thermofocus thermometers cannot reliably predict rectal temperature. Therefore the authors do not advise replacement of rectal measurement as the gold standard for detecting fever in children by one of these devices. When rectal measurement is not used, the infrared skin thermometers appear to perform less well than tympanic measurements.