Background: Our objective was to determine whether infrared tympanic thermometry is as reliable as the rectal digital thermometer. Earlier reports have given conflicting results on the issue.
Material and methods: 199 children aged 1 month to 12 years were included in the study. Rectal temperature and the temperature in at least one ear were obtained from all of the children. Children with perforated otitis media, intracranial tumours or treated with immunosuppressive medication were excluded from the study.
Results: The mean temperature in the rectum was 0.4 degree C higher than in the ear. The tympanic and the rectal readings were strongly correlated (r = 0.83, p = 0.01). With a definition of fever as rectal temperature > or = 38 degrees C the sensitivity of the tympanic thermometry was 71%, the specificity 95%, the positive predictive value 93%, and the negative predictive value 78%.
Interpretation: Infrared tympanic thermometry has obvious advantages compared to rectal measurements: It is more hygienic, faster and less painful for the child. However, its sensitivity is rather low when used in an emergency department. This means that a number of children with fever will not be diagnosed as such.