Temperatuurmeting bij kinderen met een oorthermometer: Betrouwbaarheid en validiteit van een niet-professionele trommelvlies-infraroodthermometer

Objectives: To compare tympanic infrared temperature measurement with rectal temperature measurement in children under the age of seven. To determine whether performing one tympanic measurement is as valid and reproducible as taking the higher of two measurements. To determine the influence of possible confounding factors, like the presence of acute otitis media or cerumen and suboptimal co-operation of the child, on tympanic temperature readings. Methods: Children presenting in a general practitioner's practice during approximately 5 months enrolled in the study if they were suspected of having fever or if temperature measurement seemed necessary for medical decision making. Temperatures were taken sequentially by one investigator: twice in one ear with a tympanic thermometer and once rectally with an electronic rectal thermometer. Data were registered on the presence of acute otitis media, the quantity of cerumen and the child's co-operation. Results: Data of 37 children aged 1 to 72 months of whom 16 had fever (rectal temperature >38.0°C) were analysed. The mean temperature difference between tympanic and rectal measurement was -0.1°C (95% CI -0.3°C to +0.1°C). First tympanic measurements were practically identical to the higher of two tympanic measurements. Compared to the rectal measurements tympanic measurements tended to give slightly higher readings when otitis media was present (mean +0.1°C) and lower readings in case of cerumen or suboptimal co-operation of the child (mean -0.4°C and -0.2°C to -3°C respectively). The sensitivity, specificity, positive predictive value and negative predictive value of tympanic measurements to detect fever were 94% (95% CI 70-100%), 95% (95% CI 76-100%), 94% (95% CI 70-100%) and 95% (95% CI 76-100%) respectively. These figures were the same when the higher of two tympanic readings was used. Conclusions: Tympanic temperatures did not differ much from rectal temperatures in this homogeneous population. Performing one tympanic measurement yields results as reproducible and valid as taking the higher of two measurements. For a definite judgement on the discriminative power of infrared tympanic temperature measurement and the factors that influence its outcome, larger studies with various subgroups are required.