Accuracy of infrared ear thermometry in adult patients

Objective: To assess (1) the agreement between infrared ear thermometry and core reference temperature (in the pulmonary artery), (2) the agreement between measurements in the right and left ears, and (3) the screening validity of infrared tympanic thermometry in detecting rectal fever. Design: Temperatures were measured in both ears with an infrared thermometer, in one group of patients by simultaneous measurements with thermistors inserted in the pulmonary artery, esophagus, and rectum, and in the other group with a rectal glass-mercury thermometer. Setting: An intensive care unit and a department of internal medicine in a secondary care hospital. Patients and participants: Two samples: 16 adult patients admitted to the intensive care unit and 103 consecutive patients admitted to the department of medicine. Measurements: The major outcome measures were (a) the agreement between infrared ear thermometry and thermistor pulmonary artery temperature and (b) the sensitivity and specificity for detecting fever, using rectal measurement as reference. Results: Both rectal and esophageal thermistor measurements showed better agreement with the pulmonary artery reference temperature than single ear tympanic thermometry. The sensitivity and specificity of ear thermometry for detecting fever (≥38.0°C rectal reference) were 0.58 and 0.94, respectively. Double ear thermometry had a sensitivity of 0.61 and a specificity of 0.95, when using the mean value. Conclusions: Both rectal and esophageal thermistor measurements showed better agreement with pulmonary artery temperature than single ear thermometry. Using the mean of two ear measurements improves the agreement and screening validity for detecting fever by rectal temperature. If temperature measurements are critical, esophageal measurements achieve excellent agreement with pulmonary artery temperatures. Received: 15 November 1995 Accepted: 28 September 1996