Accuracy of infrared ear thermometry and other temperature methods in adults.

OBJECTIVE: To compare the accuracy of infrared ear-based temperature measurement in relation to thermometer, ear position, and other temperature methods, with pulmonary artery temperature as the reference.
METHODS: Ear-based temperature measurements were made with four infrared thermometers, three in the core mode and two in the unadjusted mode, each with tug and no-tug techniques. Pulmonary artery, bladder (n = 21), and axillary temperatures were read after each ear-based measurement and oral temperature was measured once when possible (n = 32). Subjects consisted of a convenience sample of 50 patients with pulmonary artery catheters who were in adult critical care units of a university teaching hospital.
RESULTS: Ear-based measurements correlated well with pulmonary artery temperature (r = .87 to .91), although closeness of agreement differed among thermometer-mode combinations (mean offsets = -0.7 to 0.5 degree C) and had moderately high variability between subjects (SD = +/- 0.5 degree C) with all instruments. Use of an ear tug either made no difference or resulted in slightly lower readings. Bladder temperature was nearly identical to pulmonary artery temperature values (r = .99, offset = 0.0 +/- 0.2 degree C). Oral readings were slightly lower (r = .78, offset = -0.2 degree C) and axillary readings much more so (r = .80 to .82, offset = -0.7 degree C); both were highly variable (SD = +/- 0.6 degree C) and affected by external factors.
CONCLUSIONS: Infrared ear thermometry is useful for clinical temperature measurement as long as moderately high variability between patients is acceptable. Readings differ among thermometers, although several instruments provide values close to pulmonary artery temperature in adults. Readings are not higher with an ear tug. Bladder temperature substitutes well for pulmonary artery temperature, whereas oral and axillary values may be influenced by external factors in the critical care setting.