The patient: a novel source of error in clinical temperature measurement using infrared aural thermometry

INTRODUCTION: The vital signs are an important component of patient assessment. With respect to body temperature; there has been a move away from mercury-in-glass thermometers toward the relatively inexpensive, safer to use infrared tympanic, auricular, or ear thermometer. Although already in widespread use, the reliability of these devices has increasingly been called into question. Few studies, however, have considered that the problem might reside outside the device itself. OBJECTIVE: To investigate the correlation between left ear temperature and right ear temperature using an infrared tympanic thermometer (ITT). DESIGN: A prospective, single-blind trail with randomized assignment of the first ear to be assessed. SETTING/LOCATION: Welsh Institute of Chiropractic (WIOC), University of Glamorgan. SUBJECTS: One hundred and thirty two (132) asymptomatic subjects who were opportunistically sampled on entry into the WIOC. None of the subjects had any neurologic deficit, or any known underlying pathologic problems. All were students at the University of Glamorgan (age range, 18-48 years). A smaller sample of the same cohort was used on a separate occasion to compare ITTs from 2 different manufacturers using the same protocol. INTERVENTION: Measurement of ear (ear canal/tympanic) temperature bilaterally using an ITT. OUTCOME MEASURES: Bilateral asymmetry in ear canal temperature. Analysis consisted of nonparametric statistical tests, using Spearman's correlation for comparison and the Wilcoxon matched pairs test for differences. RESULTS: In 132 subjects who completed the main study, although a degree of correlation was found, the relationship did not extrapolate through zero. At temperatures more than 0.4 degrees C above or below 36.7 degrees C, there was a significant difference in temperature in the left compared with the right ear. Additionally, the left ear tended to register a lower temperature than the right ear at temperatures below 36.7 degrees C, whereas the left ear tended to read a higher temperature than the right ear above 36.7 degrees C. CONCLUSIONS: The authors suggest that clinicians consider measuring both ears, and take the greater of the readings to be a more accurate reflection of the patient's core temperature. The difference may perhaps even be related to the person's health. This change in methodology should increase the reliability of the measurement and go some way toward restoring confidence in the use of ITT devices.