Accuracy and Precision of Noninvasive Temperature Measurement in Adult Intensive Care Patients

Abstract

Background Research on the accuracy and precision of noninvasive methods of measuring body temperature is equivocal. Objectives To determine accuracy and precision of oral, ear-based, temporal artery, and axillary temperature measurements compared with pulmonary artery temperature. Methods Repeated-measures design conducted for 6 months. Sequential temperature measurements on the same side of the body were obtained within 1 minute, with measurements repeated 3 times at 20-minute intervals. Accuracy, precision, and confidence limits were analyzed. Results In 60 adults with cardiopulmonary disease and a pulmonary artery catheter, mean pulmonary artery temperature was 37.1°C (SD 0.6°C, range 35.3°C–39.4°C). Mean (SD) offset from pulmonary artery temperature (with the mean reflecting accuracy and SD reflecting precision) and confidence limits were 0.09°C (0.43°C) and −0.75°C to 0.93°C for oral measurements, −0.36°C (0.56°C) and −1.46°C to 0.74°C for ear measurements, −0.02°C (0.47°C) and −0.92°C to 0.88°C for temporal artery measurements, and 0.23°C (0.44°C) and −0.64°C to 1.12°C for axillary measurements. Percentage of pairs with differences greater than ±0.5°C was 19% for oral, 49% for ear, 20% for temporal artery, and 27% for axillary measurements. Intubation increased oral measurements compared with pulmonary artery temperatures (mean difference 0.3°C, SD 0.3°C, P = .001). Conclusions Oral and temporal artery measurements were most accurate and precise. Axillary measurements underestimated pulmonary artery temperature. Ear measurements were least accurate and precise. Intubation affected the accuracy of oral measurements; diaphoresis and airflow across the face may affect temporal artery measurements.