RESUMO
BACKGROUND: There is concern that the values provided by devices using infrared thermometry in emergency departments (EDs) do not reflect body core temperature accurately. OBJECTIVES: Evaluation of three thermometers commonly used in the ED. METHODS: Two infrared ear thermometers and an infrared forehead thermometer were evaluated using 1) the Voltcraft IRS-350 calibration device, 2) comparing temperature values to a rectal end-exercise temperature (T-RECT) of 38.1°C in 12 participants, and 3) comparing temperature values to rectal temperature in 133 ED patients. RESULTS: Calibration across the human core temperature range revealed that the ear thermometers underestimated radiant temperature by 0.77 ± 0.39°C and 1.84 ± 0.26°C, respectively, whereas the forehead thermometer overestimated radiant temperature by 0.90 ± 0.51°C. After cycling exercise, all thermometers underestimated T-RECT (0.54 ± 0.27°C and 1.03 ± 0.48°C for the ear thermometers and 1.14 ± 0.38°C for the forehead thermometer). In the ED, the ear thermometers underestimated T-RECT by 0.31 ± 0.37°C and 0.46 ± 0.50°C, whereas the forehead thermometer exhibited a nonsignificant overestimation of 0.04 ± 0.46°C. If the threshold for fever in all systems had been set to 37.5°C instead of 38.0°C, the sensitivity and specificity of the systems for real fever (T-RECT ≥ 38°C) are, respectively, 71% and 96% (ear thermometer 1), 57% and 97% (ear thermometer 2), and 86% and 90% (forehead thermometer). CONCLUSION: We conclude that the investigated thermometers are not reliable as devices to measure radiant temperature, cannot be used to assess body core temperature during exercise, but may be used as a screening device, with 37.5°C as a threshold for fever in emergency care settings.