LACK OF AGREEMENT BETWEEN TYMPANIC AND ORAL-TEMPERATURE MEASUREMENTS IN ADULT HOSPITALIZED-PATIENTS

Background: The purpose of this study was to compare temperature measu rements obtained by tympanic thermometers with those obtained by oral electronic or mercury-glass thermometers in adult hospitalized patient s. Methods: A prospective study of 406 nonintensive care unit adult pa tients hospitalized during an 8-month period in a tertiary care commun ity medical center.Results: Poor agreement was observed between tympan ic versus electronic thermometer readings, with 95% limits of agreemen t of -2.11 degrees F to +2.81 degrees F. Similarly, poor agreement was observed between tympanic versus oral mercury-glass temperatures, wit h 95% limits of agreement of -1.72 degrees F and +2.64 degrees F. Of 2 7 oral electronic temperatures 100 degrees F (37.7 degrees C) or highe r, 10 (37%, 95% confidence interval 19% to 58%) readings were 99.5 deg rees F (37.5 degrees C) or lower, and six (22%, 95% confidence interva l 9% to 42%) measured lower than 98.6 degrees F by tympanic thermomete rs. Conclusions: Temperatures measured by tympanic thermometers genera lly have poor agreement with those measured by oral electronic or merc ury-glass thermometers in adult hospitalized patients. We recommend th at tympanic thermometers not be used for routine screening for fever i n this patient population.