Email updates

Keep up to date with the latest news and content from BMC Family Practice and BioMed Central.

Open Access Highly Accessed Research article

Accuracy of parents in measuring body temperature with a tympanic thermometer

Joan L Robinson*, Hsing Jou and Donald W Spady

Author Affiliations

Department of Pediatrics and Stollery Children's Hospital, 2C3 Walter MacKenzie Centre, 8440-112 St., Edmonton, Alberta, T6G 2B7 Canada

For all author emails, please log on.

BMC Family Practice 2005, 6:3  doi:10.1186/1471-2296-6-3

Published: 11 January 2005

Abstract

Background

It is now common for parents to measure tympanic temperatures in children. The objective of this study was to assess the diagnostic accuracy of these measurements.

Methods

Parents and then nurses measured the temperature of 60 children with a tympanic thermometer designed for home use (home thermometer). The reference standard was a temperature measured by a nurse with a model of tympanic thermometer commonly used in hospitals (hospital thermometer). A difference of ≥ 0.5 °C was considered clinically significant. A fever was defined as a temperature ≥ 38.5 °C.

Results

The mean absolute difference between the readings done by the parent and the nurse with the home thermometer was 0.44 ± 0.61 °C, and 33% of the readings differed by ≥ 0.5 °C. The mean absolute difference between the readings done by the parent with the home thermometer and the nurse with the hospital thermometer was 0.51 ± 0.63 °C, and 72 % of the readings differed by ≥ 0.5 °C. Using the home thermometer, parents detected fever with a sensitivity of 76% (95% CI 50–93%), a specificity of 95% (95% CI 84–99%), a positive predictive value of 87% (95% CI 60–98%), and a negative predictive value of 91% (95% CI 79–98 %). In comparing the readings the nurse obtained from the two different tympanic thermometers, the mean absolute difference was 0.24 ± 0.22 °C. Nurses detected fever with a sensitivity of 94% (95 % CI 71–100 %), a specificity of 88% (95% CI 75–96 %), a positive predictive value of 76% (95% CI 53–92%), and a negative predictive value of 97% (95%CI 87–100 %) using the home thermometer. The intraclass correlation coefficient for the three sets of readings was 0.80, and the consistency of readings was not affected by the body temperature.

Conclusions

The readings done by parents with a tympanic thermometer designed for home use differed a clinically significant amount from the reference standard (readings done by nurses with a model of tympanic thermometer commonly used in hospitals) the majority of the time, and parents failed to detect fever about one-quarter of the time. Tympanic readings reported by parents should be interpreted with great caution.