Reliability of body temperature measurements in hospitalised older patients.

Aims and Objectives: To compare different body temperature assessment methods in older people and to assess the role of cognitive and functional characteristics in temperature recordings.
Background: Axillary gallium-in-glass thermometers are commonly used. Their accuracy depends on the proper placement of the device and their permanence in place for eight minutes. With adequate instruction, well-functioning patients can measure their axillary temperature by themselves, while in cognitively and functionally impaired older people, inadequate understanding of instructions and misplacement of the thermometer might determine significant recording errors. Electronic ear and axillary temperature measurements are faster, but their accuracy has not been demonstrated convincingly with older people.
Methods: Patients (n = 107; aged 65-104 years) were recruited. Barthel Index and Short Portable Mental Status Questionnaire (SPMSQ) scores were obtained for each patient. Temperature readings were obtained using: the axillary gallium-in-glass thermometer, with (T(nurse)) and without (T(self)) the nurse's assistance; the electronic axillary thermometer (T(el)) and the infrared tympanic thermometer (T(tymp)). The T(nurse) was considered as the reference method.
Results: Mean difference and standard deviation (mean +/- SD) in temperature recordings between the different techniques and T(nurse) differed significantly from zero for T(self) (-0.40 SD 0.42) and T(tymp) (+0.19 SD 0.48). No significant differences in temperature recordings emerged between T(nurse) and T(el). In simple linear regression models, the difference between T(self) and T(nurse) significantly correlated with age, gender, SPMSQ score and Barthel Index. Multiple linear regression analysis showed an underestimation of body temperature in older patients with cognitive impairments.
Conclusion: Unassisted gallium-in-glass axillary temperature assessment is inadequate, in older patients. The differences between T(self) and T(nurse) are significantly influenced by age and mental decline. T(el) provides adequate accuracy. Relevance to clinical practice. In geriatric settings, the electronic axillary thermometer is a safe and accurate alternative to the more traditional gallium-in-glass thermometer, with the advantage of saving time (five seconds in recording vs. eight minutes).