Detection of Fever Using Continuous Skin Temperature Monitoring in Hospitalized Older Adults

Background: The continuous monitoring of body surface temperature has been proven to help detect potential fever events in hospitalized patients. However, the efficacy of using body surface temperature to detect fever in older adults remains unclear due to the relatively low and slower-to-change body surface temperature in this population. Purpose: This study was designed to investigate 1) the relationship between changes in body surface and routine tympanic temperatures, 2) the correlation between body surface temperature measurement frequency and detection of fever, and 3) the factors related to the incidence of fever in hospitalized older adults. Methods: A prospective study was conducted on 33 hospitalized older adults aged 65 years or older who were suspected to have or diagnosed with an infection in an infectious disease and medical ward at a medical center in southern Taiwan from March to November 2020. Demographic, routine tympanic temperature, and heart rate data were collected by reviewing the participants’ medical records. Body surface temperatures were monitored continuously using HEARThermo every 10 seconds until one of the following conditions were met: hospital discharge, no fever for three continuous days, and HEARThermo was removed. Descriptive analysis was used to compare the variations in body surface temperature and routine tympanic temperature measurements. Pearson correlation was used to analyze the correlation between different measurement frequencies and fever events. Finally, mixed effects logistic regression was used to analyze the factors significantly related to fever events. Results: Seven hundred and twenty routine body temperature measurements were taken, with 209 (29.0%) fever events detected in 23 (69.7%) of the participants. The body surface temperatures were more closely correlated with tympanic temperatures during fever events than non-fever events (r = .260, p < .001). More fever events were detected using body surface temperature monitoring frequencies of every 10 seconds and every 1 minute. After controlling for demographic factors, the results of the mixed effect model indicate that body surface temperature and heart rate are significant factors related to fever events in hospitalized older adults (odds ratio, OR: 1.74, p < .001; OR: 1.11, p < .001). Conclusions/ Implications for Practice: The continuous monitoring of body surface temperature may improve the detection of fever events in hospitalized older adults. The application of wearable devices and cloud platforms may further facilitate the real-time assessment and care capabilities of nurses, thus reducing their workload and improving care quality.