INTRODUCTION: Body temperature measurement is a fundamental requirement for clinical decisions in nursing care, medical diagnosis, and treatment. Therefore, it is pivotal that body temperature measurements are accurate and precise. AIMS: To test the diagnostic accuracy of an ear temperature screening procedure among adult hospitalized patients. Further aims were to test the precision of the measurements being carried out by trained registered nurses compared with daily routine practice and to investigate patients' preferences for different measurement methods. METHODS: In Aalborg University Hospital, 274 patients were included in a cross-sectional comparative study. Each patient had four temperature measurements and responded to a survey regarding their preference for measurement. Bland-Altman analysis was used to evaluate the difference between ear- and rectal measurements. Sensitivity and specificity were evaluated at different cut-off points. RESULTS: The ear temperature was 0.1 to 0.2 °C lower than rectal temperature. At a cut-off point at 37.5 °C an ear thermometer is accurate and can be used for screening, while higher cut-off points risk missing patients with fever. There was no significant difference in the mean temperature measured by a trained registered nurse or other staff members and patients. Patients preferred ear measurements to rectal measurements. CONCLUSION: The tested ear thermometer is accurate for screening fever in an adult population during hospital admission. Using ear measurement as a screening tool can contribute to a less resource-demanding care activity and a more convenient alternative to rectal measurements in hospitalised patients.