Agreement among glass mercury thermometers, infrared thermometers and digital thermometer temperature recordings in febrile under-five children

Background: This study aimed to assess the degree of agreement between the temperature readings of a glass mercury thermometer (GMT), a digital thermometer (DT) and an infrared thermometer (IR) in paediatric patients. This was a comparative study conducted at a tertiary institution in Nnewi, Anambra State, Southeast Nigeria. One hundred febrile children <5 years of age were recruited via convenience sampling. Methods: Temperatures were measured via an axillary GMT, an axillary DT and a noncontact forehead IR thermometer. Data were compared among the three methods via a 2×2 contingency table, receiver operating curve (ROC) and Bland Altman plot with GMT as the standard. Results: There was a greater significant correlation between GMT and DT (r=0.901, p=0.001) than between GMT and IR (r=0.695, p=0.001). Bland Altman plots revealed that IR and DT could be used interchangeably with GMT with an arithmetic mean of 1.1°C, confidence interval of 0.9359 to 1.1581, limits of agreement of -0.1–2.1 (37.4°C–39.6°C), p<0.001, outliers 3% at 1.96 SD and arithmetic mean of -0.1°C, confidence interval of -0.1337-0.00427, limits of agreement of -0.7–0.5 (36.8°C–38.0°C), p=0.037 and outliers 4% at 1.96 SD with GMT as the standard against IR and DT, respectively. The area under the receiver operator curve values were 0.811 and 0.923 for GMT versus IR and GMT versus DT, respectively. Conclusions: Digital and infrared temperature readings are in reasonable agreement with a glass mercury thermometer and both can be used interchangeably with a glass mercury thermometer. Compared with infrared thermometers, digital thermometers are more closely related to glass mercury thermometers.