An initial study on the agreement of body temperatures measured by infrared cameras and oral thermometry

Abstract The COVID-19 pandemic has led to the rapid adoption and rollout of thermal camera-based Infrared Thermography (IRT) systems for fever detection. These systems use facial infrared emissions to detect individuals exhibiting an elevated core-body temperature, which is present in many symptomatic presentations of Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2). Despite the rollout of these systems, there is little independent research supporting their efficacy. The primary objective of this study was to assess the precision and accuracy of IRT screening solutions in a real-world scenario. The method used was a single-centre, observational study investigating the agreement of three IRT systems compared to digital oral thermometer measurements of body temperature. Over 5 days, 107 measurements were taken from individuals wearing facial masks. During each entry, two measurements of the subject’s body temperature were made from each system to allow for the evaluation of the measurement precision, followed by an oral thermometer measurement. Each participant also answered a short demographic survey. This study found that the precision of the IRT systems was wider than 0.3 °C claimed accuracy of two of the systems. This study also found that the IRT measurements were only weakly correlated to those of the oral temperature. Additionally, it was found that demographic characteristics (age, gender, and mask-type) impacted the measurement error. This study indicates that using IRT systems in front-line scenarios poses a potential risk, where a lack of measurement accuracy could possibly allow febrile individuals to pass through undetected. Further research is required into methods which could increase accuracy and improve the techniques viability