Background: During the 2003 severe acute respiratory syndrome and 2009 pandemic influenza A (H1N1) outbreaks, infrared thermal detection systems (ITDS) were used at international ports of entry and in hospitals to screen for fever. However, evidence is limited to support the accuracy of ITDS and their benefit over self-reported fever for mass screening. In this study, we compared three different ITDS to self-reported fever.Methods: A cross-sectional study of 2986 patients (age≥18 years) was conducted in three hospital emergency departments. Patients were asked if they felt that they had a fever (selfreported fever). We measured patient skin temperatures by using three ITDS (FLIR A20 M, OptoTherm Thermoscreen, Wahl HSI2000S) and oral temperatures (≥100oF = confirmed fever) by using digital thermometers. ITDS temperature measurements and self-reported fevers were compared using oral temperatures as a reference. Data were analyzed using simple and multiple linear methods.
Results: Of 2873 patients with an oral temperature recorded, 64 (2.2%) had a confirmed fever. Fever was reported by 476 (16.6%) patients and 48 (10.1%) of these were confirmed. Self-reported fever had 75.0% sensitivity and 84.7% specificity. At optimal cutoffs for detecting fever as found in this study, the OptoTherm Thermoscreen and FLIR A20 M temperature measurements had greater sensitivity (85.7% and 79.0%) and specificity (91.0% and 92.0%) than self-reported fever. Of the three methods evaluated (ITDS, self report, and a combination in which a signal on either ITDS or self report indicated a fever), ITDS (OptoTherm Thermoscreen and FLIR A20 M) had the highest total sensitivity and specificity for fever detection. Correlations between ITDS measurements and oral temperatures were similar for the OptoTherm Thermoscreen (r = 0.43) and FLIR A20 M (r = 0.42), and significantly lower for Wahl HSI2000S (r = 0.14, p)0.001 for both comparisons).
Conclusion: When compared with oral temperatures, two ITDS (FLIR A20 M and OptoTherm Thermoscreen) were reasonably accurate in detecting fever and were better predictors of fever than self report. These findings may be particularly important in the context of travel in which fever may not be reported or cannot be measured using contact thermometers. In such settings, ITDS could provide an objective means for detecting fever as part of a comprehensive public health screening strategy.