Mass screening for fever in children: a comparison of 3 infrared thermal detection systems.

Selent, Monica U; Molinari, Noelleangelique M; Baxter, Amy; Nguyen, An V; Siegelson, Henry; Brown, Clive M; Plummer, Andrew; Higgins, Andrew; Podolsky, Susan; Spandorfer, Philip; Cohen, Nicole J; Fishbein, Daniel B

OBJECTIVES Infrared thermal detection systems (ITDSs) have been used with limited success outside the United States to screen for fever during recent outbreaks of novel infectious diseases. Although ITDSs are fairly accurate in detecting fever in adults, there is little information about their utility in children. METHODS In a pediatric emergency department, we compared temperatures of children (<18 years old) measured using 3 ITDSs (OptoTherm Thermoscreen, FLIR ThermoVision 360, and Thermofocus 0800H3) to standard, age-appropriate temperature measurements (confirmed fever defined as ≥38.0°C [oral or rectal], ≥37.0°C [axillary]). Measured temperatures were compared with parental reports of fever using descriptive, multivariate, and receiver operating characteristic analyses. RESULTS Of 855 patients, 400 (46.8%) had parent-reported fever, and 306 (35.8%) had confirmed fever. At optimal fever thresholds, OptoTherm and FLIR had sensitivity (83.0% and 83.7%, respectively) approximately equal to parental report (83.9%) and greater than Thermofocus (76.8%), and specificity (86.3% and 85.7%) greater than parental report (70.8%) and Thermofocus (79.4%). Correlation coefficients between traditional thermometry and ITDSs were 0.78 (OptoTherm), 0.75 (FLIR), and 0.66 (Thermofocus). CONCLUSIONS Compared with traditional thermometry, FLIR and OptoTherm were reasonably accurate in detecting fever in children and better predictors of fever than parental report. These findings suggest that ITDSs could be a useful noninvasive screening tool for fever in the pediatric age group.