Long-wavelength infrared (IR) detectors have a more than 20-year history in medical thermometry and have been used widely for febrile screening. However, over the past year and a half, public health entities, medical professionals, and the general public have begun to question the claimed accuracy of non-contact body thermometry. The standard assessment of a device's performance relies on clinical testing with febrile individuals, yet this practice may have inadvertently allowed the approval of IR systems that are unable to detect moderate fevers. The ability to test device performance without relying on febrile test participants would have important ramifications for public health, especially if this test discovered undisclosed differences in accuracy in widely used devices.

The aim is to examine the effect of the local environment and the physiology of the human body on the relationship between core body temperature and inner canthi (region near tear duct) skin temperature measured using non-contact thermal imaging and to use of this relationship to test actual device performance at detecting simulated elevated temperatures, without requiring volunteers having actual elevated temperatures. The overall goal of this research study is to validate and improve the science of non-contact core body temperature measurement.

It is difficult to assess the performance of non-contact thermometry at detection of febrile subjects unless febrile subjects are included in the test, which severely limits development and independent testing of such devices. The underlying mechanisms of non-contact thermometry might be used to develop a test without requiring the inclusion of febrile subjects, and the development and demonstration of such a test is the object of this study.

Literature review and pilot study revealed the offset between skin temperature and core (oral) body temperature is dependent on air temperature by a linear relationship. In a 70F room, the inner canthi temperature of a human with core (oral-reference) body temperature of 98.6F is 94.4F, a difference of 4.2F. This same individual in an 87F room (after waiting several minutes) will have an inner canthi temperature of 97.0F, a difference of only 1.6F (the actual values may depend on sensor confounds not addressed here). Therefore, one could use a several-minutes exposure to elevated air temperatures (equilibration environment) to simulate an elevated body temperature, by continuing to operate the device under test from a non-elevated air temperature environment (test environment), while taking care to limit the effects of temperature changes and gradients due to the mixing of these two environments.

Subjects will be recruited from a college campus via a recruitment protocol to participate in an up to 2 hour study session. The study session will involve collection of demographic information on a non-identifiable form, five collections of oral thermometry, four 10 minute periods of equilibration inside 4 elevated temperature environments during repeat continuous non-contact thermometry, and measurements via several non-contact test devices operated from within a non-elevated environment through a window into each elevated environment.

Inclusion Criteria:

• Age > 18.
• Have received a clinical oral thermometry measurement in the past.

Exclusion Criteria:

• Contraindication to oral temperature measurement.
• Inability to follow instructions.
• Any diagnosis associated with abnormal core thermoregulation, e.g. autonomic dysfunction.
• Obvious signs of tissue damage present in either inner canthus.
• Facial artery abnormality if previously diagnosed.
• Experienced discomfort while receiving most recent clinical oral thermometry measurement.

Participants will be selected from those responding to electronic advertisement on St Olaf College internal email lists and message boards. It is expected that the majority of the study population will be students, with a small proportion of Staff and Faculty.

No identifying information was collected except for thermal images of faces, which were extracted to only numeric inner canthi temperature values per subject. This data is being shared via a github repository.

The first set of 28 subjects' data is available as of Jan 24, 2022