Temperature Checks Are Everywhere Now. They Can Be Wildly Inaccurate.

Temperature check stations are everywhere now—concert venues, bars, airports—and you might think that clocking in with a fever at one of them would mean you should go take a COVID test, or at the very least, exit the premises. But Brooklyn federal court officers apparently have a different approach: instruct the person who received a too-high temperature reading to bask in the cool winds of an air conditioner vent for a few minutes, then reenter, according to a New York Daily News reporter who experienced this method himself.

On the other end of the thermometer, while checking in for a doctor’s appointment, Slate senior managing editor Megan Wiegand clocked in at a cool 94 degrees—within the hypothermic range—even after a sweaty bike ride. Unconcerned, the receptionist handed over insurance paperwork and waved her through.

Temperature checks have become popular as an easily implemented, quick measure to combat the spread of COVID by detecting individuals with a fever. But as these anecdotes illustrate, temperature checks are not a very solid screening method for infections. While the Occupational Safety and Health Administration suggests that “temperature screening may play a part” in COVID detection, and the Centers for Disease Control and Prevention classifies them as an “optional strategy” for combating the virus, many public health experts are against temperature screenings entirely. “The temperature check is of no value,” Eric Topol, executive vice president of Scripps Research, told Popular Science. “It should be abandoned.” Bruce Aylward, senior adviser at the World Health Organization, told the L.A. Times that temperature checks are “the bane of my existence.”*

A main critique of the temperature check is that someone actively contagious with COVID-19 is hardly guaranteed to have a fever, given that the virus is often transmitted in the absence of symptoms. (One study estimated that just 4.6 percent of people with the virus “are likely to have a fever while contagious.”) But another issue is that temperature fluctuates—or appears to fluctuate—for a number of reasons that have nothing to do with illness at all.

First, the temperature threshold used to determine if a patient has a fever—100.4 degrees Fahrenheit—may be inaccurate for some people. What constitutes a “normal” temperature for a human being is a little fuzzy in the first place. The figure of 98.6 was established in the 1800s, while more recent research suggests that the humans of today are running around one degree cooler. But there is significant variation: Older adults tend to have lower temperatures; bodies in general might be one or two degrees warmer in the afternoon than in the morning. Air temperature can drive up body temperature, as can exercise. “Normal” body temperature is individual, and also more of a range than a single number. This makes using fever to screen for potential COVID-19 infections all the more iffy. An analysis of data gathered from the ZOE COVID Study app found that COVID cases among the elderly—a high-risk group—may have gone underdetected by temperature checks, because fevers in that age group are often lower than the 100.4 degree threshold.

And then, there can be issues with the thermometer itself. Non-contact infrared thermometers, or NCITs, have become the go-to thermometer for COVID screening. The small, hand-held, gunlike devices don’t require direct contact with the patient/shopper/restaurantgoer, just a quick scan of the forehead. They are less likely to transmit disease and are better suited to testing large numbers of people quickly (just imagine having to get a communal thermometer placed under your tongue before entering a place of business). However, the limitations to these devices are significant. These thermometers are measuring skin temperature, which can fluctuate independently of the core body temperature. NCIT readings can be thrown off by anything that warms or chills the skin: direct sunlight, drafts, and even sunburn. Sweat, the body’s way of cooling the skin, can lead to a misleadingly low temperature reading, (as may have been the case for our 94-degree Slatester, who, so far, has not died of hypothermia). The list of factors that complicate NCIT temperature readings is long and potentially incomplete, as there have been few comprehensive studies on the devices. The only sure conclusion from the few studies on NCITs that exist is that “Evidence for the accuracy of infrared skin thermometers is equivocal.”

So, while an employee armed with an infrared thermometer in the doorway to a building is a good way for a business to signal that they are taking COVID seriously, it’s OK to take the readout with a grain of salt—or at least wait a few minutes and try getting scanned again.