US states are instituting temperature checks at offices and restaurants, but a quarter of coronavirus patients don't develop a fever

• The majority of coronavirus patients develop a fever, according to research from China and the US.
• But using thermometers or thermal cameras to check people's temperature isn't a foolproof way to identify cases.
• Temperature checks don't identify patients who are asymptomatic, pre-symptomatic, or experiencing symptoms other than a fever.

Since early in the coronavirus pandemic, airport screeners have relied on temperature checks to detect potential cases among travelers. The method is also now being used in restaurants, stores, hospitals, theme parks, and office buildings as states relax lockdown restrictions across the US.

Amazon, for example, has been using thermal cameras to temperature check more than 100,000 employees per day at warehouses since April 2. Anyone with a temperature above 100.4 degrees Fahrenheit is sent home and asked not to return until the fever has been gone for three days. Disneyland also plans to institute temperature screenings for visitors when the theme park reopens on July 17.

The US Centers for Disease Control and Prevention lists temperature checks as an "an optional strategy" for workplaces to decrease the risk of outbreaks. So individual states have instituted a variety of policies and recommendations with regard to the practice. Temperature checks are mandatory at gyms and fitness centers in Utah, as well as in manufacturing facilities in Michigan, restaurants in New Jersey, and retail stores in Washington. Colorado, Kentucky, and Vermont require daily temperature checks for all employees going to work, while other states merely recommend that. Some Southern states — including Mississippi, Georgia, and Arkansas — don't require temperature checks for workers.

But fever screenings still miss a large swath of infections.

Anthony Fauci, director of the National Institute of Allergy and Infectious Disease, estimated in April that between 25% and 50% of people infected with the coronavirus never show symptoms. More recently, the CDC estimated the portion to be around 35%. Plus, a person's symptoms generally don't appear for two to 14 days after their exposure to the virus. This leaves a window during which a person can be contagious but still have a normal body temperature.

And of course, not all coronavirus patients develop a fever. That means temperature checks are far from foolproof.

A quarter of US hospitalized cases didn't have a fever

According to one of the largest studies of coronavirus patients' symptoms — a February report from the World Health Organization — around 88% developed a fever out of nearly 56,000 laboratory-confirmed cases in China. The virus' two other main symptoms, cough and shortness of breath, were less common: Coughs developed among 68% of patients, and around 20% of patients experienced shortness of breath.

However, a preliminary, internal Trump administration report obtained by the New York Times found that a fever wasn't as ubiquitous among hospitalized patients in the US.

Out of more than 2,500 US patients hospitalized from March 1 to May 1, around 75% presented a fever and chills, according to the document. The same percentage of patients developed a cough, and 70% experienced shortness of breath.

Another study of 5,700 hospitalized patients in New York City found that only one-third of the patients developed a fever. This suggests that fevers may be weaker indicators of a coronavirus infection than scientists previously thought.

Jim Collins, a bioengineering professor at MIT, told Business Insider that temperature checks are "an easy-to-use means" of screening, but they have a "very low" capacity to identify most coronavirus cases.

"You can have a fever for a lot of reasons," he said. "Maybe you ran to work or you rode a bike and your temp goes way up, or you have any number of strains of flu or a cold or strep throat."

Medications like aspirin or ibuprofen can also lower a person's temperature, presenting another opportunity for an infected patient to slip through the cracks.

"Fever screening can be one part of a wider system, but it's still unclear how much marginal benefit there is," Amesh Adalja, a senior scholar at the Johns Hopkins Center for Health Security, recently told Health.com.

Thermometers aren't always accurate, but alternatives are limited

The effectiveness of temperature checks also depends on the device and the conditions under which it's used.

Chinese officials often use infrared thermometers, which can produce a false reading if they're too close or too far from a person's forehead. That's because the device measures surface temperature — the amount of heat emanating from a person's skin — as opposed to internal body temperature. As a result, a person's temperature reading may rise or fall if the check is performed outdoors in particularly hot or cold climates.

The device used by Amazon, thermal cameras, can offer more accurate readings from farther away, but they're more expensive and there's a risk of supply shortages, Reuters reported.

Collins' lab is working to create a more direct approach to identifying cases: a face mask with sensors that could detect coronavirus particles when a person speaks, coughs, sneezes, or breathes. But the product is still in the research phase and hasn't been manufactured yet.

For now, temperature checks offer a feasible method for identifying some infected people, but they're unlikely to catch every case.