Developing Evidence-Based Symptom Criteria for Testing

The research could lead to new protocols for determining who will be tested and how best to use testing resources.

One compelling finding of the study is that a symptom only recently added to the Centers for Disease Control and Prevention’s list of COVID-19 predictors ─ loss of taste or smell ─ had the highest level of correlation with an eventual positive test result.

Clemency is lead author on the study — titled “Symptom Criteria for COVID-19 Testing of Health Care Workers” — that was published May 12 in the journal Academic Emergency Medicine. 

The limited availability of testing resources nationally and a lack of consensus on who should be tested have been major challenges to health care systems.

Under-testing of patients with COVID-19 can lead to further spread of the disease. Over-testing of patients who do not have COVID-19 potentially takes time, money and resources away from patients who need them most.

The goal of the UB study was to develop an evidence-based symptom criteria for COVID-19 testing, based on the probability of positive test results, in order to facilitate appropriate use of testing resources.

Little of the existing literature covers ways to identify COVID-19 in people with early or mild symptoms, which constitutes the majority of cases. 

The UB study looked at which COVID-19 symptoms were predictive of positive test results among a group of health care workers tested through the testing program run by Kaleida Health, Western New York’s largest health care system.

The research consisted of an observational study of COVID-19 testing of 961 health care workers.

They were asked about the presence of 10 symptoms:

• fever
• fatigue
• dry cough
• loss of appetite
• muscle pain
• difficulty breathing
• coughing up phlegm
• sore throat
• diarrhea
• loss of taste or smell

Subjects were then tested using nasopharyngeal or oropharyngeal swabs at drive-through testing sites.

The symptom data was then compared to test results to find links between specific symptoms and an eventual positive test result.

In the test group, 225 subjects — 23 percent — had positive COVID-19 test results.

“The data analysis showed that the loss of smell or taste, which has only recently been included in the list of COVID-19 symptoms, was the symptom with the highest likelihood of a positive result,” Clemency says, adding that fever was also predictive of positive test results.

Fever or loss of taste or smell was reported by 61 percent of the heath care workers and 89 percent of the health care workers with positive COVID-19 tests in the study. 

If a health care worker in the study had neither of these findings, the chance they would test positive for COVID-19 was only 7 percent.

This, Clemency says, indicates that “an evidence-based approach to COVID-19 testing that at least includes fever and loss of taste or smell should be utilized when determining which health care workers should be tested.”

The data in the study were collected from Kaleida Health between March 26 and April 16.

According to the Erie County Department of Health, on March 26 the county had 166 confirmed cases of COVID-19 and two confirmed deaths. Three weeks later, when the study ended, the county had 1,951 confirmed cases and 115 confirmed deaths.

“When adequate testing resources are available, a testing criteria based on fever, cough, difficulty breathing and/or loss of taste or smell is a reasonable approach and is superior to a testing criteria based on only fever, cough and/or difficulty breathing,” says Clemency, a physician with UBMD Emergency Medicine.

“When testing resources are limited, a strategy of testing based on fever and/or loss of taste or smell should be considered.”

Co-authors on the study from the Department of Emergency Medicine are:

• Brian Ludwig, DO, a trainee in the emergency medicine residency program
• Robert F. McCormack, MD, professor and chair
• Danielle K. Scheafer, DO, a trainee in the emergency medical services fellowship program
• Renoj P. Varughese, MD, a trainee in the emergency medical services fellowship program
• Jacob V. Welch, research associate

Other UB co-authors are:

• Chang-Xing Ma, PhD, associate professor of biostatistics in the School of Public Health and Health Professions
• Nan Nan, a student in the biostatistics doctoral program
• Thomas A. Raab, MD, clinical professor of medicine

Registered nurse Theresa Giambra, Department of Medicine, Catholic Health System, is also a co-author.