Automated Online Triage System Helps Determine COVID Severity

Editor's note: Find the latest COVID-19 news and guidance in Medscape's Coronavirus Resource Center.

To provide the most appropriate care for people with symptoms of COVID-19 ― and to minimize the number of patients with milder cases coming to acute care centers ― researchers at the University of Pennsylvania, in Philadelphia, Pennsylvania, devised an online, automated triage system to help initially manage patients remotely.

How and why they developed the tool, which is publicly available, and how well triaging COVID-19 patients with the technology accords with how physicians and nurses triage the same patients are described in an article published November 3 in Applied Clinical Informatics.

The technology is not meant to replace practitioner judgment "but rather to offer a way to safely offload clinical volume," co–lead study author Elana Meer, a researcher and an MD/MBA student at the Perelman School of Medicine and the Wharton School of the University of Pennsylvania, told Medscape Medical News.

"We hope that by improving our screening and triage mechanisms, we would actually help physicians have more time to spend with patients that need the physician's knowledge, experience, and clinical judgment," she said.

Meer and her colleagues found that triage using the tool and triage by practitioners were in accord 29% of the time. But that's not the full story.

The technology was designed to be conservative and to err on the side of patient safety. "We wanted a false negative rate ― meaning the chance that the tool would suggest a patient stay home when in reality they needed to seek care to be basically zero," Meer said.

In 70% of cases, the symptom checker indicated higher clinical severity than did the clinician who later managed the patient, she said. In these cases, the practitioner was able to safely "down-triage" the patient to a lower level of COVID-19 severity.

"From our standpoint, this was a win," Meer said. "The symptom checker successfully flagged almost all patients of concern and made reasonable and appropriately conservative referrals to expert clinicians in over 99% of cases."

Only six patients were "up-triaged" by a physician or nurse to a more severe category.

The study included 782 people who completed using the online triage assessment and later had a consultation with a practitioner via a traditional phone triage line.

A Good First Step

There is a "huge need for an accurate triage tool," and this is a "good beginning," Theresa Cullen, MD, FAMIA, American Medical Informatics Association board member and public health director from Pima County, Arizona, said when asked to comment on the study.

"This tool can be perceived as an initial step in developing an accurate triage tool," she added.

Dr Theresa Cullen

Cullen pointed to the 29.2% concordance rate between the triage tool and clinical assignment. "The down-triage rate was 70.1%. Ideally, there would be closer concordance between the tool and the provider," she said. Down-triaging "would imply that the individual was asked to seek a higher level of care than might be needed."

She noted that if 70% of patients who were triaged to a higher level of care than was needed sought such care, it could strain resources and have an overall "potential negative impact on accessibility to clinical care."

Majority of People Asymptomatic

In the study, the comparison was limited to people whom the triage tool determined had a moderately severe level of COVID-19 illness. People in this group were instructed to call the phone triage line at the University of Pennsylvania.

People whom the tool indicated had disease of a low or high level of severity received other instructions. For example, those classified as having severe disease were instructed to call 911 or seek emergency care in person.

There were 30,321 completions of use of the triage tool by 20,930 unique patients between its launch on May 4, 2020, and January 31, 2021.

Of the 30,321 completions, the tool classified 51.7% of people as being asymptomatic, 15.6% as having disease of low severity, 21.7% as having disease of moderate severity, and 11.0% as having disease of high severity.

Not Your Typical Online Symptom Checker

The tool differs from online symptom checkers, Meer said.

"Internet resources can often triage patients' needs inaccurately, generating worry among patients and unnecessary visits to the emergency room or hospital," she said.

By contrast, well-designed triage tools "could facilitate the patient receiving care from the right provider at the right time from the right location, or appropriately managing their symptoms at home."

The online tool, along with a questions page, includes an automated, online chatbot that can ask further questions of the patient.

The technology saves information using a unique code, so when a patient subsequently calls or visits the health system, practitioners do not need to start from scratch.

About 65% of the online triage users were female, and about 60% were younger than 40 years.

Cullen would have liked to have seen a greater number of older patients participate in the study. She pointed out that "only 13% of patients were over 61, even though a high risk for COVID-19 morbidity and mortality is increasing age, so utilization of this tool with those over 65 would be important."

The researchers note that "racially, demographics of the tool's users nearly matched those of the Philadelphia metropolitan area, which was reassuring in the context of concerns that lack of access to internet resources may disproportionately impact people of color."

Triaging Other Patients Possible

The findings of this study could apply beyond COVID-19, the researchers note. "Similar triage tools could really be applied in any clinical setting," Meer said. "For example, a tool based on a similar premise was applied to the ophthalmology department to help triage patients presenting to acute care clinic with concerns."

Rigorously validated triage tools can offer multiple advantages, she said. "Patients may benefit by gaining immediate access to accurate, vetted information. Health systems may benefit by offloading call centers and triaging patients so providers can focus their time, which is becoming increasingly limited, on patients who are most likely to benefit from a conversation or their care."

"More work is needed," Cullen added, "and more science knowledge that we are gaining on a regular basis."

Appl Clin Inform. Published online November 3, 2021. Abstract

Damian McNamara is a staff journalist based in Miami. He covers a wide range of medical specialties, including infectious diseases, gastroenterology and neurology. Follow Damian on Twitter: @MedReporter.

For more news, follow Medscape on Facebook, Twitter, Instagram, and YouTube.

Comments

Commenting is limited to medical professionals. To comment please Log-in.

Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.