Design, Implementation, and Validation of an Automated, Algorithmic COVID-19 Triage Tool

Area of Science:

• Medical Informatics
• Public Health
• Digital Health

Background:

• The COVID-19 pandemic necessitated rapid development of tools for patient assessment.
• Algorithmic triage tools offer potential for efficient and scalable patient management.

Purpose of the Study:

• To design, implement, and validate an online, publicly available tool for algorithmic triage of patients with symptoms similar to SARS-CoV-2.
• To compare the tool's triage assignments with clinician assessments and evaluate patient outcomes.

Main Methods:

• A chart review was conducted on patients using the triage tool and contacting a phone triage hotline.
• Data collected included tool- and clinician-assigned triage codes, demographics, COVID-19 test results, and healthcare utilization.
• Concordance, down-triage, and up-triage rates were calculated.

Main Results:

• Over 30,000 triage tool completions were analyzed from over 20,000 unique patients.
• The tool assessed 51.7% asymptomatic, 15.6% low, 21.7% moderate, and 11.0% high severity.
• Concordance between tool and clinician was 29.2%, with a 70.1% down-triage rate and minimal up-triage.

Conclusions:

• The online COVID-19 triage tool demonstrated safe patient triage capabilities when compared to clinicians.
• This validates the potential for automated triage systems to improve patient experience and access to care.
• The approach may be applicable to triaging other medical conditions beyond COVID-19.