Owner system

A validated risk stratification system for managing COVID-19 outpatients

A rampant increase in the number of coronavirus disease (COVID-19) cases has made it difficult for healthcare providers around the world to manage the overwhelming number of patients arriving at hospital outpatient departments.

Although several scoring systems have been developed to analyze mortality risks in critically ill, hospitalized individuals, no such measure to monitor the severity of illness in outpatients has been developed so far to help treatment and management decision-making processes. .

To study: Simple risk scores for predicting hospitalization or death in outpatients with COVID-19. Image Credit: janews / Shutterstock.com

Current COVID-19 Scoring Systems

Scoring systems such as the 4C Risk Score, ABCS Risk Score, and COVID-GRAM Risk Score typically require laboratory testing and imaging. Therefore, these types of facilities are not always available in all outpatient or telehealth settings.

The COVID-NoLab Risk Score for Inpatient Mortality is an exception that only requires oxygen saturation, age, and respiratory rate with no imaging or lab testing needed; however, this scoring system is not available to outpatients.

The OutCoV score predicts the likelihood of hospitalization in outpatients without the need for laboratory tests and takes into account various factors, including age, fever, dyspnoea, hypertension and chronic respiratory disease. Unfortunately, this scoring system has not been externally validated.

In an effort to overcome the limitations of current COVID-19 scoring systems, a group of researchers developed a risk stratification system from existing scoring methods that are exclusively for outpatients. To this end, the researchers collated a dataset with patients who had been diagnosed with COVID-19 at any Lehigh Valley Health Network (LVHN) outpatient or ExpressCARE primary care clinic and discuss their findings in a recent medRxiv* pre-publication study.

About the study

The researchers wanted to prospectively validate the OutCoV and COVID-NoLab risk scores for predicting hospitalization or death in outpatients. By doing so, they hoped to develop new risk scores based on simple measurable parameters without the need for imaging or laboratory testing.

Outpatient data was collected from electronic health records of primary care and ExpressCARE outpatient clinics at LVHN who had tested positive for COVID-19 by polymerase chain reaction (RT-PCR) between March 13, 2020 and September 30, 2021.

A total of 13,418 outpatients diagnosed with COVID-19 were considered for analysis and, after excluding patients under the age of 12 and patients with missing data, the final dataset consisted of 9,649 outpatients with COVID-19.

The early cohort with data obtained before March 1, 2021 included 5,843 patients, while the late cohort included data obtained after March 1, 2021 and included a total of 3,806 patients. These cohorts were also called early and late cohorts, respectively.

Study results

A total of 89 of 641 (13.9%) patients were hospitalized on the same day as their outpatient visit. A total of 641 patients were hospitalized, 55 of whom died. No non-hospitalized patient died following a diagnosis of COVID-19.

The overall probability of hospitalization or death was lower in the late cohort than in the early cohort at 5.5% and 7.4%, respectively. Increasing age, increasing respiratory rate, decreasing oxygen saturation, complaint of dyspnea and all comorbidities were associated with an increased likelihood of hospitalization.

The researchers used three regression models that assumed that oxygen levels, respiratory levels, or both could be missing among the top five data points collected. The precision of these three data points was then assessed, which then led to the identification that the AUC was between 0.772 and 0.785, indicating good precision. The researchers also found hospitalization rates of 12/1,199 (1.0%), 23/519 (4.4%), 32/259 (12.4%) and 15/49 (30.6%) ) in the very low, low, moderate and high categories. -risk groups, respectively, in the late cohort.

OutCoV and COVID-NoLab risk scores in the late cohort identified more patients in low-risk groups. This group also had higher hospitalization rates, 3.8% to 4.0% in the early cohort and 2.8% in the late cohort, compared to the new models developed by the researchers.

Consequences

In the validation group, the new risk scores developed by the researchers identified a very low risk group that included 53% to 57% of the general population with a lower probability of hospitalization of 1.7%. or less. This implied that these patients could potentially be initially managed on an outpatient basis with advice to contact their primary care physician if symptoms worsen.

The low risk had a probability of hospitalization of 5.2% to 5.9%, which is similar to that of the population as a whole. About one in six patients were classified in the moderate risk category with 14.7% to 15.6% chance of hospitalization or in the high risk groups with 32.0% to 34.2% chance of hospitalization . These patients needed advice on regular oxygen monitoring if they were being cared for on an outpatient rather than an inpatient basis.

Such models, if implemented in practical scenarios, would help stratify risk in patients and reduce unnecessary burden in hospitals. This could lead to more efficient management of the health system and allow these services to remain available only for emergencies.

*Important Notice

medRxiv publishes preliminary scientific reports that are not peer-reviewed and, therefore, should not be considered conclusive, guide clinical practice/health-related behaviors, or treated as established information.