Demographics, clinical evaluations, and basic clinical test results can be used to categorize patients who are at a higher risk of contracting coronavirus disease 2019 (COVID-19) and should, therefore, be prioritized for testing and isolation, according to a study published in Clinical Infectious Diseases.
The aim of this retrospective, case-controlled study was to create a model for use in care facilities with limited resources to assess the epidemiologic and clinical risk factors correlated with COVID-19. Data for this study came from the National Centre for Infectious Diseases treatment facility in Singapore. Patients were self-referred, referred by another medical care facility, or referred through contact tracing efforts from a national agency. Demographics, comorbidities, symptoms, exposure risk, clinical evaluation variables, and blood laboratory results were collected on all patients. The risk score model grouped variables into exposure risk factors, demographic variables, clinical findings, and clinical test results, and Akaike’s Information Criterion was used to calculate a prediction score.
Of the 788 patients included in the study, 54 patients tested positive for COVID-19, and 734 patients were included as controls. Of the positive cases, the median age was 42 years, 53.7% were men, and 88.9% were Chinese. Of the negative cases, the median age was 34 years, 51.7% were women, and the majority were Singapore citizens (52.5%) or Chinese nationals (18.4%). Positive cases were more likely to be older individuals (P <.001), to have had contact with a known patient with COVID-19 (P <.001), have an elevated temperature (P =.003), and have radiologic findings suggestive of pneumonia, as well as lower counts of white blood cells, platelets, neutrophils, lymphocytes, eosinophils, and basophils (all P <.001). Positive cases were not more likely than controls to have comorbidities.
Four multivariable models were created to assess covariate risk estimates. Model 1 was the only model to use the “traveled to Wuhan, China” variable; the other models excluded exposure-associated risk factors. Elevated temperature was the strongest predictor across all models; gastrointestinal symptoms, elevated respiratory rate, and absence of sore throat and of sputum production were strong predictors in the models in which they were selected. Radiologic evidence of pneumonia and blood parameters of low neutrophil and eosinophil counts were the most predictive clinical tests results.
Results also demonstrated model 1 showed that incorporating all easily ascertainable data at presentation for COVID-19 testing performed exceptionally well. The performance of model 2, which removed exposure risk factors, suggested that clinical findings and tests can identify people at high risk for COVID-19. Furthermore, exclusion of radiologic evidence of pneumonia in model 3 did not significantly affect the performance. However, model 4 showed that when basic blood test results, such as complete blood count were excluded, predictive accuracy was reduced substantially.
Limitations of this study included the limited data set and the need for further validation.
The researchers concluded “[p]rediction models which include rapidly ascertainable clinical findings and clinical tests, especially basic blood tests, have sufficient predictive value to identify individuals with a higher probability for COVID-19 and should be considered to stratify at-risk populations for laboratory testing (where available), isolation and contact tracing measures.”
Disclosure: Several study authors declared affiliations with the pharmaceutical industry. Please see the original reference for a full list of authors’ disclosures.
Sun Y, Koh V, Marimuthu K, et al. Epidemiological and Clinical Predictors of COVID-19 [published online March 25, 2020]. Clin Infect Dis. doi:10.1093/cid/ciaa322