This study presented a novel scoring system which can be used to predict progression of Covid–19. Identifying Covid–19 patients who are at a higher risk of developing severe symptoms in early stages can inform better medical resource allocation and patient care during massive outbreaks. We found age, CRP, LDH and hemoglobin as independent high-risk factors associated with progression of Covid–19 infection. Some of our findings are consistent with previous studies that identified different risk factors to be associated with poor clinical outcomes in patients with Covid–19.6,7,9,10 Older age, high Sequential Organ Failure Assessment (SOFA) score, and d-dimer greater than 1 µg/mL at admission were described as potential risk factors for mortality of in-hospital patients with Covid–19.6 Wang et al.9 reported older age, dyspnea, lymphopenia, comorbidities (e.g., cardiovascular disease), and acute respiratory distress syndrome (ARDS) as predictors of fatal outcomes in elderly Covid–19 patients. Age, sex, CRP, LDH, lymphocyte count, and features derived from CT images were most reported predictors of severe Covid–19 progression.10
To our knowledge, KDDH scoring system is one of few scoring systems to predict Covid–19 progression at early stage. Ji et al.11 created a CALL score model with age, comorbidities, lymphopenia and LDH to predict Covid–19 progression at early stage. This is a relatively simple scoring system that only require basic tests for laboratory parameters. However, it has limitations of smaller sample size and no validation. There exists difference in definition of progression to severe Covid–19. CALL model defines deteriorated chest radiologic findings in progression, but the present study only defines patients requiring oxygen therapy as progression.11 We did not include progression of radiologic findings without hypoxemia in the progression group, as deterioration of radiologic finding tends to be reflected later than clinical course. Gong et al.12 presented a prognostic nomogram based on seven factors (older age, higher LDH and CRP, direct bilirubin, red blood cell distribution width (RDW), BUN, and lower albumin) to identify patients who are likely to develop severe Covid–19 infection. The nomogram was conducted with multicenter patients and was validated. However, its scoring system shows limited applicability due to score model complexity and greater number of clinical parameters. Liang et al.13 developed a clinical risk score to predict risk of developing critical illness in Covid–19 patients, which was not associated with progression to severe pneumonia.
The KDDH scoring system has advantages of high sensitivity and specificity together with strong calibration that ensures no statistical difference between quintile of predicted risk. Moreover, this is uncomplicated and can predict progression to severe pneumonia by using a simple blood test that can be conducted in outpatient clinics.
High negative predictive value (98.1%) of the presented scoring system demonstrates efficacy in identifying low risk patients, who can be managed at other treatment facilities with minimal monitoring or through self-quarantine. High-risk group who are likely to require oxygen therapy can be assigned first to hospitals and receive priority care in early stage. With this scoring system, hospitals can prevent unnecessary overuse of medical care in limited resource settings (e.g., during massive outbreaks), and may reduce mortality through effective allocation of medical resources. During the peak of the Covid–19 outbreak in South Korea, several patient deaths occurred in their homes while waiting to be hospitalized after receiving Covid–19 diagnosis due to shortage of beds. By triaging patients using the KDDH scoring system, such mortality cases could be reduced and be prevented in future outbreaks.
This study has limitation of being a retrospective, single center cohort study conducted in Daegu, South Korea. However, this hospital treated the largest number of patients with Covid–19, thus its patient data is well-representative of the entire country. Moreover, it is the first study conducted outside of China to develop and validate a risk scoring model for Covid–19 patients. A substantial number of patients were enrolled and followed up for more than 4 weeks without any change observed in final patient outcomes, and the scoring system was well-validated.
In our experience, based on the KDDH scoring system, the low-risk patients received care in the mild patient ward where oxygen therapy was unavailable. The high-risk patients were admitted in the main ward where oxygen therapy and close monitoring could be provided. By concentrating medical personnel who can provide critical care at the main ward, we could treat the critically ill patents effectively.
In summary, the presented KDDH scoring system was validated to be a highly informative and successful risk stratification tool to identify Covid–19 patients at higher risk of progression to severe pneumonia. Early adoption of this scoring system can assist optimal usage of limited medical resources in different health facility settings that are undergoing rapid Covid–19 outbreaks.