COVID-19 Mortality Risk Score Prediction Strategies

Does a new model allow us to predict mortality in those hospitalized with COVID-19?

Can we predict mortality with a risk score approach? Understanding those predictive features of patients admitted to a hospital as a result of a SARS-CoV-2 infection can have far-reaching benefits for their treatment. Moreover, as we learn more about long-haul COVID-19, could it be possible to identify those at risk for more severe disease and work to lessening it?

A new research study evaluating patients across 260 hospitals in England, Scotland, and Wales, sought to test a predictive mortality risk score for those patients admitted to hospitals with COVID-19. The study, coined the International Severe Acute Respiratory and emerging Infections Consortium (ISARIC) World Health Organization (WHO) Clinical Characteristics Protocol for the UK, was performed by utilizing a cohort of patients recruited from February 6 to May 20 of this year, for model training. A second cohort of patients helped validate this model after it was development, from May 21 to June 29.

Such patients were those aged at least 18 years and admitted to a hospital with COVID-19 a minimum of four weeks prior to the final data extraction. Evaluation of in-hospital mortality was the main outcome of measure in this study performed by ISARIC Coronavirus Clinical Characterisation Consortium- ISARIC 4C.

The authors noted that “demographic, clinical, and outcome data were collected by using a prespecified case report form. Comorbidities were defined according to a modified Charlson comorbidity index. Comorbidities collected were chronic cardiac disease, chronic respiratory disease (excluding asthma), chronic renal disease (estimated glomerular filtration rate ≤30), mild to severe liver disease, dementia, chronic neurological conditions, connective tissue disease, diabetes mellitus (diet, tablet, or insulin controlled), HIV or AIDS, and malignancy.”

In the first cohort derivation dataset there were over 35,000 patients with a mortality rate of 32.2%, whereas there were over 22,000 in the validation group, with a mortality rate of 30%. The use of the 4C Score showed what the authors noted as a “high discrimination for mortality (derivation cohort: area under the receiver operating characteristic curve 0.79, 95% confidence interval 0.78 to 0.79; validation cohort: 0.77, 0.76 to 0.77)”.

The research team found that the predictive model revealed that patients with a score of 15, which was roughly 19%, had a 62% mortality compared to the 1% of those with a score of 3% or less. Ultimately, this risk stratification score is not only quite pragmatic, but utilized easy to employee hospital parameters, such demographics and comorbidities that are already collected. Variables like increased age, is widely known as one that increases risk for mortality with COVID-19, which is just one of many variables utilized.

The research team emphasized that this approach varies from existing tools due to the combined use of variables but also their weighting approach. Variables such as chronic neurological conditions, renal disease, and various other comorbidities will increasingly require awareness and research into their role in COVID-19 mortality. Predictive models that are easy to employ only serve to improve medical provider awareness, but also targeted prevention efforts. Enhancing these strategies and raising awareness in those communities will be pivotal for reducing not only transmission, but also more apt medical intervention.