COVID-19 mortality prediction models: How reliable are they in predicting death during the pandemic?

The calculator has been developed on the basis of clinical data and demographics gathered from seven weeks of COVID-19 patient care during the early days of the pandemic at five different hospitals in Washington DC.

Myupchar September 23, 2020 23:47:53 IST
content powered by
COVID-19 mortality prediction models: How reliable are they in predicting death during the pandemic?

Representational image

Research teams at Mount Sinai and Johns Hopkins have released two different prediction models to predict disease severity and chances of death in COVID-19 patients.

The one by Mount Sinai used machine learning and clinical dataset (over 3,800 patients) including three major clinical features including the age of the patient, minimum oxygen saturation and type of patient encounter (inpatient, outpatient or telehealth visits).

The study, published in The Lancet, suggests the model is highly accurate.

The team at Johns Hopkins have developed a COVID-19 risk calculator (available online) that include various COVID-19 risk factors including patient age, BMI, presence of chronic disease and lung health along with the vital signs of the patient and the symptom severity at the time of admission to the hospital.

The calculator has been developed on the basis of clinical data and demographics gathered from seven weeks of COVID-19 patient care during the early days of the pandemic at five different hospitals in Washington DC.

The Johns Hopkins study, published in the journal Annals of Internal Medicine, was a retrospective cohort analysis that included 832 patients.

Previous models

These two are not the only prediction models for disease severity and mortality in COVID-19. Over the last few months, researchers have proposed various mathematical models to predict the magnitude of the disease to aid in assessing disease severity, prognosis and need for diagnosis.

Some models are based on the SIR (susceptible, infected and recovered), while some are based on the SEIR (where the added E stands for exposed). Some use AI and online tools while some need more complex mathematical calculations.

Some give short term predictions based on the continuation of preventive steps while some give a more generalised prediction regardless of the preventive measures.

Though all of the prediction models have their own perks, experts have over and again questioned the credibility of these models for one reason or another.

Questions on reliability

A study, published in the journal Humanities and Social Sciences Communication, discussed three different prediction models, used by the UK, the USA and Australia to guide their health policies, and pointed out that all these models don’t clearly outline their uncertainties (for example, the projections change every time you input new data) and hence their predictions should be considered with caution.

The Centers for Disease Control and Prevention, USA is now using multiple prediction models to get an ensemble forecast.

A review article published in the Medical Journal Armed Forces of India, a peer-reviewed journal published by Elsevier, indicated that short term projections have more importance than long-term projections and the former need to be revised over and again as new data emerges.

A letter published in the European Respiratory Journal, a group of researchers from the UK and Netherlands mentioned 66 prediction models and suggested that all these models are prone to bias due to data quality, reporting and statistical analysis.

The authors of the study explained how missing data (patient information which is bound to be missing for all the factors needed to make a prediction) may affect the accuracy of the model and how artificially balancing data is not a good predictor of real outcomes.

For more information, read our article on COVID-19.

Health articles in Firstpost are written by myUpchar.com, India’s first and biggest resource for verified medical information. At myUpchar, researchers and journalists work with doctors to bring you information on all things health.

Updated Date:

also read

Italy going back to games with no fans in stadiums due to rising coronavirus cases
Sports

Italy going back to games with no fans in stadiums due to rising coronavirus cases

The new decree says professional games are allowed in outdoor venues “without fans in attendance.”

Study suggests aspirin may help lower risk of death from severe COVID-19 complications
Health

Study suggests aspirin may help lower risk of death from severe COVID-19 complications

Scientists stated that aspirin can reduce the production of immune cells in patients with pre-existing cardiovascular disease, thus helping in reducing the incidence of a cytokine storm