Predicting COVID-19 prognosis in hospitalized patients based on early status

Area of Science:

• Medical Prognostics
• Infectious Disease Epidemiology
• Health Resource Management

Background:

• COVID-19 is a leading cause of death, necessitating effective patient risk stratification.
• Efficient allocation of critical care resources (ventilators, ICU beds) is vital during surges.
• Accurate prognosis aids clinicians in managing patient care and difficult conversations.

Purpose of the Study:

• To develop and validate predictive models for COVID-19 patient mortality risk.
• To create tools that utilize readily available clinical data for rapid risk assessment.
• To support clinical decision-making and resource management in COVID-19 care.

Main Methods:

• Development of the PLABAC and PRABLE predictive models.
• Utilized patient age and five common laboratory test results as input variables.
• Validated model generalizability across diverse external patient populations in the US.

Main Results:

• The PLABAC and PRABLE models accurately predict COVID-19 patient risk of death.
• Models require only basic demographic and laboratory data for prognosis.
• Demonstrated consistent performance and generalizability in external validation cohorts.

Conclusions:

• PLABAC and PRABLE provide practical, efficient tools for assessing COVID-19 prognosis.
• These models can serve as real-time aids for clinicians discussing patient outcomes.
• The models facilitate better communication and resource allocation for COVID-19 patients.