Early Oxygen Treatment Measurements Can Predict COVID-19 Mortality: A Preliminary Study

Area of Science:

• Medical research
• Clinical informatics
• Pulmonology

Background:

• Arterial hypoxemia is a critical challenge in COVID-19 patient care.
• Current methods assess oxygenation at single time points, lacking predictive power for mortality.
• The predictive value of prolonged oxygen supplementation on COVID-19 mortality requires investigation.

Purpose of the Study:

• To develop a predictive model for hospital mortality in COVID-19 patients.
• To utilize data on oxygen supplementation modalities for mortality prediction.
• To enhance clinical decision-making for high-risk COVID-19 patients.

Main Methods:

• Analysis of hospitalization data from 545 COVID-19 patients.
• Initial assessment of oxygen supplementation modalities in 182 patients.
• Development of a Random Forest prediction model using clinical and blood test data.

Main Results:

• Patients requiring reservoir oxygen masks within the first two days had a 75% mortality rate.
• Patients not requiring oxygen had a 4% mortality rate.
• The Random Forest model achieved AUC scores between 81.3% and 83.0% for mortality prediction.

Conclusions:

• Oxygen supplementation patterns are strong predictors of COVID-19 mortality.
• The developed model can aid in identifying high-risk patients.
• Optimized treatment and management strategies can be implemented based on these predictions.