Objective:

To evaluate the predictive validity of the updated SOFA-2 score against 30-day mortality in ICU admissions and compare it with the original SOFA-1.

Approach:

• Validation Study: Independent external validation of SOFA-2 in nearly 30,000 ICU admissions.
• Comparative Analysis: Systematic comparison of SOFA-2's predictive validity against SOFA-1.

Key Findings:

• SOFA-2 showed a small discriminative advantage on ICU day 1, but the net reclassification improvement was minimal.
• The predictive validity of SOFA declined significantly over time, influenced by competing risks.
• Predictive validity varied significantly by case-mix, with trauma patients showing higher AUROC than sepsis patients.
• SOFA-2 components lacked manual validation, and the scoring system may obscure prognostic imbalances.

Interpretation:

The advantages of SOFA-2 are primarily observed on ICU day 1 and may not apply uniformly across different patient populations or severity levels.

Limitations:

• Reclassification rates may not reflect true prognostic improvements.
• SOFA-2's predictive validity is influenced by patient demographics and underlying conditions.
• Lack of manual validation for SOFA-2-specific components raises concerns about accuracy.

Conclusion:

While updating the SOFA score is conceptually valuable, the benefits of SOFA-2 appear limited and require cautious interpretation.

Sources:

• Helleberg et al.