Intensive Care Decision Making: Using Prognostic Models for Resource Allocation.

Area of Science:

• Critical care medicine
• Health services research
• Biostatistics

Background:

• Accurate clinical outcome prediction is vital for healthcare planning and resource management.
• Existing scoring systems like APACHE II and SAPS II are widely used but require validation in diverse populations.

Purpose of the Study:

• To evaluate the performance of APACHE II and SAPS II scoring systems in predicting mortality among Iranian patients.
• To assess the applicability of these models for clinical prioritization and quality control in Iran.

Main Methods:

• A prospective cross-sectional study involving 1799 patients across four tertiary care centers in Iran.
• Performance evaluation using Brier score, Area Under the Receiver Operating Characteristics Curve (AUC), and Hosmer-Lemeshow goodness-of-fit test.

Main Results:

• Observed mortality (24.4%) exceeded international rates, particularly within APACHE II categories.
• Both APACHE II (Brier score: 0.17, AUC: 0.745) and SAPS II (Brier score: 0.196, AUC: 0.751) showed acceptable discrimination.
• Neither model demonstrated adequate goodness-of-fit to the Iranian patient data (H-L ρ < 0.01).

Conclusions:

• APACHE II and SAPS II show moderate predictive ability but do not fit the Iranian population well.
• Recalibration of these scoring systems is essential before their clinical application in Iran.
• Further research is needed to develop or adapt prediction models for the local context.