Time-dependent error in the APACHE II scoring system

Area of Science:

• Critical Care Medicine
• Health Services Research

Background:

• The APACHE II score is a widely used severity of illness scoring system in intensive care units (ICUs).
• Accurate prediction of patient outcomes is crucial for resource allocation and clinical decision-making in critical care.

Purpose of the Study:

• To evaluate the predictive accuracy of the APACHE II scoring system in different intensive care unit settings.
• To assess how the duration of patient stay in the ICU affects the predictive performance of the APACHE II system.

Main Methods:

• Retrospective analysis of 189 patients from Wanganui ICU and 194 patients from Harare ICU.
• Comparison of actual versus predicted outcomes using APACHE II scores.
• Stratification of patients based on length of ICU stay (less than six days vs. six days or more).

Main Results:

• The APACHE II system demonstrated varying predictive accuracy based on patient length of stay.
• Predictive error was significantly higher for patients with longer ICU stays (38%) compared to shorter stays (15%, P < 0.01).
• A statistically significant increase in predictive error was observed with extended ICU duration.

Conclusions:

• The predictive accuracy of the APACHE II scoring system diminishes with longer patient stays in the intensive care unit.
• Clinical interpretation of APACHE II scores should consider the duration of ICU admission.
• Further research may be needed to refine scoring systems for prolonged critical illness.