Prognostic Accuracy of Delirium Prediction Models

Area of Science:

• Critical Care Medicine
• Neurology
• Geriatrics

Background:

• Delirium is a common complication in intensive care unit (ICU) patients.
• Accurate prediction of delirium is crucial for timely intervention and improved patient outcomes.
• Existing prognostic models for delirium in the ICU have shown variable accuracy.

Purpose of the Study:

• To evaluate the prognostic accuracy of different models for predicting delirium in ICU patients.
• To identify the best-performing model for delirium prediction in this population.
• To assess the generalizability of the best model across different patient subgroups, including those on mechanical ventilation.

Main Methods:

• Systematic review and meta-analysis of studies evaluating delirium prediction models in ICU settings.
• Comparison of the predictive performance (e.g., AUC, sensitivity, specificity) of various models.
• Subgroup analyses based on mechanical ventilation status.

Main Results:

• The prognostic accuracy of models for predicting ICU delirium varies significantly.
• The PRE-DELIRIC model demonstrated superior performance compared to other evaluated models.
• The PRE-DELIRIC model's accuracy was consistent in both patients receiving and not receiving mechanical ventilation.

Conclusions:

• The PRE-DELIRIC model is a reliable tool for predicting delirium in ICU patients.
• This model can be effectively applied to a broad range of ICU patients, irrespective of their ventilation status.
• Implementation of the PRE-DELIRIC model can aid in early detection and management of delirium in critical care settings.