Association between delirium prediction scores and days spent with delirium

Purpose: To determine the correlation and discriminative value of the E-PRE-DELIRIC and PRE-DELIRIC scores with delirium exposure to evaluate the prognostic value of both models.

Methods: A secondary analysis of a randomized clinical trial enrolling 1506 delirium-free, critically ill adults with an anticipated ICU stay of ≥2 days. Days spent with delirium (≥1 positive CAM-ICU) or coma (≥1 RASS ≤-4) in the 28-days after ICU admission were calculated. Patients were categorized into four groups: no delirium, short-exposure (1 delirium day), moderate-exposure (2-5 delirium days), and long- exposure (≥6 delirium days) to determine the correlation and discriminative value of the E-PRE-DELIRIC and the PRE-DELIRIC with days spent with delirium.

Results: The correlation between the overall E-PRE-DELIRIC and PRE-DELIRIC scores and days spent with delirium were: R = 0.08 (P = .005) and R = 0.26 (P < .001), respectively. The correlation between both prediction scores and days spent with coma or delirium were R = 0.21 (P < .0001) and R = 0.46 (P < .0001), respectively. The highest Area Under the Receiver Operating Characteristic for both E-PRE-DELIRIC [0.57 (95% CI:0.51-0.62)] and PRE-DELIRIC [0.58 (95% CI:0.53-0.62)] was found in the long delirium exposure group.

Conclusion: The E-PRE-DELIRIC and PRE-DELIRIC model each poorly correlate and discriminate with days spent with delirium in the 28 days after ICU admission.