Related Experiment Video
Updated: Aug 5, 2026
A Machine Learning Approach to Design an Efficient Selective Screening of Mild Cognitive Impairment
Published on: January 11, 2020
Early Risk Stratification of Delirium in Intermediate Care Units and Its Impact on 30-Day Mortality: A Decision-Tree
Fabrizio Lucente1, Lucia Filippi1, Arian Zaboli2
1Intermediate Care Unit, Department of Internal Medicine, Hospital Alto Vicentino (AULSS7), 36014 Santorso, Italy.
Background:
Delirium is an under-recognized complication in hospitalized acute care patients, associated with worse outcomes including increased length of stay, higher ICU admission rates and greater mortality. Although screening tools exist for early diagnosis, the absence of admission-based tools limits early risk stratification and timely prevention strategies.
Aim:
To identify early risk factors for delirium development in IMCU patients and quantify its prognostic role on 30-day mortality.
Methods:
In this prospective single-center study, 651 consecutive IMCU patients without delirium at admission were enrolled. Admission variables were analyzed using multivariable logistic regression and Classification and Regression Tree (CART) analysis to identify clinically relevant risk phenotypes. The association between delirium and 30-day mortality was assessed in adjusted models.
Results:
Delirium developed in 18.6% of patients within 96 h. Key independent risk factors included age (OR 1.04), male sex (OR 1.75), alcohol use disorder (OR 3.59), cognitive impairment (OR 3.35), COPD (OR 1.68), NEWS (OR 1.18), and need for NIV (OR 3.83). CART analysis identified NIV as the dominant early discriminator, followed by cognitive vulnerability and acute severity. 30-day mortality was significantly higher in patients with delirium (22.3% vs. 9.8%, p = 0.001). Delirium remained an independent risk factor after adjustment (OR 2.53). CART analysis further corroborates delirium as a significant determinant, enhancing prognostic stratification beyond its role as a mere surrogate of disease severity.
Conclusions:
Admission-level clinical variables and exploratory CART analysis identified clinically interpretable delirium risk phenotypes in IMCU patients. Delirium was independently associated with increased 30-day mortality. These findings provide a preliminary framework for future validation studies and development of IMCU-specific risk-stratification approaches.