Validation of a delirium predictor scale in hospitalized older people.

Abstract

Delirium is an acute confusional state characterized by cognitive impairment not attributable to another disorder. Globally, its incidence ranges from 10% to 31% upon hospital admission, increasing mortality risk in older hospitalized adults by 2 to 4 times. Up to 60% of cases remain undiagnosed, and the annual cost of care is estimated at approximately USD 164 million. At our institution, the frequency of delirium is around 10%. Objective: To develop and validate a predictive scale for delirium in hospitalized patients at CMDLT. Materials and methods: A prospective cohort study was conducted between July and October 2025 in older adults. The study applied a new clinical prediction scale based on associated variables: advanced age, female sex, cognitive impairment, polypharmacy, and comorbidities. A specific score was assigned to each variable, and the 4AT was applied upon admission for delirium diagnosis. Statistical analysis included logistic regression, ROC curve analysis, and odds ratio. Results: The cohort comprised 36 patients, aged 76 ± 9 years; 61% were women and 39% men. The frequency of delirium was 11%. Patients scoring ≥ 4.5 points on the scale (AUC 0.85) had an odds ratio (OR) of 2.99 (95% CI: 1.45–8.42, p = 0.0125), demonstrating a significant increase between the 4AT score at admission and during hospitalization. Conclusions: This predictive model is easy to implement and effectively highlights risk factors for delirium. Increasing the sample size is essential to identify a greater number of patients requiring close monitoring and preventive measures during hospitalization.