A study to assess the best predictive model for ICU delirium among the patients admitted in critical care units of a selected tertiary care hospital in Udupi district, South India. The purpose of the study is to compare the predictive performance of two delirium prediction models and to find out the best predictive model for ICU delirium. The findings of the study may help the nurses in screening of patients in ICU for the development of delirium so that timely treatment can be provided. Prior to data collection, Participant Information Sheet (PIS) will be given to all the participants and purpose of the study will be explained to them. Also informed consent will be taken. Data for each delirium predictor will be collected from consecutive patients, immediately after and within 24 hrs of admission to ICU by the researcher using E-PRE-DELIRIC and PRE-DELIRIC model respectively. The delirium will be evaluated once a day daily by the trained research nurse with prior experience in using delirium assessment scales till 7days or ICU discharge. The CAM-ICU will be used for bedside assessment of delirium and patients will be confirmed to have ICU delirium with at least one positive assessment of delirium using the CAM-ICU scale. To avoid false negative delirium screenings, patients who were administered haloperidol, dexmeditomidine or an atypical antipsychotic drug for delirium treatment anytime during their ICU stay while in the enrollment period will also be considered to have delirium.The patient’s level of sedation using the Richmond Agitation-Sedation Scale and current intravenous sedative therapy will be documented at the tie each delirium assessment is completed. Delirium will be assessed when patients are maximally awake (RASS above -3). Patients in coma (i.e., RASS =  -4 or -5) will be labelled as unable to be assessed for delirium. All patients enrolled in the study will be followed until 7 days or till ICU discharge (whichever is early).