Development and validation of a delirium risk prediction model using a modified version of the Delirium Elderly at Risk (mDEAR) screen in hospitalized patients aged 65 and older: A medical record review.

Most delirium risk prediction models are cumbersome to use, time consuming to complete, and require education ensure accuracy. The purpose of this study was to develop and validate a risk prediction model using routinely assessed risk factors predictive of delirium including: cognitive impairment, ≥80-years old, functional dependence, sensory impairment, and chronic substance use. This retrospective study included 7999 encounters of hospitalized patients aged 65-years or older admitted from 1/1/2019 to 12/31/2019. Various models were compared, with the best tested for validation. A model, where cognitive impairment was worth 2-points and a threshold of 3-points to predict delirium, was determined to be the best model and was validated with an area-under Receiver-Operating-Characteristic curve=0.7126. Management of delirium could be enhanced by integrating a nursing admission delirium risk screening process into the workflow, triggering initiation of prevention interventions and prompt assessment for signs and symptoms of delirium for those at high risk.