Abstract
Delirium results in adverse outcomes for hospitalized older adults but is often missed. Use of a predictive algorithm embedded within the electronic medical record (EMR) may provide additional information regarding risk for delirium. The purpose of this study was to develop and evaluate a delirium prediction rule to two cohorts and compare this rule to previously developed delirium prediction rules. The population included randomly selected hospitalized veterans from the Veterans Affairs (VA) External Peer Review Program at 118 VA medical centers with inpatient facilities. Risk factors associated with the National Institute for Health and Clinical Excellence (NICE) delirium rule were used as a base for development of the prediction rule. Delirium at admission as identified by trained nurse reviewers was the main outcome measure. A total of 27,625 patients were included in the derivation cohort and 11,752 in the confirmation cohort. Delirium at admission was identified in 2343 patients (8.5%) in the derivation cohort and 882 patients (7.0%) in the confirmation cohort. Random forest methods were used to identify accurate risk factors for prevalent delirium resulting in the consolidated NICE Prediction Rule, which was compared to the electronic NICE and Pendlebury NICE. Pre-existing cognitive impairment, infection, sodium level, and age of 80 years or older as delirium screening targets within Consolidated NICE. Use of this algorithm in an EMR could allow for targeted delirium assessment and management efforts towards those at highest risk.