Development and validation of a delirium risk prediction preoperative model for cardiac surgery patients (DELIPRECAS): An observational multicentre study

To develop and validate a delirium risk prediction preoperative model for patients undergoing cardiac surgery.Observational prospective multicentre study.Six intensive care units in Spain.689 patients undergoing cardiac surgery consecutively, aged ≥18 years.The primary outcome measure was the development of delirium, diagnosed using the Confusion Assessment Method in Intensive Care Units (CAM-ICU), during the stay in the intensive care unit after cardiac surgery.The model was developed with 345 consecutive patients undergoing cardiac surgery at six hospitals and validated with another 344 patients from the same hospitals. The prediction model contained four preoperative risk factors: age over 65 years, Mini-Mental State Examination (MMSE) score of 25-26 points (possible impairment of cognitive function) or < 25 (impairment of cognitive function), insomnia needing medical treatment and low physical activity (walk less than 30 min a day). The model had an area under the receiver operating characteristics curve of 0.825 (95% confidence interval: 0.76-0.89). The validation resulted in an area under the curve of 0.79 (0.73-0.85) and the pooled area under the receiver operating characteristics curve (n = 689) was 0.81 (0.76-0.85). We stratified patients in groups of low (0%-20%), moderate (> 20%-40%), high (> 40%-60%) and very high (> 60%) risk of developing delirium, with a positive and negative predictive value for the very high risk group of 70.97% and 85.56%, respectively.The DELIPRECAS model (DELIrium PREvention CArdiac Surgery), consisting of four well-defined clinical risk factors, can predict in the preoperative period the risk of developing postoperative delirium in patients undergoing cardiac surgery. An automatic version of the risk calculator is available.