Recalibration of the delirium prediction model for ICU patients (PRE-DELIRIC): a multinational observational study

Abstract

This is the accepted version of the paper.This version of the publication may differ from the final published version. Methods: A prospective multicenter cohort study was performed in eight intensive care units (ICUs) in six countries. The 10 predictors (age, APACHE-II, urgent and admission category, infection, coma, sedation, morphine use, urea level, metabolic acidosis) were collected within 24 hours after ICU admission. The confusion assessment method for the Intensive Care Unit (CAM-ICU) was used to identify ICU delirium. CAM-ICU screening compliance and inter-rater reliability measurements were used to secure the quality of the data. Permanent repository linkResults: 2,852 adult ICU patients were screened of which 1,824 (64%) were eligible for the study. Main reasons for exclusion were length of stay <1day (19.1%) and sustained coma (4.1%). CAM-ICU compliance was mean (SD) 82±16% and inter-rater reliability 0.87±0.17. The median delirium incidence was 22.5% (IQR 12.8%-36.6%). Although the incidence of all ten predictors differed significantly between centers, the area under the receiver operating characteristic (AUROC) curve of the 8 participating centers remained good: 0.77 (95%CI:0.74-0.79). The linear predictor and intercept of the prediction rule were adjusted and resulted in improved recalibration of the PRE-DELIRIC model. Conclusions:In this multinational study we recalibrated the PRE-DELIRIC-model. Despite differences in the incidence of predictors between the centers in the different countries the performance of the PRE-DELIRICmodel remained good. Following validation of the PRE-DELIRIC model it may facilitate implementation of strategies to prevent delirium and aid improvements in delirium management of ICU patients.4

Making advances in delirium research: coupling delirium outcomes research and data sharing

“…Of the 25 potential risk factors, ten were retained (Table 1). The area under receiver operating characteristics curve (AUROC) of pooled data varied for development (AUROC = 0.86), temporal validation (AUROC = 0.89), and external validation (AUROC = 0.84), resulting in an AUROC of 0.85.In a second study [4], the PRE-DELIRIC was recalibrated in eight ICUs from six countries; the median delirium incidence was 22.5 % (IQR 12.8-36.6 %). The incidence of the ten predictors differed significantly between centers, but the AUROC of the eight participating centers remained good at 0.77 [95 % confidence interval (CI) 0.74-0.79].…”

“…In a second study [4], the PRE-DELIRIC was recalibrated in eight ICUs from six countries; the median delirium incidence was 22.5 % (IQR 12.8-36.6 %). The incidence of the ten predictors differed significantly between centers, but the AUROC of the eight participating centers remained good at 0.77 [95 % confidence interval (CI) 0.74-0.79].…”

Recalibrating our prediction models in the ICU: time to move from the abacus to the computer

“…Validation of prediction models in different study populations is necessary to increase generalizability, because a prediction model can show myriad patterns of invalidity when tested in new settings and also have its performance decline over time. Therefore, in this issue of Intensive Care Medicine, the same group reports the results of a study that aimed at recalibrating the PRE-DELIRIC model using a large multinational data set [2]. Their main result is that the distribution of the predictors included in the score highly differed between centers and thus recalibration resulted in better prediction performances.…”

“…In their updated study, van den Boogaard et al chose to recalibrate the PRE-DELIRIC model without revising the original model [2]. They used four approaches to recalibrate the model, working on the concept of the slope and the intercept in a calibration plot.…”

“…It should be noted that the former PRE-DELIRIC model already had an acceptable performance in the multinational sample, as a result of a very well conducted score development in the first study [1]. However, the recalibration study [2] suffers some weaknesses. The most important is the lack of external/internal validation procedure.…”