Recognition of Delirium in Postoperative Elderly Patients: A Multicenter Study

Objectives: To evaluate to what extent delirium experts agree on the diagnosis of delirium when independently assessing exactly the same information and to evaluate the sensitivity of delirium screening tools in routine daily practice of clinical nurses. Design: Prospective observational longitudinal study. Setting: Three medical centers in the Netherlands. Participants: Elderly postoperative adults (n = 167). Measurements: A researcher examined participants daily (Postoperative Day 1-3) for delirium using a standardized cognitive assessment and interview including the Delirium Rating Scale Revised-98 as global impression without any cut-off values that was recorded on video. Two delirium experts independently evaluated the videos and clinical information from the last 24 hours in the participants' record and classified each assessment as delirious, possibly delirious, or not delirious. Interrater agreement between the delirium experts was determined using weighted Cohen's kappa. When there was no consensus, a third expert was consulted. Final classification was based on median score and compared with the results of the Confusion Assessment Method for Intensive Care Unit and Delirium Observation Scale that clinical nurses administered. Results: Four hundred twenty-four postoperative assessments of 167 participants were included. The overall kappa was 0.61 (95% confidence interval = 0.53-0.68). There was no agreement between the experts for 89 (21.0%) assessments and a third delirium expert was needed for the final classification. Delirium screening that nurses performed detected 32% of the assessments that the experts diagnosed as (possibly) delirious. Conclusion: There was considerable disagreement in classification of delirium by experts who independently assessed exactly the same information, showing the difficulty of delirium diagnosis. Furthermore, the sensitivity of daily delirium screening by clinical nurses was poor. Future research should focus on development of objective instruments to diagnose delirium.