Objectives: To develop and validate a screening strategy for delirium within the inter RAI acute care comprehensive assessment system.
Design: Prospective validation cohort study.
Setting: Acute general medical wards in two acute care metropolitan hospitals in Brisbane, Australia.
Participants: Two hundreds thirty-nine subjects with and without delirium, aged 70 and older.
Measurements: Trained research nurses assessed subjects within 36 hours of hospital admission using the inter-RAI acute care (AC) system which includes four observational delirium items: Acute change mental status from baseline (ACMS), mental function varies over the course of the day (MFV), episode of disorganised speech (EDS), and easily distracted (ED). Geriatricians assessed subjects face to face within 4 hours of nurses' assessment using the Diagnostic and statistical manual of mental disorders (DSM IV) criteria and clinical judgement to determine delirium presence. Based on the performance of each delirium feature and to achieve highest predictive accuracy, a combination algorithm of either ACMS or MFV was developed and compared with the reference standard diagnosis determined by geriatricians.
Results: Geriatricians diagnosed delirium in 52 of 239 (21.7%) subjects aged 70-102 years. The area under the receiver operator characteristics (AUC) for interRAI-AC delirium screener algorithm was 0.87 (95% CI; 0.80, 0.93), sensitivity 82%, specificity 91%, positive and negative predictive value of 0.72% and 95%, and likelihood ratio of 9.6 achieving the highest predictive accuracy of all possible combination of 4 delirium features. Underlying pre-morbid cognitive impairment did not undermine validity of the screening strategy, AUC 0.85 (95% CI; 0.74, 0.95), sensitivity 90% and specificity 69%.
Conclusion: The interRAI AC delirium screening strategy is a valid measure of delirium in older subjects in acute medical wards.