Performance of the Delirium Rating Scale-Revised-98 Against Different Delirium Diagnostic Criteria in a Population With a High Prevalence of Dementia

Background: Delirium diagnosis in elderly is en complicated by underlying dementia. Objective: We evaluated performance of the Delirium Rating Scale-Revised-98 (DRS-R98) in patients with high dementia prevalence and also assessed concordance among past and current diagnostic criteria for delirium. Methods: Cross-sectional analysis of newly admitted patients to a skilled nursing facility over 6 months, who were rated within 24-48 hours after admission. Interview far Diagnostic and Statistical Manual of Mental Disorders, 3rd edition-R (DSM)-III-R, DSM-IV, DSM-5, and International Classification of Diseases 10th edition delirium ratings, administration of the DRS-R98, and assessment of dementia using the Informant Questionnaire on Cognitive Decline in the Elderly were independently performed by 3 researchers. Discriminant analyses (receiver operating characteristics curves) were used to study DRS-R98 accuracy against different diagnostic criteria. Hanley and McNeil test compared the area under the curve far DRS-R98's discriminant performance for all diagnostic criteria. Results: Dementia was present in 851125 (68.0%) subjects, and 361125 (28.8%) met criteria for delirium by at least I classification system, whereas only 19136 (52.8%) did by all. DSM-III-R diagnosed the most as delirious (27.2%), followed by DSM-5 (24.8%), DSM-IV-TR (22.4%), and International Classification of Diseases 10th edition (16%). DRS-R98 had the highest AUC when discriminating DSM-III-R (92.9%), followed by DSM-IV (92.4%), DSM-5 (91%), and International Classification of Diseases 10th edition (90.5%), without statistical differences among them. The best DRS-R98 cutoff score was >= 14.5 for all diagnostic systems except International Classification of Diseases 10th edition (>= 15.5). Conclusions: There is a low concordance across diagnostic systems for identification of delirium. The DRS-R98 performs well despite differences across classification systems perhaps because it broadly assesses phenomenology, even in this population with a high prevalence of dementia.