ABSTRACT
INTRODUCTION: Delirium is often underdiagnosed, resulting in adverse clinical outcomes. The goal of this study was to identify how patients correctly diagnosed with delirium differ from those who are misdiagnosed. METHODS: A retrospective chart review was conducted using a database of 1,000 consecutive psychiatric consultation requests. Patients were identified based on a diagnosis of delirium made by the consultation team. Charts were then reviewed for data on race, gender, age, time and month of the consultation, documented diagnosis of mental illness, and information that would help establish a delirium diagnosis based on DSM-IV-TR criteria. Univariate and multivariate analyses were performed. RESULTS: Cases were judged to be diagnostically concordant (consultation requested for delirium or encephalopathy, n = 30) or discordant (n = 81). The two groups did not differ significantly in age, sex, race, time and month of the consultation, or documentation of mental illness. The concordant group had a significantly greater number of identifiable diagnostic criteria compared to the discordant group (mean 3.0 ± 0.8 criteria vs. 1.9 ± 1.3 criteria, P < 0.001). Identification of individual diagnostic criteria was greater in the concordant group, with significant differences for two of four categories, namely acute onset (100.0% vs. 50.6%, P < 0.001) and fluctuating course (93.3% vs. 66.7%, P = 0.004). Multivariate analysis suggested increased odds of identifying delirium if more diagnostic criteria were identifiable (OR: 2.355, P < 0.001, confidence interval [CI] 1.502-3.690), and increased likelihood of the delirium diagnosis being missed if there was documentation of psychiatric illness (OR: 0.387, P = 0.049, CI: 0.151-0.995). CONCLUSION: This study highlights the need for educational programs and easy to implement screening tools to ensure delirium is not overlooked.