Emergency physician recognition of delirium.

OBJECTIVE: To compare the recognition of delirium by emergency physicians based on observations made during routine clinical care with concurrent ratings made by a trained researcher after formal cognitive assessment and to examine each of the four individual features of delirium separately to determine the variation in identification across features.
METHODS: In a prospective study, a convenience sample of 259 patients, aged ≥65 years, who presented to two urban, teaching hospital emergency departments (EDs) in Western Pennsylvania between 21 June and 29 August 2011, underwent paired delirium ratings by an emergency physician and a trained researcher. Emergency physicians were asked to use their clinical judgment to decide whether the patient had any of the following delirium features: (1) acute change in mental status, (2) inattention, (3) disorganised thinking and (4) altered level of consciousness. Questions were prompted with examples of delirium features from the Confusion Assessment Method. Concurrently, a trained researcher interviewed surrogates to determine feature 1, conducted a cognitive test for delirium (Confusion Assessment Method for the intensive care unit) to determine delirium features 2 and 3 and used the Richmond Agitation and Sedation Scale to determine feature 4.
RESULTS: In the 2-month study period, trained researchers identified delirium in 24/259 (9%; 95% CI 0.06 to 0.13) older patients admitted to the ED. However, attending emergency physicians recognised delirium in only 8 of the 24 and misidentified delirium in a further seven patients. Emergency physicians were particularly poor at recognising altered level of consciousness but were better at recognising acute change in mental status and inattention.
CONCLUSIONS: When emergency physicians use routine clinical observations, they may miss diagnosing up to two-thirds of patients with delirium. Recognition of delirium can be enhanced with standardised cognitive testing.