Screening for delirium on a general medical ward: the tachistoscope and a global accessibility rating.

With a psychiatrist's standardized clinical diagnosis as the criterion, the "Hand-Held Tachistoscope" was 100% sensitive, but only 45% specific in detecting delirium among hospital patients on a general medical ward. For each true positive in this sample of 97 patients, there were almost 5 false positives. The 10 patients with clinically diagnosed delirium could not see the stimulus. This was also true for 24 of the 87 nondelirious patients. Performance on the tachistoscope was related to age and education. This was not true for another method of delirium case detection, the global accessibility rating. This rating was 90% sensitive and 95% specific when compared with the psychiatric diagnosis, and was stable across two days of ratings. The global accessibility rating warrants further evaluation as a simple screening test for delirium.