The frequency of full syndromal and subsyndromal delirium is understudied.

We conducted a point prevalence study in a general hospital.

Possible delirium identified by testing for inattention was evaluated regarding delirium status (full/subsyndromal delirium) using categorical (Confusion Assessment Method (CAM), DSM-IV) and dimensional (Delirium Rating Scale-Revised-98 (DRS-R98) scores) methods.

In total 162 of 311 patients (52%) screened positive for inattention. Delirium was diagnosed in 55 patients (17.7%) using DSM-IV, 52 (16.7%) using CAM and 58 (18.6%) using DRS-R98⩾12 with concordance for 38 (12.2%) individuals. Subsyndromal delirium was identified in 24 patients (7.7%) using a DRS-R98 score of 7–11 and 41 (13.2%) using 2/4 CAM criteria. Subsyndromal delirium with inattention (v. without) had greater disturbance of multiple delirium symptoms.

The point prevalence of delirium and subsyndromal delirium was 25%. There was modest concordance between DRS-R98, DSM-IV and CAM delirium diagnoses. Inattention should be central to subsyndromal delirium definitions.

Delirium is a common neuropsychiatric syndrome with considerable heterogeneity in clinical profile. Identification of clinical subtypes can allow for more targeted clinical and research efforts. We sought to develop a brief method for clinical subtyping in clinical and research settings.

A multi-site database, including motor symptom assessments conducted in 487 patients from palliative care, adult and old age consultation-liaison psychiatry services was used to document motor activity disturbances as per the Delirium Motor Checklist (DMC). Latent class analysis (LCA) was used to identify the class structure underpinning DMC data and also items for a brief subtyping scale. The concordance of the abbreviated scale was then compared with the original Delirium Motor Subtype Scale (DMSS) in 375 patients having delirium as per the American Psychiatric Association's Diagnostic and Statistical Manual (4th edition) criteria.

Latent class analysis identified four classes that corresponded closely with the four recognized motor subtypes of delirium. Further, LCA of items (n = 15) that loaded >60% to the model identified four features that reliably identified the classes/subtypes, and these were combined as a brief motor subtyping scale (DMSS-4). There was good concordance for subtype attribution between the original DMSS and the DMSS-4 (κ = 0.63).

The DMSS-4 allows for rapid assessment of clinical subtypes in delirium and has high concordance with the longer and well-validated DMSS. More consistent clinical subtyping in delirium can facilitate better delirium management and more focused research effort.