Previous studies showed different classification systems lead to different case identification and rates of delirium. No one has previously investigated the influence of different classification systems on the outcomes of delirium.

To determine the influence of DSM-5 criteria vs. DSM-IV on delirium outcomes (mortality, length of stay, institutionalisation) including DSM-III and DSM-IIR criteria, using CAM and DRS-R98 as proxies.

Prospective, longitudinal, observational study of elderly patients 70+ admitted to acute medical wards in Sligo University Hospital. Participants were assessed within 3 days of admission using DSM-5, and DSM-IV criteria, DRS-R98, and CAM scales.

Two hundred patients [mean age 81.1 ± 6.5; 50% female]. Rates (prevalence and incidence) of delirium for each diagnostic method were: 20.5% (n = 41) for DSM-5; 22.5% (n = 45) for DSM-IV; 18.5% (n = 37) for DRS-R98 and 22.5%, (n = 45) for CAM. The odds ratio (OR) for mortality (each diagnostic method respectively) were: 3.37, 3.11, 2.42, 2.96. Breslow-Day test on homogeneity of OR was not significant x2= 0.43, df: 3, P = 0.93. Those identified with delirium using the DSM-IV, DRS-R98 and CAM had significantly longer hospital length of stay(los) compared to those without delirium but not with those identified by DSM-5 criteria. Re-institutionalisation, those identified with delirium using DSM-5, DSM-IV and CAM did not have significant differences in discharge destination compared to those without delirium, those identified with delirium using DRS-R98 were more likely discharged to an institution (z = 2.12, P = 0.03)

Assuming a direct association between delirium and examined outcomes (mortality, los and discharge destination) different classification systems for delirium identify populations with different outcomes.

The authors have not supplied their declaration of competing interest.