Objectives: To examine delirium, chronic medical problems and sociodemographic factors as predictors of activities of daily living (ADL), basic ADL (BADL) and instrumental ADL (IADL).

Methods: A prospective cohort study of four groups of elderly patients examined in the emergency department (ED): those with delirium, dementia, neither, and both. All were aged 66 years or older and living at home. Delirium was assessed with the Confusion Assessment Method and dementia with the Informant Questionnaire on Cognitive Decline in the Elderly. Demographic variables and chronic medical problems were ascertained with questionnaires. Outcome was ADL at 6, 12 and 18 months, measured with the ADL subscale of the Older Americans Resources and Services instrument.

Results: Univariate analyses suggested significantly poorer ADL, particularly IADL, at 18 months in the delirium versus the non-delirium group, in the absence of dementia only. Statistically significant independent predictors of poorer ADL at 18 months in the non-dementia groups were poorer initial ADL, stroke, Parkinson's disease, hypertension and female sex. Independent predictors of poorer BADL at 18 months in the non-dementia groups were poorer initial BADL, Parkinson's disease, stroke, cancer, colds/sinusitis/laryngitis, female sex and hypertension. Independent predictors of poorer IADL at 18 months in the non-dementia groups were poorer initial IADL, stroke, never-married status, colds/sinusitis/laryngitis, arthritis and hypertension, with Parkinson's disease showing a non-significant but numerically large regression coefficient.

Conclusion: Rather than finding delirium to be a predictor of poorer functional outcome among survivors, we found an interaction between delirium and dementia and several plausible confounders, primarily chronic medical problems, although we cannot rule out the effect of misclassification or survivor bias.