The problem of definitions in measuring and managing ICU cognitive function.

Epidemiology and clinical trials require valid, repeatable definitions that ideally dichotomise patients into having, or not having, a clinical condition. • Some conditions are clearly dichotomous, such as pregnancy; others such as hypertension or obesity rely on defining a threshold on an objective scale. • Defining delirium and "adequate" sedation and analgesia in the intensive care unit is more difficult, as there is no universally agreed scale that quantifies the relative importance of various diagnostic features, distinguishes features merely observed from those actively sought, quantifies severity or fluctuation over time, or accounts for the variable approaches of clinicians and the effects of assessment environment and pharmacological treatment. Definitions of delirium and adequate sedation and analgesia therefore vary by assessment method and context, making studies using different methods and personnel not necessarily comparable. • Although there is no simple solution, we suggest better awareness of these problems will be helpful. Further, we propose a simplified categorisation to facilitate clinical communication and treatment in the ICU.