Delirium is a serious neuropsychiatric disorder characterised by acute onset of fluctuations in arousal and cognition. This condition affects up to 80% of patients admitted to Intensive Care Units (ICUs), where it is associated with numerous adverse outcomes including risk of mortality and prolonged hospitalisation. Despite this, ICU delirium remains undiagnosed in up to 75% of all cases, with the detection and assessment of delirium in this setting hindered by the high proportion of patients that are sedated or rendered non-verbal by mechanical ventilation; and with available delirium screening instruments limited by their reliance on subjective observations. The present study aimed to adapt an existing measure of sustained attention (the Edinburgh Delirium Test Box; EDTB) to the ICU setting (EDTB-ICU) through consultation with ICU nursing staff, and by tailoring task difficulty and response methods to allow for the assessment of this heterogeneous patient population. Pilot investigations (n = 20) revealed that EDTB-ICU tasks were suitable for the assessment of sustained attention among ICU patients with a range of physical and cognitive capabilities. Longitudinal investigation of 30 ICU patients revealed that EDTB-ICU sustained attention tasks were able to sensitively (100%) and specifically (92%) detect delirium (as diagnosed by the Confusion Assessment Method for the ICU), with patient performance on this measure fluctuating over time in line with delirium diagnosis. These results indicate that the EDTB-ICU assessment provides an objective and suitably sensitive means of detecting and assessing delirium in the ICU