Introduction: There is a lack of objective tools for the diagnosis of delirium. Previous studies have utilised tasks of sustained attention to identify delirium. This study used a newly developed smartphone application to administer tasks of sustained attention to elderly hospital in-patients. Test performances are compared in patients with delirium, dementia and cognitively-normal controls. The effect of depression on test performance is also investigated. Methods: In phase 1, 20 participants were recruited and their scores on the EDTB, a computerised neuropsychological testing device, were compared to their smartphone application scores. In phase 2, 159 participants were recruited (51 with delirium, 54 with dementia and 54 controls). Each completed the DelApp tests of sustained attention, as well as a measure of cognitive impairment (OMCT), traditional test of attention (BAT) and a brief depression screening tool. Delirium was diagnosed using the Confusion Assessment Method (CAM) and delirium severity assessed using the Delirium Rating Scale-Revised98 (DRS-R98). Arousal was measured using the Observational Scale of Level of Arousal (OSLA). Results: Patients with delirium showed significantly poorer performance on all tasks of attention, in comparison to the dementia and control patients (p<0.001). Delirium and dementia patients were matched in level of cognitive impairment. Those who were depressed had lower DelApp scores than those who were not depressed (P<0.01). Discussion: Patients with delirium demonstrated significant deficits in sustained attention, as indicated by their performance on the objective DelApp tasks of sustained attention. This highlights the potential use of sustained attention to identify delirium and differentiate between delirium and dementia.