Objective: To identify the flow charts and discriminators of the Manchester Triage System that are most likely to identify the onset of delirium in older people.Background: Delirium is an underdiagnosed geriatric syndrome, and up to 80% of all cases of delirium go undetected in emergency departments. Patient triage seeks to manage clinical risk with a view to safely and appropriately managing patient flows.Design: A case–control study was performed according to the STROBE checklist.Setting: The emergency department of a secondary hospital.Participants: Older adults aged ≥65 years and admitted from 1 January to 31 December 2020.Methods: Older patients were identified from the emergency department research database. Cases were defined as patients diagnosed with delirium (n= 128), excluding cases of delirium due to alcohol or substance abuse. Controls were randomised from the remaining patients (n= 128).Results: A total of 29.35% of the subjects admitted to the emergency department were older adults with an incidence of delirium of 0.7%. The flow charts with the highest probability of delirium were ‘unwell adult’ [OR = 3.04 (95%CI:1.82–5.1)] and ‘behaving strangely’ [OR = 16.06 (95%CI:3.72–69.29)], and the discriminators were ‘rapid onset’ [OR = 3.3 (95%CI:1.85–5.88)] and ‘new neurological deficit less than 24 h old’ [OR = 4.76 (95%%CI:1.01–22.5). The area under the curve for ‘unwell adult’ in the presence of dementia, previous stroke and fall in the previous 30 days was 0.73 (95%CI: 0.67–0.79), and that for ‘behaving strangely’ in the presence of diabetes was 0.75 (95%CI: 0.69–0.81).Conclusions: Knowing which flow charts, discriminators and risk factors are most likely to predict delirium allows the identification of the older population at risk for triage screening in emergency departments.Relevance to clinical practice: Risk factors such as diabetes, dementia, previous stroke and recent fall among ‘unwell adult’ or ‘behaving strangely’ triaged older per-sons should be assessed for the probable presence of delirium.