Monitoring for delirium: an open question
Commentary. These are the author’s interpretations of the cited evidence. Disclosures.
Recognise a change when it happens. The scheduled score should not decide when care starts.
A patient can have no delirium on admission and develop it later. Monitoring is the process of watching for that change. It can include routine observation, information from families and carers, and repeated structured assessments. These are different activities, even when they sit in the same care pathway.
There is no single settled approach across hospital settings. Tools such as Nu-DESC, DOSS and RADAR organise observations. SQiD brings a report of recent change into the process. Some services use an observation to trigger a bedside assessment such as the 4AT. Others repeat a structured assessment on a schedule. ICU and paediatric services have their own requirements. NICE recommends at least daily observation for indicators of delirium, with assessment when indicators are identified (NICE, 2023).
The evidence does not establish one ideal tool, frequency or sequence for every setting. Nor does a validation study of one assessment settle the question of its repeated use over a long admission. Patient burden, staff time, test completion, diagnostic accuracy and the response to a concerning result all matter.
A shift-based tool can summarise what happened during care. The problem arises if recognition or action waits until the score is entered at the end of the shift. Delirium may then have been present for hours. Handover can add further delay. This is a foreseeable weakness of the process, not an inevitable feature of every observation tool. Staff should respond to a concerning change when they notice it.
Penfold et al. (2024) found wide variation in completion and positive-score rates in routine general-hospital care. Those measures need to be read with the population, timing and denominator. A completed form is not proof that delirium has been detected. A low positive rate is a reason to investigate, but it is not a direct measurement of sensitivity.
My practical preference is to make the response explicit: new concern prompts assessment and clinical review; a scheduled observation does not replace that response. Services should also check a sample of negative and uncompleted assessments. Otherwise, a system can appear efficient while overlooking the patients who need it. The same standard applies to every tool.
The patient journey diagram shows the distinction. A separate note on CAM explains why omitting its preceding cognitive assessment is a specific problem. Properly repeating a structured assessment is different from removing part of it.
References
MacLullich, A. (2020) Delirium detection in routine clinical care: two basic processes. Delirium Words.
NICE (2023) Delirium: prevention, diagnosis and management, CG103, recommendations 1.5.1 and 1.6.1.
Penfold, R.S. et al. (2024) Delirium detection tools show varying completion rates and positive score rates when used at scale in routine practice in general hospital settings: a systematic review. Journal of the American Geriatrics Society, 72, pp. 1508–1524.