Background Delirium is characterised by acute onset of fluctuating confusion and altered conscious level. It is common in palliative patients and associated with worse outcomes. Recognition and management of delirium is poorly supported in palliative care inpatient units.

Aim To produce a sustainable improvement in prevention, recognition and management of delirium in St. Gemma’s Hospice Inpatient Unit, Leeds.

Methods This mixed-method service improvement project uses a behaviour change and normalisation-process theory based approach formed of three evaluation and two implementation stages. The evaluation stages use a mixed-methods approach to gather data from several sources:

• Retrospective case note audit

• Staff survey

• Staff interviews

The first implementation stage modified the delirium guidelines including:

• Introduction of the 4AT rapid clinical test for delirium.

• Delirium severity assessment replaced by agitation assessment, using the Richmond Agitation–Sedation Scale for palliative care inpatients (RASS–PAL).

Members of staff were appointed as ‘delirium champions’ to address sustainability and ongoing training. The second implementation stage will design and implement a behaviour change intervention:

• Data from evaluation sources will be collated to identify behaviours to be targeted.

• The delirium guidelines and strategies for its implementation will be modified using behaviour change techniques.

Results In the first evaluation stage 77 patient admissions were audited. 58 delirium episodes were retrospectively identified, of which 19% were diagnosed during admission. 27% of admitted patients were screened for delirium. No patients were risk assessed for delirium and no delirium prevention measures were taken. 17% of delirium episodes had appropriate non-pharmacological management while 88% received pharmacological management as per guideline.

Conclusion This baseline audit shows that prevention, recognition and management of delirium is poorly conducted. The first implementation phase has completed, and the second evaluation stage is in progress. I anticipate results from this phase will be available to present at The Palliative Care Congress.