No one dies unassessed: comfort documentation in the last days of life on a dedicated end-of-life ward versus matched wards

Introduction
NICE NG31 requires regular, documented assessment of symptoms in the last days of life, yet dying patients on acute wards frequently go unassessed. We audited last-days-of-life (LDOL) flowsheet documentation on a newly opened nurse-led end-of-life care ward against matched deaths on acute wards.

Methods
The first 50 consecutive ward deaths were matched to 50 acute ward deaths (exact specialty; age within ten years; nearest length of stay). Electronic flowsheet entries for the final 72 hours of life were extracted for all comfort domains (pain, agitation, respiratory symptoms, mouth care and others), and compared against trust audit standards. Observation intensity was compared using paired Wilcoxon tests.

Results
An individualised care plan was in place for 100% of ward patients (target 95%). Each comfort domain was documented in 98–100% of ward patients versus 65–67% of matched controls. Observation intensity was higher (median 9.7 versus 8.7 entries per domain per 24 hours; p=0.042). No ward patient went through their final 24 hours without an agitation or respiratory assessment, versus 29% of controls. Identified uncontrolled pain was acted on within four hours in 64% versus 26%. Honest gaps remain: pain reviewed in the final six hours in only 60% (target 80%), a named lead clinician documented in 40%, and prognosis discussion documented in 56%.

Conclusions
A ward-level model built around comfort observation eliminated unassessed dying and doubled the rate of acted-upon symptoms. Residual gaps are driving an electronic documentation redesign and re-audit at six months.