Can the impact of an acute hospital end-of-life care tool on care and symptom burden be measured contemporaneously?

Objective: To determine the utility of a screening question to identify patients who might die during hospital admission and feasibility of scoring symptoms in dying patients within a study assessing the impact of a brief end-of-life (EOL) tool.

Methods: Between March 2008 and July 2010 patients admitted to five wards of an acute hospital were screened using the question 'Is this patient so unwell you feel they could die during this admission?' Once 40 patients were recruited, the brief EOL tool was introduced to the wards and a further 30 patients were recruited. Symptom scoring using the Edmonton Symptom Assessment System (ESAS) began when the patient was recognised as dying. Relatives were asked to complete the Views of Informal Carers-Evaluation of Services questionnaire to validate the results of the contemporaneous symptom assessments and assess the impact of the tool.

Results: The sensitivity of the screening question was 57%, specificity 98% and positive predictive value 67%, so the question was useful in enrolling study patients. There were limitations with the ESAS but core EOL symptoms were scored more frequently after the tool was introduced. Questionnaire responses suggested relatives perceived aspects of care improved with the EOL tool in place.

Conclusions: It is possible to identify dying patients and study care given to them in hospital in real time. Outcome measures need to be refined, but contemporaneous symptom monitoring was possible. We argue interventions to improve EOL care should be unequivocally evidence-based, and research to provide evidence of impact on the patient experience is possible.