Addressing resource allocation for advance care planning discussions in hospital.

16 Background: Advance care planning (ACP) is fundamental in quality palliative care. There is no data detailing the demand for ACP discussions in hospitals, nor any standard objective criteria to decide urgency of such discussions. This cross-sectional study addresses this by comparing the ability of two commonly used instruments to detect palliative patients, and to predict death in the current admission.

METHODS: All inpatients were censored from the largest tertiary hospital in the state on a single inpatient day. 475 patients were followed for 3 months or until discharge or death, whichever earlier. Quan-modified Charlson score (QCS) and the Palliative Prognostic Score (PaP) was identified for each patient. These instruments were chosen based on external validity and ease of scoring.

RESULTS: See Table. 134 patients were identified as palliative (using Gold Standards Framework Indicators). 40 patients died. Both instruments were highly specific in detecting palliative patients and deaths but not sensitive, meaning they predicted patients who were palliative or died that admission. PaP>5.6 was most sensitive test to detect death in current admission.

CONCLUSIONS: Despite poor sensitivity, the most sensitive instrument (QCS) detected many patients requiring ACP discussion on a single day in hospital, highlighting the demand for ACP-trained staff. Secondly, streamlined usage of these instruments may assist in prioritising resource allocation. The QCS could identify patients needing ACP discussions (despite poor sensitivity, still identified many patients), whereas the PaP > 5.6 could refine the prioritisation of such discussions. This strategy could assist in reaching as many patients as possible using current staffing levels. [Table: see text].