Advanced Dementia End‐of‐Life Assessment measures: A systematic Review

Abstract

Background: Advanced dementia is a terminal condition characterised by profound cognitive decline, other significant and progressive symptoms, yet there is no consensus on which end‑of‑life assessment measures best capture all symptoms in this population. This study aims to identify and synthesise end‑of‑life assessment measures used in people with advanced dementia and to evaluate their measurement properties and clinical feasibility for nursing practice. Methods: This systematic review adhered to PRISMA 2020 and COSMIN guidance for measurement instruments. PubMed, ScienceDirect, Embase and the Cochrane Library were searched for peer‑reviewed quantitative studies from 2016 to 2025 using terms related to advanced or severe dementia, end of life or palliative care, and assessment or measurement. Ten studies met the inclusion criteria and were charted in a review matrix; because of heterogeneity in designs, settings and outcomes, findings were synthesised narratively across four domains: global end‑of‑life and prognostic measures, cognitive tools, behavioural pain scales and documentation‑based assessments. Results: The included studies showed that multidomain tools (EoLCAD, CADEOLD, SMEOLD, SWCEOLD) and the ADEPT prognostic score generally have acceptable internal consistency and basic construct validity but only modest prognostic discrimination and limited evidence on responsiveness. SIRS offers a brief severe‑stage cognitive assessment with minimal floor effects, while behavioural pain measures (PAINAD, PAINAD‑K, PATCIE) demonstrate good reliability, although PAINAD may partly reflect general distress in acute settings. Conclusion: The combination of end‑of‑life tool, SIRS and a behavioural pain scale is suggested to be more reliable than depending a single measure, but further COSMIN‑informed validation and implementation research is needed to improve end‑of‑life assessment in advanced dementia.