Predicting six-month mortality in people living with dementia in care homes: a systematic review of prognostic tool

3890
E West 1,2; L Mulligan 2; P Paudyal 3; TJ Quinn 1,2; JK Burton 1,3;
1 Academic Geriatric Medicine, School of Cardiovascular and Metabolic Health, University of Glasgow; 2 Glasgow Royal Infirmary, NHS Greater Glasgow and Clyde; 3 University Hospital Hairmyres, NHS Lanarkshire;
Scientific Research
SP - Other medical conditions
Dementia

Abstract

BACKGROUND:

Dementia is a leading cause of death among people living in care homes. Identifying palliative care needs in individuals with dementia remains challenging, often resulting in frequent hospital admissions and burdensome medical interventions during the final months of life. This review aimed to synthesise existing literature on prognostic tools designed to predict six-month mortality in individuals with dementia residing in care home settings

METHODS:

This systematic review was registered with PROSPERO (CRD42024623940). A comprehensive search of Embase, CINAHL, PsycINFO, and MEDLINE databases was conducted in December 2024. Studies evaluating prognostic tools for predicting six-month mortality in individuals with dementia living in care homes were included. Risk of bias of included studies was assessed using the PROBAST tool.

RESULTS:

Thirteen studies involving 503,501 participants were included from a total of 5,423 records screened. Eleven prognostic tools for six-month mortality were identified. Most tools were developed using either logistic regression or Cox proportional hazards models The number of predictive factors per tool ranged from 4 to 12 (median = 7). Nutritional status was the most commonly included factor, followed by co-morbidity related factors and age. Discriminative performance was primarily reported using the area under the receiver operating characteristic (AUROC) curve statistic , which ranged from 0.41 to 0.74 (median = 0.65), indicating modest predictive power.

CONCLUSION:

The identified prognostic tools exhibited substantial heterogeneity. While there was general consensus on the importance of nutritional status and comorbidity in predicting six-month mortality, the methods for assessing comorbidities and the specific conditions considered varied widely. This variability, along with the limited predictive accuracy of the tools, underscores the ongoing difficulty in reliably identifying individuals with advanced dementia approaching the end of life. These findings reinforce the importance of recognising and supporting palliative care needs in this population living in care homes. 

Comments