Abstract
BACKGROUND: Dementia is a leading cause of death among residents in care home settings internationally. Proactive identification of those approaching end-of-life can support future care planning, respecting preferences. We aimed to synthesise existing literature on prognostic tools designed to predict 6-month mortality in individuals with dementia residing in care homes including their content and prognostic performance. METHODS: A systematic review was undertaken, searching Embase, CINAHL, PsycINFO and MEDLINE in December 2024. Studies evaluating prognostic tools for predicting 6-month mortality in individuals with dementia living in care homes were included. Risk of bias was assessed using the PROBAST tool. Data were extracted and narratively synthesised. RESULTS: Thirteen studies involving 503 501 participants were included from a total of 5438 records screened. Eleven prognostic tools were identified. Only three studies were assessed at low risk of bias. Predictive factors per tool ranged from 4 to 12. Age, change in cognitive function, functional decline/dependency and concerns around nutritional intake/weight were the commonest domains of predictors across included tools. Reporting around the discriminative performance of tools was limited and only two tools (Mortality Risk Index Score and Advanced Dementia Prognostic Tool had acceptable discrimination (area under receiver operating curve >0.70), which was not replicated in validation studies. CONCLUSIONS: While identified predictors have clinical congruence, there is significant variability in how these are assessed and operationalised and in the application of prognostic tools. Specific prediction of mortality remains challenging and would benefit from further research to adapt and validate dynamic prognostic tools for use in practice.