[Fall prediction models in older adults with dementia: a rapid review].

INTRODUCTION: Falls in older adults with dementia are common and have multiple consequences for their health and quality of life. Fall prediction models are the core of the development of interventions to prevent falls, especially in institutional settings. OBJECTIVES: To identify and describe existing clinical and technology-based models to predict falls in older adults with dementia. METHODS: A rapid literature review was conducted in Scopus, Ovid Medline and Pubmed. Included studies developed fall prediction models in older adults with dementia and were published up to 2024. Excluded studies consisted of reviews, studies conducted in other populations or without explicit predictive models. RESULTS: 2.293 references were identified, and, after the full-text review, 22 studies were included. Of these, 54.5% were conducted in Europe, 18.2% in Japan, 18.2% in North America, and 9.1% in other contexts. The developed models recruited participants from residential, home, and hospital settings and used sociodemographic, clinical, functional, cognitive, and mobility-related predictors. 22.7% of the identified models used artificial intelligence algorithms. Only three studies used internal validation m ethods. CONCLUSIONS: Many prediction models are available to predict falls in older adults with dementia, but their clinical use is limited due to poor validation and predictor heterogeneity. Future studies should develop robust and valid fall prediction models tailored to support clinicians in providing adequate patient-centered care.