Extracting geriatric syndromes from electronic health records: a scoping review

INTRODUCTION: Geriatric syndromes are multifactorial ageing-related presentations. There is limited consensus regarding which presentations should be considered geriatric syndromes. Clinical codes generated following healthcare contacts are commonly used to identify diverse medical conditions. It is well demonstrated that geriatric syndromes are harder to identify using codes, possibly partly due to varying definitions, although identifying information is often located in unstructured text. Accurately identifying geriatric syndromes in healthcare records may therefore require different approaches compared to more firmly defined conditions. We review how studies using Electronic Health Records (EHRs) and derived databases align or differ in terms of geriatric syndromes selected for study, and how the syndromes are identified in the data sources. METHODS: The review was registered on the Open Science Framework ( https://osf.io/9ejhq ). Databases were searched in October 2024. Of 8086 titles screened, 74 studies had full text review, and 12 studies were included. RESULTS: Twelve studies from 7 countries identified 31 geriatric syndromes from primary databases (N = 6) or secondary databases (N = 6) using diagnostic codes, screening tools, or free text. The five most common geriatric syndromes were falls, functional decline, incontinence, delirium and malnutrition. Thirteen geriatric syndromes were included in two or more studies. Eight studies used more than one identification method. One study used two methods for all included syndromes. KEY CONCLUSIONS: There is heterogeneity in what presentations studies include as geriatric syndromes, and how they identify them in EHRs. It is essential that researchers clearly describe their included geriatric syndromes, reason for selection, and identification methods.