| アブストラクト | BACKGROUND: Stroke and transient ischaemic attack (TIA) survivors frequently experience multiple long-term conditions (multimorbidity) placing substantial demands on healthcare systems. Moving away from a single-disease approach could lead to more efficient and effective care for stroke survivors with multimorbidity. However, it is unclear how to categorise the stroke population to achieve this. To inform care, this population-based study identified and described clusters of stroke survivors with multimorbidity. METHODS: Using the Clinical Practice Research Datalink GOLD database, we identified 69,372 adult stroke/TIA survivors who were currently registered with a general practice on the 1st July 2017. We defined multimorbidity as the co-occurrence of >/= 2 of 36 long-term conditions (including stroke/TIA) and divided patients into four age strata (< 45, 45-64, 65-84, >/= 85). Within each stratum, Latent Class Analysis identified classes of co-morbid stroke survivors based on statistical diagnostics, clustering interpretation and clinical input. We investigated the validity of our findings, using a training and a test set. We described clusters according to prevalence of long-term conditions, demographic factors (age, gender, ethnicity, smoking, BMI) and health outcomes (mortality, hospital admissions, primary care consultations, prescriptions). RESULTS: In the UK, 94.7% of adult stroke/TIA survivors live with at least one additional long-term condition, with a median of five conditions per patient. We identified 12 clusters. Among 45-64-year-olds, the "alcohol and substance misuse" (6%) and the "established cardiovascular disease" (5%) clusters, have the highest mortality, while the "lower morbidity, better outcomes" cluster included 49% of patients. In ages 65-84, the "poor mental health" cluster (25%) exhibits the highest mortality. Among patients [Formula: see text]85, the "dementia-dominant" cluster had the highest rates of mortality, whereas the "established cardiovascular disease" cluster had the most hospital admissions. In each age strata, a cluster with increased mental health needs, and another with lower rates of multimorbidity and the best health outcomes could be identified. CONCLUSIONS: Internally validated clusters of stroke survivors can be identified with distinct patterns of multimorbidity, healthcare utilisation, and mortality. By understanding these clusters, more targeted, efficient, and integrated models of post-stroke care can be designed to better address the overall healthcare needs of stroke survivors. |
|---|
| 組織名 | The Primary Care Unit, Department of Public Health and Primary Care, University;of Cambridge, Cambridge, UK. em716@medschl.cam.ac.uk.;of Cambridge, Cambridge, UK.;Novo Nordisk R&D Digital Hub, 1 Pancras Square, Pancras Rd, London, N1C 4AG, UK.;Department of Computational Biology and Medical Big Data, Faculty of Computer;Science and Artificial Intelligence, Shenzhen University of Advanced Technology,;Shenzhen, 518107, China.;Center for AI in Medicine, Artificial Intelligence Research Institute, Shenzhen;University of Advanced Technology, Shenzhen, 518107, China. |
|---|
