Abstract
Introduction: Providing complex care for older adults poses significant challenges, especially when systems are not integrated and information is lacking or is not being shared. To address this gap, the iCAREtool - an IT-platform based on machine learning models - was developed. This research aims at providing decision support for better prognostication of health trajectories and treatment impact in older care dependent persons, using real world high quality data.
Methods: The study used interRAI assessment data as well as register data from Canada, Italy, Finland, The Netherlands, Belgium, New Zealand and the USA. First, Latent Class Analysis (LCA) was used to classify individuals according to their underlying disease patterns, starting from a list of 19 chronic conditions. Then, AI models were constructed and applied to large datasets.The AI tool can predict health care paths and intervention outcomes for people receiving complex care.
Results: The LCA on data from 102,000 individuals yielded a 5-class solution as the best model for all countries: (1) Alzheimer/dementia; (2) psychiatric diseases; (3) cardio-pulmonary diseases; (4) stroke/hemiplegia; (5) other dementias. The groups of cardio-pulmonary disease pattern and the stroke/hemiplegia disease patterns showed the highest complexity and impairment, especially on Activities of Daily Living (ADLs). AI models were built and an IT-platform is being tested in home care and residential settings to enhance clinical decision-making and foster information sharing.
Conclusions: Prognostic algorithms are being validated to better predict various health outcomes and the modifying impact of pharmacological and non-pharmacological interventions. After testing this tool in seven countries, the IT-platform will be optimized and will be available to be used by healthcare professionals worldwide.
Acknowledgements. The I-CARE4OLD project has received funding from the European Union’s Horizon 2020 research and innovation programme under Grant Agreement number 965341 and from the New Frontiers Research Fund, grant number NFRFG-2020-00500. The study was approved by authorized medical ethical committees in each of the participating countries, and will be conducted in compliance with both local and EU legislation.