Project 460393

My thesis applies emerging machine learning methods to identify common prescribing patterns over time to better describe potentially inappropriate prescribing and psychotropic medication exposures in older adults with dementia. Older adults newly diagnosed with dementia represent a vulnerable group where optimizing medication therapy throughout disease progression is critical. Inappropriate prescribing for this population has been associated with adverse health outcomes, including hospitalizations, falls, and mortality. Psychotropic, or central nervous system active, medications are particularly associated with falls in older adults. Two gaps exist in current research: 1) a lack of methods which capture broader prescribing context over time and 2) patient and care partner perspectives are seldom integrated in research studies on inappropriate prescribing, despite importance of patient engagement. This project will apply machine learning approaches, such as network analysis and optimal matching algorithms, to discover prevalent but understudied medication combinations, to better define exposure to medications over time, and compare the impact of different exposure definitions on the rate of fall-related hospitalizations. The Engagement of People with Lived Experience of Dementia (EPLED, https://www.epled.ca/), a Canadian Consortium of Neurodegeneration in Aging (CCNA) program which brings those with lived experience of dementia and their care partners, will be consulted during all stages of the research process.