CFE 2019: Start Registration
View Submission - CMStatistics
Title: Identifying dementia prognostic factors among diabetes individuals Authors:  Aamna AlShehhi - MIT and Khalifa University (United States) [presenting]
Abstract: Dementia is degenerative neurodegenerative diseases. It destroys the brain normal functionality such as memories access, and decision. The failure rate of dementia therapies accounts by 99.6\%. Those failures are relevant to various factors such as heterogeneous of the patients. There is increasing interest to study the cognitive impairment in Diabetes mellitus population since 60\% of diabetes population are at risk of developing it. A UK population with Diabetes was identified in Clinical Practice Research Datalink. We examine dementia associated with different risk factors and comorbidities by applying statistical modeling of censored time-to-event data, namely the cox model. The cox model incorporates with inverse probability of treatment weighting using propensity score. Finally, we evaluate the model on the holdout dataset using Concordance-index and Brier scores. The results show Metformin is usually associated with a significantly lower risk of dementia in London, but not in the other regions. The results also show being male to significantly lower the risk of developing dementia in most of the cities. Also, COPD and CKD are two morbidities associated with dementia risk. The dementia risk factors and its associations vary based on the region of care. These results point to the need of adapting different care practice to different regions to account for the hidden lifestyle and environmental factors.