Since dementia can affect functional abilities, we aimed to compare dementia and non-dementia patients according to their disabilities using propensity score matching. We also determined if developing a model for predicting disability trained on people with dementia (Specialized) would perform better than model trained on dementia and non-dementia people (Generic). Although dementia group revealed higher level of disability based on propensity score matching, the models were not statistically different in terms of performance.

Learning Objective: 1. Comparing how dementia and non-dementia people differ in terms of different types of disabilities such as bathing, grooming etc.
2.Comaring how the machine learning-based models that predict people’s disability differ in terms of performance for people with dementia.

Authors:

negin asadzadehzanjani (Presenter)
George Mason University

Janusz Wojtusiak, George Mason University
Allison Williams, Department of Veterans Affairs
Cari Levy, Department of Veterans Affairs

Presentation Materials:

• Abstract/Manuscript - 1