Premium
P4‐402: USING 3D SKELETON MOVEMENT DATA AND MACHINE LEARNING TO EVALUATE DISEASE PROGRESSION AND THE IMPACT OF TREATMENTS OR INTERVENTIONS IN PEOPLE WITH DEMENTIA, MCI OR PHYSICAL IMPAIRMENT
Author(s) -
Chen Zizui,
Czarnuch Stephen M.,
Dove Erica,
Astell Arlene J.
Publication year - 2019
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1016/j.jalz.2019.06.4074
Subject(s) - physical medicine and rehabilitation , dementia , computer science , session (web analytics) , psychological intervention , movement (music) , artificial intelligence , medicine , disease , pathology , psychiatry , aesthetics , philosophy , world wide web
Background: Social media platforms (i.e., Twitter and online health forums) are increasingly used as a novel data source for understanding many clinical conditions, including dementia. We explore user-generated dementia-related content on these platforms, and compare their utility for researchers conducting patient-oriented dementia research, and clinicians and policymakers hoping to improve understanding of dementiarelated needs. Methods: We collected 78, 265 Twitter posts through Twitter’s application program interface over 55 hours, and 15, 538 posts from an online forum (Alzheimer’s Association UK) from its inception until April 2018. All posts contained the keywords ‘dementia’, ‘Alzheimer’ or ‘Alzheimer’s’. We performed content analysis on 10% of Twitter posts and 5% of forum posts to generate the top ten most common themes from each platform. We summarized the unique features of each platform, including types of users, data quality, and data collection needs. Results: Twitter posts sampled were most frequently posts using ‘dementia’ as a derogatory term (4.8%; n1⁄4376), followed by posts raising awareness for dementia (3.9%; n1⁄4302). Preliminary results show online health forum posts were most frequently personal experiences of both caregivers and people with dementia, caregiving strategies, and requests for information and support. Further, Twitter and health forums provide different data quality (e.g., signal to noise ratio, richness) and vary in user demographics. Conclusions: Health forums are valuable for gaining insight into the perspectives of people with dementia and their caregivers, whereas Twitter data largely reflects societal perceptions of dementia. Dementia-related research, policy setting and clinical practice can be improved by harnessing these novel data sources.