Abstracts

Social Conversations Regarding Epilepsy and Driving: Big Data Techniques

Abstract number : 3.307
Submission category : 11. Behavior/Neuropsychology/Language / 10A. Adult
Year : 2016
Submission ID : 198206
Source : www.aesnet.org
Presentation date : 12/5/2016 12:00:00 AM
Published date : Nov 21, 2016, 18:00 PM

Authors :
Jane von Gaudecker, Indiana University- Perdue University; Janice Buelow, Indiana University- Perdue University; Doyle Groves, Indiana University- Perdue University; and Wendy Miller, Indiana University- Perdue University

Rationale: In the United States, nearly 3 million people live with epilepsy. Among the many challenges that people with epilepsy (PWE) face, loss of independence is a significant one. One key reason for loss of independence in PWE is restricted driving privileges. Driving restriction laws for PWE vary from state to state, and it remains challenging and confusing for PWE to determine when it is safe to drive. The purpose of this study is to understand PWE's social media discussions about driving-related issues in epilepsy. Generation of this knowledge can provide a truly patient-centered description of PWE's ongoing concerns and questions about driving. Methods: We analyzed data from the www.epilepsy.com user forums using a Word Adjacency Graphing (WAG) technique. The initial data set, spanning 12 years, contained 50,000 individual posts, and we used key words such as "drive," "driving," "transportation," "car," and "license" to filter for driving-related discussions, which narrowed the data set to 1,463 posts. Each post was split into token words, and stopwords (e.g. and, the, of) removed. Remaining words were considered in sequence to build a summary table of words. Word pairs occurring at least 10 times were used to build a network graph model. The resulting network graph of 2-word phrases represented all the frequently-occurring adjacent word pairs in the original content. Applying Louvain modularity to this network organized it into optimally-sized communities, appearing in this application as clusters of words around a common topic (driving). A 2-dimensional view of this reorganized network helps visualize the overall topic landscape in this content (Diagram 1). Results: The study describes the concerns and discussions regarding epilepsy and driving among epilepsy.com users. Although laws related to epilepsy and driving are well known for each state, confusion among PWE and loved ones exists. Out of 11 separate words clusters generated by WAG analysis, thematic analysis yielded five themes: 1) Concerns and advice regarding driver's license; 2) Issues with public transportation; 3) Frustrations about 'no-driving'; 4) Losing driving privileges; 5) Comparison of driving with epilepsy to that in other conditions. Conclusions: Initial findings suggest that epilepsy.com users, both PWE and loved ones, have significant questions, concerns, and peer-to-peer discussions about driving issues in epilepsy. Follow-up analyses will involve categorizing specific questions and concerns from the five current themes. Funding: None
Behavior/Neuropsychology