- myAURA: Personalized Web Service for Epilepsy Management The development of personal health libraries is particularly important for those with chronic conditions such as epilepsy, since information abundance makes it difficult to understand best treatment options or even relevance of information to each personal case. Qualitative and quantitative studies show that persons with epilepsy and their caregivers (PWEC) need visually engaging, easy-to-use, online tools to: (1) extract, classify, organize and personalize information and (2) provide automated, proactive recommendations in support of evidence-based decisions about treatment and self-management. There are currently no online tools that integrate relevant information for PWEC. They must conduct a large number of separate searches in different resources and manually comb through often irrelevant results and a confusing array of medical and self-management options. The proposed research aims to directly address this problem with innovative data- and network-science methods to securely integrate multiple resources into myAURA, a personalized, easy-to-use web service. Novel contributions include computation of a large-scale epilepsy knowledge graph together with a network inference method to remove redundant information, which are used to identify, visualize, recommend and personalize relevant information. The interdisciplinary team of experts in biomedical informatics, text and social media mining, visualization, user interface design, and epilepsy self-management will collaborate with patients, caregivers, and their advocates to design the tool according to their needs. Furthermore, stakeholders such as the Epilepsy Foundation of America, have granted an exclusive use agreement to obtain PWEC data from their website, discussion groups, and social media presence. MyAura will integrate practical, location- and patient-specific health-care information with targeted scientific literature, biomedical databases, social media, and epilepsy-related websites with information about specialists, clinical trials, drugs, community resources, and chat rooms. It will build on innovative data and network science methods pursued in the research aims.
Aim 1 : Produce an epilepsy knowledge graph by integrating sources of large-scale data such as social media, electronic health records, patient discussion boards, scientific literature databases, advocacy websites, and mobile app data;
Aim 2 : develop recommendation and visualization algorithms based on the automatic extraction of backbones of the knowledge graph, which are likely to contain information that is relevant to specific user interests derived from electronic health records, web searches, social media and activity on epilepsy.com.
Aim 3 : User-centered development and pilot testing of myAURA with end- user studies, to validate if and how it improves patient activation. The established partnership with epilepsy stakeholders is an excellent opportunity to build a pilot personalized health library that responds to patient needs. Toward a generalized framework, we will document and share our pipeline to serve patients with other chronic conditions or general users to track broad health interests.
- myAURA: Personalized Web Service for Epilepsy Management Epilepsy provides an ideal case-study for developing a personalized health library, as it affects more than three million Americans who face a bewildering array of medical and self-management options. Experts in informatics and epilepsy self-management come together with engaged stakeholders to develop novel data- and network- science methods, which can sift through vast amounts of information from medical research, social media, electronic health records, mobile apps, and websites, and extract information best-suited to an individual patient's needs. Completion of the project will yield the initial version of an easy-to-use web service (myAURA) to aid patients and caregivers in making decisions about epilepsy care and management, as well as result in a novel methodology that can be used to develop similar tools for other chronic conditions.
| Azoulay, Pierre; Graff-Zivin, Joshua; Uzzi, Brian et al. (2018) Toward a more scientific science. Science 361:1194-1197 |
