Presentation Authors: Yuliya Zekster, Gabriela Gonzalez*, Los Angeles, CA, Carine Khalil, Kristina Vaculik, Corey Arnold, Brennan M.R. Spiegel, Los Angeles , CA, Jennifer T. Anger, Los Angeles, CA, Christopher V. Almario, Los Angeles , CA
Introduction: Characterizing women&[prime]s perceptions of urinary tract infections (UTI) helps patients achieve a better understanding of their disease and increases the likelihood of treatment success. Focus groups, interviews, and surveys are designed by providers and limited by their sample size. Online-community research allows for an enormous, international and unbiased sample, allowing researchers to hear authentic patient experiences. Analyzing thousands of posts, we identified shared emotions, prevention strategies, alternative therapies, and knowledge gaps regarding UTIs.
Methods: We collected 83,572 posts from 863 websites used internationally by collaborating with Treato, a social media data mining service, which combines search terms with Java-based natural language processing. From the randomized dataset, 200 posts were qualitatively analyzed using grounded theory to identify preliminary themes. To complement the qualitative analysis and validate a new computational technique, we concurrently applied Latent Dirichlet Allocation (LDA) probabilistic topic modeling to the full dataset for theme discovery in the entirety of the posts.
Results: Seven preliminary themes with sub-themes were identified. Users trust online communities and posts are not subject to the Hawthorne effect. Online forums share a dynamic source of information and encouragement as peers validate symptoms and feelings and advise on treatments. The diversity of posts reflects the breadth of variation in UTIs from symptoms to therapies. Regarding recurrent UTIs especially, women are frustrated and feel unheard by their providers, turning to homeopathy for possibilities. There is much information shared on behavior modification for prevention: post-coital practices, hydration, and hygiene. Women share experiences with various antibiotics and discuss diagnostic tests. There is a strong understanding of complications of untreated UTIs and fear surrounding lack of treatment. Barriers to accessing care and confusion about care permeate in many posts. There is a desire for knowledge acquisition in all aspects
Conclusions: In social media networks, patients feel supported online, users validate feelings and confirm symptoms. Providers can use this data to identify common misconceptions and improve knowledge-sharing with patients leading to improved patient satisfaction and better care through optimal communication and genuine shared-decision making.
Source of Funding: A pilot grant from NIDDK Prevention of Lower Urinary Tract Symptoms (PLUS) Research Consortium.