Annie T Chen1. 1. School of Information and Library Science, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599-3360, USA. atchen@email.unc.edu
Abstract
OBJECTIVE: This study sought to characterize and compare online discussion forums for three conditions: breast cancer, type 1 diabetes and fibromyalgia. Though there has been considerable work examining online support groups, few studies have considered differences in discussion content between health conditions. In addition, in contrast to the extant literature, this study sought to employ a semi-automated approach to examine health-related online communities. METHODS: Online discussion content for the three conditions was compiled, pre-processed, and clustered at the thread level using the bisecting k-means algorithm. RESULTS: Though the clusters for each condition differed, the clusters fell into a set of common categories: Generic, Support, Patient-Centered, Experiential Knowledge, Treatments/Procedures, Medications, and Condition Management. CONCLUSION: The cluster analyses facilitate an increased understanding of various aspects of patient experience, including significant emotional and temporal aspects of the illness experience. PRACTICE IMPLICATIONS: The clusters highlighted the changing nature of patients' information needs. Information provided to patients should be tailored to address their needs at various points during their illness. In addition, cluster analysis may be integrated into online support groups or other types of online interventions to assist patients in finding information.
OBJECTIVE: This study sought to characterize and compare online discussion forums for three conditions: breast cancer, type 1 diabetes and fibromyalgia. Though there has been considerable work examining online support groups, few studies have considered differences in discussion content between health conditions. In addition, in contrast to the extant literature, this study sought to employ a semi-automated approach to examine health-related online communities. METHODS: Online discussion content for the three conditions was compiled, pre-processed, and clustered at the thread level using the bisecting k-means algorithm. RESULTS: Though the clusters for each condition differed, the clusters fell into a set of common categories: Generic, Support, Patient-Centered, Experiential Knowledge, Treatments/Procedures, Medications, and Condition Management. CONCLUSION: The cluster analyses facilitate an increased understanding of various aspects of patient experience, including significant emotional and temporal aspects of the illness experience. PRACTICE IMPLICATIONS: The clusters highlighted the changing nature of patients' information needs. Information provided to patients should be tailored to address their needs at various points during their illness. In addition, cluster analysis may be integrated into online support groups or other types of online interventions to assist patients in finding information.
Authors: Michelle L Litchman; Heather R Walker; Ashley H Ng; Sarah E Wawrzynski; Sean M Oser; Deborah A Greenwood; Perry M Gee; Mellanye Lackey; Tamara K Oser Journal: J Diabetes Sci Technol Date: 2019-03-10
Authors: Nicola Luigi Bragazzi; Howard Amital; Mohammad Adawi; Francesco Brigo; Samaa Watad; Gali Aljadeff; Daniela Amital; Abdulla Watad Journal: Clin Rheumatol Date: 2017-05-01 Impact factor: 2.980
Authors: Alexis Koskan; Lynne Klasko; Stacy N Davis; Clement K Gwede; Kristen J Wells; Ambuj Kumar; Natalia Lopez; Cathy D Meade Journal: Am J Public Health Date: 2014-05-15 Impact factor: 9.308