Online health forums are used by patients and caregivers as community and information resources, especially for chronic disease management, and could help determine user needs for digital health app design. This study aims to assess the feasibility of using online forum posts on Alzheimer's disease to inform user needs in mobile health application design and whether this process can be automated through text clustering methods. A total of 413 posts were analyzed manually through thematic coding and yielded three themes and nine subthemes for patient and caregiver needs. The external evaluation showed fair to substantial similarity between the automatically and manually derived labels. Four personas were developed to assess the validity of forum-generated needs. These results establish that health forum data can provide sufficient information to understand user needs. However, further refinement of the analysis process and algorithm is necessary to generalize this method to other disease conditions and types of forum data.
©2023 AMIA - All rights reserved.