Centering the Disabled User Experience of Health Information in a World Driven by Artificial Intelligence: A Mixed Methods Investigation

Abstract

Disabled people are more susceptible to infectious diseases (like COVID-19) than nondisabled people; finding accurate, relevant health information is especially pressing. Internet search technologies are often touted as empowering disabled people who seek healthcare information online, but does this depiction reflect reality? This dissertation encompasses three studies assessing disabled user experiences in finding health information online: a structured literature review, a cross-sectional survey, and a concurrent think-aloud study. The literature review shows that, across 11 articles, disabled people were often not considered in public health messaging surrounding COVID-19. We then analyze 142 cross-sectional survey responses about usability and satisfaction regarding web-based COVID-19 information. Usability and satisfaction were both lower in people who had developmental or mental health disabilities (r=-0.21, p=0.0131 for usability; r=-0.24, p=0.0111 for satisfaction). Satisfaction was also lower among screen magnification or closed caption users (r=-0.21, p=0.0262). In the concurrent think-aloud study, ten participants were asked to internet search four prompts and narrate their experiences in real-time. Themes included concerns about accessibility/usability, AI-generated information, peer-reviewed articles, hospital or government webpages, news/advertising, and sentiment/trust. Participants also reported physical fatigue (n=5) and distracting layouts (n=5) while searching online. All participants encountered AI-generated information in their searches. This dissertation ends in reflection on how future work by scholars in health and information sciences should be shaped by the input of disabled people.