HUMAN CENTERED DESIGN AND EVALUATION OF AI APPLICATIONS IN MEDICAL HEALTH

Abstract

Through this research study, the study aims to facilitate AI-based medical health applications both local and international, more accessible and useable to semi-literate users of Pakistan. At the initial stage, systematic literature review (SLR) identified critical factors regarding AI-based medical health applications and semiliterate users. Twelve factors are identified within three categories: Usability, User Experience, and Content & Design. These factors included Ease of Navigation, Memorability, Learnability, Efficiency, Effectiveness, Accessibility, User Satisfaction, Interactivity, Readability, Ease of Use, Content, and Design. MCDM techniques like AHP and Fuzzy AHP are applied to prioritize the factors. From the results, the top six significant factors are selected for further analysis; these are Ease of Navigation, Learnability, Effectiveness, User Satisfaction, Memorability, and Readability. Then, a further SLR is conducted for choosing applications, which include ratings from the application, downloaded statistics, presence of AI features, and user reviews. Next, two validation methods are applied for this research. First, we will consider the match of 10 selected AI-based health applications to prioritized factors with expert assistance using the Google Form survey. Applications in compliance with these factors are further selected for the next method: heuristic evaluation. With this heuristic evaluation, experts draw attention to several usability issues and provide recommendations to enhance the applications' usability. Among the selected three applications that are WebMD, Healthwire, and Dawaai, least user experiences were satisfied through these applications, and for this reason, they are selected in the next phase of heuristic evaluation. A prototype is created with the feedback, which included all the changes it planned to make based on the issues. The same experts reviewed this prototype to check whether the new design met their expectations and whether they fully agreed with the proposed improvements. This iterative process highlights a need for AI-based health applications to be culturally sensitive and accessible, especially for semi-literate users in Pakistan. The research focuses on critical usability factors and iterates the design based on expert feedback. The improvement achieved is in terms of usability, user experience, and accessibility of AI-based health applications to cater to diverse populations more effectively.