As Artificial Intelligence (AI) adoption expands, mental health chatbots are increasingly used for emotional support and self-management, yet their roles and responsibilities remain ethically ambiguous in everyday use. Prior research often treats these boundaries as externally defined, overlooking how users interpret and negotiate them in practice. We analyzed a large set of app store reviews from four mental health chatbots using topic modeling and qualitative analysis. We identify three recurring processes: how users infer roles through interaction, how they negotiate ethical tensions during use and breakdown, and how they collectively define boundaries through reviews. These processes indicate that users infer roles from how the chatbot communicates and extend trust cautiously, setting boundaries on what they rely on it for. Breakdowns prompt reassessment, while reviews serve as sites where users define acceptable behavior. These findings suggest that ethical boundaries are actively constructed through interaction and platform participation, indicating the need for designs that better support how users interpret, negotiate and share judgments about AI behavior over time.

Despite growing interest in mobile games for older adults, there is limited exploration of older adults’ gaming behaviors, perceptions, and experiences as they engage with casual puzzle games over a period of time. To address this, we conducted a 9-month study with 20 older adults, examining training needs, in-situ experiences, and preferences. Participants were trained on tablet PCs and ten selected games. During the study, participants documented their experiences and attended technology workshops. Gaming behaviors were logged and analyzed using descriptive and inferential statistics, revealing patterns and statistically significant differences in play frequency and duration over the course of the study. Thematic analysis identified facilitators and barriers to engagement such as customization, co-play experiences, and health issues. Based on these findings, we recommend incorporating educational elements, enhancing user control, leveraging identity and nostalgia, supporting social interactions, designing for tangible interaction, and emphasizing the importance of learning aids. Future research should test the effectiveness of these recommendations in increasing older adults’ engagement with casual games.

Personal health informatics (PHI) aims to provide individuals with personalized, actionable health insights through digital platforms. Despite advancements in digital health, a standardized framework to evaluate the personalization of digital health solutions remains absent. Developers often rely on "free-form" or "weak" personalization, contributing to high user attrition and non-personalized "cold start" experiences. This paper examines the evolution of personalization in digital health, covering early web-based systems, data-driven advancements, behavioral theory integration, and AI-driven personalization. We propose the development of PERSEID (Personalization Index for Digital Health) to measure and enhance personalization in digital health. This index, guided by a comprehensive ontology, will assess multiple personalization dimensions, adapting to users' evolving behaviors over time. By addressing current gaps and establishing a validated personalization framework, PERSEID aims to enhance the effectiveness of personalized digital health systems, ultimately advancing the field of PHI and supporting improved health outcomes.

This paper presents a conceptual framework for designing relational agents (RAs) in healthcare contexts, developed through the findings from multiple user studies on RAs about their acceptance, efficacy, and usability. The framework emphasizes a user-centered design (UCD) approach that takes into account the unique needs and preferences of patients, non-patient users, and healthcare professionals (HCPs). Based on the results of these studies, we analyzed and refined the RA designs and proposed a UCD-based conceptual framework for designing effective and user-friendly healthcare RAs. The paper aims to provide an initial resource for researchers, designers, and developers interested in developing RAs for healthcare contexts by thinking of UCD techniques.

Tools that can help older adults self-manage multiple health goals in collaboration with their care managers are rare to find. Informed by the Self-Determination Theory, Goal-Oriented Care paradigm and our prior findings, we used an iterative, user-centered process to design a tablet application to facilitate Goal-Oriented care in community-dwelling low income older adults with chronic (multi)morbidity. A formative in-situ evaluation was conducted in which 20 participants used the app to set and track health and wellness goals for 24 weeks, while participants' interactions with the app were logged. At the end of the study, semi-structured interviews were administered to understand how the app was used. Thirteen participants used the app throughout the study, while the remaining abandoned after short usage. Thematic analysis of the qualitative feedback shows that participants who used the app increased their commitment towards their goals and adopted healthy behaviors. Health issues, time constraints, lack of technical know-how and doubts about goal-setting paradigm were identified as primary reasons for low app usage and abandonment. Tools for Goal-Oriented care should support personalized goal exploration, build trust in the care paradigm, support collaboration, design for motivation, lower barriers to tracking and support re-engagement after abandonment. Carefully designed mobile apps have the potential to support Goal-Oriented care for older adults.

User personas play a crucial role in user-centered design (UCD) methodology-based technology development. In this way, appropriate user personas improve the design of interaction interfaces between a relational agent (RA) and its users, particularly when the RA is about delivering health services to the general population. This article discusses the approach for establishing patient personas to help in designing and developing interaction scenarios and dialogues between a health service RA and its target users during various COVID-19 stages. We employed a mixed-method research that was completed in two phases involving both healthcare professionals (HCPs) and individuals who have encountered COVID-19. Our approach resulted in the identification of three COVID-19 related user personas i.e., experiencing symptoms, COVID-19 positive with mild symptoms, and recovering from COVID-19.

Due to the increasing rates of mobile phone adoption in low-income communities in the United States, mobile apps can be used to increase access to prenatal care in this population. But, design and evaluation studies in this area are rare. Using existing guidelines and needs assessment results, we developed, MomLink, that aligned with the needs of the target population and their community health workers. Nine women from a low-income community evaluated the app and provided suggestions to improve its design. After making suggested changes, we deployed MomLink and its corresponding provider app for a pre-pilot evaluation. The evaluation met an unexpected barrier resulting in low usage of the system by both women and providers. Based on our findings, we discuss opportunities to improve mHealth apps for the target population and better ways to collaborate with community-based partners.

Portion size estimation is important for managing dietary intake in many chronic conditions. We conducted a 6-week field study with nine varying literacy dialysis patients to explore the usability and feasibility of a dietary intake mobile application that emphasizes portion size estimation. Seven participants demonstrated sustained use of the application and improved their self-efficacy, knowledge, and ability to estimate portion sizes in pre- and post-study assessments. Participants reported moments when portion size information in the application differed from their prior understanding, challenging them to reconcile dissonant information. Although participants acquired new knowledge about portion sizes, they struggled to accurately estimate portion sizes in situ for most foods. Despite using the application consistently, rating it highly, and exhibiting learning, we found that self-efficacy and knowledge are not sufficient to support improved behaviors in everyday life.