Exploring Self-Tracking Practices of Older Adults with CVD to Inform the Design of LLM-Enabled Health Data Sensemaking

Abstract

Wearables and mobile health applications are increasingly adopted for self-management of chronic illnesses; yet the data feels overwhelming for older adults with cardiovascular disease (CVD). This study explores how they make sense of self-tracked data and identifies design opportunities for Large Language Model (LLM)-enabled support. We conducted a seven-day diary study and follow-up interviews with eight CVD patients aged 64-82. We identified six themes: navigating emotional complexity, owning health narratives, prioritizing bodily sensations, selective engagement with health metrics, negotiating socio-technical dynamics of sharing, and cautious optimism toward AI. Findings highlight that self-tracking is affective, interpretive, and socially situated. We outline design directions for LLM-enabled data sensemaking systems: supporting emotional engagement, reinforcing patient agency, acknowledging embodied experiences, and prompting dialogue in clinical and social contexts. To support safety, expert-in-the-loop mechanisms are essential. These directions articulate how LLMs can help translate data into narratives and carry implications for human-data interaction and behavior-change support.

Figures (4)

• Figure 1: Overview of the research process. The study combined a seven-day self-tracking diary with follow-up semi-structured interviews to explore how older adults with CVD engage with self-tracking technologies and interpret their personal health data. Diary entries (text + screenshots) provided contextual insights and informed interview prompts. Interview transcripts formed the primary dataset, which was analyzed using reflexive thematic analysis to develop themes grounded in participants’ lived experiences.
• Figure 2: Eight older adults with cardiovascular disease engaged in a seven-day diary study and follow-up interviews. The images represent the study data. Online/In-person interview sessions (A, B, C). Photos uploaded by participants during the 7-day self-tracking diary (D, E, F, G). During one of the online interviews, a participant demonstrated a feature in a mobile health application to illustrate why they found it useful (A). In the in-person interviews, screenshots that participants had submitted during the 7-day self-tracking diary were printed and discussed. Some participants also used their wearable devices alongside the screenshots to elaborate on their experiences (B). As part of the interview procedure, participants were invited to interact with the WHO Health Agent “Sarah” SarahWHO and to engage in brief conversations about physical activity knowledge (C). Throughout the 7-day diary study, participants uploaded screenshots from both the wearable interface (D, E) and the mobile health application (F, G). Participants expressed appreciation for the quick daily activity overviews provided by their devices (D) and for motivational feedback on achieving activity goals (E). They also highlighted the value of health metric dashboards for providing a concise overview of key indicators (F), and found anomaly summaries, such as low heart rate notification, particularly useful for monitoring and interpreting potential health changes (G).
• Figure 3: The diary study data were organized and summarized using a visual mapping approach on a Miro board (left). The horizontal axis represented the tracking day number (Day 1–7), while the vertical axis represented the participants by participant IDs. Each diary entry was assigned to the corresponding coordinate, yielding a structured overview of responses across time and participants. Examples of participants’ uploaded screenshots are shown on the right.
• Figure 4: Study procedure. Participants were registered electronically and completed a seven-day self-tracking diary with daily reminders. Prompts asked them to reflect on daily activities, describe how they accessed and interpreted health data, and upload screenshots of device interfaces. These diaries were then used to guide one-hour semi-structured interviews, which also included general reflections on tracking and discussion of AI-supported chatbot tools. All interviews were recorded with consent and transcribed verbatim for analysis.