Personal Informatics in Interpersonal Contexts: Towards the Design of Technology that Supports the Social Ecologies of Long-Term Mental Health Management
This blog post is a summary of our CSCW 2018 paper “Personal Informatics in Interpersonal Contexts: Towards the Design of Technology that Supports the Social Ecologies of Long-Term Mental Health Management” by Dr. Elizabeth Murnane, Caitie Lustig, Tara Walker, Beck Tench, Dr. Stephen Voida, and Dr. Jaime Snyder.
Background
Recent years have seen a significant increase in the availability of personal informatics (PI) technologies for helping people capture various details of one’s life and reflect on personal information. These tools range from Fitbits and other wearables to health and wellness apps for tracking exercise, food, mood, and more. For many people, the motivation to use these technologies is a desire to gain self-knowledge, get healthier, or manage a chronic condition. In our study, we focused on people who use PI tools for tracking the symptoms of bipolar disorder (BD), a chronic mental health condition that causes fluctuations in mood, energy, and activity levels.
Research shows that tracking everyday activities like sleep, social interaction, and exercise can improve the outcomes of BD treatment. Often, this tracking is done in collaboration with others, such as family members and clinicians. For instance, a family member might watch out for patterns of significantly less sleep (a symptom of a manic episode), and clinicians sometimes discuss tracking records with patients in an effort to help make sense of acute changes in behavior or mood.
However, most PI tools are focused on the individual, rather than facilitating such collaborative forms of care. Furthermore, these tools are often built around normative assumptions about what is desirable and expected when it comes to tracking practices and metrics, fueling a sense of inadequacy in users whose own experiences and outcomes don’t align with such norms. One of our participants illustrated how the perception of what is “normal” is highly personal:
“Pushing it into categories is so difficult to know: are you high, medium, or low/high? What does that mean?! […] It also makes me feel like everyone else has such a different perception of how the world works. That can be further alienating in some ways.”
A main motivation of our work is therefore to enhance the design of more socioculturally-aware PI tools for people managing conditions like BD.
Study
This research was conducted at the University of Washington (UW) and the University of Colorado Boulder (CU). At UW, we interviewed 14 people with BD; at CU, we conducted focus groups with 8 people who either had BD or were involved in the care of those individuals. Participants used a wide variety of methods (e.g., lists, journals, handwritten charts and graphs, sticky notes, calendars, spreadsheets, paper-based or online BD forms, smartphone apps, and wearable devices) to track various indicators and behaviors, including mood, medication adherence and dosing, sleep, exercise, weight, finances and spending, alcohol/drug/tobacco use, and general routines.
Sharing such data as well as the responsibility of tracking it was important to participants. For instance, we found that it is not unusual for trusted friends or family members to take over the monitoring and capture of personal information when an individual is unable to log data themselves. Participants also described using tools like a shared calendar to collaborate with caregivers. With respect to clinicians specifically, participants’ perceived advantages of sharing data include an ability to provide “hard evidence” of a change in symptoms or ground questions about trends or surprises revealed through the data. At the same time, participants did raise concerns about privacy and agency, to which collaborative care setups must be sensitive.
Application and Extension of the Ecological Systems Theory Model
Nearly forty years ago, Bronfenbrenner developed the influential Ecological Systems Theory model (EST) to demonstrate how “macro” level factors, such as social norms and laws, co-exist with “micro” level influences, such as family relationships and peer support, to create individual experiences. We applied and expanded this model based on our interviews and focus groups to characterize the complex and dynamic social relations that exist in and around the long-term management of serious mental illness (SMI). In particular, we added a temporal dimension, signaled the mediating role of technology in an informatics layer, and identified a series of properties to describe the ways in which multiple layers interact.
Design Implications
Based on this model, we developed some specific design recommendations for the creation of PI technologies that might be used in this context:
- PI technologies for collaborative care must be designed to protect the privacy and agency of people managing illnesses, and these systems must take into account that relationships between those involved in care are dynamic. Connections will form, break, and change over time. The interactions between the microlayer and temporal layer may change the valence, intensity, direction, and dynamism of a relationship).
- It is important that PI systems let people compare their data to others’ and avoid normative framing of expectations for how these data should “look.” In other words, concerted attention should be devoted to understanding the relationship between the societal macrolayer and the more user-centered informatics layer.
- For people managing SMI, self-tracking is about crisis mitigation and crisis management. An impending crisis was not typically signaled by just one event, but by a series of cues. This is represented by multiple tensions and intersections among the person managing the SMI, the informatics layer, and the temporal level.
Careful system design is essential, particularly when introducing and/or limiting access to personal data within the challenging and dynamic social context of managing SMI. Ultimately, these design implications point to ways in which PI systems can enable individuals to bridge and maintain connections across the multiple layers of their social ecosystem, while remaining sensitive to potential unintended consequences.
