In order to understand the experiences of people living with Type 1 diabetes, we employed human-centered design methods of needfinding and user interviews, and facilitated participatory workshops on topics related to aging with type 1 diabetes. In total, 27 people with T1D (PWT1D), 5 loved ones (partners of PWT1D), and 7 healthcare providers (HCPs) were engaged through this research.
For recruitment, volunteers with expertise in Type 1 diabetes and aging, and staff from the University of California San Diego (UCSD) Diabetes Design Initiative (DDI), identified people with Type 1 diabetes as well as subject matter experts and clinicians in diabetes and aging to participate in the various stages of the project. This research was reviewed by UCSD’s IRB and found not to be human subjects research.
Prior to our initial user interviews (n = 13) with PWT1D, we sought to collect quantitative and qualitative data on lifestyles and diabetes management of the participants. The survey was a short Google Form [Google, Mountain View], with questions on their profession, age, confidence in utilizing diabetes technology, if participants have caretakers or caregivers, existing health conditions, active medications, and experiences with healthcare. In this context, a caregiver of a PWT1D may share in the responsibilities of and decisions with the PWT1D, ranging from daily assistance to occasional and as-needed help. Examples of caregivers include anyone from a spouse or adult child to auxiliary HCP or a paid caregiver, whether occasionally or full-time. The initial survey questions were created with the intent of establishing a contextual foundation of participants’ medical conditions, healthcare, and diabetes management. Out of the 13 participants who participated in the interviews, 11 also responded to the survey. Survey results were analyzed manually, with facilitation from Google Form’s automated response summary function.
Interviews were conducted in a semi-structured, one-on-one format. We employed the Expert-Apprentice model approach to interviews, which is a form of contextual interviews that involves the assumption that the target user is an expert at a subject. The goal of this method of inquiry is such that interviewees tell stories from which researchers can distill important facts and key insights. Interview topics covered technology usage, diabetes management, planning for aging, and community support. Participants were also asked about hypothetical scenarios since our research interest involved planning for the future and the topic of aging, and not all participants had necessarily faced aging or addressed the issues of their own aging directly. The goal for these topics was to allow participants to reflect on what they perceive would be future challenges in aging and Type 1 diabetes, and to identify possible gaps in the area of aging with Type 1 diabetes. The interviews were recorded through Zoom [Zoom, San Jose] cloud recordings with the permission of the participants, and coding of the interviews was performed following the completion of all interviews.
In addition to the primary research with PWT1D (n = 13), we also initially interviewed an expert from the Joslin Diabetes Center to gain additional insights to the experiences of PWT1D when they interact with the healthcare system.
While interviewing people with T1D was imperative to our initial progress, we additionally conducted an online participatory workshop, hosted on Zoom, for a new group of PWT1D (n = 4) and their loved ones (n = 5). Our research goal rested on our need to find recurring patterns among our participants in order to find connections between the needs of PWT1D, their overall community, and their loved ones. We structured our workshop to have two separate breakout rooms where we separately explored the needs of our two user groups: PWT1D and their loved ones. We utilized Miro [Miro, San Francisco] as our collaborative tool during the virtual workshop, where activities were conducted with the help of facilitators. For PWT1D, participants were asked to complete a milestone activity to allow researchers to understand how life trajectory and events might influence health and diabetes care. The milestone activity asked participants to identify major life events related or unrelated to their diabetes journey and it included a display of a timeline with empty sticky notes for participants to fill out. Participants filled out their stack of sticky notes with life events and placed them onto the timeline. For the loved ones, they were asked to participate in a Mad Libs activity, similar to a fill-in-the-blank. This activity was conducted to provide the research team with a clearer understanding of how loved ones approach providing care and support to their partners living with T1D.
From the interviews, each interview was recorded through Zoom and later transcribed by Otter.ai [Otter.ai, Mountain View]. Each team member made notes for each interview, and each key idea was grouped roughly to generate themes. We summarized these findings into an affinity map, where qualitative data was grouped together thematically to create a digestible summary of the primary data. We also developed two personas  based on our data synthesis.
Moving from the discovery phase of this research, we sought to further understand the specific relationship between PWT1D and their health care providers (HCPs). We conducted additional one-on-one user interviews with 6 PWT1D (5 new participants, and one prior participant from the original n = 13 exploratory interviews) and 3 HCPs. Recruitment of the participants was again made possible through UCSD DDI, who provided researchers with an additional list of possible participants. These interviews focused on addressing the key theme that emerged from the discovery phase of research, which was the concern of people with T1D about diabetes management and perceived control of management when interacting with the healthcare system and different types of healthcare providers, plus the concern of differing levels of experience that various HCPs have with T1D. These interviews focused on gathering insights about PWT1D experiences and their early opinions on developing a design solution that would support an identified gap between HCP and PWT1D: a diabetes advance directive. Interview topics also ranged from the relationship between PWT1D and HCP, conflicts PWT1D had experienced in past healthcare encounters, struggles of being a PWT1D and facing aging. Like the discovery phase interviews, these interviews were recorded through Zoom cloud recordings with the permission of the participants.
From these interviews, we again used identical user research methods for coding the data using transcripts (recorded through Zoom, then transcribed with Otter.ai) then inputting and organizing using affinity diagrams and clusters (e.g., actions, beliefs, histories, etc.) into a Miro board. These clusters were grouped around goals that relate back to the previously identified gaps and insights. After developing higher-level themes, potential solutions were then assessed. Using brainstorming activities (such as Crazy 5, Worst Possible Ideas, and Brainwriting), we identified that one of the potential solutions would be a diabetes-specific health care directive, or “diabetes advance directive”.
After developing an initial diabetes-specific care directive document, which we call a diabetes advance directive, we performed one-on-one user interviews with an additional group of 6 PWT1D (5 new participants, and another prior participant from the original n = 13 exploratory interviews) and 3 more HCPs. With our interviews, we prepared a workflow of an illustrated hypothetical and a proposed diabetes advance directive. The workflow is a comic-based spread that visualizes how the diabetes advance directive would be used in a hospital setting, from being admitted with the diabetes advance directive in hand, to the HCP understanding the needs of the PWT1D and bringing in their endocrinologist into the conversation. Within the interview, we also tested using two different tones for the diabetes advance directive, with either formal/legal or patient-centered informal language. We made iterations to the diabetes advance directive after each interview, including changing sections of the diabetes advance directive (i.e., removing list of medications, etc.) and employing additional A/B testing  with ideas like bolded or unbolded and with-highlights or without-highlights. Our usability test  included questions on general thoughts on a diabetes advance directive, the length of the document, the tone used, etc.
Similar to the previous user interview processes, we analyzed these usability interviews by making notes after the recordings (Zoom) were transcribed with Otter.ai. However, for the usability testing interviews, the transcript was also copied into Dovetail to highlight quotes from interviewees and categorize insights that were important or recurring points. Using the ‘Insights’ option on Dovetail [Dovetail, Sydney], we reviewed our highlighted quotes to group together into an ‘Insight’ that we thought was a major theme. An example of a theme: “In general, PWT1D prefer one-page documents over lengthier and wordier documents because they believe that HCPs will not spend a lot of time reading them.” Direct quotes from the interviews were tagged to these major themes as supporting evidence.
Additionally, an affinity map  was developed to formulate our ideas visually, where our core themes that we developed had quotes and ideas from the user interviews. The data gathered from interviews was also used to create two personas, one each representing a person with Type 1 diabetes (see Fig. 5, Appendix) and a typical healthcare provider (see Fig. 6, Appendix) who might encounter a person with Type 1 diabetes. For this set of personas, the person with Type 1 diabetes is someone who is troubled or concerned by their interactions with a new healthcare provider, whereas the healthcare provider is concerned about the lack of available time for a healthcare encounter with a person with Type 1 diabetes, as well as realistic about their limited experience supporting people with Type 1 diabetes. To expand on the possibilities of uses that follow the illustrated hypothetical workflow used during user interviews, we used an experience map  to detail the user journey and the ideal use cases. It is designed to be concise with specific locations that are touchpoints in which the diabetes advance directive can be used, showing the entire process from learning about diabetes advance directive to using the diabetes advance directive in a healthcare encounter.