DiaTrek
What can simulations teach healthy people about living with chronic disease?
UX
Curriculum design
Healthcare
My Role
Dreamer, Designer, Developer
This was a solo project I did during my time at the UCSD design lab, when I was a UX lead and project manager for the Diabetes Design Initiative. I drove the research, design and development needed for the project implementation.
Background
The Diabetes Design Initiative (DDI) gives design students a chance to partner with major diabetes technology companies through client projects.
Many of the students start these projects excited for exposure to the design industry, only to find that there is a huge learning curve because diabetes is a complicated disease to understand.
As a result, our industry partners find themselves spending a lot of time educating students about diabetes and helping them build empathy for people living with the condition, rather than focusing on the design project itself.
The Challenge
How might we teach students about diabetes prior to the beginning of their industry projects?
Students in DDI traditionally started learning about the disease and building empathy with their target audience when user research begins.
Speaking with past students and industry partners revealed that there is a desire to kickstart this process and start thinking empathetically earlier.
I feel like it was a little overwhelming in the beginning because it's just a lot just thrown at you. Since I don't have family members with diabetes and I don't have diabetes, it was difficult for me to really understand everything.
Leslie, Past DDI Designer
The Solution
DiaTrek is an onboarding experience aimed at simulating the cognitive burden of diabetes.
The features of this experience can be broken down as follows:
01
A bot is your diabetes toolkit.
During this week, students interact with the DiaTrek Slackbot, which creates a blood sugar value for each individual.
Students are able to test their blood sugar levels, take insulin, exercise, and log other events via commands to the bot, which will cause their blood sugar value to change over time.
Figure 1. DiaTrek Slackbot during a simulation.
02
Reflections as a learning opportunity.
Throughout the week, students also actively reflect upon the feelings, thoughts, and experiences they had simulating living with diabetes, both individually through reflection journals, and in group reflection activities.
03
Understanding the challenges of wearables.
A few students additionally were chosen at random to take on the challenge of “wearing” a fake insulin pump for 24 hours and going about their day as normal. This introduced the idea of how physical devices can also come with cognitive burden.
Figure 3. Fake pump wearable made of modeling clay. Excuse the background, I made the first one in my car!
The impact
The DiaTrek Project was first fully implemented Summer of 2022, and was used across 12 industry projects before the Diabetes Design Initiative sunset in 2024.
The implementation of DiaTrek as a portion of the DDI onboarding experience allowed our students to start thinking empathetically before even beginning their research in the space (Fig 4). This enabled them and our industry partners to focus on deeper and more focused design discoveries.
I would worry if I am calculating my insulin wrong and making a mistake. Will it lead to spikes or drops? How bad would that be if I had diabetes in real life?
I experienced a lot of frustrations related with eating out… there isn’t enough data online for me to know the carbs. And also looking these things up when I’m out with friends is just awkward.
What do you do when you’re going to friends houses to eat? Do you need to check in with them with every food they provide?
Figure 4. Quotes from student reflections during DiaTrek onboarding, demonstrating they’re thinking beyond just the physiological and also branching into the social complications that come with living with diabetes.
Deep Dive
Research
11 literature reviews and 15 designers interviews revealed opportunity for experience-based learning.
Reviewing existing literature helped me survey the existing techniques other companies and healthcare settings used to introduce people without chronic conditions to the disease space. Speaking to designers who had done DDI projects before helped me to dig deeper into places where students struggled onboarding to their industry projects.
This work led to three key findings:
01
There is an opportunity gap when it comes to teaching diabetes to people who are not directly involved with the disease (e.g. people who create the technology).
02
Students want to be able to see and interact with different diabetes technologies.
03
Simulation techniques have been effective for teaching parents and healthcare professionals how to care for people with chronic conditions.
I was particularly inspired by my third finding: simulation techniques have been used to teach caretakers how to care for chronic conditions.
So I wondered, what if we tried simulations, not for the purpose of caretaking, but to inspire empathy? After all, the knowledge you need to design good medical technology overlaps with what caretakers need to know.
I decided to code an interactive bot that simulates everyday diabetes tools.
Figure 5. Articles that inspired the project!
Figure 6. The slackbot had a unique command for each of these diabetes technologies. During the experience, we called the bot your “diabetes toolbox.”
Proof of concept
My first round of testing used a Wizard of Oz MVP where I acted as the bot for one day.
The goal of this test was to help me validate my assumption that simulations would help people learn about the disease before I went and coded anything.
I chose a student who had done a DDI project before, so I had gone in with the assumption that she would know a lot about how diabetes worked. And well, I was proved wrong.
Figure 7. Me, acting as a bot and responding to the queries from my test subject!
Despite having been in the diabetes tech space before, the participant I was testing this with thought carbohydrates and calories were the same thing, so she was constantly taking insulin for a calorie count instead of a carb count. For context, the amount of calories in foods are usually at least 10x the amount of carbs. This meant she was taking way too much insulin for the whole day!
This mistake exposed not only the failure of the previous DDI curriculum, but also the need for us to teach more about diabetes pathophysiology.
Luckily, not all was lost– my participant did demonstrate the effectiveness of the simulation in getting her to think about the mental strain living with chronic diseases creates.
Finding
Designers need more support around the details of how diabetes works.
Iteration
I added a Diabetes 101 video course as a resource to the simulation experience.
Figure 8. Screenshot from the Diabetes 101 Course, showing the available topics for students.
Validation Round #2
After cobbling together a buggy slackbot and a proper curriculum, I kicked off a second trial run.
I tested with a bigger group this time— 2 DDI staff, who are familiar with diabetes, and 2 student designers who know nothing about it.
This testing round lasted a full week and tested all of my curriculum artifacts, including daily reflections and the iterated Diabetes 101 course.
Figure 9/ Group reflection activity with round 2 testing participants.
Figure 10. Breakdown of the DiaTrek onboarding experience, which was slowly taking shape.
This round of testing was highly successful, as even the DDI staff who had been around diabetes tech for quite a while learned new things, especially about how frustrating the condition is.
I only made two minor tweaks to the program from their feedback:
Finding
I had asked participants to do daily VLOG reflections, but they were time consuming so almost no one did them.
Iteration
Considering the success of the group reflections, I changed the format from VLOGs to optional written reflections.
During this test, only one participant filmed VLOGs, and she only kept this up for two days. Looking back on this request of mine is a bit laughable because in hindsight, of course people are going to be struggling to do this on top of life responsibilities and the bot interactions!
Finding
Bots are already confusing, and the names of my commands made them even more confusing.
Iteration
I made some tweaks to the bot commands to make them clearer, and created physical command cards for participants to reference.
At this point, I was really starting to want to build a simulation app to replace the slackbot experience. Unfortunately this was a solo project, and I wasn’t experienced enough at coding at this point to build an entire app. I sadly shoveled this into my wishlist and did what I could to make the bot commands closer to natural language (Fig 11).
Figure 11. Command cards with updated commands for participants to use during the week.
Following these changes, DiaTrek was officially implemented for the first time with two groups of students for DDI Summer 2022.
Figure 12. Workshop during the first official round of DiaTrek, prior to working with Dexcom industry partners.
Reflections
Using UX to design a non-app-based experience
With the proliferation of the double diamond strategy across bootcamps and college courses, I feel a tendency of a lot of people (myself included) is to jump to thinking about app screens for mobile and web when they hear “experience design.”
This project however is proof that the UX process can be applied to contexts outside of just mobile applications. We can use the process of talking with stakeholders and doing in depth research to design rich, tangible experiences that have impact, even if we never create a single screen.
That being said, I still do hold onto my private wishlist item of turning this project from just a bot-focused experience into an app– mainly because I think it could turn it from something used in just UCSD design lab to something that can is more scalable. Perhaps that is something I will design one day!