Overview

Adults with type 1 diabetes must manually take doses of insulin at every meal based on estimated carbohydrate content of their food. On top of the typical everyday problems of adult life, they must also constantly evaluate their own physical state, monitor glucose levels, and manage diabetes as just one part of their daily routine. Additionally, they must be able to recall and communicate data and trends of their own health to medical professionals. Current solutions need improvement: existing apps are a mixed bag feature-wise, and users want an app scaled back to essentials.

Design Question

How might we use mobile technology to help independent adults with type 1 diabetes balance their busy lives and their health?

MEalTIME is a mobile app to help with external cognition, allowing diabetic adults the ability to focus on the things they want to focus on. Features include meal, exercise, and mood trackers, a carbohydrate calculator, and day-by-day health reports.

Understanding Diabetes

We started our journey with a rapid round of research. 2 contextual inquiries were conducted with adults with diabetes. An interview was also conducted with a medical professional. I was responsible for one of the contextual inquiries.

Research Questions

How can mobile technology help users keep track of their medication schedules?
How do adults with diabetes manage their condition day-to-day?
What are the most difficult parts of managing diabetes?
What is the mealtime routine for people with diabetes?
How can technology help users with diabetes in general?

Affinity diagramming

After reviewing our research, we compiled and sorted data to discover the trends, needs, and problems of our user group.

Key Findings

Users often simply guesstimated the insulin amount they needed at every meal. Sometimes this results in an over or underdose.
Users were constantly acutely paying attention to their mood and energy levels, but may not always be able to trust their body’s responses.
Different users have different ways of monitoring blood sugar levels or injecting glucose.
Daily routine revolves around blood sugar levels.

Sketching an Early Prototype

Using our findings, we created a list of design goals the app must fulfill.

Design Goals

Allow users to manage the demands of diabetic treatment within the constraints of independent college or workforce settings.
Facilitate users to conveniently calculate insulin levels needed for the food they're about to eat with a greater degree of confidence than the rule-of-thumb methods they might use at present.
Monitor mood and energy-level changes in the user in a manner that requires as few steps as possible so as to minimize disruption to daily life.
Encourage & reinforce a daily routine and habits that bolster good moods and balanced energy levels based on historical data.
Give users more peace of mind with respect to their treatment and the daily habits that they are forming & continuing around their insulin and dietary needs.

Rapid prototyping

Using Balsamiq, we quickly iterated on the features needed to satisfy our requirements. These core features were tracking meals, exercise, mood, and glucose levels. We also calculated and tracked insulin intake based on meal content.

User Testing and Interactivity

4 users were asked to test the first prototype. This included one of the contextual inquiries, as well as the medical professional and two proxies.

Usability Testing Insights

The app should limit the amount of advice it gives-- we are not doctors. Focus on data management and tracking.
Features needed increased discoverability.
Data graphs should be front and center.

No more advice

The biggest takeaway from our usability testing is that we should not give medical advice. As such, we removed our second design goal and instead largely refocused our efforts on data management.



Coding interactivity

The second prototype was developed in Javascript using a wizard of oz technique with dummy data. Because we were required to program the frontend ourselves, we focused primarily on interactions rather than visual fidelity.

Final prototype

Prototype designed for mobile dimensions.

View Prototype

Reflections

Next steps

Next steps could include another round of user testing on the higher fidelity prototype as well as a more visually polished iteration of the prototype.

Takeaways

In the research phase, we could have increased our number of data points, but we were limited to the students and professors around us. Similarly, testing could have used more users who were a part of our target group. Including techniques such as user flows and personas could have focused the solution better. This project was also constrained by the requirements of the course, such as mandating our prototype be programmed by the team. As such, we were not able to have as much freedom in our final deliverable as we were limited by what we could express through HTML/CSS/Javascript. I worked as a part of the course staff the following year to remove this requirement and focus more on the design process.