Project

This was a collaboration with IBM where a small group of people were asked to design healthcare applications using IBM Watson’s machine learning capabilities.

Duration: 5 months

Methods: Bodystorming, Concept Mapping, Design Workshops, Evaluative Research, Interviews, Mind Mapping, Personas, Prototyping, Scenario Maps, Sketching, Storyboards, User Journey Maps

Tools: Adobe Acrobat, Adobe After Effects, Adobe Illustrator, Adobe InDesign, Adobe Photoshop, Adobe Premiere, Google Suite, Sketch, Zoom

Research

Key Insights:

• Information should be tiered to reduce cognitive load

• Having relevant touchpoints in one place is critical

• Visualization of data is typically cliché and unimaginative

• User input should be qualitative over quantitative. They shouldn’t feel overwhelmed.

Methods:

• Bodystorming: After establishing the context for our investigation, our group set about meeting and bodystorming scenarios with one another to help ground where we wanted this digital artifact to take place. We used a series of prompts generated by our professor, Helen Armstrong, exploring potential futures within the context of a health application using machine learning.

• Concept Mapping: This method was used after a solid core idea had been established. We looked at our main core idea and expanded into what kind of relationships can be drawn between the core idea and the design space.

• Design Workshops: After establishing the main idea of creating a healthcare application that gives users agency over their information and appointments, several small groups worked together as a whole to come up with ways machine learning might be used. This was done by filling in a rather large matrix and sticky noting potentials that could later serve as inspiration for establishing context.

• Evaluative Research: In order to keep our project grounded in reality for which IBM could actually use, evaluative research was executed to look at precedents, general capabilities of the technology we were working with, and what resources were available to us.

• Interviews: After establishing our target audience, interviews took place. One with someone who had been diagnosed with Crohn’s since childhood, one where they are an adult and just found out they have Crohn’s disease, and one from a parent’s perspective of watching over their child who has Crohn’s disease. These interviews were recorded and transcribed, and when cross-referenced with precedents and research material, ended up being the foundation of what pain points we wished to address.

• Mind Mapping: This was an early use stage where we discussed, as a group, ideas that came to mind when thinking about a topic and mapped their relationships. 

• Personas: Our personas are based strongly on the individuals we were able to interview on the matter. We had three running ones but later decided to focus on two, and have each one engage in a unique scenario that was established through the user journey map and later in a final video. We wanted to look at where these people were in life, who they were interacting with, what level of agency they have, and how they are related to our context of Crohn’s disease.

• Prototyping: Digital low-fidelity screens were used and strung together to test what interactions may take place, how gestures may work, and what the overall ease of the screens was. 

• Scenario Maps: Visualizing the scenario’s process worked in creating the final user journey maps. After establishing the scenarios we wished to look at, we fleshed them out through scenario maps.

• Sketching: This was done rather early in the process in order to quickly get ideas across before moving to wireframe construction. Oftentimes this could be tied to stages of concept and mind mapping, as well as storyboarding the scenarios we wished to focus on.

• Storyboards: Narrowing down to two scenarios was desired in order to better focus and strategize around the personas we had created. We looked at what was needed in any given scenario using scenario and user journey maps and sketched out ideas of where pain points could be lessened through a machine-learning healthcare application.

• User Journey Maps: Taking from our scenario maps, we used that information and applied our personas to the given situations, tweaked information based on interactions that would take place, and made a user journey map. This stage was integral in deciding what screens and interactions would be taking place for our storyboard and wireframes.

Retrospective

Wider Implications:

• Building an empathetic relationship toward the self as a healthcare strategy.

• Effective proxy management leads to a correlation in a positive health agency.

• The role of delight and playfulness in de-stigmatizing sensitive healthcare topics.

Reflection:

• How does the role of a virtual companion develop over time in response to potentially growing feelings of codependency?

• What kinds of information should be collected? What information would the patient want to be collected?

• How could the design of the user experience be both friendly and lighthearted while still maintaining sensitivity to serious medical topics?