MediGuide

Inspiration

Customers with a health concern often turn to Google or an AI chatbot instead of Baylor Scott & White's own Help Me Decide tool. They can't navigate to the tool in the MyBSWHealth app, don't know whether to trust it, and get recommendations like "go to urgent care" with no explanation. We designed MediGuide to fix those four problems: discovery, trust, repetitive questions, and recommendations that lead nowhere.

What it does

MediGuide is a redesign of Help Me Decide. It recommends a level of care, explains why, and connects the user to that care. It never diagnoses.

Discoverability: a "Not sure where to go?" card on the app home screen, a symptom search bar, and quick-tap symptom chips.
Trust: a "How this works" screen states what the tool does, what it does not do, and how information and photos are used.
Efficient symptom sharing: type, speak, or tap. Every question can be skipped, and earlier answers carry forward so no one answers twice.
Explained results: a care ladder (self-care, schedule a visit, urgent care, emergency) shows where the user lands. A "Why this?" panel ties the recommendation to the user's own answers.
Care access: one-tap booking, urgent care directions, nurse line, or virtual visit, with a summary passed along so the user doesn't repeat themselves.
Safety: warning signs trigger an emergency screen, a nurse is reachable from every screen, and uncertainty defaults to the higher level of care.
Inclusion: plain language, large text, voice input, and [Spanish / caregiver mode, only if shown in the prototype].

How we built it

Designed in Figma within 24-hour window; version history is linked. We used the free Material 3 Design Kit for base components and built all screens, flows, and content from there.

Defined the problem and four breakdown points from the challenge brief.
Researched target audience, diversity & inclusion, AI usage, easy navigation methods.
Built three personas to cover different ages, health literacy, and tech comfort.
Prototyped one complete flow: home screen, how it works, symptom input, result with explanation, next steps.

Challenges we ran into

The prototype explain recommendations using the user's own answers and warning signs, never possible conditions.
Limits and data use are shown in short, layered text so they don't deter less tech-savvy users.
Designing for very different users without adding a learning curve.
First time using Figma, so we scoped to one complete flow instead of many partial ones.

Accomplishments that we're proud of

A complete, clickable path from discovery to care access.
The "Why this?" explainer, which answers the main complaint about unexplained recommendations.
Safety and accessibility built into the core flow.
Yet to test with real users.

What we learned

Trust is a design problem. People choose search and AI because they get clarity and an explanation.
A recommendation without a reason is a dead end.
Stating the tool's limits increases credibility in healthcare.
Testing early with people unlike us exposed problems we would have missed.

What's next for MediGuide

Test with older adults, people with disabilities, and Spanish speakers, and measure findability and completion rates.
Validate the care ladder and warning-sign questions with BSWH clinicians.
Review photo and health data handling with BSWH privacy and compliance teams.
Expand language support and screen-reader compatibility.