simonthesun | UW Medicine

Overview

Context

UW Medicine’s Digital Experience team partnered with the Human-Centered Design & Engineering program at the University of Washington to sponsor a master’s capstone project, seeking a holistic redesign of the online carefinding experience.

Problem

Navigating all the necessary details to book an appointment is overwhelming to patients, with information being spread across various difficult-to-navigate channels, sometimes even providing conflicting data.

Goal

Deliver a single online source of truth and simplify the process for users trying to find care on the UW Medicine website.

Timeline

January 2023 - May 2023
(18 weeks)

Role

UX Design
UX Research

Team

1 Researcher
2 Designers

Deliverables

High-Fidelity Prototypes
UX Report

Solution 2

CareCorner

CareCorner is a persistent navigation tool available to anticipate needs and provide clarity. It is available on every page and uses aggregated user trends to generate appropriate suggestions based on what page the user is on and their navigation history.

A comparative analysis grid filled with notes and screenshots

Tools

Research: Refining the Problem

The initial design brief was extremely open-ended and may lead to “boiling the ocean” given the complexities of the healthcare space. As such, we needed to solidify a specific design question using our research.

Methods

Site Mapping
Competitor Analysis
Literature Review
Usability Studies
Surveys
Individual Interviews

Sample size

39 survey respondents
5 user interviews
2 SME interviews

Tools

Qualtrics
Google Sheets
FigJam

Insight 1

Participants preferred not to have to call to book appointments if they can avoid it.

Insight 2

Location and connection were top priorities users looked for in a provider.

Insight 3

Lack of clarity in regards to insurance was the biggest frustration brought up.

Insight 4

Participants often felt lost and consistently wanted “all the pieces in one place.”

Screenshots of webpages and components of the UW Medicine project are overlayed over a blue background with a pill bottle.

Design: Insights to Solutions

Our insights formed a clear picture: users’ key need is information processing, rather than access.

Going Digital-First

A pop-up window titled "book appointment online." The bold text reads "what is your primary concern?" A text box has text typed into it, reading "I have a rash that isn't going away." Three suggested searches are clickable buttons, labeled "finding primary care," "nausea and fever," and "cold and cough." A pagination tracker is at the bottom, with labels for "primary concern," "insurance," "find a provider," and "book an appointment."

Insight

Participants preferred not to have to call to book appointments

solution

Open-ended symptoms field using NLP to suggest who to see, reducing manual searching and setting up an online-first booking experience

Rationale

We wanted to bring the online experience on par with the flow of a call, to provide guidance without need for speaking to someone on the other end

Making a Match

A pop-up window titled "book appointment online." The bold text reads "We found 6 providers who can address your concerns" at the top of the page. Three cards each featuring a provider with portrait and details are laid out, with an arrow to scroll for more on the right. A filters column lies along the right edge, showing location, gender, and specialty. A pagination tracker is at the bottom, with labels for "primary concern," "insurance," "find a provider," and "book an appointment."

Insight

Location and connection were top priorities

solution

Provider suggestions and expanded filters

Rationale

The current system required a manual search, which created huge cognitive load. By meeting the user halfway with suggested matches, we hoped users can put their focus onto their location and connection priorities.

Simplifying Coverage

A pop-up window titled "book appointment online." The bold text reads "Does your insurance cover it?" at the top of the page. A card prompting the user to upload an image is below. A divider line reads "OR" and below that is a text box to manually type in the insurance name. A pagination tracker is at the bottom, with labels for "primary concern," "insurance," "find a provider," and "book an appointment."

Insight

Lack of clarity in regards to insurance

solution

Optional insurance verification step including automatic card scan to provide a clear, definitive answer to coverage

Rationale

Users previously needed to manually check if they were in-network. The data is on the website, so a straightforward yes or no can provide clarity.

Narrowing the Haystack

The billing and insurance page of the UW Medicine website. The quarter-circle of CareCorner is in the lower right, reading "It seems like you are trying to pay a bill." A button is underneath reading "bill pay."

Insight

Participants often felt lost.

solution

Predictive navigation guide to point users to relevant information when outside the booking process

Rationale

The UW Medicine website is massive and analytics show that users enter the site in more ways than the homepage. We wanted to provide guidance while avoiding a chatbot feature.

Step by Step

A pop-up window titled "book appointment online." The bold text reads "Confirm primary concern" at the top of the page. The body text asks the user to confirm that they are dealing with a rash and suggests seeing a primary care provider. Two buttons are laid out, one to go back and one to see next steps. A pagination tracker is at the bottom, with labels for "primary concern," "insurance," "find a provider," and "book an appointment."

Insight

Participants consistently wanted “all the pieces in one place.”

solution

Progressive disclosure to break down the booking process and avoid overwhelming the user

Rationale

Interviews revealed that users wanted a streamlined step-by-step system. Additionally, competitive analysis showed that other hospital systems use a similar pattern.

An animated gif showing the scrapped card sorting feature. Instructions at the top read "use this space if you wish to sort through providers." Below that, a button reads "open notes." Five cards lie underneath, each showing a provider portrait, name, and details. The mouse drags the first card between the second and third, shifting the cards and placing the card down. A pagination tracker is at the bottom, with labels for "primary concern," "insurance," "find a provider," and "book an appointment."

Testing and Changes

Due to time constraints, we utilized iterative concept testing through the prototyping phase rather than multiple rounds of formal usability testing. This helped us rapidly evolve our designs in each pass from wireframes to high fidelity.

This process also served as design-as-research, producing new findings. For instance, users needed information provided to them, but did not necessarily need help processing it. Originally we had a spatial sorting feature, but users found it confusing and unhelpful, and thus the feature was scrapped in favor of more familiar patterns.

Key Learnings

As the capstone to my master's degree, this project allowed me to apply my classroom knowledge to a real-world project.

Learning 1

Problem Comes First

Given the vastness of the healthcare space, it's important that we solved for a very specific problem. The right design starts with the right question.

Learning 2

Stay Flexible on Process

When facing constraints such as tight deadlines and losing team members, staying adaptable let us keep trucking forward when we would've otherwise been paralyzed by a "perfect process."

Learning 3

Hats Change Easily,
Morale Does Not

When we lost a researcher early in the project, I stepped up and took on multiple roles, embracing a startup mentality. However, team morale was low and took a while to recover. I learned that wearing many hats and doing more work is often necessary, but to truly come back from a setback as a team requires more than just a redelegation of tasks.