Senior UX Designer | Oak Street Health | 1 Month

A simple yet effective feature that reduced hospital admissions by 35%

Case Collaboration

How can we using AI advancements to create an egg freezing assistant to improve women's fertility success rates?

How can we create a new feature to easily flag and monitor high-risk patients, preventing them from avoidable ER visits?

Design Process

I believe in front-loading on user research because it's the fastest way to build a product people actually want. I strife to provide the evidence needed to support every design decision.

Role

As Senior UX Designer, I was the sole designer - I led the end-to-end design lifecycle, collaborating closely with Product Management and stakeholders from concept through launch with UI Designers and front/back-end developers.

Goal

Design an embedded “flag patient” feature that proactively surfaces high-risk cases, streamlines team oversight, and integrates within existing systems

Outcome

Delivered a feature in one month that reduced hospital admissions by 35%, cut identification time by 50% and lightened provider workload

Project Details

Meet Dr. Katz

A dedicated physician buried in admin work, taking time away from patients who need her most.
“I need a straightforward way to keep tabs on my most vulnerable patients without jumping through hoops.”

Meet Our Users: Providers Overwhelmed with Admin Tasks

“I want to feel like I’m not falling behind in life just because I haven’t figured everything out yet. I need support that gets me—something that helps me make smart decisions without scaring me.”

Tech Behaviors

Tracks cycle using Clue app

Uses Headspace for guided meditation

Shops online, values personalized recommendations

Comfortable with chatbots if they feel intelligent and friendly

  • Manual patient tracking is slow and error-prone

  • Existing tools aren't intuitive for flagging patients

  • Communication gaps hurt coordinated care

MadLib Post It Interview

I chose this type of interviewing style because the more relaxed envinorment tends to produce more honest, open feedback


Card Sorting

By separating into common patterns + themes, I was able to easily identify the overarching pain points + user needs

Analogous Research

My favorite type of research! By taking the pressure off of only analyzing the direct problem at hand, I am able to more freely ideate and brainstorm other possibilites.

Allow providers to “flag” patients - like favoriting a song in Spotify, creating a shared alert list visible across care teams without leaving the EMR

Research Methods

Must be accessible inside EMR

Must be accessible inside EMR

Must be accessible inside EMR

Cannot add cognitive overhead

Cannot add cognitive overhead

Cannot add cognitive overhead

Sparks useful peer conversations

Sparks useful peer conversations

Sparks useful peer conversations

Must fit in existing workflow

Must fit in existing workflow

Must fit in existing workflow

Unified patient data in 1 place

Unified patient data in 1 place

Unified patient data in 1 place

Quick flagging with minimal clicks

Quick flagging with minimal clicks

Quick flagging with minimal clicks

Challenges / Users' Needs

  • Quickly identify and track high-risk patients

  • Prevent hospital admissions

  • Help care teams work together better

  • Easy-to-use tool for flagging and categorizing patients

  • Real-time updates and notifications

  • A central place for team collaboration

  • Overwhelmed

  • Anxious

  • Distracted

"I don’t want to miss a high-risk patient just because I’m juggling a dozen charts at once"

Mapping User Flow

While this is not the most complex user flow, it's interesting because there are three possible starting points. This means the way to enter the flow needed to be universal and easily accessible within different parts of the EMR. We didn't have an icon like this, so I submitted a request to have it added within the Design System.

Lofi

This is only a small portion of the lofi but it strategically addressed all the user needs:

  1. Easily accessible from any point

  2. Automatic filter

  3. Reduced opportunity for human error

  4. Supports collaborative nature

  5. Very few steps/clicks

  6. Lives directly in the EMR

Translating User Research + Requirements

Design Viablility

Reasons for Success

Reasons for Success

Reasons for Success

2 Main Pieces of Feedback

2 Main Pieces of Feedback

2 Main Pieces of Feedback

Yes! Positive Feedback

Yes! Positive Feedback

Yes! Positive Feedback

"The system lets us be proactive, so nothing falls through the cracks. It's a huge relief."

Optimizations based on Feedback

Refining CC Filters

Problem
It was confusing to refine the tier level (not all CC patients are VIP — some are serious or good); some patients are in-active, but still appear on the list

Solution
Restructure the toggle to be an overall page filter (see problem 1 above), then remove tier as a tab, but instead have it within "More Filters" tab

Design Reasoning
This was a way to consolidate the tabs into more over-arching categories, such as: View, Provider, Center instead of having more minor details be a complete tab (ie: tier status)

Remove Toggle Button

Problem
It was confusing what was actually being switched "on" so we needed a way to create more clarity

Solution
To combat this, I changed the structure of the page. I added "Case Collaboration" as a new page filter instead of a toggle within the existing page

Design Reasoning
This solution provided more clarity and uniformity from a systemic level

Final Design

Key Call Outs

Key UX Feature Breakdown

Icon to "Like" "Favorite"

Research
Users need a quick button that's easily accessible

Design Strategy
Include a button that's similar to favoriting a song on Spotify that will live next to the patient's name that can be accessed anywhere within the EMR


Quick Summary

Research
Providers need a way to get top line data - reason for high-risk, who last saw the patient and when

Design Strategy
When the "up arrow" is filled in, a summary appears when hovered. Users can click to edit/archive.

Filter for Collaboration

Research
Users need a way to easily locate high risk patients to help other providers collaborate on a plan

Design Strategy
Have a quick filtering menu

Measuring Success

Improved continuity of care and fewer emergency visits

Improved continuity of care and fewer emergency visits

Improved continuity of care and fewer emergency visits

50% faster identification of high-risk patients


50% faster identification of high-risk patients


50% faster identification of high-risk patients


35% reduction in hospital admissions


35% reduction in hospital admissions


35% reduction in hospital admissions


Lessons Learned

The Good

With just 3 clicks, patients lives are being saved!! It's so simple

The Bad

The limitations of the EMR software make the UI look very out of date and not as sleek as I wish it could be

Opportunities

If I had more time, I would want to make the filter easier to access (less clicks). But it would mean changing it from the ground up, which was a larger scope than what we had the resources for at the time