PathFinder: patient board to facilitate discharge planning

Actionable Data Delivered in a Delightful Humane Way

Please contact me for the password to access project content.

Brief


Timeline
Jan. 2018 - Oct. 2018
Team
PM, FE and BE
Role
Lead UX/UI Designer
Targeted Users
Hospital in-patient care team
Design Scope
User Research IA Interaction Design Redesign UI Specs
Outcome
Flagship product used at 20+ hospitals with widely positive feedback

Why PathFinder


Today’s healthcare systems are increasingly focused on reducing patient flow efficiency to save costs and capital spending. Managing patient flows are shouldered by frontline care teams, who are already overwhelmed with increasing patient demands, complex dashboards and reports.

Qventus offers PathFinder solution to optimize patient flow in real time while reducing the cognitive burden of care teams.

Product Goals


When I got onboard with PathFinder project, PM and I set up meetings with key stakeholders within the company to understand both business and product strategies. I also initiated the competitive research to understand the dashboards and existing solutons of patient data display. Below are the main goals we summarized based on the internal input and market landscape.

Understand Users


We shadowed multiple hospitals to observe how frontline nurses handled the patient flows in their day to day work. Onsite interviews and observations helped uncover touchpoints where PathFinder could facilitate care teams to improve discharge process. I consolidated our learnings into the persona and flow diagram so that the whole team can leverage these user data to understand problems with enough context.

Stakeholder mapping and key persona

Patient discharge flow diagram

Targeted use case: IDR (Interdisciplinary Rounding)

How might we support charge nurse to review patients during IDR, like daily standup by the TV and to make a decision, so that patient flow and discharge are streamlined and move patients forward in the process to drive the outcome of reduced length of stay.

Design Challenge


How might we redesign the patient board to enable proactive discharge planning?

Gap of the previous patient board

Before I joint the team, there was already a patient board for IDR use case. After talking to customer success teams who have experiences implementing the board, we identified a few gaps:

1. Not configurable

Data columns on the board were manually written with SQL and it took a long time to configure every board and test the schemas.

2. Poor affordance

Some columns are editable and require user entry. However the previous UI didn't show clearly interactive fields and users couldn't locate where they need to enter data.

3. How to differentiate from other healthcare products?

Heathcare IT software mostly share a corporate blue hue which tends to look alike. How might we make PathFinder visually appealing and differentiate from other competitors?

Experience Goals

Based on these insights, I started the design project by establishing explicit experience goals for the new patient board:

Easy to configure and reuse

Actionable and transparent

Delightful and refreshing

Solution Framework


In terms of approaches to solve the design challenge and meet experience goals, my design processes can be summarized into three themes:

- Modularized and scalable board

- Actionable component

- UI exploration for stellar visual

Modularized and scalable board

Patient data taxonomy - IA

To make sure the board design meets configuration requirements, I worked closely with BE and FE to layout the information architecutre, so that each column and field aligned with how data would be presented to support user interaction.

The taxonomy table below helps the team catalog different categories of data, whether it requires user input or read only, or whether it supports special style or data validation.

Expand board space for more data

Previous table UI was challenging for adding more columns. I started exploring different board layouts to expand the horizontal space.

Focus on list view design

After the first iteration, we narrowed down to three options because of the project timeline and technical constraints. The list view became the unification point where data was presented in a progressive way, and also with the advantage to sort and filter. We also constructed the low-fi mockup covering data structure, so that moving forward we could validate both layout and data with end users.

User testing findings

When we presented three options above to end users, 4 out of 5 users chose Option Two because of it being clean and familiar. Expanding the row rather than replacing/relocating information was preferred, which made it easy to scan and compare data across rows.

After I brought these findings back to the team, we decided on constructing the list view first. Later we would investigate further into the expansion view as V2.

Actionable Component

PathFinder boards serve as the single data source to prompt care teams to enter patient dicharge plan, and collaborate on resolving barriers. User-entry data on the board is critical to patient discharge plan and also to train Qventus machine learning models to improve prediction. How might we design the data columns to encourage user input?

EDD (estimated discharge date): deliver predictive leaving date

Qventus developed machine learning models to predict when patient is going to leave. We surface these prediction results next to EDD field for entry reference. Among the solutions I explored, we chose Option D for the short term MVP solution based on the data availability.

Disposition flow indicator

When the care team are trying to decide where the patient is going post hospital visit, there are usually dependency on the order status as required by disposition. To visualize order status up front for the care team, it could help prompt the care team to check for orders so that patients will not be delayed for disposition.

Barrier design

Design problem of managing barrier is: how could care teams quickly identify important barriers and easily manage an updating list? After I explored different barrier list UI, our team decided on the solution to combine Option B and C to show an easy-to-scan and scalable list.

Outcome

During our internal feedback with clinical team, we got a lot of positive feedback. Also from tracking usage data at Amplitude, we found completion rate of these fields general high along with increasing DAU. These indicators helped validated the design improvements to prompt user input and enhance engagement.

UI Exploration

Visual identity criteria

We started exploring the visual identity by collaboraing on the moodboard, identifying a few dimensions to explore different themes and finally converge to one design that the whole team is happy with. Dimensions of visual divergence include color, layout and data highlight patterns.

Review and feedback

Along the exploration process, our product team do daily reviews to help visual direction move fast. The team arrived at the concensus that we want to convey intelligent and energetic feelings. Among these, blue wave pattern provides a lot of possibilities for animation and user delights.

Final Product UI

After three months of design work focused on modular, actionable and aesthetic boards, we put together all the interaction and UI pattterns into one demo board. Sales team have been resuing the demo board to pitch to potential clients for various use cases.

UI Delivery


During the PathFinder development, I worked closely with our frontend and backend engineers to deliver the UI specs and understand technical constraints. Thanks to the close collaborative team, I was able to conduct design QA of the product early and ensure that the final delivered board aligned with the design solution.

Responsive design

In the later stage of introducing PathFinder to more users and use cases, there is a prominent request to view the list on mobile devices. I started cataloging all the patient data by priority groups to guide the wireframe design. Next step is to validate with care team, and explore unique use cases for mobile usage.

Outcome


Widely deployed at client hospitals

Since the new PathFinder was launched in June 2018, 20+ leading hospitals and health systems are streamlining patient flow with the product, achieving results including:

0.5+ day reduction in in-patient length of stay

20%+ decrease in postop transfer delays

20%+ increase in patient satisfaction

Frontline user feedback

“With Qventus you have a snapshot of a patient without digging through any charts and without having to ask a lot of questions. You know what the plan is, where they are going, and what's keeping them here.”- Registered Nurse

“Finally, I have a technology that makes my job easier.” - Charge Nurse

Design system development

Because PathFinder UX/UI received a lot of positive feedback, our design team proposed to establish the design system to generalize the stellar UX across products and provide an enterprise suite experience. Below is the diagram representing our approaches to design system, which includes cycles of auditing, unification and inspiration.