PORTFOLIO

Medicaid and CHIP Financial (MACFin) Product:
Expenditure Submissions

Overview

MACFin is an enterprise application that is replacing an out-dated Medicaid enterprise financial system within the Centers for Medicaid and Medicare Services (CMS). Expenditures Submission is one of several modules in MACFin. States will use Expenditures Submissions to submit their quarterly Medicaid Expenditures to CMS for oversight and monitoring.

To prepare their submissions, states use external financial systems to gather and prepare their data with the end result being complex Excel spreadsheets that can contain up to 100 tabs for some states. These spreadsheets are used to manually input data into the system by state submitters. Their submissions are approved by state reviewers and certifiers. CMS reviewers must also certify the submission.

Challenges

Designing a complex data-driven system that can contain over 100 integrated forms, and feeder forms that "feed" line level details into each of them.
Planning user hierarchy and roles to accommodate different types of users: state submitters, state reviewers, state certifiers, CMS reviewers, CMS analyzers and Auditors.

My Roles

Assisted with User Discovery Research
Lead UX Designer

Tools Used

Zoom
Dovetail
Figma
Confluence (for creating/documenting user research plans, interview guides, links to conceptual models for easy team access, etc.)
Jira (for maintaining a project backlog and stories needed to complete design and development)

UX Design Methodology and Process

The following UX methodologies and processes were used for this project. Additional information for each process, what was involved, and outcomes are detailed below.

Discovery Interviews
Analyzing data obtained during contextual interviews to determine key findings using Dovetail
Creating artifacts and reports to share findings
Creating low-fidelity conceptual models in Figma
Validate conceptual models with users
1. Iterate as needed
Creating high-fidelity prototypes and screens using conceptual models and requirements
1. Iterate as needed
Usability Testing

Discovery Interviews

Discovery interviews were planned and conducted with 2 types of users

State users who submit expenditures for their state or territory
Internal CMS analysts who review expenditures submitted by states

Discovery Goals

Learn about the users: their role, amount of experience in their role, where they work, how they work, their level of technical experience
Learn about the current process state users take to prepare and submit their CMS expenditure packages
Understand when the current CMS application fits into their workflow and processes, and how
Identify user needs and pain points; what they like or dislike about the current process
Identify opportunities for improvements to create a better user experience

Study Methodology

Conducted remote 1- to 1.5-hour (video) interview/observation sessions via Zoom
User to demonstrate and explain their current process, preferably through walking through key aspects of the process and completing some tasks

As facilitator I used a prepared Research Guide that contained semi-structured questions to explore other topics and issues using the following moderation techniques
Concurrent Think Aloud (CTA)
- CTA encourages participants to keep a running stream of consciousness by thinking aloud during the session. CTA can help moderators understand participants’ thoughts as they interact with the prototype and complete. Observing an individual’s behaviors while thinking aloud gives us the closest representation of what their experience would be in real life.
Concurrent Probing (CP)
- CP requires that the moderator ask follow-up questions of participants as they work on tasks. When they say something interesting, or do something unique, or seem confused, etc., the moderators will ask follow-up questions to probe for additional information.

Data Analysis and Synthesis

Interviews were recorded via Zoom and imported into Dovetail whereupon detailed analysis was performed including affinity mapping. Critical thinking was applied for mapping data into several logical categories including:

Data Synthesis in Dovetail

Key Findings

Data is keyed in manually using complicated spreadsheets involving over 100 tabs
- User Quote: "It takes probably three to four solid days of keying into [the current system]. That's not just one person, my whole team gets in there."
It’s easy to make errors while keying; takes several days to input all data into [the current system]
Variances between a state’s expenditures and their previously submitted budget aren’t identified until after all data has been keyed in and submitted which causes additional work.
The system provides very little guidance on where to find identified variances or other errors. State users spend considerable time searching for the errors.
The list of forms available is extensive. They are not listed in any order nor can they search for forms or filter them in any way. They manually hunt for a form they need which is complicated because many forms share a similar name. User quote:

"You can see how many forms there are, they're going all the way down the page. They’re in no particular order and there is no way to sort it. It drives you crazy."

Although there are numerous reports available within the system, not all reports are useful nor is it clear what information a report will produce. User quote:

"There are a lot of reports in here that I'm slowly going through, painstakingly going through, and saying, well, what does that report give me? Well, it says this, but when I run it, it really only gives me this."

Discovery Artifacts

Using data obtained during discovery interviews, the following artifacts were created: User Personas, Journey Maps, Presentation of Research Findings

Personas --

Using data obtained during discovery interviews, two user personas were created to represent a state user and a CMS analyst/reviewer. The personas allowed us to identify users as “real” people and to understand their needs and wants to help inform design decisions. Note: I contributed data to these personas but another member was responsible for the design and layout.

Persona for Aadya, a State User

Journey Maps --

I was solely responsible for designing two Journey maps for our key users: state submitters and a CMS analyst/reviewer. Using Adobe Illustrator, I created a visualization of the process users must complete to prepare and submit expenditures, or for a CMS analyst to complete a review. In order to have a complete picture of the journey, I included steps in the process, pain points along the way, user quotes, and possible areas for improving the experience and functionality. The maps were well received and very useful to the team. A quote from one user:

“This is a delightful document. I've seen a lot of similar documents with swim lanes, but this is really quite intuitive.
It also sort of takes you on an adventure. It's very pleasingly done.”

Journey Map for Aadya, a State Submitter

A presentation of Discovery findings was also created and shared with CMS stakeholders and the product team: PowerPoint of Research Findings

Identified Improvements

Working with user discovery findings my UX team mates and I identified opportunities to improve functionality and the user experience. A few opportunities identified were:

Ability to upload/import data to reduce the time spent manually keying in data, and to reduce errors
Maintain the ability to do manual entry for corrections and small modifications
Make forms easier to find using search functionality with sorting ability to filter using form names or keywords
Functionality that prevent errors during the submission process instead of at the end
Functionality to identify and flag variances between the state’s budget and expenditures before submission is completed

Conceptual Models

With the completion of user discovery and analysis of findings, the next step was to brainstorm and design conceptual models. The team prepared several screens that included ideas for how to solve pain points such as the ability to upload forms and the ability to search and filter forms. We then shared our concepts with some of the same state users to validate our conceptions and obtain additional feedback. The feedback was positive:

Ability to upload data via templates or do manual entry. Supporting quote from user:

“If we have multiple forms and we can just upload those forms. Yes, definitely that would be very beneficial and would definitely help expedite the data entry process.”

Ability to easily find and select a specific form; improved sorting and presentation of Expenditure forms being submitted. Supporting quote from user:

“It looks good to me. If you can make it like this, that it has the filter and maybe an archive option. I think that's perfect for what we need.”

Designing High-Fidelity Prototyping

High fidelity, interactive prototypes are currently being designed for new Expenditure Submissions module. There are over 70 screens needed for the entire submissions module so it's an on-going process! For this case study, I’m sharing selected screens created for the initial submission steps.

Expenditure Submission Dashboard Prototype

When a state user logs in to submit expenditures, they first see a

Dashboard that lets them create a new Submission, resume a

submission already in progress, or view a completed submission.

Expenditure Submission Landing Page Prototype

If a user is creating a new submission, the next screen will be a submission

screen without any forms added. To get started state users can select to

download form templates (to populate and then upload) or add forms

to their submission page for manual entry.

Add a Submission Form for Manual Entry

If a user choses to add forms manually, a dialog opens. They can

use filters in the dialog to narrow form selections by name or

type (this addresses a common pain point learned during discovery).

Highlighted Design Challenge

As one would expect for enterprise systems with an abundance of data, there are going to be design challenges. Here is one of many that have caused me to put my thinking cap on and do some brainstorming. The links below will open a Figma prototype that showcases the legacy UI, explains the problem, and 2 design concepts for improving compliance tracking in the new MACFin system.

Table Redesign for State Compliance Tracking

As stated previously, design and development is currently on-going. Additional research and usability testing is now being completed by a dedicated HCD researcher who was added to the team in 2024. The current roadmap for the MACFin project has a projected launch date of January 2026 whereupon integration testing begins before the final launch.