The Department of Veterans Affairs (VA) helps veterans transition into civilian life by providing benefits and services. VA disability compensation offers a monthly tax-free payment to veterans who got sick or injured while serving in the military.
The current disability benefits rating system is outdated and not automated. It requires manual calculation and data entry by hand, which could result in an inaccurate rating decision. It also takes time. VA is experiencing a surge in backlogged claims. As of January 2022, there were about 260,000 backlogged claims. On average, the claims process takes about 100 days.
I was the lead UX designer for the project, responsible for discovering user needs, ideating site features, and creating all UX deliverables. The team includes two UX designers and one UX researcher.
To understand adjudicators' needs and pain points when evaluating a veteran's disability, we have conducted user interviews.
The UX researcher and I conducted the user interviews to understand users' pain points and needs. We have recruited two participants who are working as Rating Veteran Service Representatives (RVSRs) at VA.
The RVSRs’ time is wasted when a claim is not ready for a decision due to insufficient exam data and has to be deferred. They manually have to go through all the data, which slows down the rating process. It also leads to a “hurry“ to meet production standards, which increases the chance of making mistakes.
The rating process requires the RVSR to review the veteran’s medical and service treatment records, as well as any submitted Disability Benefits Questionnaires (DBQs) for blood pressure readings. In some cases, the RVSR has to manually aggregate blood pressure readings.
Continuous medication is used to control hypertension. Participant 1 said that they relied on their knowledge of hypertension medication names, instead of looking up relevant information on the system.
"I physically count the blood pressure measurements and how many meet the specific criteria. Then I determine the percentage of them that meet that criteria."
"Medication names are really just knowledge. As you progress with institutional knowledge, you just start knowing which medications are used for hypertension. If I don't know, I google the medication name itself. And Google is a great resource. You use Google or WebMD search.”
Taking all the insights from the research, I created personas that represent our users. It helped me identify the key strategies to tackle the issues the users are facing.
Nicholas is a seasoned rater with 15 years experience at VA. He is old-fashioned and prefers to use pen and paper while working on claims. It normally takes about an hour to process a single claim. His major pain point is that he depends a lot on his knowledge and experience, so he makes mistakes here and there. It takes extra time to look up regulations and medications information.
David is a young rater who has just joined VA as a rater a year ago. He is very motivated and eager to learn. However, he finds it frustrating that he's not familiar with all the federal regulations and how to evaluate claims. He goes to Google or WebMD to find necessary information for rating claims.
A system that allows faster claim review and accurate evaluation
After synthesizing users' pain points and needs, our team came up with the three key strategies to improve their rating process.
The system automatically calculated the highest rating for veterans, which reduces human errors caused by manual data entry
Display the blood pressure readings in a chart and table to provide insights into a vast amount of blood pressure data
Display Code of Federal Regulation (CFR) and medication information reduces the time to look them up outside of the system
The goal of the design is to reduce human errors and rating time by streamlining the rating portal and automating the rating system.
First of all, I created a wireflow of the rating system to help me visualize the claim rating process. It helped me understand the system as a whole and how the user gets to the screen I’ll be designing.
Based on the initial findings from the user interview, I came up with the wireframes with a dashboard interface. I decided to use a dashboard layout since there is a variety of information RVSRs need to look up when rating a claim. It includes a list of blood pressure readings, evidence source files, Code of Federal Regulation (CFR), and so on.
I presented my wireframes during the weekly Product Refinement meetings to gather feedback from the design team. After 10 repetitions of the feedback session and iteration, some of the team members were still questioning the way they navigate the screen. There was a lot of information on a single page, which confused users as to where to look at first. I decided to go back to pencil and paper to reimagine the page layout.
While the dashboard UI displays everything at once, I wanted to guide the users eyes' from more important sections that need their manual entry to the final evaluation page. I created the simple userflow of the screen, which pinpointed three sections where important actions happen. First of all, the user reviews the necessary information for rating. Then the user gets to review all the available blood pressure and medication data. Lastly, he/she grants the evaluation based on the evaluation criteria and medical data.
Based on the userflow, I came up with a sidebar menu for more streamlined navigation. This layout was well accepted by the team and I developed it to mid-fidelity wireframes to gather insights from actual users.
Based on the userflow, I came up with sidebar menu for more streamlined navigation. This layout was well accepted by the team and I delveloped it to mid-fidelity wireframes to gather insights from actual users.
Our final product allows users to review a wide range of data at a glance using visualized chart and table. It also suggested the highest possible benefit ratings for veterans, which reduces the processing time and risk of manual entry process.
The new system suggests the highest possible evaluation based on the data from all the evidence. It also provides the reason why the system is suggesting the specific percentage. Users no longer need to calculate a large number of data manually and can easily rate complicating cases.
For hypertension rating, the system displays the blood pressure chart that shows the change of blood pressure. Instead of going through multiple blood pressure reading data, users can track the change of blood pressure and medication data by looking at the chart.
The visualized blood pressure data table shows the list of blood pressure readings in hierarchy. It groups the data that results in the same evaluation and highlights the group of data that will yield the highest rating.
It is common for users to have multiple windows on their laptop for references. To tackle this pain point, the users can easily look up necessary legal information for rating inside of the system, without navigating out of the system.
Users mentioned they would depend on Google or WebMD to gather information on medical information. We have implemented the tooltip interface on the name of the medication, so that users can quickly gain information within the same page without having to refer to 3rd party website.
We conducted usability testing of web prototype to validate our design decisions. We turned to the two rating specialists we interviewed in the beginning. They gave 5/5 stars to our design and provided valuable feedback.
It normally takes an hour to process a single claim. Users expect that he new system would be able to reduce the process time in half.
The system-suggested rating could significantly reduce the human error and increase confidence of the users.
Users find it intuitive and easy to learn to see the federal regulation and checklist that shows how the rating is achieved.
Rating criteria displayed with visualization of blood pressure readings will help newer raters learn the process of rating hypertension.
My biggest lesson from this project is not to get too invested in the initial sketch. Developing design with initial sketches resulted in spending extra time and effort to change entire layout in the future. Next time, I would create as many as sketches and gather feedback to explore possibilities and select the optimal layout for my design.
One of the factors that attributed to the successful design was to listen to the right users. Since the system is specifically for the users who works for VA and work as rators, we wouldn't been able to gather relevant insights if we didn't hear from them. Knowing the users is the key!