by.

Anthem Claims Processing system

UX/UI Design | User Research
Research and Design

01.

Introduction

Overview

Anthem is a health insurance company, and this project is for the Federal Health Products & Services (FHPS), which is a part of the Anthem Federal Program. They are using an enterprise system called Streamline to manage the structure and process of care management and its supporting systems for health insurance. Streamline is a multifaceted and multiyear project. Anthem envisions leveraging advanced technology to transform the existing legacy system into a new system named "Fusion".

Client

Anthem Insuarance

Role

UX/UI Design

Tool

Figma

Timeline

May - July 2019

Problem

The insurance claims processing system was complex and tedious. Anthem claims processors had to use a legacy system that was 20 years old, with around 30 supporting applications and completely different interfaces and logins.

02.

Research

Plan

Throughout the project, I maintained a close feedback loop with the claims processors and business team to gather their input and insights on the existing system that they had been using for many years. Understanding their values and expectations was crucial in designing the new system to meet their needs and the broader business requirements.

Methods

Survey

A total of 218 respondents participated in the survey, which aimed to gather feedback on the usability and user experience of the Streamline system. The survey consisted of two sections: the first section collected information about the participants' roles and demographics, while the second section asked questions about their experience using the Streamline system.

Interview

Conducting user research was crucial to gain a comprehensive understanding of the needs and challenges of the users of the existing system. It was essential to gather insights about the context and mindset of the users, as well as the issues they had faced over the years.

To achieve this, I conducted six contextual interviews with Claims Processors and Operation experts, as well as six interviews with Product Owners, in a semi-structured manner.

Findings

My role was  conducting and overseeing the entire research process, and I communicated the findings to all the key stakeholders.

Existing Work Flow

Users had to utilize 30 different applications and navigate through 20 streamline screens in order to process a single claim. The lack of consistency in the design and user interface of these applications was a major issue.

Individual User Journey


By looking at individual user journeys, I uncovered issues where users were wrongfully converting, getting lost in the site, and abandoning products that they signed up for

User Behavior

Using a combination of data from interviews, user tests, and surveys, I gained insight into how users preferred to use and learn about existing products. For example, one finding was that users appreciate the ability to open multiple screens to compare and edit, even though the existing system may be less efficient.

User Motivators

User interviews and surveys helped to identify the features that resonated the most with the users, which were later cross-referenced during the testing phase to determine what worked best for them. While the steps for processing claims may vary depending on the type of claim, Claim Processors typically reference manuals and paper documentation to determine the necessary steps. This requires them to memorize a significant amount of information, which can be challenging for new hires.

"I use technology and apps for my personal which are way advanced and I don’t understand why I need to got back and learn something which is so old and not leveraging on any of the new technologies - A New Claims Processor "
Enthusiastic People
Manual Performance tracking
lengthy Instructional Manual Books
A desk full of sticky notes to remember passwords and commands
Screen with lot of apps in use
Slower Claims adjudication

I interviewed claims processors, some of whom were new hires, while others had been working with Anthem for more than a decade. Those who had been working for more than 10 years had memorized the commands and developed a process over time to handle the claims. I observed that most of them had adapted their own personal styles to track their performance, use manuals, and remember all the commands and logins.

Task Analysis

To do the task analysis I break down the process in small task and list activities that user has to perform . With each task I evaluated the problems user face while performing these tasks. Through my research process I discovered that no single person had end-to-end  complete process knowledge. Through the in-depth analysis of the process I found numerous redundancies and areas prone to error.

Understanding Claims Types

During the secondary user research it was found that their are many different kind of claims types which comes for settlement for the and each claim need a different treatment. It was imperative to divide these several

Automating Claims

There were several claims that required simple matching of certain informations those can be easily automated to reduce the workload on CPS.
I made a list of all such claim types with the help of claims processors and operations experts

Error Prone Claims

There were few types of claims which are more prone to error because of nature of verification it requires.
I made a list of all such claim types with the help of claims processors and operations experts.

Claims which need a letter to be sent to the provider to process

Some of the claims needs a letter to be sent to the provider and pricing claims needed to be treated in a different manner

Ideas on Paper

The first step was to jot some ideas on paper to get a quick feedback and iterate for low and medium fed wireframes.

Initial Designs- Getting User Feedback

Putting high fidelity prototypes in front of our users allowed us to gather more accurate user feedback.

Key Findings

list of Claims in a tabular format
Minimised Menu
Complicated Filters and No Notifications
Weekly Claim info in graphs
Table Cutomization
Too many features for table View

03.

Iterations

             Additional Insights

a. Shortcuts: Ctrl + First Letter

The claims processors are trained to use the keyboard and prefer it over the mouse, as it speeds up their overall performance. Shortcuts for each action point are provided to enable faster processing.

b. Multiple Window View

The claims processing screen consists of multiple windows that display the information necessary for the claims processor to review.

c. Custom Attributes

In order to ensure that users only view the necessary information, we delved deeper into product usage and user behavior, and discovered valuable insights. As a result, when a user clicks on a claim, they will only see information specific to that claim type. We were able to achieve this by utilizing claims documentation, which outlines the steps required for each type of claim.

d. Customizable Task list

To meet the needs of different styles of users, the design includes customization options in all menus.

Out Comes

Improved Performance Monitoring

Based on the user feedback, the designed prototypes improve the overall performance monitoring.

90 % Increase in User Satisfaction

User test shows 90% increase in overall user satisfaction and ease of use.

Potential Increase in Claims Adjudication

With new designs user expect improvement in speed and increase in claims adjudication

80% Decrease in time to process every claim

No more Juggling between many applications helped reducing process time

08.

Conclusion

Reflections

This project presented many challenges, such as understanding the insurance process and Anthem's claims handling methods. I learned that breaking down a large task into smaller ones can be very helpful.

Another challenge was finding a balance in the design to address the emotions of end users who have been using the system for over a decade and the needs of the business team. Integrating ideas and creating a cohesive flow was also challenging.

I would have liked to spend more time correcting user flow discrepancies and designing for edge cases to provide a more comprehensive solution. Despite these challenges, I am proud of what we achieved, especially considering where we started in the first iteration.

Thanks For Reading