Case Study

Veterans who file claims to receive benefits to cover certain health issues can sometimes get stuck in a waiting pattern. Nava is working with the U.S. Department of Veterans Affairs (VA) to build tools that reduce the amount of time it takes to process a Veteran's disability benefits claim. Together with the VA’s Office of the Chief Technology Officer’s (OCTO) Benefits Delivery Team, we developed a prototype to speed up the handlings of some of these claims. 

We developed a prototype to process one specific type of disability claim—hypertension—and released it for VA staff to use on a small number of claims. We found that the prototype, called the Rapid Ready for Decision prototype, allowed many of these Veterans’ claims for hypertension disability benefits to be processed in days instead of months.

We have now expanded the prototype to asthma claims and are evaluating new claims for suitability as well. The positive outcomes we have observed are a testament to the strength of our small scale, iterative approach and the use of prototypes when introducing new digital software.

Approach

Veterans currently experience long wait times regarding benefits claim decisions from Veterans Affairs. As of Fall 2021, Veterans could wait as long as over 100 days to hear back about a claim that impacts their health. 

VA staff tasked with these claims decisions–called claims benefits adjudicators–need access to relevant medical information, such as blood pressure readings, in order to make that call. Sometimes, that means asking a Veteran to complete a medical exam in order to get that necessary medical information. This requirement contributes to longer wait times for Veterans. 

But based on research by the Veterans Benefits Administration (VBA), we know that in some cases, necessary medical information already exists in the VA’s system. However, accessing this information is a laborious process for adjudicators. We worked from a hypothesis that by surfacing this relevant medical information to adjudicators, we could help reduce unnecessary medical exams for Veterans when possible. This in turn would help decrease the time Veterans spend waiting for a claim decision.

We approached this project by deciding to narrow the scope dramatically to quickly release a prototype to a small number of claims. Our prototype addressed one disability type for a subset of one type of claim, resulting in a few claims per day to be handled by our two pilot claims adjudicators. By starting with such a dramatically slow rollout, we could learn and iterate the prototype based on how it performed. This is a crucial step in an evolving technology landscape like the VA’s where systems are continually upgraded. Both OCTO and the Veterans Benefits Administration (VBA) supported and fostered the kind of environment where we could release and continually improve these small prototypes.

We went on to expand the prototype to asthma claims, and are evaluating new claims for suitability as well. The measurably positive outcomes we have observed are a testament to the strength of our small scale, iterative approach and the use of prototypes when introducing new digital software.

Outcomes

Our initial prototypes have allowed some Veterans’ claims for hypertension and asthma disability benefits to be processed in days instead of months. So far, processing times using our prototypes have averaged 3 to 6 days for claims that don’t need exams and 25 to 35 days for those that do. That’s compared to 80 to 100 days on average. Even with our low volume of claims, we saved Veterans over 5,000 days of waiting for a decision in the first quarter.

These are small outcomes, but that’s intentional. Our approach is to deliver value more quickly by starting small and learning through prototyping and iteration. Keeping this in mind, these outcomes are good indicators that are helping us learn and evolve our prototype. Our work is both serving Veterans and providing knowledge generation for sustainable, repeatable processes in the VA’s future.

Process

Nava collaborated with 18F, part of the General Services Administration (GSA) of the United States Government, on designing and building a PDF that surfaced relevant medical information to help claim benefits adjudicators make a decision. We pulled medical information, which includes blood pressure readings and medications related to hypertension, from Veterans’ existing VA medical records, reducing the need for Veterans to undergo unnecessary medical exams.

We used collaborative, continual user research with a small group of claims adjudicators throughout the building process, making sure our prototype fit into their existing workflow. For example, we chose to display information in a PDF because adjudicators already work with that format. This would not require a change to the interface for the system that adjudicators work with, easing the process for all. 

We worked around a core design principle that showing the right information at the right time would reduce unnecessary noise and lessen cognitive load. For the hypertension prototype, we knew that claims were decided based on existing blood pressure readings and medication data. So we worked closely with adjudicators to determine how we could present this information to make decision making easier for them. We conducted usability testing–or evaluating a design by testing it out on representative users—on several PDF designs with adjudicators, working iteratively to make new designs based on their feedback. We used what we found during testing to design and build the final prototype.

Next, we used the hypertension framework for a new condition: asthma. The asthma fast tracker is decided by medication type, dosage, and frequency. This required us to find a new way to display medication data to make sure that adjudicators could quickly find the information relevant to their decision. We took a deep dive into real medical data from former claimants and conducted research with adjudicators to define what information to surface in the PDF. 

We created several design concepts that we tested with adjudicators. We called this our “Wizard of Oz” experiment, because our team worked behind the scenes manually to simulate what the software would eventually do on its own. The experiment helped us refine our design to meet user needs quickly and efficiently. Through this process, we developed a strategy to build user feedback directly into the pilot to build a broader understanding of how adjudicators use medical evidence. 

Conclusion

Our prototype now has the ability to process asthma claims as well as hypertension ones, and will soon expand to other types of disability claims as well–currently we are expanding to other conditions and to presumptive conditions. Our small scale, iterative approach helped us build with, not just for, claims adjudicators and to continually involve them in feedback.