From a month-long process to a one-day solution.
i created a tool that stops overpriced prescriptions before they reach the counter.
The Challenge
“I sure do love paying too much for my medications” said no one, ever.
CVS’s Pharmacy Benefits Management (PBM) division does so many things to ensure that every member can get the drugs they need at a reasonable, affordable price. However, what isn’t explicitly known is how complicated that process can be.
Every single day, new drugs are released into the market and it’s the job of the PBM to review that report every day and make sure there’s nothing that could bankrupt someone accidentally makes it’s way through into our system.
That process to ensure identify and flag these drugs took 30 days on average.
We needed to understand what was causing this right away. So to make that happen, our team of two designers and two engineers set out to uncover what can be done to streamline this process to make it faster, smarter and generally easier for the users doing this every day.
CVs Health
PROJECT NAME: Hyper-Inflated Assistance Portal (HAP)
DELIVERABLES: Production ready designs for platform MVP
MY ROLE: Lead Product Designer
CHALLENGE: New drugs are released to the market every day. How might we find a way to streamline the identification of hyper-inflated drugs, for our members?
SOLUTION: A user portal to both review, organize and sort new to market drugs based on drug type and cost. Furthermore, an improved, streamlined process to push the findings all the way to the pharmacy.
BUSINESS VALUE: A decrease in new to market overpriced drugs accidentally being dispensed members the day they are made available, rather than weeks to months.
USER VALUE: End to end workflow is reduced from 30 Days to 1 Day, removing manual interventions and automating data sources.
WHAT I LEARNED: Human Centered Design means more than a tangible design, it means understanding everything around the problem space and finding the true problem and not the perceived problem. Keep asking questions and see where it takes you!
THE REsults:
-97% reduction in task completion
-Members (you and I) pay less for drugs
The Discovery
To get started, we used one of the more efficient tools we as designers have in our tool belts - the user interview!
Since this was all remote, we accomplished this through the use of Microsoft Teams and Cisco WebEx. Through a number of interviews, ranging from a few minutes to over an hour, my design pair and I would facilitate a discussion with various users on various teams to get a better understanding of their day to day. One of us would take the lead on the discussion while the other one would be taking notes and recording the session.
Often times these interviews would lead to more interviews as we would uncover more nuances we wanted to fully understand before we felt comfortable starting to solution. Each of these calls would be used to generate a process flow deliverable for our stakeholders to have a good understanding of where the issues of the flow can be found, and to make sure we’re solving the actual issues and not a perceived solution.
After each interview, we would return to our virtual data dump board in Miro to unpack what we learned and further map out the whole process as well as identify the various pain points along the way.
The Solution
With the lo-fi screens mapped out and the data maps supporting the screens, we moved to a more high fidelity design. We actually started the designs in Sketch, but later migrated to Figma as the team made an executive decision to migrate our design tool (big fan). We would account for not only the MVP screens, but the edge case screens or the transition screens. We would validate and revalidate with our users, or even seek out feedback from fellow designers that could offer a fresh perspective.
Once the screens were generated, we had the luxury of a PM that would be able to act as a user as he came from the team that would be using this. His feedback helped fine tune the design iterations and ultimately got us to our MVP push to production!
The ideation
With our research synthesized, our paint points prioritized and our stakeholder’s buy-in; we started sketching out ideas. We would try to go super wide to begin. Lots of “hey this might be a crazy idea” statements thrown around, tons of sketches and lots of feedback. But what was especially fun about this process was our ability to incorporate our engineers and PMs into this phase to seek out as many solutions as we can. They, like anyone else, can offer a unique perspective that ultimately lead us to some solutions we all agreed on. More specifically some UI patterns that account for the issues identified.
Slowly but surely, the UI and flow started to take shape. There would be more consistency in our mockups and the number of scrappy shapes and napkin drawings started to diminish (but let’s be real, those never fully go away). While the designs started to get more polish, we also made sure to account for what would be occurring on the backend of this proposed platform.
So we would generate UI concepts, while also creating new proposed backend process flows that would accompany these designs. We wanted to make sure the engineers would be included in this as they’ll be the architects connecting these screens together.
Additionally, we would validate and revalidate with our users that we wouldn’t be missing anything major should we continue in a particular direction.
The Impact
Results not only improved the time needed to process these new to market drugs, but the users were thrilled with the automations in place
Automated systems reviewed the new reports once released to organize the drugs and compare them to similar drugs
Users spent less time formatting reports and could start reviewing potentially problematic drugs, immediately.
What was once a month long process could now be completed, end to end, in a single day.
the Key Takeaways
The entire database provisioning process (initial request to hand off), can take up to 30 days.
The new to market drugs are released every day through a report subscription known as the “Medispan report”.
The findings in the report are then sent to a team to review the findings, but the meeting occurs once every 7 days. Before then the findings simply sit and wait.
Once reviewed, the findings are passed to another team to have the updates “coded” into our system - this process could take weeks due to the simple fact that they are not sent in a format that can be ready by our systems.
The majority of the 30 days was spent waiting, then being reformatted
The actual coding of the findings took only a few minutes but the larger process was causing the hold up.