Background: Top-level executives and strategists decided we should improve the Meds 360 product by adding in the ability for users to look up and be credited for PDMP data. This would increase our data accuracy and provide as an additional selling point while allowing users to combine all their meds lookups into our singular platform.
Goal: Discovery research into PDMP usage and behaviors to inform the design and build into the Meds 360 product.
Process:
Discovery research - informal user interviews, desk research
Problem definition - Interviews, findings & personas
Solution discovery - user flow
Solution definition - mockup
Outcome: This project was based on an acquisition with another company that housed PDMP data. The acquisition fell through so the project was stopped until further notice.
Prescription Drug Management Program Investigation
Cureatr, 2019
UX Researcher & Product Designer
Discovery Research
Desk Research
I started out with some informative desk researching. Reading up on what the program was about, what it’s goal is and the regulations around it.
What is PDMP? PDMP stands for the Prescription Drug Monitoring Program. The goal of the program is to allow more transparency around controlled substances by empowering providers with a report of all the controlled substances a patient has ever been prescribed. A few things to note about PDMP that impacted the project:
It is regulated by state.
Not all states participate,
Each state has a different online database that users have to connect to in order to check a patient’s PDMP.
Providers and pharmacists are supposed to check a patients PDMP report before providing controlled substances (varies again by state).
Credit is granted for each PDMP lookup which is a key metric that hospitals and providers are responsible for.
Generative / Informal Interviews
I outreached a handful of providers who I had already had done interviews with before or who our customer success team was close with. I held free form conversations around what these provider’s understanding of PDMP was, how they use it (or don’t) at their practice and what their impressions of it are.
The information I got from these informal interviews helped structure a more robust research session that more directly informed designs.
Problem Definition
The generative interviews acted as a gateway to a pool of other questions we had about PDMP. At this point, we also had learned about more technical constraints that would play into the product build, so we were able to ask more pointed questions in the next round of research.
Formal User Interviews
I sent out a screener to around 50 participants and selected a population who ranged on their familiarity with PDMP and their use of it at the moment. The research questions I went into these interviews with was:
Who uses the prescription drug management program software?
Do people like PDMP?
Why do providers not use PDMP?
What’s the frequency of their usage when prescribing controlled substances?
What causes them to pull a PDMP report?
Is there repercussion for not checking PDMP before prescribing a controlled substance?
How would they expect to access this information within the current product ecosystem?
Findings
The research uncovered two personas which I will delve into in the next section. At a high level, it revealed that the prescription drug monitoring program is severely underutilized. This is mostly because of it’s lack of usability, slow loading time, and an extra step in the workflow. There are also no repercussions for providers for not using it, although in theory a hospital or provider could get audited if a situation were to arise.
Personas
Reactive Rachel, primary persona
The primary persona I named “Reactive Rachel”. This user references the program infrequently, and only in response to a request or a concern, such as drug seeking behavior. Oftentimes, they are comfortable prescribing without referening the platform. They believe the program is a good thing and it should be utilized more, but find it too much of a hassle to check for every patient.
"It's definitely an extra step so that's probably why I don't use it as much as I probably should."
Proactive Peter, secondary persona
The secondary persona is “Proactive Peter”. This is a physician who typically works in a clinical setting in which pain killers are extremely common to prescribe, such as surgery or radiation therapy. This provider is in the habit of always checking a patient’s PDMP as part of their review of the patient. They appreciate having this extra clarity and data into the patient’s situation and consider it another piece of the clinical puzzle.
“The more data, the better”
How might we make the PDMP lookup process more seamless so that our users are encouraged to look at the data more frequently?
How Might We...
Expose PDMP data in fewer steps
Help paint the holistic picture of patient medication history for providers
Make the process easier to get data
Encourage more PDMP lookups
Solution Discovery
Technical Constraints
PDMP data has a loading lag of 10 seconds.
The UI and ability to data scrape a PDMP report has to be approved on a state by state basis. We get a link to an HTML report that we can show in an iframe but we are not allowed do anything else with the data until we get state approvals
Important to note is also, users are given a PDMP credit when they request a report, even if they don’t see it. This played a factor in the decision on how to load the report.
Proposed solution
It was generally agreed upon that PDMP reports are good things, and even users who barely touch the PDMP platform expressed they should use it more.
So I advocated for a solution that automatically requests the report for a patient, when the provider pulls a med rec from our platform for the patient. This way, the loading lag is less cumbersome and the report is readily available to view. Previously, users have to go to the state website, enter the patient demographics and states to search, and then wait for it to load. This would be one click away from the task they were already completing, and in a place where they are looking to get clear medication data about a patient.
Due to the technical constraint around state verfication of the UI, I also proposed we just load an iframe with the HTML report we receive. It is not the best visual experience, but it is already what providers look at when reviewing these reports and allows us a quicker build.
Solution Definition
User Flow
User flow depicting how a provider would access the PDMP report within the product ecosystem.
Mockup
Due to the constraint of not being able to scrape the data to put into our UI until every state approved the designs, we decided to just load an iframe of the HTML report in it’s own tab next to the medication history tab. This report, while not visually stimulating or challenging design-wise, is what providers currently see when they pull a report like this. So it is comfortable for them and could actually build trust with us that this is the same information they would find on their state website. It also made the build able to happen very quickly.
Outcome
Integration of the PDMP data was coming from acquisition with a small company that worked with PDMP data. About a month after we landed on the design and engineering had a test build running, we got the news that the acquisition fell through. This meant that the project was terminated, but what we learned throughout this process was still valuable in better understanding our users and the medication reconciliation landscape.
