Discovery
The product team embarked on this project with so many unknowns. However, we did know for certain was that we were maintaining two platforms with a lot of overlapping functionality. This meant that we had two separate clinical teams under GMS and GMA, and these team members were creating records, prepping for patient visits, and conducting visits on these two separate platforms with (presumably) different workflows. This also translated in two separate tech stacks maintained by different groups of engineers who had expertise on one platform or the other. To complicate matters, each platform was built to perform specific functions. For example, while the GMA platform was build specifically to conduct patient visits, the GMI platform could facilitate test order and authorization.
We knew that this multi-level redundancy was not sustainable, but we lacked clear direction on how we should tackle this problem on a high level. Should we bet on one platform and try to build in the functionality it lacked? Should we use one platform for certain tasks (like test order) and another platform for another (like patient visits)? Should we cut or losses and start over with a new platform?
It was impossible to answer these questions without getting a better sense of where we were starting. We had very little documentation around clinical processes, both the workflows themselves as well as qualitative feedback from the clinical team. We knew we needed to capture a lot of this information before moving forward with decision making.
One key metric we knew from the start was the contrast in clinical efficiency on the GMA and GMS sides. While our gross margin for clinical services on the GMA side was 40%, the gross margin on the GMS side was -2%. Moving forward, clinical efficiency seemed to be the key area to investigate.
Research Goals & Process
- Document key workflows from clinical users on both the GMS and GMA sides by interviewing leads from the following teams: genetic counselors, genetic counseling assistants, care coordinators, intake coordinators, administrators, and customer success
- Document areas of efficiency and pain points on both platforms
- Extract findings from the data gathered and create easily digestible, sharable documentation to better facilitate decision-making
- Assess duplicative functionalities and pinpoint which platform was stronger in each area
Key Research Findings
There was a great deal of variation in workflows between GMA and GMS. While clinicians were performing similar functions, they were working with different toolsets. The systems were built in two very different ways with different central functionalities. While the GMA was built primarily for genetic counseling services, GMIP was built with more varied functionality, such as the ability to order and authorize tests.
The functionality around the following areas on Pioneer led to a great degree of clinical efficiency
- Patient scheduling: both the scheduling platform used (Acuity) and the workflows
- Clinical documentation: the structure of the templates, which affected the workflows around preparation and completion
- Availability planning: the component that allowed administrators to manage genetic counseling schedules
Project Goals
Once we identified the key workflows and efficiencies/pain points on both sides, we had to decide on a path forward. It was also clear that such a complex problem couldn’t be solved by the product team alone. We gathered a cross-functional team to participate in the decision making, including leaders from engineering and clinical operations. Though we had a great group of people who all brought valuable subject matter expertise, different leaders understandably had different sets of priorities. This made homing in on a singular, core focus challenging.
To help bring some high level clarity, I facilitated a whiteboarding session with production management to establish a set of guiding principles that distilled the most important focus areas for the project. We landed on:
- Make user experience improvements in areas that will have the most immediate impact on the company’s revenue and ability to scale efficiently.
- Deliver a cohesive, consolidated experience for clinicians and partners through a singular platform.
- Make choices that support the design and engineering teams’ ability to build, scale, and maintain the product as effectively as possible but not at the expense of delivering needed improvements in a timely manner.
- Create a single source of truth in all cases, from systems to data architecture to content management.
- Take into account the company’s long-term vision of programmatic and longitudinal care but not at the expense of more immediate revenue goals.
After the whiteboarding session, the Prod-Eng team regrouped. We agreed there was no way we could reconcile every element of the two platforms as well as all the different workflows and capabilities associated with them. To that end, we landed on a phased, component-based approach that would allow us to use the capabilities of both platforms without fully transitioning to one or the other nor immediately building something net new.
High-Level Phase 1 Goals
- Increase the allotted availability and the number of patients we see per week for our genetic counseling services
- Increase genetic services gross margin
- Standardize genetic counseling assistant, intake coordinator, outreach coordinator, and care coordinator workflows and set baseline metrics
- Conduct all genetic counseling visits on a standard platform
Outcomes
An important step for us was mapping out a workflow that took into account how all users would interact with the GMI and GMA platforms as well as key backend interactions. This comprehensive swim-lane diagram I created allowed members of both the product and engineering teams to start visualizing the approach and also made us better able to start creating concrete milestones.
Once we had a high-level overview of requirements, the team started designing and building out each feature that would allow our genetic counselors to conduct all of their visits on the GMA platform and increase their efficiency.
Towards the end of Phase 1, we wanted to ensure that our wider clinical team understood all the changes that would impact them. To this end, we created a high-level prototype to educate them before the release and serve as a springboard for discussion.
Takeaways
Wins
- Once we decided on the component-based approach and created Phase 1 milestones, we were able to align as a product team around a singular vision and reduce confusion and friction
- We aligned with the engineering team on our high-level goals as well as our granular milestones, which made estimation and execution go a lot faster and more smoothly
- We maintained a strong collaboration with clinical leadership, which led to a smooth rollout of Phase 1 since we validated all of our decisions along the way
Challenges
- Starting out, this problem was very challenging to tackle because it touched so many areas of our organization, including product, engineering, clinical, and customer success. It was really difficult to align on a high-level approach that made sense to each team and unite around a process for tackling the whole project. As a result, work at the beginning felt a bit unfocused and slow-moving before we reached a space of full alignment.
- There was a great sense of urgency due to the immediate business necessity to increase our clinical services gross margin. This led to a more pressurized exploration phase as the impetus to build was so strong.
- Different areas of the organization had familiarity and alliance to the two separate platforms, so a solution was always going to cause some friction for the teams who had to make the most transitions.
