Discover Physician Needs via Generative Design Research
This project sought to understand how physicians keep up with medical literature and help them discover potential treatment options.
Partnering with a large Health Maintenance Organization (HMO), we uncovered three critical physician pain points:
UX Strategy, Research, and (some) Design
My role on the project was to transform this fuzzy business problem in to a product:
"How can artificial intelligence improve physician efficiency and clinical outcomes into a marketable solution?"
I crafted the research plan and was hands-on during all phases of research. I also lent a hand to our design to make rapid design changes between research iterations.
My direct reports on the team included:
- Design Principle
- Senior Researcher
- Junior Researcher
Research Driven Persona: Paul the Physician
As-Is Journey Map
Product Definition
I performed a series of ethnography studies to understand how physicians within this HMO were consuming and applying the growing body of medical literature. Some of the insight uncovered:
- Physicians are handling the same literature multiple times
- Physicians need to validate and trust the source of literature
This fieldwork also found several adjacent pain points related to electronic medical records:
- Disparate data sources can lead to inconsistent reporting
- Systems aren't able to exchange information between practices
Although outside of our initial scope, these adjacent topics became critical to our design decisions, as detailed below.
Design Thinking
Next, I facilitated a three-day Design Thinking workshop with physicians, product managers, developers, and our sales team.
Meet Paul
Before the workshop, I worked with two junior researchers on my team to create a data-driven persona as the target user of this solution. Paul has a complex job, little time to keep up with emerging medical literature, and does not always trust that the electronic health record contains the most current information about his patients.
Big Ideas (Design Thinking Exercise)
We came ups with some exciting ideas to solve Paul's pain points, including:
WAZE
Kind of like you need to crowdsource with other physicians on WAZE. You need a minimum number of sets of care determinants to identify the right treatments path accurately.
Travel Guide
Like a travel guide, the physician is in uncharted territory. The guide helps him navigate, calls out important landmarks (clinical info)not to miss, low-cost hotel options (affordable plans).
MVP Prioritization
We sketched ideas for how the top feature concepts might take shape to determine which features were worth building into an MVP prototype. These sketches helped us in our further prioritization activities.
Project Hills
Finally, we crafted Hills to guide our work coming out of the workshop:
- Paul, the acute physician, can view ranked treatment options based on exhaustive research in one place for this patient.
- Paul, the acute physician, knows his choice is optimal by viewing what inputs substantiate the treatment options.
- Paul, the acute physician, by seeing evidence-based treatment options that he wasn't aware of, can continuously learn and adapt.
physcian decision support tool
To-Be Journey Map for Physician Decision Support Tool
Hill 2-User knows his choice is optimal by viewing what inputs substantiate the treatment options.
Hill 1- View ranked treatment options based on exhaustive research in one place for this patient.
Hill 3-See evidence-based treatment options that physician wasn't aware of.