Designing the Future of Primary Care with Deep Technology
98point6
Head of UX Design & Research
Employee #52 | Series A - Series C | Reported to CPO
Mission: To make primary care more accessible and affordable.
Impact:
2017: one patient per day, zero corporate clients
2019: > 1 million patients and 150 corporate clientsPatient experience received 4.8 stars from 24,000+ iOS & Android reviews
Scaled UX D&R from myself to 2 Directors leading full Member and Clinician UX teams (12 total)
The Hammond design system and clinician console from 2019 are now powering a SaaS business model
Sections
Design Challenge (45 secs)
Creating our North Star (1 min)
MVP: Patient Experience (2.5 mins)
v2: Hammond Design System (2 mins)
Clinician Console (45 secs)
Operating Tenets (3 mins)
Key Learnings (1.5 mins)
Design Challenge
In 2017, 82% of consumers weren’t using telehealth options and of those who were, 42% didn’t find video visits to be a credible primary care alternative. 98point6 was an audacious idea—a play to make quality text-based care convenient for patients and an attempt to solve the primary care crisis through the use of deep technology.
The original product vision centered around the automation of the patient visit, from check-in to exit with a treatment plan in hand. It included reducing the time clinicians spent on documentation while empowering them to focus on the most complex inbound patient cases, leaving the more obvious cases to be diagnosed and treated by the machine.
I was brought on as the Head of UX Design & Research, and charged with creating an experience that people trusted to be as good or better than an in-person visit with their doctor. My immediate focus was to get a v1 patient experience to market while scaling up my team.
Brand Design
It’s difficult for a service to deliver on a brand promise without a North Star narrative in play.
I partnered early with our CMO and engaged with my former colleagues at studio analogous to run an executive workshop for developing a story unique to 98point6.
The result was Continuity Care.
Continuity Care
Insight - There is much that happens during and between encounters, and there is much that is learned. When the health care team is present, listening, and available throughout a patient’s healthcare journey to keep people healthy—from sickness to wellness and everything in-between—better outcomes are possible.
Belief - 98point6 believes technology can provide consistent, smarter, empathetic, and life-long care on the patient’s terms.
Support - That’s why we are designing a technology that puts care in a patient’s pocket. It listens to and learns from every encounter. It provides diagnosis, treatment plans, and quality care experiences based on its growing understanding of the patient. And through ongoing education, it puts knowledge about one’s cown care in the patient’s hands.
A few months later, we reconvened to establish the brand mark, app logo, and tone & visual language—all used by my team to establish a consistent brand experience within the product and pursue evolving alignment with comm design.
v1 Patient Experience
Text based care was central to our strategy to increase patient population management through the use of AI.
Our collective hypothesis was that as we increased automation in a text-based visit—through natural language processing and AI delivering appropriate questions and responses to the patient—the clinician would be able to increase population management.
I created a vision framework to establish the tone and cadence of the bot, while also signaling our long-term aspiration to reach sentience. The artifact helped me gain alignment with marketing, and eventually a plethora of stakeholders, from the iOS engineers to product managers to our board of advisors.
The script covered a range of bot-patient interactions with conditional logic to either move the visit forward or disqualify the user prior to becoming a patient.
Machine learning automation began once the patient provided their chief complaint by serving up the most relevant medical questions asked by doctors after receiving similar CCs from patients.
As I fleshed out the ‘intro bot’ prototype on an open source SMS platform, my designer and researcher tag-teamed formative research.
Results: People felt comfortable onboarding and answering symptomatic questions from the bot.
We deployed the v1 patient experience on iOS, Android, and Web in four months.
All as I began scaling the team, bringing on 2 FTE UXDs—one with a spike in UId, the other in system design—a UXR Phd in Social Psychology, and a contract designer.
Continuity Care
All of 98point6 was inspired to transcend the episodic relationships that people experience with primary care providers.
As we considered our product approach to reach Continuity Care, we had begun an exploration to modernize the interface, moving the bot away from the standard SMS scroll.
We reconsidered the interface through the lens of extensibility, visual simplicity, and accessibility—supporting the needs of the deaf, blind, and hard of hearing & sight.
Hammond Design System was born.
With a new design system in play, Auto Follow-ups had become the initiative to begin manifesting a path towards our North Star.
By extending the bot into smart push notifications, we were delivering on our brand promise to the market—proving to our patients that 98point6 proactively engages to improve health outcomes.
Clinician Console
In year two, we prioritized deep investments in the clinician console.
From interfacing with an EMR to manage patient records to automating SOAP notes to presenting an AI driven differential diagnosis to adding a seamless prescription service to the experience to scaling a dedicated Clinic UX team, the focus was on realizing clinician efficiencies while enabling quality care.
Within six months I had hired a Director to lead the Clinic UX team, along with three more Sr. UXDs and a Sr. UXR Phd with experience testing design experiments and interviewing role-based users.
Operating Tenets
As a Product & Design organization, we valued making the right things for our customers, and then validating that we were making those things right for our users—from patients to buyers to clinicians.
I co-designed our operating tenets with the Member & Clinic UX teams once I had a grasp on the complexity of our ecosystem, the agency we held as a function, the investment the business was willing to make in research, and the role each person could play in delivering value.
-
Frame It to Make It
We believe that the work to investigate, frame, and pursue the outcomes that serve the people who use our product is the kindling that sparks innovation. Empathizing with users is critical for feeding ideation, and framing such insights points us in the right direction. We do not get paid to do the wrong things right.
-
Everything is a Prototype
We approach our day-to-day to improve how we deliver value. Impactful experiences don’t come easy; they begin messy and ambiguous. We put forth ideas, have conversations, make artifacts, observe dynamics, and rapidly iterate to advance forward. We’re precious when we refine, not when we mine.
-
Data Are Our Friends
We synthesize data as numbers by investigating the context of data as stories (and vice-versa). We socialize our hypotheses, transforming information into knowledge, and lean on each other to gain collective wisdom. Data shapes our approach, and we align deeply to make the proper bets to move forward.
-
All the Things, Intentionally
We engage with intent, from meetings to sketching to interacting with peers. We understand that our approach to work models best practices for those with less experience. This takes concentrated effort, and we know that nothing worthwhile is ever easy, especially within challenging and ambiguous ecosystems.
-
Friction as Necessary
We are gracious partners, share learnings, and believe value is born through healthy discourse. We dig in deep within difficult conversations, always providing grace. We exemplify courage by asking pointed questions, which may slow progress. We’re commit to explore value across a multitude of human-centered dimensions.
-
Multilingual Modus Operandi
We actively listen as we investigate the intersection of people, technology, and business. We speak multiple languages to understand diverse organizational drivers while negotiating trade-offs, ensuring short-term value as we steer towards long-term effectiveness. We inform the business, as we too are the business.
-
Range of Scenery
We leave our comfort zone, and ask open ended questions. We observe behavior so that we can design our service to deliver value for real people. We look for cues that spark our next “aha!” moment. As novel mental models become apparent, we lean in, prototype, and test our hypotheses.
-
Scaling Expectations
We operationalize for effectiveness as the company grows. Our commitment to deliver human-centered value remains unwavering AND we recognize the importance of aligning with business strategy that unlock scale. As a design function, we support each and every stage of the journey.
Key Learnings
Clearly frame AI outcomes within the context of a desired experience
The product experience itself was intrinsically tied to the decisions we made with deep technology. Without clearly articulated and aligned upon long-term product vision, we would’ve made short-term decisions to serve the loudest voice in the room. Deep technology is so costly that a complete business model pivot may occur if framing isn’t pointed, curated, and pursued diligently.
Design needs more than intermittent collaboration with Data Science
My teams actively kept DS within our sights as we worked, but the organization model (CTO > DS = PM < Product) worked against seamless collaboration much of the time. When AI becomes a technological pursuit, in the loop and external users may operate within scenarios that aren’t fully supported. DS and UX collaborated at key inflection points, but there was room for deeper alignment.
People are finicky, and these moments can prove to be valuable
Each doctor had a divergent approach for how they communicated with patients, which created friction in our automation efforts. Early UX involvement may have helped focus on personalized automation OR the design of a longitudinal model to support the needs of clinicians in the loop while training the machine to reach discrete inflection points of automation. In MedTech, quality care is everything and the data are small—factoring in longer-term goals when it comes to sensitive aspects of autonomous agents is pragmatic and ethical.
Design craft is progress, not perfection
As the team focused on quality making, I coached them up on their quality thinking and collaboration—to craft artifacts that furthered the conversations that they needed to have (e.g. a journey map to propose an analogous area of improvement with stakeholders vs. a quick no-code prototype to get in front of engineers). The complexity of the healthcare ecosystem, particularly virtual care, and the fast pace of a technology startup made me a firm believer in spending just enough time in an artifact to move the conversation forward.