Case study · Product design · Patient Flow Manager

The Missing
Handoff

Role Staff Product Designer Timeline 1 quarter (2025) Scope Doctolib Pro (500K users) · Germany & France · Multi-doctor practices
My role in this initiative

I partnered with the Lead Designer throughout discovery, definition, and validation of the the end-to-end design of Patient Flow Manager as Staff Designer. This meant running research in German practices, benchmarking the competitive landscape, and iterating through two rounds of user testing before a single line of production code was written. We designed a coordination system that had to serve two distinct user groups simultaneously: Medical Assistants (called MFAs - Medizinische Fachangestellte) managing patient movement and doctors prioritising their queues, across practices of vastly different sizes. I also drove the post-launch analysis that reframed the challenge: from "did we build the right product?" to "how do we design the path from activation to adoption?"

The problem

German practices had no coordination tool. They'd built workarounds instead.

In German healthcare practices running multiple doctors, rooms, and simultaneous patients, medical assistants and doctors had no shared view of where patients were, who needed what, or what tasks were pending. The result was staff running between rooms asking "Where's my patient?", an invisible coordination tax applied to every working day, at every practice.

60% of German practices confirmed they relied on their Electronic Health Records'(EHR) waiting room functionality as a critical tool. Yet Doctolib had nothing. Practices were forced to use third-party solutions, fragmenting their workflow and creating friction at every step. The competitive landscape was split in exactly the way that signals an opportunity: Medatixx offered high customisability but suffered from a poor NPS of −4.8. Tomedo had strong satisfaction (NPS 76.5) but limited functionality. Neither had solved the simplicity-power trade-off that multi-doctor practices actually needed.

"The waiting room is a key feature of our current EHR software and the view I interact with most in my daily work."

German Paediatrician and Medical Expert
Screenshots of competing waiting room features across Medatixx, Tomedo, and other EHR providers
Competitive benchmark: powerful but painful vs. simple but insufficient. No player had solved the trade-off.

The initiative

A coordination layer for practices of all sizes, replacing tools they'd relied on for years.

The mission wasn't simply to add a waiting room view. It was to design a real-time coordination system that could work across the full spectrum of German practice complexity, from a two-doctor family practice (even though small-sized practices are rare in Germany) to a 50-doctor specialist centre, and replace third-party tools with years of user habit embedded in them. That combination of scale, variety, and entrenched competition defined every design decision that followed.

Core goals

  • Handle practices of vastly different sizes (2 doctors vs. 50+ doctors) without two separate products
  • Serve two distinct user groups simultaneously: MFAs orchestrating patient movement and doctors prioritising their queues
  • Replace deeply entrenched third-party tools that users had relied on for years, without demanding relearning
  • Integrate seamlessly into existing Doctolib workflows and the broader platform ecosystem without adding complexity

Research foundation

Two rounds of testing to understand what we didn't know yet.

2
Rounds of user testing: on-site Berlin, then remote validation
2
Distinct user groups served: MFAs managing flow, doctors prioritising queues
60%
Of German practices cite the waiting room view as their most-used EHR feature

The critical research insight early on was that patient flow is not one problem: it's many. Small practices with 2–5 doctors needed simple visibility. Large practices with 10–15 doctors required sophisticated task coordination. MFAs orchestrated the full patient journey across rooms. Doctors needed a focused queue view and the fastest possible path to starting a consultation.

Rather than build and hope, we ran two structured testing rounds. Round 1 was on-site with MFAs and doctors in Berlin, revealing what resonated immediately (drag-and-drop, integrated task management) and what caused friction (ambiguous column labels, unclear status icons). Round 2, conducted remotely, validated iterations and surfaced the final refinements before development. Each round informed the next, making it possible to fail fast and refine before any production investment.

In parallel, a comprehensive benchmark of existing solutions (from legacy EHRs to standalone coordination tools) confirmed that no market player had solved the simplicity-power trade-off. That gap was the opportunity.

The solution

Three tools in one view: dashboard, coordination layer, and consultation launchpad.

The Patient Flow Manager became three things simultaneously: a real-time status dashboard showing the full practice at a glance, a coordination tool enabling patient movement and task assignment without leaving the view, and a consultation launchpad letting doctors start appointments directly, bypassing the agenda navigation entirely.

The Patient Flow Manager interface showing patient cards across care step columns with task management integrated
The Patient Flow Manager: status visibility, coordination, and consultation start in a single, configurable view.

Key features shipped

  • Drag-and-drop patient movement: grab any patient card, drop it where they need to go. No forms, no menus, no friction. The interaction model that tested best in Round 1 and held up through Round 2.
  • Patient cards at a glance: name, insurance type (critical in German healthcare billing), waiting time, colleague notes, and visit reason, all visible without a single click.
  • Customisable care steps: each practice defines their own workflow (Reception → Lab → Doctor A → X-Ray → Checked-out). Each user then customises which steps they see, hiding irrelevant ones from their personal view.
  • Integrated task management: tasks appear alongside patients in the same view. No more "Did someone handle that prescription?" Everyone sees what's pending, assigned to whom, and what's done.
  • One-click consultation start: doctors launch consultations directly from the waiting room without navigating through the agenda, eliminating one of the most common sources of daily friction.

Navigating complexity

What seemed simple in wireframes became complex in German practice reality.

User research surfaced complexities we hadn't anticipated. These weren't edge cases, they were structural realities of how German practices actually operate, and each one forced a meaningful architectural decision before we could ship.

Care steps: space, doctor, or stage?
The care step concept seemed straightforward until we tested it with real practices. Unlike the linear "waiting → consultation → checkout" model we'd assumed, German medical structures are fluid: a single doctor might see patients across multiple rooms throughout the day. Users weren't sure whether a care step represented a physical location, a doctor assignment, or a workflow stage. Ownership and assignment added another layer we hadn't designed for. Resolving the mental model ambiguity required reworking the information architecture before we could validate the interaction design on top of it.
Walk-in patients don't follow scheduled logic
Walk-in patients represent roughly 30% of appointments in Germany, too large a share to treat as an edge case. Our initial design assumed all patients entered through the calendar feature before being assigned to doctors, matching scheduled appointment logic. But users expected walk-ins to appear directly in the relevant doctor's queue, not in a generic holding area. That single insight forced a separation of "today's scheduled visits" from "active patient queues" and a complete rethink of walk-in patient behaviour to match real-world expectations rather than system constraints.
One view, two fundamentally different needs
MFAs orchestrate the full patient journey across rooms and need the broadest possible view. Doctors need a focused queue, not the complete coordination picture. Designing a single view that served both roles without cluttering either required careful prioritisation of what each persona saw by default, and a customisation model that let each user hide what wasn't theirs to manage.

These discoveries reinforced a principle that held throughout the project: what seems simple in a wireframe becomes complex in reality. The only way to uncover these nuances was to test with real users in real workflows, which is exactly why the two-round research structure was non-negotiable from the start.

Impact

Strong satisfaction. A harder adoption curve than we anticipated.

78%
CSAT, above the 70% target, with 60–100% consistency over 90 days
195
Organisations adopted of 348 enabled, across Germany and France in less than a month
66%
Adoption rate among enabled French organisations, the fastest-moving market
78,26% CSAT, above the 70% goal.
Users who adopted the Patient Flow Manager loved it. Satisfaction was not only high, it was consistent: holding between 60% and 100% throughout the 90-day monitoring period. The product worked. The problem was getting users to the point where they could discover that.
Premium orgs adopted at nearly twice the rate of orgs without the EHR solution.
38,1% adoption among Premium (C&F) organisations versus 21,17% in Classic. Practices with fewer entrenched third-party tools and more openness to new workflows adopted significantly faster, a direct signal about where to focus activation efforts next.
The adoption-satisfaction paradox.
High CSAT but 9,6% user adoption in non-premium organisations. The gap between organisation adoption (21,17%) and user adoption (9,6%) revealed a critical dynamic: practices had champions, but not team-wide buy-in. For a coordination tool, partial adoption limits value, which reframed the entire next design challenge.
Only 56% of enabled organisations created care steps.
Even with a deliberately simplified setup design, initial configuration remained a meaningful hurdle. Onboarding friction was the single biggest barrier between "enabled" and "active", and the clearest signal that great product design alone isn't enough without a great activation path.

France and Germany moved at different speeds, for different reasons. French organisations adopted at 66,67%, a market less entrenched in third-party tooling. German non-premium organisations moved more slowly, held back by years of embedded habit with existing tools. Both trajectories pointed to the same next question: how do you design the activation experience, not just the product itself?

What's next

Building the right product and driving adoption are two different design challenges.

78,26% CSAT proved we solved the right problem. Users who reached the Patient Flow Manager loved it. But the adoption numbers in Classic organisations revealed the gap: getting users to start required more than an intuitive interface. We designed for usability within the tool, and underestimated initial setup friction and the team adoption dynamics that are essential for any coordination system. A tool that requires whole-team usage can't succeed with individual champions alone.

The foundation is strong: high satisfaction, proven value for adopters, and clear differentiation across markets. Now the design focus shifts from building to activating. Short-term priorities include guided onboarding flows, practice-type templates that eliminate the cold-start configuration barrier, and in-app contextual help that meets users at the moment of friction. Medium-term, team activation workflows that encourage collective adoption, and case studies from satisfied practices that give hesitant ones a credible reason to try.

We set out to replace fragmented third-party tools with an integrated solution. We succeeded in building something users love. But this project is a reminder that the best products don't just work well, they're easy to start working with. We built the first part. Now we design the second.