EXPLORATION MEDICAL CAPABILITY · NASA HUMAN RESEARCH PROGRAM · JOHNSON SPACE CENTER
A hospital that exists only as a model.
Before any Artemis medical system is built, it exists as a SysML model: the Long Duration Foundation — a Level of Care IV medical system for nine-month lunar orbit and surface missions, complete with ConOps, 700+ clinical capabilities, and traced requirements. As UI Architect embedded with the MBSE team, my job was the layer where humans meet that model: the web reports clinicians, engineers, and managers actually use.
The mission
The design reference mission is demanding: nine months in cis-lunar space, including three on the surface, with a crew of four and a physician-level Crew Medical Officer. Evacuation to definitive care takes days, not hours. NASA's IMPACT tool ran 100,000 probabilistic simulations of this mission to estimate medical risk — loss of crew life, need for evacuation, crew task time lost — and to recommend a medical capability set within hard mass and volume constraints.
The system behind the system
The Foundation is built by two teams in lockstep: systems engineers (systems, biomedical, aerospace, human factors) and the Clinical & Science Team — pharmacists, nurses, and physicians across emergency, internal, family, rehab, and aerospace medicine. Their shared artifact is a SysML model in MagicDraw, stored in Teamwork Cloud, with requirements configuration-managed in Cradle as the source of truth.
Because the model is the single source of truth, the web reports were the only view of it most stakeholders ever had.
The mapping effort alone shows the scale: three CST groups re-ordered the requirements into clinical-workflow order, then mapped 700+ clinical capabilities to 80 functional requirements in reverse — surfacing redundant and orphan requirements along the way. All of it published as HTML reports: one internal (with model-linked commenting), one public on the HRP website.
The UI problem
The model's HTML reports were generated straight from the model's structure — which meant users got the engineer's view: deep trees of SysML packages, dense icon sets, and navigation that leaned on the browser's back button. Clinicians reviewing requirements drowned in information overload; managers couldn't find the ConOps.
A clinician opened the report and saw a SysML package tree when she was looking for a patient scenario. Closing that distance was the redesign.
The strategy came from two classics — Jakob's Law (users spend most of their time on other sites; meet their existing habits) and Shneiderman's mantra (overview first, zoom and filter, details on demand). I restructured the reports to read like the documents NASA stakeholders already live in: a simplified landing page routing each audience, breadcrumb navigation replacing the back button, and a standardized, minimal icon set (Material Design, later Apache) to cut visual clutter — all built so the MBSE team could maintain it without a designer on staff.
Traceability, made legible
Every requirement in the Foundation carries a rationale, with traces running from NASA standards down to specific clinical conditions. The UI's job was to make that chain easy to follow:
The outcome
The Long Duration Foundation was vetted through the ExMC Control Board and presented across JSC's medical boards (MOG, SMOCB, HRPCB) — with the web reports as the vehicle stakeholders actually reviewed, comment threads anchored directly to model elements. The reports went into review as the working interface, and the redesign was kept simple enough for the MBSE team to maintain without a designer on staff.
The hardest constraint was also the most ordinary: the MBSE team had to keep the interface running after the designer was gone. Building for that is why it outlived the engagement.