Pulse
A point-of-care diagnostic reader for rural clinics — built around a sixty-second workflow, one-handed cartridge loading, and a housing that survives being dropped from a tailgate.
- Year
- 2021
- Category
- Medical device
- Role
- Industrial Designer & Human Factors
- Client
- Confidential — diagnostics company
- Timeline
- 18 months, including summative validation
- Team
- 2 ID, 5 ME, 3 EE, regulatory & HF specialists
- Status
- Cleared and deployed
The brief
The client had validated assay chemistry and a benchtop reader designed for laboratories. The commercial opportunity was elsewhere: mobile clinics and rural health posts, where the user is a community health worker with an afternoon of training, not a laboratory technician with a degree.
Everything about the benchtop unit assumed conditions that do not exist in that context — mains power, a clean level surface, climate control, unhurried attention, and a user who reads error codes. Our brief was to redesign the reader for the actual place it would be used, without touching the validated optical path.
That last constraint shaped the entire programme. The optical bench, the cartridge geometry and the thermal control loop were locked. We had the housing, the interface, the workflow and the power architecture to work with — and a hard regulatory requirement that every change be justified with human factors evidence rather than preference.
What the research changed
Three findings did most of the work. Each one closed off a direction we had been treating as obvious, which is the only reliable sign that research is earning its budget. Methods: contextual inquiry across eighteen sites, competitive teardown of six units, and two rounds of formative testing with printed rigs.
The other hand is always busy
In 30 observed tests, the operator was holding a sample, a phone or a patient record during cartridge insertion in 26 of them.
Sunlight defeats screens
Every existing device we observed in the field was being read in shade created by the operator's own body.
Errors are silent failures
Users routinely mistook a failed run for a negative result. Ambiguity in the result state was the single highest-severity use error we identified.
Four decisions that shaped it
The sixty-second workflow
We mapped the entire task from sample collection to result capture and cut it to five steps. Cartridge insertion is self-aligning through a chamfered lead-in with 4 mm of positional tolerance, so it can be done one-handed, by feel, against the operator's own body if there is no surface available.
Unambiguous result states
The result is communicated three ways simultaneously: a transflective display legible in direct sun, a colour band that does not rely on red/green discrimination, and a distinct audible pattern. An invalid run produces a different pattern from a negative in every channel — the failure mode that concerned us most is now the loudest thing the device does.
Housing and ingress
Glass-filled PC with an overmoulded TPE bumper on all six faces, validated to IEC 60601-1 mechanical requirements and a 1.2 m drop onto concrete on each face and corner. Sealed to IPX4 with a single compression gasket and a labyrinth path at the cartridge slot, so the one opening the device needs is also the one that cannot be sealed shut.
Power and service
A swappable battery gives 40 runs per charge with a solar-capable charge input. Field service was designed for a screwdriver and no training: four captive fasteners, three field-replaceable modules, and colour-coded connectors that only mate one way.
Specification
The long tail of the programme was tolerance stack-ups, part consolidation and assembly sequencing, run jointly with engineering. Headline figures from the released specification below.
Outcome
Numbers are the fastest way to make a case study credible, so include them even when they are modest. If a programme was cancelled, say so and say what you learned.
Credits
Work shown with permission. Some details altered or omitted under NDA.