The process ran on external software and whatever each doctor had put together for themselves. A patient’s history, their tests and their results lived in different places, and nothing connected the patient directly to the doctor ordering the test.

Interpretation posed a greater challenge than organization.

Tests are measuring biological markers whose meaning differs according to whether their value is above or below range and according to which markers are affected alongside them. The laboratory’s internal specifications are detailed and meant for experts only; certainly far too detailed to discuss with a patient looking at you.

There were four limitations guiding the process. The interface design was based on INDE and therefore restricted the patterns that could be applied. There was the limitation that nothing could be vague: misunderstanding a marker could cause serious repercussions, something a design error usually wouldn’t have. There had to be growth from the client’s established brand identity, not a replacement. Finally, there were two users for one set of information: the doctor interpreted the results while the patient got a report.

A number of about a hundred markers is used in common for all tests conducted in the laboratory. This particular marker is repeated in different tests.

Instead of making explanations about each marker in each individual test, we created one library where there is every marker with information about its function and what it means if it has high value and what if it has low value. Physicians get to this library from any test or from any marker.

Interpretation gets difficult when more than one marker is altered at once.

When a doctor selects two, the platform surfaces a written explanation of how they interact: the part that previously sat buried in the guidelines, or in the doctor’s own memory.

Creating a patient and ordering an exam are separate steps, on purpose.

The patient completes their medical history through a form sent from the platform. The doctor then finds a record already populated, opens it as a refresher, edits it if needed, and moves to the exam panel with the patient’s data still in view. Nothing asks them to remember a detail from the previous screen.

More than a dozen report layouts, combining values, reference ranges, charts and written explanations.

These are read by someone with no clinical training, so the hierarchy and the writing had to make the results understandable without flattening what the numbers actually say. Every marker carries its range, its position in that range, and what a deviation means, in that order.

The component library built here is the most complete we have delivered, and it is now the basis for every test the laboratory offers.

Detailed enough for developers to implement directly: every state defined, every variant drawn, built on the client’s existing brand rather than beside it.