Adaptive Part 3: Patient-Centered Architecture, Specimens and Dashboards

Project summary

This is Part 3 of the Adaptive Order Management story. Part 1 covers the research and strategy, and Part 2 shows how the team reduced cognitive load. This part covers three themes from the research: a patient-first architecture, specimen and shipment workflows, and dashboards.

Putting the patient first

Adaptive sees itself as a patient advocacy organization, so the information architecture had to put the patient first.

• Linking Subjects to Patients. Patients who arrive through a clinical care order are Patient records, with identifiable details. Most clinical trial participants are Subject records, de-identified for regulatory reasons. With no link between them, a patient could be asked for the same blood or bone marrow samples more than once. Working with the SMEs, the team drafted a Human record type that sits above both and links its Patient and Subject records.
• Live patient status. A new aggregate tile shows the status of a patient's recent tests and specimens, so analysts can answer a clinician's question without opening a stack of orders in separate tabs.
• Clear headers. A Header Bar component marks each page as an Order, Subject or Patient.

Specimens and samples

Specimen data needed to move cleanly between the OMS and the LIMS and keep teams updated on shipments. Alongside Flags, from Part 2, the team redesigned the Specimen tables:

• Expandable rows. Carbon's expandable rows hold the secondary information on a specimen and its child samples. The SMEs helped decide which attributes are primary and which are secondary, and read-only fields use a fluid style so the open row reads as a tidy grid.
• Status icons. Icons for processing, completed, moved, discarded and returned replace text, which decluttered a text-heavy table.
• A toolbar for scale. Carbon's data table toolbar brings search, filters and batch selection, plus new batch actions the team found specimens needed. These matter most in clinical trial orders, which can hold hundreds of specimens.

The design started at the whiteboard, with information architecture, workflow definitions and prioritized pain points, then moved through low-fidelity drafts and rapid revisions with the SME panel in Figma.

Dashboards

Carbon's data visualization tools made it practical to give teams real insight. The team introduced insight tiles, built on a revised information architecture, for worklists across clinical diagnostics and batch research orders. Each worklist can be queried for questions like:

• How many orders are awaiting resolution?
• What share of the worklist hasn't changed in 24 hours?
• How many orders are nearing stabilization, or may have gone 3 days without a change?
• How many patient orders are MRD-enabled, and how many fresh specimens (bone marrow, blood) are in?

A dashboard homepage shows every worklist group with live analytics, so staff can pick up where they left off and leadership can see team progress. This matters most in the lab, where samples process overnight and assignments and volumes can shift a lot by morning.

What's next

The team explored smart tiles: saved custom queries against specific views, built from the columns a user cares about, that update order metadata live and can be shared, giving leadership better reporting for decisions and planning.

Skills: Information Architecture, UX Design, Data Visualization, Design Systems, Healthcare UX.