Cutting diagnostic turnaround for a national healthcare provider

The challenge

Diagnostic reads were queuing unevenly across sites. Urgent cases waited behind routine ones simply because of arrival order, and turnaround varied widely depending on which facility a scan landed in.

What we did

• Built a triage model that ranks incoming studies by likely urgency, so the most time-critical reads surface first
• Deployed entirely inside the provider's own cloud boundary — no patient data left their environment
• Built an evaluation harness with clinician-labelled cases, scored on every model change
• Kept a clinician in the loop on every decision; the system prioritises, it does not diagnose

Outcome

• Urgent studies surfaced substantially earlier in the queue
• Turnaround variance between sites narrowed considerably
• Radiologists reported less time spent triaging and more spent reading

“The honest thing they did was tell us which parts of our original scope were not going to work. That saved us a year.”

— National healthcare provider

Client names are withheld under our confidentiality agreements. Outcomes are described qualitatively rather than with precise figures we are not permitted to publish — we would rather say less than overstate.

← All case studies