Care pathway triage for a regional hospital network
A regional hospital group asked us to reduce delays between primary care referrals and specialist appointments. Teams were juggling PDF playbooks, legacy forms, and ad hoc email threads—fast for edge cases, slow for everyone else.
What we set out to fix
We focused on the moment a referral lands: who should act first, which pathway applies, and what data is missing. The goal was not to replace clinical judgment but to make the default path obvious and auditable.
How we worked
We interviewed coordinators and specialty leads, then modelled their real decision trees—not the idealised versions in policy docs. Engineering paired with operations weekly so every release addressed a measurable wait-time or rework metric.
In eight weeks we delivered:
- A single triage API that encodes pathway rules with version history
- A coordinator console that surfaces exceptions and “needs human” cases first
- Integrations with two EHR-facing queues they already used, so training stayed minimal
Outcomes
- Median time from referral receipt to first structured action dropped by more than a third in the pilot sites
- Rule changes ship as configuration, not emergency code deploys
- Compliance and clinical leads share the same audit trail for every decision
Technical notes
We optimised for explainability: each recommendation lists the rules that fired and the data elements they depended on. When confidence or completeness was low, the system explicitly handed off to a person—with context attached.
The engagement is now expanding to additional specialties, with the same architecture reused for intake and bed-management pilots.
