The challenge
The group's chief operating officer needed to address a patient flow problem that was affecting both clinical outcomes and operating efficiency across all six facilities. Patient flow was managed by three teams — bed managers, theatre coordinators, and ED charge nurses — each using a different system with no integration between them. The consequence was the classical flow problem: ED boarding by mid-morning, theatre cases delayed by bed unavailability, and discharge decisions made too late in the day to clear the next admission cycle. Every clinician could describe the problem; no leadership team had a tool that showed the network as a single operating system.
Discovery
A four-week diagnostic study working alongside the three key roles showed that each team had part of the operating picture and was spending most of their time in phone calls reconstructing the parts they could not see. The finding shaped the programme design: the first deliverable needed to be a shared real-time view — not a model — so that calls stopped being about establishing facts and could become about making decisions.
Our approach
Network flow view design and build: designed and built a network-wide patient-flow view joining EHR, theatre scheduling, ED triage, and discharge planning into one operations surface
Discharge prediction programme: developed a discharge-prediction model trained on the network's own historical stays, with the predicted discharge time surfaced to the ward team during the morning round
Theatre risk management: designed a theatre-cancellation-risk model flagging cases at risk of delay from ICU bed availability the night before — enabling pre-emptive intervention
Operating cadence redesign: established a 7am network operations huddle using the new view, replacing per-site stand-ups with a single network-wide conversation about decisions rather than facts
The Outcome
The engagement delivered a clinical operations intelligence layer above the EHR and scheduling stack — a single network view, two operations-grade predictive models, and a redesigned daily operating cadence built around the data the layer produces. Clinicians retained their existing tools; the leadership team gained the ability to manage patient flow as one system across all six facilities.
The results
“The view gave us back hours of phone calls per day and let us discharge people earlier in the morning. The clinical impact came from that, not from the model on its own.”
