The virtual workplace that helped explain a queue of surgical instruments

At Denmark’s Den Fælles Sterilcentral, incoming instruments waited eight to ten hours in a buffer lift before hand-washing, against a two-hour target. Chief executive Thomas Bøgeskov initially thought the lift was the main bottleneck.[2]

He told diginomica’s George Lawton in June that the twin showed how changes to routines and staffing could reduce the wait. In that analysis, packing emerged as the next constraint. He did not give a measured new clearance time.[2]

A familiar room can make unfamiliar data useful

SAS built a virtual version of key parts of the Herlev facility, where instruments move through washing, drying, packaging and delivery. Its October account says the operation supports around 1,000 surgeries a day across seven hospitals. On 6 October, SAS announced recognition for the project in Fast Company’s technology awards.[1]

The twin combines SAS Viya’s analytical tools with a three-dimensional environment made using Epic Games’ Unreal Engine. Operational data feeds the model, allowing teams to investigate disruptions and try changes in simulation. Machine learning sits alongside other analytical and simulation methods in the system SAS describes.[1]

Bøgeskov said ordinary production staff could recognise their stations in the twin, extending process discussions beyond a small group of specialists.[2]

That is a practical role for the picture. A worker knows the demands of a particular station; a planner sees how jobs move across the building. A shared view gives them somewhere to connect those accounts. The conversation can move from a disappointing total to the sequence of actions that produced it.

Rehearsing a change before the next shift

A queue identifies where work is waiting. By itself, it cannot explain whether the delay comes from arrivals, the next available worker, a later packing step or several of those things together. Changing the daily plan in a model lets a team examine an explanation before changing the real routine.

SAS’s May conference description also identifies a separate use for the virtual environment: generating synthetic examples to train computer-vision models for situations that are difficult to capture. That is a second purpose for the twin, alongside investigating production flow.[3]

The distinction helps keep the project intelligible. A model used to train a camera system and a simulation used to discuss a staff plan can share a virtual workplace while answering different questions. Here, the revealing result was a new explanation for a visible queue.

The October announcement describes benefits to instrument flow, without isolating a quantified effect from AI alone. The strongest lesson in the operator’s account is more immediate: making the whole workplace visible helped people reconsider where the problem began.[1]

Report an error or suggest a correction