Batch 02 · ICU turnover
Cleaning and turnover time unmeasured between patients
-
System failure identification
A bed vacated in the morning is not occupied until afternoon. The gap is cleaning, inspection, and portering, none of which appears in any capacity figure as a duration. Placement staff are routed to beds that are not ready to receive anyone.
-
Data pipeline diagnostics
Beds are modelled with two states, occupied and free. The intermediate state has no representation, so time spent in it is silently attributed to availability the unit does not actually have. A state that is not modelled cannot be measured, staffed, or improved.
-
Predictive computational model
Introduce a turnover state with its own timestamps and measure its duration by unit and shift. Available capacity then subtracts beds that are free but not yet usable. Duration by unit shows where portering and cleaning capacity is actually short.
-
Workflow integration
Feed the corrected availability into the existing bed board rather than publishing turnover as a new metric. Bed managers need the number they already use to stop overstating capacity. Bed managers need their existing number corrected, not a second one to reconcile.