There is no quiet window in a hospital, so the work is designed as a sequence of small reversible steps rather than as an event. Nothing here is allowed to stop, which means every change has to be safe to undo at the moment it is least convenient.
What we do here
The lines that matter most in this sector, and why.
Estate assessment
Including the workarounds, which are usually undocumented and always load-bearing.
Modernization
In reversible steps, because there is no downtime to hold in reserve for a bad one.
Migration and cutover
Rehearsed three times, with a parallel run whose end date is fixed in the plan on day one.
The workaround is the system
Where a taped-up sheet has grown up beside the software, it is telling you something.
The workaround is the system
Where a physical board or a taped-up sheet has grown up beside the software, that is the requirement. Replacing the software without earning the board back changes nothing — people keep the second copy because the first one failed them, and they are right to until it does not.
What we plan around
| Constraint | How the plan absorbs it |
|---|---|
| No available downtime | Every step is reversible on its own rather than as part of a batch |
| The parallel run's end date | Fixed in the plan at the start, when it is still cheap to mean it |
| Shift patterns | Who is awake, and who is authorized at three in the morning, agreed in writing |
| The paper fallback | Must keep working throughout, and is tested as part of the rehearsal |
Questions we are always asked
Four of them, answered the way we answer them on the phone.
There is no maintenance window. How does anything change?
In small reversible steps, each safe to undo on its own. When there is no window, the plan stops being a sequence of events and becomes a sequence of positions you can hold.
What about the workaround on the corridor wall?
We treat it as a requirement rather than as a bad habit. A physical board that has grown up beside the software is a record of what the software does not do, and replacing the software without reading it recreates the same gap.
How much clinical time does this take?
Less than an interview-led approach, deliberately. We measure first and talk second, so when we do take clinical time we arrive with a traffic graph rather than a blank page.
What if a step has to be reversed halfway through?
Then it is reversed, and that is a normal outcome rather than a failure. Steps that cannot be reversed are not scheduled until the ones around them are proven.
Case studies
Programs in this area, described by what they were rather than by who paid for us.
