The idea
Equipment fails. Chillers trip, pumps wear out, boilers need maintenance, the utility grid goes down in a storm. A hospital cannot choose when those things happen, and it cannot close while they are fixed. So the systems are designed to keep working with a piece missing.
What N+1 means
The common term is N+1. N is the amount of equipment needed to carry the full load. The plus one is a spare. If the cooling load needs two chillers, an N+1 plant has three, so any one of them can be out of service and the building still gets what it needs.
Some systems go further, with full duplicate capacity or separate paths. The right level depends on what the system serves and what happens if it stops.
Where it shows up
Cooling
At the Manchester VA we designed a fully redundant chilled water system for the main medical center, so one chiller can fail without the building losing cooling.
Power
Hospitals carry emergency generators that pick up critical loads when utility power drops. At Albany, the boiler plant work added a new backup generator so the plant that supplies the campus steam keeps running through an outage.
Heat and steam
Central plants are usually built with more boiler capacity than the peak load, so a boiler can be taken down for service in the middle of winter without the campus going cold.
The cost of it
Redundancy costs more up front. More equipment, more space, more piping and controls. The return is that maintenance can happen on schedule instead of in a panic, and a failure becomes a work order instead of an emergency.
The point
Equipment will fail eventually. The hospital should not. Designing for that from the start is what lets nobody inside notice when it happens.