A standard is an answer with the question deleted

Design guidance looks like a rule and behaves like a conclusion. Somebody once had a question, reasoned about it under a particular set of conditions, reached an answer, and wrote the answer down. The reasoning and the conditions were then discarded, because a rule is easier to apply than an argument.

That works extremely well for as long as the conditions hold, which inside one country they largely do. It is why a mature body of guidance is such an efficient thing: it lets an ordinary team produce a competent building without re-deriving fifty years of accumulated learning.

It fails in a specific way when the building moves, and the failure is silent by construction. Nothing in the document says depends on a reliable grid, because when it was written that was not a variable, it was the air everybody breathed. So the assumption moves with the rule, invisibly, and is adopted by a team who never saw it.

The method that follows from this is simple to state and takes real discipline to do. For every requirement, separate the objective from the specification. The objective is what the patient or the staff actually need. The specification is one answer to it, correct under conditions somewhere else. Export the objective. Re-derive the specification.

Eight families of guidance, which between them determine most of what a hospital physically is. Read the second column, because that is the part no document states and every document depends on.

The last column is the finding, and it runs against the intuition most projects start with. Adopting a British standard feels like the conservative choice, the safe default, the thing nobody will be criticised for. In four of these eight it is the opposite: the specification was calibrated against infrastructure that does not exist here, and applying it unchanged buys less resilience than it did at home while appearing to buy the same.

Three days of fuel is a sound figure where the grid returns and tankers run on schedule. Neither holds here, so three days is not caution, it is a number carried across from a place where it meant something different. The same is true of fire compartment sizing against brigade response, and of a water strategy written for a continuous mains.

This is the layer that converts every other decision into square metres, and it is the bridge between the bed number and the parcel.

The unit everything is built from is the clear bed space: 3,600mm wide, measured centre to centre between beds, by 3,700mm deep. That dimension is not about comfort. It is about what has to be able to happen around a bed at four in the morning, which is a trolley, a hoist, a resuscitation team of five and a relative, simultaneously. It is also a rare case of a standard that transfers entirely, because the activity and the human bodies performing it are identical everywhere.

Then the maxima, which are the part Meridian has to decide rather than inherit. Four beds is the limit in a multi bed room, and the reason is observability and infection rather than area: beyond four the room cannot be read from one position. Sixty metres is the maximum from any bed to the main circulation route, and that one is really a staffing standard wearing a dimensional costume, because every metre inside it is walked several hundred times a day for the life of the building.

The twenty four bed ward is the one most often mistaken for a maximum. It is not. It is the basis the schedules are drawn around, and the guidance says explicitly that the number should be determined locally. What actually limits a ward is how far a nurse can see and how far they have to walk, which means the right ward size here is an output of the layout rather than an input to it.

The proportion of single rooms is usually argued as a clinical preference, occasionally as a patient dignity question, and almost never as what it also is, which is a floor area decision with a price.

A single room with its support area and en-suite is 23.5 square metres. A bed in a four bed room, with its share of the sanitary facilities, is 18.1. That difference, multiplied by the beds and then by the circulation and engineering allowance the schedules add, is the number in the third column.

Around nine hundred square metres separates a half single room ward from an entirely single room one at this bed count. That is roughly six per cent of the whole hospital, and on a constrained parcel it is the difference between a hot floor that fits on one level and one that does not.

None of which argues for fewer single rooms. It argues for taking the decision knowing the price, in the currency it is actually paid in, which is footprint, and therefore parcel, and therefore the first gate.

Take one requirement and follow it across, because the abstract argument becomes concrete very quickly.

An operating theatre needs a high air change rate. That number is clinical, it is about keeping the air over a wound clean, and it does not change with latitude. The standard exports intact, exactly as it should.

What changes is the air being changed. In a temperate climate the outside air arrives most of the year needing modest conditioning. Here it arrives hot and saturated, and every kilogram of it has to be cooled and dried before it reaches the room.

Roughly four times the load, for the same specification and the same patient benefit. Continuously, because a theatre suite cannot be allowed to fall out of specification overnight. And on the essential supply, which means that number is not merely an energy cost, it is generator capacity, fuel consumption and the reason the infrastructure page reads the way it does.

This is what re-deriving looks like in practice. Nothing about the clinical requirement was questioned. The engineering answer to it changed completely, and a project that simply adopts the ventilation specification without redoing this arithmetic has under-sized its plant, its generators and its fuel store in one move.

There is a comfortable assumption at the start of most projects of this kind, and it is worth stating so it can be rejected properly. The assumption is that adopting a developed country's standards is the safe, conservative, defensible position, and that any deviation from them is a compromise to be justified.

Half of that is right. Where a standard encodes something about the human body, deviation downward is a compromise and it should be treated as one. Bed space, sterility, shielding, oxygen purity, the observability of four patients from one position: these are the same requirement everywhere and the only honest deviation is upward.

The other half is wrong, and expensively. Where a standard encodes an assumption about infrastructure, adopting it unchanged is not conservatism. It is inheriting somebody else's operating conditions along with their answer. Three days of fuel, a fire compartment sized against a fast brigade, a water strategy that assumes a continuous mains, a plant room sized for a temperate air state: every one of those is a concession wearing the costume of a standard.

So the position is asymmetric and it needs saying in one sentence, because it is the sentence that has to survive into a design team. Objectives never fall. Specifications are re-derived. Where re-derivation makes the requirement harder, it is adopted; where it would make it easier, the original stands unless a named person has written down a clinical reason.

All of which leaves a gap. If the specification is re-derived locally, what exactly is the external definition of best in class that the sale condition depends on?

This matters commercially rather than philosophically. The group sells only when a hospital is fully mature and demonstrably best in class, and the study has already argued that a gate a company grades itself against drifts toward whenever the board would like to sell. A re-derived standard is exactly the kind of thing that could drift, because the re-derivation is done by the people who benefit from the answer.

The answer is to separate the two. The local regulator sets the licence, which is a floor and must be met on its own terms. An international accreditation, chosen and named in advance, provides the external grading, because it is administered by people with no interest in this company and it assesses the operating standard rather than the specification.

And the re-derivation itself gets a register: every place the imported guidance was departed from, in which direction, why, and who signed it. That document is worth more at exit than any of the individual decisions in it, because it is the evidence that the departures were reasoned rather than convenient. A buyer's technical adviser will ask exactly this question, and the project that has the register answers it in a minute.

Three, and the first is the one that worries me most about the argument on this page.

Telling a project that an imported standard does not apply unchanged opens a door in both directions. The same reasoning that produces seven days of fuel instead of three can produce a relaxation somewhere less visible, argued in the same words, by somebody under cost pressure. The asymmetric rule in the chapter above is the only protection, and it only works if it is written into the design brief at the start rather than appealed to afterwards.