This is getting silly: even sillier than usual.

As you probably know, we reported that the NHS White Paper's rescheduled publication date of Monday 12 July is likely to be missed, and that its publication will be accompanied by five consultation documents.

In other words, where this White Paper is not blank, it's Green.

New reports emerge: in today's HSJ, Sally Gainsbury suggests that thinking is now afoot to have 50 'health authorities' each managing 10 local GP commissioning consortia and assuming accounting officer status.

No more top-down redisorganisations of the NHS, eh?

Making harlots out of GP commissioners
This is just reinventing health authorities, but worse: they will be in charge of the money, but will not be making the decisions.

Those will remain in the hands of GP commissioners - who thus remind us of Rudyard Kipling's famous line (borrowed by his cousin Stanley Baldwin) - "power without accountability - the prerogative of the harlot throughout the ages".

Making up the numbers
One of the worse trends in policy is making up the numbers.

Now clearly, I would not be suggesting here that anybody anywhere ever in the NHS falsifies or games figures (for example, leaving patients in parked ambulances to avoid 4-hour A&E breaches, rather than sorting out their badly-organised A&E.

Nor using the provided holiday dates of people on waiting lists to write and offer a date while they're away and then send them to the back of the queue.

Because those sorts of things never happen.

No - making up the numbers is about picking a new number of organisational units. It is a pseudo-science as exact as numerology, beautifully defined as "any of many systems, traditions or beliefs in a mystical or esoteric relationship between numbers and physical objects or living things".

And there are a clear set of rules for NHS numerology.

1. Thou shalt not revert to a previously-used number - Edwards' Law, named for its progenitor NHS Confederation maestro Nigel Edwards. Thus for the intermediate tier, 10 and 30 are unavailable for use. For primary care, 450, 300 and 152 are all out of play. For significant performance regulators, the holy trinity of three (NICE, Monitor and CQC) is thus unavailable.

2. Thou shalt choose a new number which is both nice and round. Halving (from 300 to c. 150, as in PCTs) is acceptable, for reasons of political rhetoric - "this government has halved the management bureaucracy that ... erm .. .we created". So now we have 500-600 consortia, and a freshly minted 50 'health authorities'. Praise The Lord (and pass the ammunition).

3. Thou shalt not have an evidence base for thy new number. PCTs came dangerously close to flaunting the Third Commandment when their numbers were halved, because co-termninosity with local government was achieved - potentially a dangerously useful and logical overlap. However, the chaos effect from a major structural reorganisation generated sufficient offsetting disruption to compensate - remember that the catchment area of Mid-Staffs saw a reorganisation of four PCTs into one.

Spurious justification for made-up numbers
More to the point, sometimes there are spurious justifications for made-up numbers. The recent example is of the 500-600 GP commisssioning consortia - presumably derived from a need to create populations of c. 80,000-100,000 for reasons of risk-mitigation against a catastrophically expensive patient.

This has a certain apparently logical appeal, but is wrong-think.

A smarter man than I pointed out why it's wrong-think. Specifically, it's an attempt to answer the wrong question. It assumes that the right question is 'what size must consortia be to mitigate risks of catastrophic-cost patients?'.

Yet the NHS has no need to think like this. It would be straightforward to have a NHS-wide risk-pooling / reinsurance / whatever-the hell system that accepts a nationally-pooled risk fund for such patients. Using structures to solve that problem is ridiculous. 80,000-100,000 is the answer to a question - but not to the right question.

Asking the right question
My very intelligent interlocutor suggested (and I've run this past lots of other clever people, who all think he's right) that the correct question as regards GP commissioning consortia is this: what is the correct size at which we can get GPs in consortia to behave corporately and efficiently?

From the answer to that, all else should follow.

What are the odds it won't?