We are in the throes of debating the structure of the NHS for, I think, the 20th time in my professional life, but structure is, in my experience, of little importance compared with the two other components of an organisation – systems and culture, and what happens in the minds of individual clinicians and patients. If structure were important why does every health service that I know or read about face the same problems of safety, waste, variable quality, equity, and failure to prevent disease?

For my support of systems and culture and dismissal of the significance of structural change I have been accused of being, among other things, a post-structuralist and an anarchist, but the truth is I am a fan of good bureaucracy, whose characteristics, which are their merits, are described by the excellent Charles Perrow as being:

1. Equal treatment for all employees.
2. Reliance on expertise, skills, and experience relevant to the position.
3. No extraorganizational prerogative… that is, the position is seen as belonging to the organization, not the person. The employee cannot use it for personal ends.
4. Specific standards of work and output.
5. Extensive record keeping dealing with the work and output.
6. Establishment and enforcement of rules and regulations that serve the interests of the organization.
7. Recognition that rules and regulations bind managers as well as employees; thus employees can hold management to the terms of the employment contract.

Without bureaucracies the weak are at the mercy of every despot and petty tyrant but if they have good features why do people hate them?

Firstly they do sometimes go over the top, Kafkaesque is the adjective for the bewildering hierarchy. The second reason why bureaucracies have had a bad press is that they don’t do, indeed they can’t do, what many people want them to do, namely solve complex such as problems such as “How can we make care for all people with epilepsy in a big city better, with no new money?”, or “How can people with bipolar disorder be better supported?” Bureaucracies are good at ensuring the open and fair recruitment and promotion of staff and the administration of money without bias or corruption but not for solving complex healthcare problems. Neither is there strong evidence that markets do that better for complex problems that require a complex adaptive system, and ant colonies are the best example. Ant colonies solve problems and learn, and they are altruistic, that is why they flourish. The recent film made them too much like humans, too selfish, if you want to read about altruism consult the classic, massive book by Holldobler and Wilson called simply The Ants.

So two cheers for bureaucracy, it would get three were it not for the fact that we always use it tasks for which it is not suitable, and four cheers for the ants.

Muir Gray is visiting professor of knowledge management, Nuffield Department of Surgery, University of Oxford.