Although some uneven competition exists in primary and secondary care, there is absolutely none (and therefore no possibility of Darwinian-style evolution) in the 'central NHS': the DoH, the SHAs and the PCTs.
So why not introduce competition here? Darwinian evolution only occurs when autonomous entities compete for finite resources with no external interference. Therefore, make PCTs autonomous; their budgets dependent on the population they serve. Allow any company to set up a PCT, provided it adheres to uniform conditions applied without favouritism across the NHS.
Crucially, make the salaries of all managers proportional to the population their PCT serves, and ensure that every area is within the remit of several overlapping PCTs.
Each year every practice would choose which of the local PCTs to join: undoubtedly this would be influenced by the various PCTs' historical attitudes to primary care. Any PCT that bullied its practices, failed to commission effectively, or squirrelled away primary care money would soon find itself the poorer (and its managers earning much less) as aggrieved practices moved to a competing PCT.
In this way PCTs would be held to account. Lean, effective and humane PCTs would gain practices, serve an increasing population and make more money for their managers. Conversely, managers of PCTs that behaved badly or inefficiently would lose out - personally - as their salaries dropped. The motivational effect would be enormous.
The same principle could apply to SHAs. Make SHAs autonomous, perhaps more numerous, and overlapping, and require them to compete for client PCTs. SHAs that were inefficient or draconian would earn less; those that were run well would succeed.
The Achilles' heel of the NHS has always been the lack of managerial incentives to create lean, efficient systems to support front-line staff. Introducing competition in the central NHS would at long last cause Darwinian evolution to occur in these wasteful and over-managed bodies.
