It takes the NHS a scarcely believable two and a half years to buy a computer system. This is nearly twice as long as the rest of the public sector, and is a prime cause of the parlous state of NHS computing.

Worse, just at the point when the NHS, suppliers and advisers have learned to deal with procurement bureaucracy, and have brought information technology purchasing back to manageable timescales, yet another NHS Executive initiative threatens to plunge health service IT into renewed chaos.

The problems with long procurements are clear. First, they are expensive. Participation can easily cost trusts and short-listed suppliers six-figure sums.

Second, the protracted timescales increase the risk of failure. The process drains the enthusiasm necessary to make complex infrastructure projects work. Rapid technology advances during the procuring period make much of the assessment work obsolete before the contract is even signed.

Changes in the buyer's own circumstances - inevitable over a long period - can make the project less relevant. The Audit Commission summed up the situation well when it reported: Compliance with formal procurement processes can become a goal in its own right, to the detriment of other goals such as having systems that add value.

The NHS is desperate for the new systems and technologies that will help deliver clinical and efficiency benefits. Procurement must not be a barrier to their implementation.

Information for Health, the government's NHS IT strategy document, promised a review of procurement and this is now under way. Draft recommendations include national standards, supplier accreditation, centralised business cases and top-sliced central funding of some projects. All worthy ideas, but fraught with danger for the embryonic IT revolution. Most risky of all is an attempt to impose a central procurement structure on 23 trusts in the South West, a project which history suggests will fail in a spectacular manner .

The NHS's sheer scale makes centralised initiatives of this type slow and difficult to manage - having to take the individual needs of each organisation into account can result in a design nearer to a camel. What's more, this procurement review is responding to urgent needs identified by the Audit Commission and hammered home in a series of ministerial briefings as early as 1994.

The Executive is now trying to solve a problem that may no longer exist.

It is the current procurement methodology used for NHS IT purchases, POISE (procurement of information solutions effectively), which is responsible for the protracted timescales that the review is now tackling. POISE was introduced as a response to an edict by a former health secretary to stem the flood of high-profile IT disasters.

It achieved this aim only by virtually stopping IT purchasing.

The facts are that between 1993 and 1998, only 40 hospital trusts managed to negotiate the procurement minefield and purchase hospital-wide clinical and administrative systems. The impact is devastating. Only 10 per cent of trusts can tell you which patient is on which ward at any given time, and IT support for clinicians is minimal. The dearth in purchases had the added effect of producing chaos in the supplier market. Many companies sold out to overseas interests, went out of business or slashed research and development investment.

But at last, the importance of IT has been recognised and the NHS is starting to invest in clinically based systems. POISE has been tamed by a number of NHS purchasers. Many procurements are now planned to complete within 12 months of being announced in the Official Journal of the European Community. Local initiatives involving a manageable number of partners, such as the North West consortium, are simplifying the buying cycle and reducing costs by creating standard documentation and sharing information. The lack of national interference is keeping things streamlined and effective. And of course, the process is assisted greatly by the new rules allowing regions to approve contracts up to£20m without Quarry House or Treasury approval.

The key lessons learned and which are being put to such effective use by more enlightened NHS purchasers are:

Don't reinvent the wheel: trusts that have already bought successfully can provide standard documentation for OJEC advertisements, output based specifications, supplier demonstration scripts, contracts etc.

Keep timescales to a minimum: this maintains enthusiasm for the project, reduces the impact of changed circumstances and keeps cost down.

Concentrate on what works: the key criteria for system selection are whether you can work in partnership with the supplier, whether the system meets your requirements, whether it is (and stays) affordable and whether it meets business objectives. In short, will the system work in your organisation?

In the last year or so, creative and devolved NHS management has tackled the procurement problem in a way that could never have been achieved by central bureaucracy .

The failure of projects such as HISS and Read have shown us all the dangers associated with central control. The NHS Executive can help by providing specialist support, streamlining mandated processes and making it easier for organisations to share information.

What we must all fear is that an overenthusiastic central push to resolve a problem, which is more or less beaten at a local level, will plunge IT procurements back into the chaos of the mid-1990s.

Will those responsible for this latest review avoid this classic error or will NHS chief executive Sir Alan Langlands find himself defending his South West project in yet another uncomfortable public accounts committee grilling?