Before the advent of league tables - and even before the dawn of the Thatcher era - hospitals were required to publish figures for costs per inpatient week. High-spending institutions were exhorted to investigate reasons for their supposed poor performance. The naivety of this approach was exposed by an economist who showed that hospitals running up the highest cost per inpatient week typically boasted the lowest cost per case treated. Nobody paid much heed to his conclusion that the NHS had, perversely, been pursuing a quite misguided objective. He went on to become Ronald Reagan's economic adviser.

When Reagan's transatlantic soulmate introduced her purchaser-provider split at the urging of yet another American academic, it was anticipated that this new regimen would necessarily trigger more meaningful cost figures, which would become the currency of the new free-market mechanism. Extra accountants were recruited to cost clinical episodes, and contracts were placed on the strength of the emergent figures.

Under New Labour's variant of the purchaser-provider model, cost data now appears destined to take on the older mantle - insights into efficiency rather than prices. But who actually takes the mass of data recently released on the reference-cost CD seriously? Does anyone really believe that the costs of straightforward surgical procedures vary by thousands of per cent? Costing depends crucially on how overheads are apportioned, and one suspects that this probably explains much of the entirely incredible variability reported. Activity-based costing is but a distant NHS dream.

As meaningful indicators of performance, these new reference cost figures are almost certainly doomed to be as valueless as their cost per inpatient week predecessors. Perhaps this whole labour-intensive effort is a complete waste of time. Why not be really radical? Simply determine the average cost for all clinical procedures, and base a trust's allocation next year on what it actually delivers this year. In other words, a complete shopping list of quantities multiplied by average prices. Such a blunt instrument would generate more real efficiency gains than would the minute dissection of the material on this and subsequent CDs.

Doubtless, some aspiring young presidential adviser is already analysing our latest outpouring of data in anticipation of being called on to assist the second President Clinton devise what she has so famously failed to deliver: a half-decent healthcare system.