John Appleby's short article ('Measuring efficiency', data briefing, 11 March) was interesting, but it concentrated too much on the apparent huge anomalies in cost to address some of the real issues. A few more points to consider...

Very large variations in costs, as shown in the tables, often result from errors in costing or anomalies in coding. One can spend many happy hours teasing these out, but it's unlikely to help in reducing the overall cost of providing healthcare.

The way activity is counted across the NHS is inconsistent (especially so once reference costing moves beyond surgical activity). It is in providers' interests to count as much activity as possible. Hence commissioners don't only need to compare unit costs, they need to compare the number of finished consultant episodes they buy for their population. Furthermore, if trusts have to start counting things in the same way, this may mean doing things in the same way; not for the first time, a new NHS initiative has disguised implications for prized clinical freedoms.

John Appleby is right that, in any cost comparison, providers will seek to justify their costs or discredit other people's. There are enough real anomalies in the costs to give them cause. One has to question whether commissioners have the 'analytical skill and determination' to help trusts understand or change their costs.

Reference costs mark a major change, and one which moves in the opposite way to industry. Generally, large industries have tried to create profit centres, leaving units free to control costs in any way they choose so long as they deliver profits. The NHS internal market, for all its faults, was trying to do the same.

To create efficiency by costing and comparing in detail is not easy; and reference costs are being done on the cheap, with little or no investment in staff or systems. The caution in the article is well founded.

Steven Bliss

Assistant director finance

Scunthorpe and Goole Hospitals