Hospital-by-hospital death rates for four main types of surgery were published by the health department on Thursday as the first step in a revolution in the publication of the results of NHS treatment.

The move came as Lord Darzi, health minister, said that improving the quality of care had to be the NHS's "organising principle" over the next few years. Publication of treatment outcomes and patient satisfaction were key drivers, he said.

The first figures published on the NHS Choices website on Thursday, however, are likely to leave much of the public bewildered.

They are adjusted for case-mix - how ill patients are from other conditions - and then adjusted again for volume, because one death in a small number of cases can distort the data.

The result means that hospitals whose death rates are all within the expected range nonetheless appear to have wildly differing figures - the adjusted standard mortality ratios range from 0 to 372 for hip and knee replacements, for example.

"Any attempt to publish a league table of these figures would be totally misleading," said Paul Robinson, a senior executive with CHKS, a leading analyst of hospital data. "There must be a better way of presenting them in a form the public can understand."

However, professor Sir Bruce Keogh, NHS medical director, was unrepentant, saying he had been "pleasantly surprised" that all hospitals had turned out to be within the normal range or better than expected. That had not been the case when the same exercise was undertaken for joint replacements in the US, he said.

He had "fully expected" to find wide variation in death rates for abdominal aneurysms. "It is a tribute to the NHS that there is so little variation," he said.

For most conditions, Sir Bruce conceded, death rates were "a pretty blunt instrument" for measuring quality - only two or three patients in 1,000 die after joint replacement surgery, for example. That is why, he said, more sophisticated measures were to be developed and published over the coming months and years.

"But I thought it was important to get these figures out there to provide a very clear message to the service that we are now in the business of measuring quality and there is no going back on that," he said.

Professor Nick Mays, a leading researcher at the London School of Hygiene and Tropical Medicine, said the chief effect of publishing the data was to harness the competitive instincts of doctors and managers.

"They will not want to lose reputation by having poor results," he said.