A leading health services researcher said this week that he had killed patients as a result of a technique he had been advised to use at medical school.

Iain Chalmers, director of the UK Cochrane Centre, which assesses medical research, told a House of Lords committee that his experience demonstrated the poor quality of medical reviews and textbooks, which were based on opinion rather than on rigorous assessment of the evidence.

"The general public finds it quite astonishing that we have let our house get into such a muddle," he told the Lords select committee on science and technology.

"All of the health professions are in a similar mess. I was fairly horrified to find out about four years after I was qualified that I was actually killing my patients," he said.

ÁñÁ«ÊÓƵ

He laid much of the blame for ignorance of medical research at the door of scientists who write reviews, editorials or chapters for textbooks. "It's quite extraordinary that people who are good scientists, writing protocols for primary research, when they are asked to write a chapter . . . jettison those principles". He said that reviews should include a description of the methods used.

He cited a review in the second edition (1987) of the Oxford Textbook of Medicine, which said: "The clinical benefits of thrombolysis . . . remain to be established." This was written more than a decade after clot-busting drugs had clearly been shown to be useful during a heart attack, he said. "Goodness knows how many hundreds of thousands of people have read lethal information," he said.

ÁñÁ«ÊÓƵ

Centres such as his were attempting to review areas of medical research.