Richard Smith thinks that the way medical journals make their money, by publishing scientific papers, is immoral. He also says they are little more than a marketing tool of the drug companies. That's harsh talk from anybody -- but even more remarkable from Smith, who was editor of the British Medical Journal (BMJ) for 25 years until his departure last summer.

Smith, like his outspoken former counterpart, Lancet editor Richard Horton, is unafraid of controversy. The two of them used to do a lively double-act now and again, exposing the sharp practices of ambitious or desperate academics and single-minded drug companies bent on publication in prestigious peer-reviewed journals. But now it appears that Smith's thinking has moved on to question the very existence of medical journals in their present form.

This conversion occurred in a fabulous 15th century palazzo in Venice, he says, where he spent two months in early 2003. He was writing a book about publication ethics at the request of Alexander "Sandy" McCall Smith, once chairman of the British Medical Association's ethics committee and now better known as the creator of Precious Ramotswe and The No. 1 Ladies' Detective Agency.

Smith, too, wanted to write a book that would appeal to a general audience, even if it did not hit the bestseller charts. He started compiling a list of incidents where the publication of a paper in a medical journal had raised ethical questions, as with the alleged link of MMR -- the measles, mumps and rubella vaccination -- to autism and the flawed study in 1990 which suggested that women who attended a particular cancer help center were just as likely to die as those who did not (it did not acknowledge that they were more ill).

"It was as I began to think about these things that this started to bother me," Smith says. "I'd always known medical journalism wasn't about the truth and I tried to write that at least once a year. It's partly because of the nature of science -- it's about provisional truths. You observe a whole lot of phenomena and produce a hypothesis and try to destroy it and there's an assumption that you will eventually destroy it."

But another thought, he said, "was that this whole business of sending original research to doctors is kind of crazy. When you talk to ordinary doctors, they are not scientists and yet here we are sending them this mass of complicated information that most of them are not equipped to critically appraise. They haven't got the time."

And then there is the question of the validity of the research being published. The BMJ is one of the four best-respected general medical journals in the world, with the Lancet, the New England Journal of Medicine and the Journal of the American Medical Association. They send out scientific papers for peer review by academics in the same field as the authors. This process gives the paper a sort of validation or stamp of authority in the medical world.

But Smith, and other journal editors, have become very disenchanted with it.

"Peer review is a very flawed practice," he says. "It is slow and expensive, a lottery, and prone to abuse and bias. Much of the time it doesn't pick up errors."

Before the Internet came along, scientific papers had to be published in journals. But now, he says, journals should give up what are effectively immoral earnings. They add no value to the scientific research, and yet it may take a year or more before they publish and they then charge people to read it. "Making money out of restricting access to research is immoral," he says.

Instead, he says, all research should be published in one large free database, with access for all. Smith has joined the board of directors of the free access online Public Library of Science (PLoS).

The biggest problem with this scenario is financial. Journals make more money from reprints of scientific papers than they do from advertising. Pharmaceutical companies strive to get their drug trials in the best-known journals, because the cachet helps sales. They will order hundreds of thousands of reprints for their reps to distribute in hospitals and doctors' surgeries. It is a huge earner, and the journals have become reliant on the money. Some journals would go bust, but Smith does not think they would be mourned.

"All this would mean is that instead of 30,000 journals or whatever, you might have 50 to 100 good ones," Smith says with equanimity.

He does not accept the arguments of medical societies, which publish some of the journals, that they need the income to support other activities.

"If these are genuine value-added things, you should find a way for them to pay for themselves," he says.

With all the research online in one huge database, it would be far easier for doctors to search for papers that were relevant to them, and for academics to collect together all the studies that have been done on one topic to carry out the sort of grand review which can lead to real changes in medical practice.

So is there any need at all for a journal like the BMJ, where Smith has spent most of his career? Was he not dismayed by the conclusions he reached during his palazzo retreat?

"I didn't think, `Oh my God, I've wasted my life.' When I think what medical journals do and what they are useful for, I think it is the same as newspapers. They are good at setting agendas and saying here are things we ought to be thinking about. They can say here is an issue we think is very important and we would like to find a way to change what happens in the real world," he says.

And yes, he does think it is possible for medical journals to change the world, although most of the time they fall short, he says, citing the BMJ's coverage of the dire state of the UK's prison healthcare.

"It is probably a bit better now than 20 years ago, but it is impossible to say that the BMJ made awfully much difference," he says.

Smith has taken the arguably more pragmatic road to changing the world now, vacating the BMJ editor's chair for the post of chief executive of United Healthcare Europe, which aims to help the British health service find more effective ways to deliver healthcare. But he insists that although he loved editing the journal, he does not regret leaving or want to go back. After a quarter of a century, it was time for a change.