Australia was one of several countries to stockpile the drugs known as Tamiflu and Relenza, believing they would reduce hospital admissions and complications of influenza, such as pneumonia, in the event of an outbreak of swine or bird flu.

But a review of data from trials involving more than 24,000 people, published on Thursday by the Cochrane Collaboration and the British Medical Journal, concluded the drugs were of only marginal benefit, shortening the duration of influenza symptoms by just half a day.

The review found no good evidence that the drugs prevented the spread of the virus between people, or any of its serious consequences, such as infections. It found Tamiflu increased the risk of psychiatric disturbances, renal problems, nausea, vomiting and headaches.

Chris Del Mar, a professor of public health at Bond University and one of the authors of the review, said governments should rethink their strategies for preparing for pandemics in response to the findings.

"Governments all around the world have invested billions in buying and stockpiling this stuff in case there's an epidemic or pandemic influenza outbreak and that doesn't look like that's going to be a cost-effective thing to do at all," Professor Del Mar said.

But Professor Del Mar said this was not to say Australia's decision to stockpile 1.9 million doses of the drugs was not ''the best decision with the available evidence".

Australia first added antiviral drugs to its national medical stockpile – a strategic reserve of drugs, vaccines and antidotes for use in a public health emergency – in 2004. A spokeswoman for the federal Health Department could not say on Thursday how much Australia had spent on Tamiflu and Relenza, but the current value of the stockpile, which includes those drugs as well as other products, was $192 million.

The US spent $1.3 billion on its stockpile of antivirals, while Britain spent £424 million on Tamiflu alone.

The review drew on more than 20,000 pages of data from clinical trials, which was not available to government agencies when they were deciding to stockpile the drugs.

''What we did that's unusual in this review is we didn't just rely on published data, we realised that this could be biased … that the less good data were being withheld in some way," Professor Del Mar said.

Data from more than half of the patients included in clinical trials was made available by drug companies Roche and GlaxoSmithKline for the first time for the review.

British Medical Journal editor-in-chief Fiona Godlee said the review was the result of many years of struggle to access trial data previously hidden from view.

"It highlights with certainty that future decisions to purchase and use drugs, particularly when on a mass scale, must be based on a complete picture of the evidence, both published and unpublished," Dr Godlee said. "Otherwise we risk another knee-jerk reaction to a potential pandemic.''

A spokeswoman for the federal Health Department said it continuously reviewed available evidence on a range of issues including antivirals and it was revising its pandemic plan.