Efforts to limit publication of controversial bird flu research could end up doing more harm than good.
Last week, a 12-year-old boy in Indonesia died after becoming infected with the H5N1 bird flu virus. His death brought the global death toll to 347 since the disease was first reported among humans in 2005.
At first sight the figure does not look too alarming compared to the many millions that die from other infectious diseases. And although the virus is usually fatal — up to 80 per cent of those infected die from it — the overall incidence of human infection remains relatively low.
This is because most people only get infected through contact with infected poultry. But what if the virus could spread between humans?
This spectre has now been raised by two teams of scientists, working at a medical centre in the Netherlands and the University of Wisconsin in the United States, respectively. Each team genetically altered the virus into a form that can pass between ferrets through the air — implying that a similar strain could evolve (or be created) that could spread between humans.
The consequences could be so disastrous that last year, a US body set up in 2005 to look at the potential biosecurity risks of laboratory-created organisms recommended that papers on the work submitted to the world's two leading scientific journals, Science and Nature, should not be published in full.
The argument of the National Science Advisory Board for Biosecurity (NSABB) was that the information could be used by terrorist groups or individuals to produce a powerful biological weapon that could spark a deadly epidemic if released into the human population.
Risks of restriction
There is a strong logic to this argument. Withholding the technical details of the steps required to produce a deadly virus would certainly make it considerably more difficult for anyone to copy the process.
And some have advocated going even further to curtail access to such information with calls for a ban on all research that could lead to new, potentially lethal viruses. Their argument is that the threat such viruses would pose if they escaped from the laboratory is so great that nothing justifies the risk of even carrying out research for them.
But both arguments have flaws. Those seeking publication of the information in heavily edited form risk denying scientists access to data that could play an essential role in preparing defences against the virus, such as developing vaccines.
A complete ban on the research could have similar repercussions. Scientific understanding of the bird flu virus, how it spreads and how it mutates, is essential to minimise the chances of another flu pandemic. The flu virus that swept across the world in 1918 killed up to 20 per cent of those infected, causing an estimated 50 million deaths.
An alternative strategy
The scientific community has had intense discussions over what to do with the papers over the past few months.
Initially the NSABB suggested a solution could lie in publishing redacted (edited) versions of the papers with some of the key scientific and technical data omitted.
Both journals to which the research was submitted have been exploring how they might do this while making full versions of the papers available to scientists who have been vetted to ensure they would use the data responsibly.
But on close examination, this option has presented difficulties. For example, much of the technical data in one of the papers has already been presented at a scientific conference so attempts to prevent its further dissemination may be ineffective.
Another significant objection raised at a meeting organised by the WHO in Geneva two weeks ago (16 February) was the difficulty of reaching an international consensus on the criteria for vetting scientists who request the full data.
Concerns in perspective
Following the WHO meeting Nature said in an editorial that the benefits of open publication outweighed the risks identified so far, and that in principle "the papers should ultimately be published in full". 
Similarly the editor of Science said in a recent interview with the BBC that "our default position is that we have to publish in complete form". 
A final decision is awaiting further discussion at the WHO. But this position is brave, and correct.
There are substantial public health benefits in as many scientists as possible having access to the data if they are to understand potential changes to the virus.
These outweigh the advantages — likely to be short-lived — of restricting access to prevent the data from falling into the wrong hands. And at present, the biosecurity concerns are "too general and hypothetical", says Nature.
Furthermore, the political sensitivities of deciding who the 'wrong hands' are, and who should make this decision, risk heightening the international tensions that already exist over attempts to limit the proliferation of weapons of mass destruction, most recently over Iran's nuclear programme.
But two things are essential if the data are to be made more widely available. First, open publication must be accompanied by an effective monitoring system. This would look out for potential misuse of the data.
Second, the issue needs as wide a public debate as possible, actively promoted by both health officials and journalists to ensure it is adequately informed.
Developing countries such as Indonesia, which has the highest bird flu transmission rates and the most fatalities, have a particular interest in the outcome of this debate. They have more to gain from new techniques to prevent virus transmission, such as effective vaccines, than from restrictions on the publication of these data.
Concerns about biosecurity must be kept in perspective. They are certainly important, but should not cloud our judgement on the urgency of developing adequate protection against an evolved form of the virus, whether natural or man-made.