Sir

Your Special Report “The 1918 flu virus is resurrected” (Nature 437, 794–795; 2005), on the debate surrounding the recreation of the 1918 (Spanish) human influenza virus, focuses solely on the question of publication, giving little attention to the possibility of regulating potentially harmful experiments before they are started.

A key element of co-operative security is trust and confidence. Only a high level of transparency will enable nations to judge others' intentions correctly, act to discourage unfounded suspicion and build confidence in compliance with arms-control treaties.

Withholding critical information on dual-use research would undermine arms-control efforts. Once a research project has been conducted, it should be published in full. The preferred option for research where potential risks are considered to outweigh potential benefits is to stop the research before it even starts. Both the UK Royal Society and the US National Academy of Sciences have proposed a systematic review of scientific research proposals before work starts, but this approach is entirely neglected in the debate about the Spanish flu virus. The consequences of this project could, and should, have been assessed ten years ago or at any time since.

Another shortcoming of the current debate is the lack of a systematic approach to risk–benefit analysis. Any intervention in the scientific process should be based on a scientifically sound assessment of both risks and benefits. Criteria to assess the potential misuse of dual-use research include its military usefulness and the level of technical expertise needed to apply this knowledge for malign purposes, taking into account a variety of potential actors. The availability of risk-management tools, such as effective arms-control measures, is also an important factor in risk assessment.

Although most commentators agree that the misuse potential of the Spanish-flu work is comparatively high, it seems that the objective assessment of benefits is the more contentious part of the equation.

Your Special Report claims that the Spanish-flu work increases understanding of virulence and pathogenicity factors, and might contribute to identifying the next pandemic strain or developing appropriate drugs. These general statements hold true for most biomedical research projects. But generalities contribute little to a systematic risk–benefit analysis, which should include an assessment of the importance of the research in a health and humanitarian context, whether alternative research avenues are available and the added value of any particular experimental approach.

Influenza pandemics are an important public-health problem, but it is questionable whether a reconstructed viral strain from 1918 is necessary to address this problem. Hundreds of other influenza strains from the past five decades, varying highly in terms of contagiousness and pathogenicity, provide an abundant research resource for comparative studies. The added value of one additional strain — and thus the concrete benefit of the reconstructed Spanish flu strain — is very limited.