If you want people to take your advice...

...befriend a civil servant. Virginia Berridge discusses how scientists can turn research into public policy.

In the recent panics caused by BSE, genetically modified food, food poisoning and foot-and-mouth disease, scientists have been blamed for not speaking out loudly enough about the risks involved in adopting policies that may have contributed to the crises. But is it really that simple?

The question is an appropriate one for the start of National Science Week (March 16-25), which is an opportunity for scientists not only to highlight the positive side of science, but also to raise awareness about how science and government come together -or don't.

The relationship between research and policy has been studied from a historical angle by two research programmes -the "Science speaks to policy" initiative and the now-completed Aids Social History project. Their findings can help us to understand the process by which research makes its way into the policy arena.

There appear to be four broad tendencies for defining the relationship between science and policy. First, evidence-based arguments, based around the concept of clinical effectiveness and the randomised controlled trial, assume that research can affect policy, directly or indirectly, and that any problems are just down to technicalities.

The second, dubbed the journalist school, also sees the relationship as positive, but when it does not work seeks to blame one or other party.

Next is what is known as the sociology of scientific knowledge. It stresses the role of "actor networks", people, organisations, other disciplines or objects and emphasises scientific process over links with government policy.

Last is a "co-production" approach, which emphasises how the same science can evoke different policy responses depending on different national and policy contexts. It sees government agencies, among others, as also involved in negotiating the meaning and boundaries of science. It talks about policy networks linking scientific communities and government in different ways.

But what do our historical case studies show? What can be learned from examples such as research into the relationship between smoking and lung cancer conducted by Sir Richard Doll and Sir Austin Bradford Hill in the 1940s and 1950s and its subsequent impact; from the continuing issue of diet and heart disease and research; and from Aids policy-making in the 1980s?

It is clear that some research has had a very direct impact, while other studies changed the climate in which policy was formed rather than affecting it directly. The most important factors are the policy alliances linking scientists and government, and change over time can be crucial in that process.

The Doll/Hill smoking research had relatively little impact in the 1950s because of many factors. By the 1970s, however, the smoking research was emblematic of a "new public health" constituency, which made the policy case strongly. Alliances had been built up.

A more direct example of this process came in the 1980s when needle exchange became a central policy response to the threat of Aids among intravenous drug users. For some time, reducing harm to users had been the unspoken objective of a new, broad-based health policy community. The Aids crisis brought funding for the evaluation of needle exchange. The results, managed by civil servants in the Department of Health, allowed deep political objections to be overcome. In this case, a series of events led to a process of mutual accommodation between scientific evidence and government policy.

Civil servants and various forms of expert committee have been important in many cases. Reports on smoking by the Royal College of Physicians gave independent authority to scientific facts and made them widely available through the media.

For diet and heart disease, scientific consensus was forged in the 1970s and 1980s through the work of the Committee on Medical Aspects of Food and Nutrition Policy (Coma) but also through "outsider" expert committees such as the National Advisory Committee on Nutrition Education, whose 1983 report turned diet and heart disease into national headlines.

Different circumstances operate in different policy arenas. Historically, areas such as diet and nutrition or smoking have involved far more interested parties than illicit drugs. This means that the illicit drugs field has had a more cohesive policy community, which has made the relationship between research and policy easier to establish.

The media has carved out a big role in the research/policy process. Since the second world war, there has been a growth in specialist health and science reporters, paralleled by the development of press offices in government departments and health organisations such as the British Medical Association. Increasingly, it is the media definition of science that matters.

What role have researchers played in these historical shifts? Has speaking out mattered? In the 1950s, the role of scientist as activist was anathema to a researcher such as Bradford Hill. He thought closeness to policy would taint science. Changes since the 1970s have drawn researchers more closely into the research/policy process -Jbut they appear rarely to have been in control of the use to which their results have been put.

Whether speaking out brings policy change depends on timing, too. In the 1980s, Aids researchers and activists successfully used the media to urge the government to tackle the disease. But the circumstances were novel: there were no defined government structures to address the situation. Later, a policy community linking to the chief medical officer was developed -creating a more traditional way of dealing with the issue.

Are there any conclusions to be drawn from history? At the practical level, researchers could be advised to build alliances with civil servants, to develop networks within government. Maybe we also need research brokers who promote research to policy-makers, or research practitioners who research with the aim of turning their data into practice. Speaking out about research can also have an effect on policy. But researchers minded to do so should choose their time with care.

Virginia Berridge is professor of history at London School of Hygiene and Tropical Medicine and head of the Science Speaks to Policy historical programme. She is the author of Aids in the UK: The Making of Policy 1981-1994 , Oxford University Press.