Buse and Walt's paper, on the other hand, draws attention to the efforts of policy communities (of scientists, policymakers and practitioners) concerned about the lack of research on vaccines and medicines for low income countries. This neglect occurred because pharmaceutical companies stood to gain little profit in countries with limited markets. The response has been a proliferation of public-private partnerships at the global level to tackle particular problems. Buse and Walt offer examples of such partnerships, born of perceived political and economic necessity and a re-alignment of actors in the health arena. Research will be needed to establish best practice in relation to agenda setting, accountability, sustainability and equity.

Another problem with the balance of research and policy is scaling up findings from research in particular settings to national or international policy. Infrastructures, human resources, cultures and health priorities all differ between countries and all affect policy choices. What works in one setting will not necessarily work in another which poses difficulties for policy makers. The Mwanza trial results were widely used to support new policies to integrate sexually transmitted disease care with primary health care (traditionally separately provided), which had also been promoted at the 1994 International Conference on Population and Development. But, as Mayhew and Lush show, scaling up to national policy was difficult to implement because of international and national politics.

In attempts to follow a rational approach to policy, policy makers not only translate small-scale research into large scale blueprints but may also insist on rigid adherence to such policies. WHO's policy to promote DOTS for tuberculosis treatment was based on several small scale studies in various settings. Even so, Ogden and Porter suggest that its application in India has had some deleterious effects, partly because the policy has been implemented without adapting it to the realities of local communities. As a result, those most vulnerable to disease and in need of care get excluded from services in case they jeopardise success rates.>

How does this knowledge help bring policy and research together for better health? And what can be done to correct imbalances in the research-policy relationship? Can acknowledging a gap between research and practice be a starting-point for developing strategies to persuade policymakers to share research results and find ways to apply them more widely?

Policymakers need to feel that they 'own' new ideas so should be involved in research at an early stage of inception and design. Findings by Brugha et al show how engaging stakeholders in decision making through technical advisory groups, combined with accreditation and steps to set treatment guidelines can clarify the roles of actors and help promote policy development.

In sum, research and policymaking proceed along different trajectories. This separation is desirable, since policy makers are motivated by political necessity while researchers should preferably be relatively independent. Sometimes their trajectories coincide, in which cases new evidence can lead to rapid policy reform. Efforts to engineer better links between the trajectories are rare but can be highly effective. Researchers and policymakers need to work harder to understand one another's priorities and communicate to compare notes wherever and whenever they can. Greater care should also be taken when research from one setting is used to develop policies to be applied in another.

Contributor(s): Louisiana Lush and Gill Walt