To see the research they fund in action, donors should work with good local governance, not informal networks, writes Justin O. Parkhurst.
Researchers who work in developing countries are increasingly being asked by funders to demonstrate that their findings are being taken up in policy and practice.
It is certainly reasonable that donor agencies want to see the impact of research they fund they too face demands from their backers, including taxpayers and philanthropists.
One way researchers can draw attention to their results is by building relationships with highly influential individuals, often referred to as 'policy entrepreneurs'. Policy analysis research has repeatedly shown that well-placed, well-networked, or highly respected people can play an important, informal role in getting research issues on the agendas of decision makers, or moving the policymaking process forward when it stalls. [1, 2]
So it may seem only natural that funders call on researchers to build such links to expedite the uptake of their findings.
But there is a key problem with this strategy it is not exactly a model of good governance.
Together with colleagues, I have described the important role that policy entrepreneurs played in getting results from drug trials of an HIV drug (cotrimoxazole) into policy in three African countries.  In each country, key individuals mobilised support or identified opportunities to get this life-saving drug onto the policy agenda.
These were physicians currently or previously involved in research about this drug. They were all in senior policy positions in well-funded programmes when the policy was being developed, but not necessarily in the same type of organisation one was in government, one at a nongovernmental organisation (NGO), and one in a donor agency.
They used their positions to engage decision-makers at the ministries of health, and often mobilised a group of supporters, such as researchers, policymakers, and/or NGO officials, who helped push for a policy change.
Mobilising influential 'entrepreneurs' was either essential to get past apparent obstacles to policy change, or enabled the proponents of cotrimoxazole to support or steer an ongoing policy process. In all three countries the processes were informal and flexible.
But just because our research showed that this is how policymaking does happen, it doesn't necessarily mean it is how policymaking should happen.
Calling for researchers to promote the uptake of their evidence by networking with a country's elite influencers risks undermining other governance structures that are more locally representative.
Scientists often idealise the notion of getting research into policy. Our work on cotrimoxazole, for example, showed how a cheap, effective and life-saving drug was rolled out to more people in need a goal with which all of us would agree.
Yet all research findings have political implications. For instance, introducing a new malaria drug might also affect how much the disease is prioritised in the local health budget; how support systems (such as drug purchasing or health worker training) are used; or how health issues are prioritised relative to other social problems.
Good governance means that policy decisions weigh up various social concerns and are made in a transparent way, by representative leaders who can be held accountable to local people.
Yes, we want to ensure that those leaders use the best possible evidence when making their decisions, but it is not for the researchers to ensure that their findings are the ones on which policy decisions are based.
If donors apply pressure to do this by awarding research funding based on policy impact it will incentivise researchers to prioritise their own ideas over those of the people in the countries where they work.
So what is the alternative?
Donors concerned with getting research findings into policy, but equally concerned with not dominating local priorities, should work to support or develop local institutions that review, process and recommend evidence to policymakers.
Rather than askingacademics to ensure that their findings change policy, donors should work to build local capacity to evaluate evidence, and ask their researchers to feed findings into this established evidence review structure.
It may seem a subtle difference in wording, but it is a critical one either we incentivise researchers to have 'outsider' influence (serving the interests of donors or the international community), or we incentivise them to work through local structures.
Meanwhile, there are attempts to provide evidence to policymakers in forms they understand, to link researchers and policymakers, and even to train policymakers in how to understand scientific evidence. The WHO's Evidence-Informed Policy Network, for instance, is working on all these fronts.
But much more limited are efforts to develop formal advisory bodies, expert committeesor technical working groups that have knowledge review and advisory roles in low-income countries. This is a logical and critical next step.
The sceptic might argue that limited capacity, political infighting or ingrained corruption could make this unlikely to succeed. Indeed, policy entrepreneurs are often influential because of the weak capacity of formal institutions.
But this is no excuse to dictate local policies from outside. Slow progress in capacity building is better than abandoning good governance ideals. Building research-synthesis structures can help to ensure that policy decisions are as evidence-based as possible, while still representing the interests of local people.
It is crucial that vanity about one's own research does not undermine efforts to establish systems of good governance and representation in low-income countries.
Justin O. Parkhurst is a lecturer in health policy at the London School of Hygiene and Tropical Medicine, United Kingdom.
 Kingdon, J. W. Agendas, alternatives, and public policies. (Harper Collins, 1984)
 Mintrom, M., and P. Norman Policy Entrepreneurship and Policy Change. The Policy Studies Journal 37: 649-667 (2009)
 Hutchinson, E. et al. Translating evidence into policy in low-income countries: lessons from co-trimoxazole preventive therapy. Bulletin of the World Health Organization 89, 312–316 (2011)