What is ‘evidence-based policy’?

Medicines are carefully tested with laboratory research and clinical trials before they can be licensed and prescribed for patients. By contrast, health policies affect whole populations and have the potential for great benefit or harm, but they are often designed and implemented without transparent and balanced use of evidence or thorough evaluation of their impact.

Evidence-based policy (EBP) making is about the transparent and balanced use of evidence in public policy making.3 The purpose of EBP is to ensure accountability and promote the conditions for securing improvements in health and well-being. So the question is: what is needed to promote EBP? Many of the answers lie in health policy research.

Science is important for effective policy making, but it is insufficient. Policy making is a complex, non-linear, and often non-rational process that uses scientific evidence, but which is also dependent on the political, economic, legal and ethical environments, diffracted through the prism of key decision-makers’ personal and political values. Successful policy making depends on three essential conditions: problem recognition, politics, and policy formulation and evaluation.4 Greater understanding of each of these components is needed to promote EBP, and policy research makes important contributions in each of these three areas.

Tobacco control provides an interesting example of the complex relation between evidence and policy making. Problem recognition happened decades ago when the harmful effects of tobacco were convincingly described in scientific terms. But policies and politics were slow to catch up, and still lag behind. Major policy changes such as the Framework Convention on Tobacco Control and the prohibition of smoking in public places were achieved, not through convincing evidence alone, but through advocacy, legal action and public pressure.5 ,6 The process of effecting policy change is indeed still ongoing, and tobacco-related illness and death remains a significant problem despite overwhelming evidence of harm.

Research for policy is about obtaining evidence, through primary or clinical research, about interventions that work in health. Translating research findings about effective interventions into actionable policy is comparatively poorly developed. Health policy research can strengthen this important step in securing health gain from basic research, for example through knowledge brokering to facilitate effective communication between researchers and policy makers.

Research about policy, or policy analysis, is about understanding the policy process and the institutions, stakeholders, ideas and political processes around policy.7 Indeed, the policy content is often not the most important factor. First, it may be the system in which that policy is implemented which may have the greatest influence on its success. Second, the effectiveness and cohesion of the various policy ‘actors’ in the process: from formal ‘policymakers’ (politicians, public sector officials), to those with pivotal system roles (healthcare workers and managers, patients) to external agencies with informal influence over policy making and implementation (eg, non-governmental organisations, charities, private sector, civil society or lobbyists).8

Research on policy is about evaluating the process of policy development and implementation, and measuring the health outcomes and wider impact of that policy. Evaluative policy research is about learning lessons from policy successes and failures, and is part of a wider comparative effectiveness agenda. Health policy research in this sense is about investigating and explaining health patterns and their link to policy. Examples include studying political, economic and social policies and their influence on health—such as describing the relations between child mortality and income inequality, family policy or the type of political system in which those policies are embedded. More familiar perhaps are examples of policy research on healthcare, such as describing the relation between health spending and mortality. Finally, there are more indirect policy research areas such as on healthcare amenable mortality, which may generate more questions than answers. For instance, where an outcome is thought likely to be influenced by a specific policy, such as access to or organisation of health services, deeper examination of trends or geographical variations in outcomes may suggest where to direct further enquiry.

What are the challenges of health policy research?

The challenges of health policy research are threefold: its complexity, which necessitates a widely multidisciplinary approach; its methodology, requiring a combination of empirical, social and anthropological approaches with quantitative and qualitative methods, and inductive with deductive techniques; and in accounting for political cycles, whereby many policies are not given sufficient time to embed and be evaluated comprehensively.

Health policy research is multidisciplinary and cross-sectoral. It involves the tools and techniques of public health and epidemiology but also health economics, and social and political science. It encompasses health services research, health systems and organisational effectiveness studies, and other emerging disciplines such as translational research; research on the prioritisation, development, and implementation of policies and their impact on population health; and implementation science (a narrower, more technical aspect of policy and systems research related to identifying and tackling implementation gaps for clinically effective interventions, within the context of complex human and organisational factors.)9

The rich multidisciplinarity of health policy research is one of its strengths, but inevitably it means that it is sometimes considered ‘dirty’ science. Attributing causality is challenging, and often impossible. For health researchers trained in biomedicine and epidemiology, the randomised controlled trial is usually the gold standard of evidence. However it is rare for a policy to be examined in a controlled environment or in isolation. More usually, policy happens as part of a complex set of inter-related measures and in a broad and rapidly changing landscape. Although randomised trials are desirable and possible in many situations, it would be simplistic given the nature of health policy always to rely on trials to substantiate policy making. Cluster-randomised trials—where the unit of allocation is a group of people for example, in a hospital, a community or a school—are frequently used to evaluate policy interventions.

Observational multidisciplinary research makes important contributions to understanding and improving health policy: they should not be considered inferior to interventional research, but rather as an important adjunct. Indeed, the recommendations that follow from health policy research findings are more likely to be a construct based on best available evidence, rather than a fully tested hypothesis with a definable margin of error around the result of controlled trials.10

Furthermore, although one should always aim towards EBP, much of policy making by necessity happens in the context of incomplete evidence. As with evidence-based medicine, absence of evidence of effect should not be confused with absence of effect. This should of course be acknowledged in line with the important concept of transparency in the policy process, but can also be useful as a driver for evaluation of the process of policy implementation. While evaluation rightly focuses on outcomes, health policy is about the human condition, and so it is important to remember that processes often matter as much as outcomes: compassionate, humane health policy must be about more than proven outcomes and cost effectiveness alone.

Finally, policy making is inextricably entwined with politics and advocacy. The impact of powerful commercial lobbies cannot be ignored, as the ongoing experience in tobacco, alcohol and food legislation attests. Economic considerations play a significant part in prioritising new initiatives, or discontinuing existing ones, based primarily on cost and short-term benefit: the reduction of the successful and evidence-based Sure Start initiative in early years’ development is one such example. It is no secret that political ideologies affect health policy making, and influences how policy makers select, interpret and use existing evidence to support or disinvest in certain policies based on political dogma. This political influence is not static but dynamic: political cycles rather than evidence can often dictate the pace and direction of policy making, which means that policies are often given inadequate time fully to develop and embed. Instead, often there is a tendency to the reactionary, with strategy, direction and implementation of policy being driven by yesterday's anecdote or headline, rather than best evidence. The all-too-frequent reorganisations of the National Health Service are an example of how policies can change tack, not with objective high quality evidence alone, but with evidence as viewed through a highly politicised and ideological lens.

Box 1 outlines three examples illustrating the impact of some of these factors on the success, or otherwise, of recent UK child health policies.

Enhancing communication between research and policy: systems and strategies

One strategy to address the need for research to assist policy is through the Policy Research Units of the Department of Health in England, of which one has a specific focus on child and adolescent health. These units carry out a mixture of functions from pursuing policy research relevant questions to providing a reactive function, that is, answering civil servant or ministerial questions.

Researchers are often frustrated at the apparent inability, or even intransigence, of policy makers to use the wealth of evidence available to inform the policy making process. While it is often true that policy makers are not as willing to engage with researchers as they might be, the onus should not simply be on them to become better versed in the language and intricacies of interpreting research evidence. Researchers can maximise their effectiveness in translating research into health policy by understanding and embracing the vagaries of the political context in which policy is being made, and by reframing the interaction with policy makers to one of collaboration.

To this end, some key principles can be gleaned among examples of successful strategies:

1. Understand the political context

   It is worth emphasising again how much of policy making, by its nature, is influenced by the political and socioeconomic climate. The impact of research findings on policy for child health, for instance, will be dependent on the priorities of the Government as a whole, and also specifically on how it might impact other policies in the government departments responsible for health, education, social care and the Treasury. The chances of successful translation of research to policy will be increased by aligning policy recommendations from research evidence to cross-sectoral priorities, and articulating the benefits in health, social and financial terms.

2. Prioritise the need for research knowledge translation from the outset

   The translation of research to policy is about more than simply communicating the results of a research study to a policy audience. Too often, it is seen almost as an afterthought, not considered until the research is near completion. To effectively influence and shape policy, researchers should frame and articulate at the outset a clear policy message that would result from their research, and plan a strategy to engage policy makers and stakeholders. Efforts to improve this are already starting to happen, and at a system level. For example, while broader policy research is the responsibility of the Policy Research Programme, the National Institute of Health Research has a very strong focus on translational research, and clearly articulating the potential policy implications of any research is key to obtaining National Institute of Health Research funding.

3. Context is everything

   In clinical trials, it is taken for granted that any results are potentially generalisable only within the context of the study parameters: they may be applied to patient groups which are similar to the study population, and that the outcome may be dependent on the intensity of treatment tested. The context in which policy evaluation occurs is equally vital. The success or otherwise of a piece of policy research cannot and should not support or preclude its widespread adoption in different settings, and this message should be explicitly communicated in the policy making dialogue. However, as exemplified by the sudden unexplained death in infancy example above, lessons in the process of policy implementation (as distinct from content) can be invaluable, and may be generalisable across different policy areas.

4. Prepare data and evidence for the policy audience

   There is much evidence to show that policy makers do not have the same understanding or training in academic research and evidence appraisal, and that much of academic research never reaches their consciousness.19 ,20 Brownson et al outline key principles for the preparation of data for a policy audience, to maximise the likelihood of creating impact to influence change21 (see box 2). Briefing papers, written largely in lay language and outlining the breadth of evidence and their relative strengths and limitations, can be a useful way to communicate with policy audiences.

5. Seeking out opportunities to engage with policy officials

   Many policy makers and officials are keen to form links with academics: reaching out from government may be as hard as reaching in. Indeed, there may be reciprocal benefits: the Department of Health's Policy Research Programme has a responsibility to consider research requests made to it from a number of arms’ length bodies in health such as NHS England and Public Health England, and opening a dialogue with policy officials may help researchers to formulate and fund policy-relevant research projects.