National and international pressure is growing for public policies and programmes to be prospectively evaluated to determine what their impacts will be on public health. This is evident from statements from bodies including the WHO,1 the European Union,2 and the British Medical Association3and may be seen as a natural development of evidence-based principles that are currently being applied in other health and non health-related fields. But where can policymakers find evidence of health impacts, and how can it be integrated with other influences on decision making? One proposed solution, which is attracting increasing interest, is Health Impact Assessment (HIA).4 However, we suggest that the current formulations of HIA, which combine both health assessment and option-appraisal, may have difficulty fulfilling both roles.

To understand why this might be true, it is worth considering the history and typical methods of HIA. HIA is the offspring of Environmental Impact Assessment and has inherited its major features. It usually begins with a screening process to exclude interventions with no significant health effects. Where health impacts might occur, the process of scoping identifies the kinds of information that need to be gathered. This allows terms of reference to be agreed, and forms the basis for assembling and appraising information on health impacts. Emphasis is given to non-quantitative information, of which community consultation is a prominent component. The resulting information about potential health impacts is then presented to decision makers to help them decide how best to mitigate harm and augment benefit to health. A range of approaches to make HIAs of policies, programmes and projects has been developed in the UK,5-7 British Columbia,8 Australia,9 New Zealand,10 the Netherlands11 and Sweden12 13 among others, and in the UK HIA has been promoted as a tool for reducing health inequalities.

One of the strengths of HIA is that it gathers a comprehensive collection of factors that influence health, based on a socioenvironmental model of health.14 Importance is placed on stakeholders' views and feelings as much as quantitative health outcomes. By engaging organisations and individuals in a debate about health impacts, HIA can function successfully as an agent of health promotion and health education. However, two problems can arise from community consultation. The first is that by opening up the field of influences on health, a variety of health concerns may be identified that the policy or project has no ability to resolve. This may be worse than ineffective and by causing public anxiety or raising unrealistic expectations for change, HIA may be harmful and contribute to “consultation fatigue”. Such activity is as unethical as screening for a disease when no effective treatment is available. A second problem is that the consultative process becomes part of the intervention and thus may act as an effect modifier or confounder when evaluating the relation between the intervention and its effect. The effects of HIAs therefore need to be evaluated.

The current appetite among policymakers and planners for evidence of effective interventions is unlikely to be satisfied by HIAs, which typically provide results of discussions with affected communities, qualitative data on projected impacts and often do not evaluate the actual outcomes. Such evidence may be drawn from many sources, including systematic literature reviews, and primary qualitative and quantitative studies. Systematic literature reviews, for example, can play an important part in synthesising the most methodologically strong evidence. Qualitative research can help to explain underlying mechanisms and to determine why and how interventions exert their effects. Qualitative studies should be rigorous and transparent in their design, sample selection, data collection and interpretation.15 And both qualitative and quantitative methods can be combined to give greater insights into health effects.16

As this evidence is assembled, it will need to be disseminated in ways that make sense to those who need to use it. The Cochrane Collaboration is an international organisation that prepares and maintains systematic reviews of the effects of interventions in health care, and receives support from public funding agencies in several countries. The Campbell Collaboration,17 which held its first planning meeting in July 1999, will play a similar part to the Cochrane Collaboration, by maintaining, and promoting access to systematic reviews of studies on the effects of social and educational policies and practices. But public health specialists must take responsibility to see that best evidence of health impacts is actually used in decision making.

How well might HIAs fulfil their role as part of this decision making process? Their effectiveness in changing plans and policies so that health is optimised has yet to be established.18 It will be a challenge for HIAs to be seen as representatives of communities' health and not as conspirators with those who fund them, political agents employed to justify decisions with health rhetoric.

We propose that HIA may benefit by clarifying its preferred function. The scientific evaluation of health effects that result from policies, plans and programmes needs to be uncoupled from the promotion of health in communities and non-health sector agencies, and from the political process of decision making. All these activities have important contributions to make to improving public health but we believe that policymakers, planners and—most importantly—the communities they serve will be better served by such refinements in the current approach.