A useful starting point for anyone wanting to elicit change is the identification of likely barriers to change. Knowing what - and where - the major hurdles are that may affect successful implementation is useful during the planning process. These challenges will often vary from policy to policy, and between different contexts. Both research findings on barriers to policy implementation in other settings, and lessons learnt from previous experiences may be informative, but they may not be sufficient.

There is no standard approach to identifying barriers to change. This process is often done informally by taking perceived barriers into account and in an implicit and unsystematic way. We propose a more structured approach to identifying barriers.

The people who will be affected by a policy - the stakeholders - are the ones likely to be best placed to foresee possible barriers to policy implementation. A number of methods can be used to explore the views of stakeholder groups about new policies including, for example, a 'mixed methods approach' for undertaking a so-called 'diagnostic analysis'. This approach may include brainstorming, focus group discussions, interviews and other qualitative methods, or a combination of these. Such activities can provide new insights into stakeholders' perceptions and identify both barriers - and facilitators - to policy implementation. Surveys can also be useful. For example, respondents could be asked to rate the extent to which a list of potential barriers actually represents obstacles to change. Practical examples of such processes are provided in Table 123.

Several frameworks and checklists have been developed to help identify potential barriers to implementing health interventions. These are often based on a combination of behavioural theories, empirical data, and common sense, and may be useful tools in guiding the process of identifying barriers. Some frameworks cover a broad range of potential barriers in various parts of the health system. For example, in one framework, barriers are categorised according to the level at which the constraints operate 4. These levels include: the household and community, delivery of health services, health sector policy and strategic management, public policies cutting across sectors, and environmental and contextual characteristics. Examples of barriers identified at each of these levels are shown in Table 2.

We have adopted a similar approach by focusing on constraints to policy implementation at three levels in the health system:

• Among healthcare recipients and citizens

• Among healthcare professionals

• At the organisational level

Once the likely barriers to policy implementation have been identified, the next step is to identify implementation strategies or interventions that can address these (Table 3 shows examples of possible links between barriers and interventions among healthcare recipients and citizens). The choice of strategies should also be guided by the available evidence of their effectiveness and costs, as well as stakeholders' views, etc. The issue of how to find and assess evidence that may be relevant is addressed in other articles in this series 56789.

The behaviour of healthcare recipients and citizens, particularly in relation to the use of health services (e.g. under-utilisation, non-adherence to recommended lifestyle changes or treatment schedules, etc.), may be a potentially significant obstacle to successful policy implementation. It is necessary to understand why the targeted recipients behave in particular ways as this will influence the choices they make in utilising health services. Well-conducted qualitative studies can provide insights into the behaviour of healthcare recipients 10.

One framework that can be used to identify factors that may influence the behaviour of healthcare recipients and citizens was proposed by a WHO working group on adherence to long-term therapies. They suggested five dimensions to consider 11:

• Socio-economic related factors

• Health system and healthcare-related factors

• Therapy-related factors

• Factors-related to the particular health conditions of patients

• Patient-related factors

As these factors are related more specifically to clinical interventions, they may be particularly useful when considering barriers to the delivery of care arrangements. For example, some of the socio-economic factors that can affect adherence to treatment among patients with tuberculosis include: a lack of effective social support networks and unstable living circumstances, cultural and lay beliefs, ethnicity, gender, age, the high cost of medication and transport, and the role of criminal justice 1112.

The Cochrane Consumer and Communication Review Group has extensively documented the effects of interventions to improve interactions between consumers and healthcare providers and systems, and has developed a taxonomy of interventions that target healthcare recipients and citizens 13. This may be helpful when conceptualising and considering what kinds of interventions to use. The taxonomy includes:

• Provision of education or information

• Support for changing behaviour

• Support for developing skills and competencies

• Personal support

• Facilitation of communication and decision making, and

• System participation

Several studies and reviews have evaluated the effects of interventions that address constraints to effective health service delivery at the level of healthcare recipients and citizens. In one review, the authors found positive effects from community participation in overcoming such constraints 14. In this instance, community participation was obtained using a variety of intervention approaches, including: health education (e.g. meetings, group teachings), encouraging a participative approach (mobilising leaders and stakeholders to understand and buy into the intervention), using an outreach strategy (targeting households and high-risk groups), and the training and supervision of providers (e.g. nurses and/or mothers). These interventions resulted in increased health-related knowledge and community empowerment and improved coverage in immunisation and sanitation practices.

Financial incentives, such as conditional cash transfers, may be worth considering if socio-economic related barriers are seen as playing an important role. This is because evidence, particularly from low- and middle-income countries, indicates that these may have an impact on the use of health services 15 (see Table 4 for details). A further illustrative example of evidence on the impacts of financial incentives is provided in Table 516.

If patient-related factors, such as a lack of information appear to be important barriers to policy implementation, interventions to improve information provision might be worth considering. A systematic review has shown that mass media interventions, for example, "can encourage increased utilisation of health services". But this finding should be approached with caution given that the study was based almost exclusively on studies from high-income countries 17 and therefore may not be applicable to other settings.

The implementation of a policy or programme will often require changes in the behaviour of those healthcare professionals responsible for implementing the policy on the ground. Changes in professional behaviour do not always necessarily happen automatically. Active and directed approaches may therefore be necessary. The identification of barriers to change or factors that influence professional practice may help to inform the design of interventions for policy implementation. Cabana and colleagues conducted a systematic review of research addressing barriers to guideline adherence among physicians 18 and identified seven main categories of barriers. These can be used as a framework for identifying barriers to policy implementation among healthcare professionals:

• Lack of awareness

• Lack of familiarity

• Lack of agreement

• Lack of self-efficacy

• Lack of outcome-expectancy

• Inertia of previous practice

• External barriers

Examples of how identifying barriers can inform implementation are provided in Table 6.

The Effective Practice and Organisation of Care (EPOC) Review Group in the Cochrane Collaboration has developed a taxonomy of provider-targeted interventions which provides an overview of the types of interventions that may be considered for implementation purposes 19. These are:

• Educational materials

• Educational meetings

• Educational outreach visits

• Local opinion leaders

• Local consensus processes

• Peer review

• Audit and feedback

• Reminders and prompts

• Tailored interventions

• Patient-mediated interventions

• Multi-faceted interventions

Several strategies aimed at achieving behavioural change among healthcare professionals have been rigorously assessed 20212223. Typically, these have taken the form of evaluations of guideline implementation strategies targeted directly at healthcare professionals. Most, but not all, have been conducted in high-income settings 24. The findings demonstrate that many interventions can influence professional behaviour effectively to a modest or moderate extent. But passive interventions, such as the circulation of guidelines or the hosting of educational meetings, seem only to have smaller impacts. Educational outreach visits and multi-faceted interventions that specifically target identified barriers to change are among the more promising approaches.

Financial incentives may be used as a means of influencing professional behaviour but these have been evaluated almost entirely in high-income settings. These can be effective in influencing individual healthcare professionals when simple and well-defined behavioural goals are provided, such as increases in the delivery of immunisations - at least in the short term 25. However, several potentially negative consequences of such programmes have been identified and the use of financial incentives may not necessarily be cost-effective.

Regulatory measures are inexpensive and potentially effective means of eliciting changes in professional behaviour but may be poorly received by professional groups 26. The impact of regulations per se as a means of achieving behavioural change among healthcare professionals has not been reviewed systematically, therefore available knowledge about their effectiveness is limited 27.

See Table 7 for further illustrative examples of evidence on the effects of interventions to achieve behavioural change among healthcare professionals 2829.

Many organisational change strategies see the measures that should be taken as steps in a process that leads to change. Defining why there is a need for change and identifying barriers to change are tasks that are typically included in this process.

Pexton has proposed a list of the most common barriers to organisational change and this can also be used as a framework for barrier-identification 30:

• Cultural complacency (resistance or scepticism)

• Lack of communication

• Lack of alignment and accountability

• Passive or absent leadership

• Micro-management

• An overloaded workforce

• Inadequate systems and structures

Ways to address each of these types of barriers are suggested in Table 8.

Examples of the tools and approaches often recommended to organisations assessing preparations for change include 31:

• Analytic models for understanding complexity, interdependence and fragmentation (such as Weisbord's six-box organisational model, the 7S model, and process models)

• Tools for assessing why change is needed, such as SWOT analysis

• Tools for determining who and what can change, such as force field analysis and total quality management

• Tools for making changes, such as organisational development, action research and project management

Most commonly used organisational change strategies are based almost entirely on theory, or else on one-off applications and opinion. Sometimes these are supplemented with case studies or anecdotes, mainly from high-income settings 31. Evidence about the effectiveness of these strategies is hard to come by, making it difficult to predict whether or not a specific method is likely to lead to the desired organisational change.

Although the impacts of such change management strategies are uncertain, they may still be useful as processes allowing for active reflection on how change in an organisation can be facilitated.

When a policy is to be implemented, changes at the general level of a health system may be necessary. These may include changes to governance arrangements, financial arrangements and delivery arrangements 32. For example, when considering the financing of a policy option, should all costs be incurred by the government, or are additional sources of funding needed? Can the current system cope with the additional bureaucratic or logistical workload, or is a new mechanism needed to deliver the service? The body of evidence on how to implement such changes is small: those making decisions will usually have to draw on case studies and experiences in other jurisdictions. For particular policy implementation issues systematic reviews may be useful, such as those related to the costs of scaling up interventions 33 or factors that may affect the sustainability of health programmes 34. In a recent overview, the authors summarised the evidence from systematic reviews on the effects of governance, financial and delivery arrangements, and implementation strategies that have the potential to improve the delivery of cost-effective interventions in primary health care in LMICs 27.

When identifying the need for system changes it may be useful to review the components of a health system and to identify where changes are required. Table 9 shows a framework that can be used as a starting point for such analyses 35.

Shared decision-making in health care. Achieving evidence-based patient choice (2^nd edition, Edited by Edwards A and Elwyn G). Oxford University Press, 2009.

Changing Professional Practice (Edited by: Thorsen T and Mäkelä M) Copenhagen: Danish Institute for Health Services Research and Development, 1999. http://www.dsi.dk/projects/cpp/Monograph/DSI9905.pdf

Grol R, Wensing M, Eccles M. Improving Patient Care: The Implementation of Change in Clinical Practice. Oxford: Elsevier, 2005.

Fretheim A, Schünemann HJ, Oxman AD. Improving the use of research evidence in guideline development: 15. Disseminating and implementing guidelines. Health Research Policy and Systems 2006, 4:27. http://www.health-policy-systems.com/content/4/1/27

NorthStar - how to design and evaluate healthcare quality improvement interventions. The ReBEQI Collaboration 2005: http://www.rebeqi.org/?pageID=34&ItemID=35

Iles V, Sutherland K. Organisational Change. A review for health care managers, professionals and researchers. 2001. London, National Co-ordinating Centre for NHS Service Delivery and Organisation R & D http://www.sdo.nihr.ac.uk/files/adhoc/change-management-review.pdf.

Links to websites

Cochrane Consumers and Communication Review Group Resource Bank: http://www.latrobe.edu.au/chcp/cochrane/resourcebank/index.html - The Cochrane Consumers and Communication Review Group is part of the Cochrane Collaboration, an international, non-profit organisation that aims to help people make well-informed decisions about healthcare. The Consumers and Communication Review Group co-ordinates the production of systematic reviews of interventions which affect consumers' interactions with healthcare professionals, services and researchers. This resource bank is a list of Cochrane systematic reviews relevant to people's health communication and participation needs, and has been produced by manually searching The Cochrane Library.

Cochrane Effective Practice and Organisation of Care (EPOC) Review Group: http://www.epoc.cochrane.org/en/index.html - EPOC is a Collaborative Review Group of the Cochrane Collaboration and produces systematic reviews of educational, behavioural, financial, regulatory and organisational interventions that are designed to improve healthcare professional practice and the organisation of health care services.