Country-level governance of global health initiatives: an evaluation of immunization coordination mechanisms in five countries of Asia

Results The evaluation has identified that there have been significant changes recently in the strategic environment for immunization, including developments in new vaccines, increasing GAVI investment, trends towards health system integration and decentralization, and institutional development of the nongovernment sector. This evaluation found that ICCs are functioning well in relation to information sharing and GAVI application processes. However, they are performing less well in the areas of evaluation, strategic gap analysis and coordination of immunization technical co-operation.


Introduction
Background on the Global Alliance for Vaccines and Immunization (GAVI) GAVI is an alliance of developing country governments, donor governments, the vaccine industry, multilateral and bilateral agencies, the Bill and Melinda Gates Foundation, the World Health Organization (WHO) and UNICEF, civil society and private individuals. It was established in 2000 for the purpose of increasing access to vaccine products and immunization services for the poorest countries in the world. The Alliance Strategy for 2007-10 identifies four strategic goals of: (1) strengthening the capacity of the health system to deliver immunization; (2) accelerating uptake of new and underused vaccines; (3) increasing the predictability and sustainability of long-term financing for national immunization programmes; and (4) adding value to GAVI as a public-private global health partnership through improved advocacy and innovation (GAVI 2006a, p.4).
The GAVI Alliance Board, represented by the previously mentioned constituencies, governs policy development and implementation, and monitors and oversees all programme areas. A secretariat, based in Geneva, coordinates Alliance activities including policy development and support to countries (GAVI 2006b). Recommendations to the Board are provided by Independent Review Committees (IRC) for vaccines and health system strengthening, and ongoing funding is recommended through an IRC for monitoring. There is no formalized organizational structure (i.e. GAVI employees) based regionally or in countries. However, Regional Working Groups operate through the WHO Regional Offices, and are the intermediate level of global GAVI governance, and generally take up the role of coordination, proposal review and coordination of technical assistance for their region.
At the country level, this function is taken up by Inter-Agency Coordination Committees (ICCs). ICCs were established in the mid-1990s by WHO and national governments in order to support polio eradication efforts. Subsequently, the ICCs were adopted by GAVI and development partners as the body whereby GAVI support is 'planned for, implemented, and monitored' (GAVI 2006c).
Specific targets of the GAVI programme include a 66% reduction in child mortality rates in 36 designated countries by 2015, and introduction of hepatitis B vaccine in 72 countries by 2010. As of 2008, GAVI has committed to provide US$3.7 billion to the health programmes of 75 developing countries between 2000 and 2015. New and underutilized vaccine candidates for prevention of rotavirus, pneumococcal and Hemophylus Influenzae Type B disease, and Japanese Encephalitis are likely to result in the expansion of GAVI programmes in developing countries (GAVI 2006c). New funding windows for health systems and civil society strengthening through GAVI have been initiated for Phase 2 (2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015). One review has established that the success of immunization service support funding in countries with baseline DTP3 coverage of 65% or less provides evidence that a public-private partnership can work to reverse a negative trend in global health (Lu et al. 2006b).
Nine of the 75 countries of GAVI are based in the South East Asia Region of the World Health Organization (SEARO). In terms of results, eight out of nine countries in SEARO have introduced underutilized vaccines (hepatitis B at a total value of US$164 047 892), and all have implemented injection safety practices (at a total value of US$51 057 200). Supported by GAVI investments in these two programme areas, all GAVI eligible countries in SEARO have either increased or maintained DTP3 from the baseline figures of the year 2000 (see Figure 1).

Governance, GAVI and ICCs
Health sector governance is concerned with the actions and means by which society organizes itself for the health of the population (Siddiqi et al. 2009). Recently international,

KEY MESSAGES
Recent significant changes in the strategic environment for immunization in Asia include new vaccines and technologies and trends towards health system integration and decentralization.
Immunization Coordination Committees, the principal coordination mechanism for GAVI immunization investments, are functioning well in relation to information sharing and proposal application processes, but less well in areas of evaluation, strategic gap analysis and coordination of technical co-operation.
In an era of more pluralistic and complex health systems, it will be increasingly important for immunization planners to link to a wider coordination system that is inclusive of the health sector, regulatory authorities, civil society and private sector interests.
Governance of global health initiatives at country level needs to extend beyond the concept of 'management by committee' to more systematic and inclusive engagement with health systems, private sector and civil society interests. multilateral and bilateral agencies such as WHO (WHO 2007), the World Bank and the Global Fund for AIDS, Tuberculosis and Malaria (Global Fund) have developed a renewed focus on health governance and have identified complementary functions or principles of governance for guiding health system investments. WHO identifies the common functions of all health systems as policy guidance, intelligence and oversight, partnership and coalition building, regulation, system design and accountability (WHO 2007, p. 23).
GAVI principles are stated to operate according to guiding principles of ownership, alignment, harmonization and accountability, based on the Paris Declaration of Aid Effectiveness (OECD 2009) and the International Health Partnership. 1 Key measures of aid effectiveness described by the Paris Declaration include: (1) alignment of aid flows to national priorities, (2) use of country financial management mechanisms, (3) use of national procurement mechanisms, (4) avoidance of parallel implementation structures, (5) aid predictability, and (6) use of common arrangements or procedures. GAVI indicates that 'increased pressure for programmatic and financial accountability is also driving a need for increased vigilance and transparency in the decision-making, oversight and monitoring mechanisms of GAVI' (GAVI 2005a).
There are a range of eligibility procedures for GAVI Grants. Countries must have a per capita gross national income of less than US$1000 and must have a multi-year plan for its immunization programme. More recent health system strengthening guidelines stipulate the requirement for a health sector plan or health sector analysis to be eligible for the Health System Strengthening window of funding (GAVI 2007). Finally, and of most relevance to this study, is that it is a requirement for GAVI funding that countries have a functioning country-level ICC.
A range of global evaluations in recent years has pointed to key issues affecting country performance as a result of limitations in global health initiative (GHI) governance. A review of immunization system support programmes globally indicated that 'where allocation processes seemed less thoughtful and transparent, the country ICCs seemed less coherent and functional' (Chee et al. 2004). A review of lessons learned from GAVI Phase 1 (GAVI 2005b) concluded that a weak ICC was said to have contributed to breakdown of policy dialogue and lack of partner investment. In countries with sector-wide approaches, there was a tendency to consider ICCs as 'parallel organizations'. A review of GHIs in 20 countries has concluded that they do not provide adequate technical support to implementers, leading to lack of absorptive capacity, and furthermore, that countries are overburdened with parallel and duplicative processes of multiple GHIs (McKinsey & Co and Bill & Melinda Gates Foundation 2005).
Given the expanding scope of work of the GAVI programme, and the scale of investment to date, it was considered timely by immunization planners at WHO SEARO to undertake a regional evaluation of the effectiveness of the ICC mechanism both as a governance mechanism for GAVI and as a mechanism for national immunization planners to support the coordination of international health technical assistance.

Aim and objectives of the ICC evaluation
The evaluation aim was to determine the utility of and future perspectives on the role of ICCs in strengthening national authorities' capacity to improve immunization coverage and access to new vaccines and technologies in the SEARO Region. The specific objectives of the study were: (1) to determine stakeholder perceptions of the current utility of ICCs in achieving the stated mission and goals of GAVI, and (2) to identify functional, structural and development strategies for ICCs in order to improve their capability to effectively monitor the impact and set directions for GAVI programmes and resources.

Methods
The evaluation approach was based on principles and methods of 'Policy-maker Surveys' as outlined by DeRoeck (2004). These key approaches include: (1) face-to-face interviews that enable more in-depth responses with senior policy makers than do questionnaires or telephone surveys, and (2) use of a semistructured interview format that is appropriate for interviewing high-level informants as well as providing an atmosphere more conducive to expression of in-depth ideas.
Sources of data included 65 interviews (using a broad question guideline) with ICC members in five countries: India, Bangladesh, Nepal, Sri Lanka and Indonesia. The question guideline was structured around the main evaluation objectives and included the following areas: Strategic directions in immunization and health system planning; Current structure and function of the ICC; Effectiveness of the ICC in supporting the goals of the national programme; Support and recommendations required for strengthening the function of the ICC and of immunization coordination more generally.
Strategic directions are outlined in the first part of the results section of this paper, and current structure and function and effectiveness of ICCs in the second. Recommendations are outlined in Box 1 in the discussion section.
Prior to the in-country visit, the terms of reference of the evaluation were circulated to National Immunization Programmes and in-country UN partners. The office of WHO in-country facilitated engagement with the main ICC members to arrange the schedule of interviews. In all cases, a prearranged list of interviewees was made available. Additional to this approach, other key informants were identified in-country during interviews. That is, during each interview, the respondent was asked to nominate a further key informant who was well placed to provide insights on the practical functioning of the ICC. As Figure 2 demonstrates, this technique (a prearranged list of interviewees supplemented by guidance in-country of key informants) enabled a wide representation of constituencies in the evaluation. This balance in the study sample and the mixed method of respondent selection reduced the risk of bias in the study.
ICC members interviewed included senior representatives of Ministries of Health, UN partners, bilateral and multilateral agencies, EPI management and civil society representatives, and country GAVI focal points. Interviews were supplemented by group discussion with stakeholders, and in one country (Indonesia) the convening of an ICC meeting. International literature on GAVI and global health initiatives, GAVI databases on financial commitments and WHO/UNICEF coverage data were also utilized for background analysis. Search fields for the literature review utilized the terms ''GAVI'', ''Governance'' and ''Global Fund''. The search was undertaken through PubMed database. Documentation at the country level that was reviewed included the terms of reference and minutes of the ICC, and national health planning documents including national health plans (MOH Nepal 2004;MOH Sri Lanka 2007;MOHFW 2008) and multi-year and financial sustainability plans for immunization. 2

Analysis
Findings on structure and function from interviews, group discussion and documents were tabulated in Excel format according to categories based on commonly identified functions and structures of ICCs. Key findings from each country were then described in country case studies according to the categories of strategic direction, current structure and function of ICCs (including their strengths and weaknesses), and identifications of major options for coordination strengthening. Themes from case studies were then used for generating an overall regional picture of immunization and GAVI coordination, including identification of overall strategies for ICC and immunization coordination strengthening.
The themes that are analysed in the discussion were further elucidated by the 'principles of governance' that are described in the introduction to this paper. The recommendations, based on suggestions from study respondents, were prioritized by the author based on the main study theme areas of immunization strategic directions, ICC structure and function, and ICC effectiveness and support.

Immunization strategic direction
Trends in social development, health system strengthening and the new vaccine agenda in the Asian region all point to a significant shift in immunization programming direction and strategy in the next 5 to 7 years. The direction of this shift is towards more systems-orientated management styles and more integrated service delivery strategies.
In terms of social development, it is apparent that in all countries there is an emergence of civil society and the private sector in advocacy and service delivery, and the academic and scientific communities in research and policy development. Decentralization is a key political and social trend in most countries that is presenting formidable challenges for health system and programme planners. These social developments are adding significantly to the complexity of the immunization coordination environment. Governments are supporting the emergence of civil society and health sector forums, as well as scientific advisory bodies in Indonesia, Sri Lanka and Bangladesh in order to manage this increasing social complexity.
In terms of health systems, the developments of sector-wide approaches (as in Bangladesh and Nepal), essential packages of services (most countries) and health system strengthening programmes through GAVI (all countries) are all factors creating increasing pressures for immunization programmes to link more closely to health sector development. A review of directions in immunization and health systems planning indicates that health planning is shifting towards more integrated health system and maternal and child health care (MCH) frameworks. The National Rural Health Mission in India (MOHFW 2008) and the Health Nutrition and Population Support Program in Bangladesh demonstrate this clearly.
In terms of immunization programme development, it is probable that there will be a significant expansion of national and international investments in immunization and new vaccines in the coming years. All countries have plans on the table for research and development of programmes to introduce new or underutilized vaccines (Hib, pneumococcal vaccine, rotavirus, Japanese Encephalitis) between now and 2010. GAVI, through a new policy requiring co-financing of new vaccine introduction, is accelerating this planning process. This will therefore create additional pressures for development of management and coordination capacity for national programmes.
The structure and function of the ICC system

Structure of Inter-Agency Coordination Committees
In all five cases, the ICC is chaired at a high level by a Minister, Secretary of Health or a Director General. Membership is broad in most instances, although academic and private sector interests are not as well represented as in the Global Fund Coordination Committees. All five countries have civil society, bilateral and multilateral representation. System specialists were less well represented; three of the five countries had a planning representative, and only one of the five had a finance specialist. None of the five ICCs had written governance procedures, and only one had a written work plan.
In three of the five countries, a technical advisory group on immunization with policy recommendation powers has been established (henceforward referred to in the SEARO generic term of National Committee for Immunization Practice, or NCIP). All countries have some form of formal or informal operations or technical working group which focuses more on day to day coordination issues of immunization programme management. In most instances, respondents identified these policy and operational groupings as more vital to immunization coordination than the ICCs. The reason commonly given is that NCIPs have the technical expertise to provide advice on immunization policy, whereas ICCs do not (quite commonly ICC members are donor representatives and not immunization experts). Secondly, technical working groups include EPI managers and their staff and technical advisors, who have a more ready grasp of the main issues affecting programme implementation.

The function and perceived effectiveness of ICCs
The ICCs in SEARO GAVI eligible countries function predominantly for the purpose of GAVI application processes and information dissemination. For this reason, most respondents considered the ICC to have an important role in international collaboration. There have been many instances where the ICC mechanism has also been successfully tasked to review the immunization multi-year plan and financial sustainability plan. In some instances, the ICC has been used more strategically for problem-solving issues of logistics and resource gap analysis for financing of immunization campaigns. These were viewed by respondents as the main strengths of the ICC mechanism. However, in most cases, respondents reported that the ICC is essentially operating for GAVI endorsement processes and information dissemination. What was commonly absent from the review was any sense that the ICC was functioning in any strategic planning, problem solving or analytic way (with some exceptions as described above). There was very little evidence that the ICC was consistently and effectively addressing the core issues of coordination and resource gap analysis. Although in many countries there are significant challenges associated with reaching the remaining 20-30% of unimmunized population, there was limited evidence that the ICC was being used as a coordination mechanism to mobilize resources to reach this population. Although unspent immunization service support funds are an issue in some countries, once again, there was little evidence to support the view that the ICC had any oversight function in relation to expenditure tracking or programme review (aside from the annual progress report). These limitations in function were described by respondents as the main factors constraining the effectiveness of the ICC as a coordination mechanism.
Various reasons were commonly provided for this limited implementation of coordination function of the ICC. These reasons included the following: 1. Structural constraints due to under-representation by key constituencies (academia, private sector, civil society, health systems). 2. Governance constraints due to lack of rules or procedures, including systems for recommendation and follow-up actions. 3. Functional constraints, such as lack of agency (especially bilateral agency) orientation to the functions and scope of the ICC and of GAVI, and in particular on the authority of the ICC to recommend on policy and strategy.

Strengthening immunization coordination
One of the main findings from this study is that the need for coordination encompasses a range of areas that includes policy, operations and system coordination.

Policy coordination
The scientific, academic and private sector are now significantly larger players in setting the immunization policy agenda. This is particularly the case for immunization programmes in Sri Lanka, India and Indonesia, where these bodies were referred to as Technical Advisory Groups. These advisory groups are typically headed by academic and research institutions, with participation from government, epidemiologists, vaccine manufacturers, regulators and government public health managers. The Regional Office of WHO has recently published a general framework for establishment of a National Committee for Immunization Practices (NCIP) (WHO SEARO 2007). Terms of reference cover such functions as immunization schedules, country decision making for new vaccines, scientific research findings, and establishment of norms and standards for adverse events and surveillance systems. These scientific bodies are the principal agencies for recommendation of new policy to government in three of the five countries. In Sri Lanka for example, the Secretary of Health will not make a decision on new vaccine introduction in the absence of a recommendation from the National Advisory Group on communicable disease control (NCIP equivalent).
Immunization programme operational coordination All countries have formal or informal technical working groups. These groups are the principal bodies for operational coordination, in terms of routine immunization programme strengthening, campaign implementation, and problem solving of logistical issues. Some countries recommended formalizing the technical working groups in order to streamline the operational coordination. Formalizing these groups could act as a stepping stone to development of NCIPs, although it should be noted that both operational and policy coordination will in the long run still require distinct forums.
Health system coordination Health sector coordination forums are established in Bangladesh and India, and are emerging in Sri Lanka, Indonesia and Nepal. This adds an additional dimension to immunization coordination that is beyond the scope of the ICC, although it was suggested that the addition of health system competencies to the ICC could help to strengthen links between immunization and the health system. In Indonesia, Sri Lanka and India, there are long-established systems for licensing of vaccines through national regulatory authorities. India and Indonesia have established vaccine manufacturing industries. Vaccine manufacturers have a particular interest in coordination in view of their need to put in place long-term market planning (which in turn supports planning of research and development). Regulators have a need to be in touch with the latest national and international science and research on new vaccine and technologies to inform their role in standard setting and licensing.
Coordination with civil society and the private sector are increasingly presenting challenges and opportunities for immunization planners. The private sector is capturing an estimated 5-10% of the immunization market (mostly in urban areas). This sector is particularly active in pioneering the introduction of new vaccines. In Bangladesh, the private sector is becoming increasingly active through 'corporate social responsibility' investments in MCH services for the very poor. In most countries, NGOs are very active in primary health care provision. As a result, the need was expressed for NGO forums, in order to coordinate their investments in a targeted way towards the hardest to reach. This was considered to be particularly important in Nepal, with a very large number of NGOs but limited mechanisms for coordination.

Immunization programme integration with health systems
Sector-wide approaches Sector-wide approaches are designed to address the efficiency and coordination problems often associated with parallel funding and vertical programming. During the interview process in the five countries, sector-wide approaches were described in the following ways: Reforming institutional processes in order to promote more integrated planning systems; 'Pooling of ideas and plans' in joint annual programme reviews, health sector strategies; and Funds pooling.
In the two countries where the sector-wide approach mechanisms are most advanced (Bangladesh and to a lesser degree Nepal), there was some interest expressed in closer coordination of immunization and health sector planning. In Bangladesh in particular, GAVI investments are analysed and financial commitments tabulated in the annual programme review of the Ministry of Health. One option on the table in Bangladesh was the idea of integrating the ICC function into the wider health sector coordination forum, with a TWG taking up the more specialist role of advising the sector forum on operational aspects of immunization coordination. There was as yet no suggestion that global health initiatives such as the Global Fund or GAVI should proceed to funds pooling. The sector-wide approach mechanism is considered to be too early in its development, and managers viewed the current Global Fund and GAVI mechanisms as flexible and efficient funding and programmatic strategies that provide the best prospects for assisting governments to scale up programmes in order to reach the Millennium Development Goals.

Health system strengthening
Despite the hesitancy about aspects of the sector-wide approach, it was nonetheless recognized by immunization planners and investors that the trend towards broader-based health system strengthening approaches to solve intractable programmatic problems is gathering momentum. Countries are responding to this trend by conceptualizing health system strengthening programmes for immunization in terms of an 'essential package' of MCH services. Managers from countries expressed the need for a wider mix of competencies in the ICC (planning, finance and human resource management) in order to assist in broadening the perspective and approach to issues of sustaining and improving immunization and MCH coverage. Most health system proposal development teams demonstrate a good mix of immunization, MCH and health system competencies. Additionally, in some instances, ICCs are being expanded to include civil society and health system specialists, and, as is the case in Bangladesh, health sector forums are being strengthened in response to the trend towards more systematic approaches to solving programme problems.
One manager commented that in previous years, the government could more easily centrally direct implementation of national policy and plans. Now that sub-national health managers are no longer governed by central command, but more so by decentralized or devolved health authorities as is the case in India, Sri Lanka and Indonesia, it is now necessary to engage a wider range of interests in decision making in order to 'get things done'. These interests increasingly include civil society, the scientific and academic community, regulators and the private sector. This increase in the complexity and diversity of the decision-making environment is consequently exerting additional pressures on national programmes to broaden their coordination network.

Strengthening technical coordination
Monitoring and evaluation In many countries, there were successful examples of evaluation and studies that included assessments of multi-year planning, financial sustainability planning, coverage surveys, cold chain and external campaign monitoring, many of which were supported through the main country technical support partners and international ICC members, WHO and UNICEF. However, in most countries, there was lack of clarity amongst GAVI Alliance partners on their role in ongoing monitoring and evaluation. Although Alliance partners have limited decisionmaking power in terms of being signatories to GAVI applications and progress reports, there is still a sense that 'accountability for performance' for this international investment is far from clear. Major areas of lack of clarity of function were in financial tracking, monitoring and evaluation of programme effectiveness, oversight of technical cooperation, and management of communications with GAVI.
This lack of clarity, particularly in relation to monitoring and evaluation, is reflected in practice in the lack of in-depth discussion in ICCs on strategic gap analysis, evaluation and immunization service support expenditure tracking. Although respondents indicated that performance in Phase 1 was still successful (as judged by sustained or improved coverage and new vaccines introduced), with management developments such as multi-year planning and financial sustainability planning, there is concern that increasing levels and diversification of investment in Phase 2 will increase pressures for more structured and in-depth monitoring and evaluation studies and in-country technical and secretariat support.

Technical coordination
Management capacity was often identified as a major constraint in relation to the monitoring and evaluation and implementation of GAVI programmes. The capacity issue is EVALUATION OF GAVI COORDINATION MECHANISMS not just related to systems development; it was most often referred to in the context of inadequate staff numbers to support programme development and administration (i.e. 'absorptive capacity'). Additionally, some countries argued that the system of technical assistance is intermittent and ad hoc. There is lack of clarity as to who is responsible for providing technical assistance for which aspects of GAVIrelated programming.
In some countries (India, Nepal), GAVI focal points have been established to support administration and technical matters. However, in some cases, there seemed to be lack of clarity regarding terms of reference and accountability, and levels of sustainable support for such roles. In Indonesia, an internal GAVI secretariat has been established to support administrative tasks and the ICC and Technical Advisory Group. The secretariat works under the direction of the immunization manager (through contracting in of support staff). This system was reported to work well in terms of easing the management burden. The alternative is to recruit more government staff, but this is considered to be extremely difficult in countries such as India and Indonesia where there are strict quotas and lengthy high-level procedures in order to gain approvals for appointments to new civil service posts. This being the case, 'contracting in' is considered highly preferable to new civil service recruitment or establishment of a 'contracted out' external GAVI country secretariat. No country suggested this was a viable option because of the risk of developing non-sustainable and parallel management systems.

Discussion
From managing by committee to managing through systems This survey has identified areas for strengthening in GAVI governance mechanisms at a country level. Governance functional capacity within ICCs, particularly for policy guidance and 'intelligence and oversight' (WHO 2007, p. 23), was notably lacking in this review of their function in five countries. These findings are consistent with recent reviews of the governance capacity of GHIs. Although not specific to GAVI, one review highlighted the finding that GHIs have adopted 'poor governance habits' including: 'inadequate performance monitoring; insufficient oversight of corporate partner selection and management of conflict of interest; and a lack of transparency in decision making' (Buse and Harmer 2007). The governance arrangements of the health system strengthening window of funding of GAVI have also been reviewed, with the review concluding that at country level, development partners have expressed concern about the oversight of health system strengthening funding and lack of clarity for accountability of strategy (Naimoli 2009).
A more recent survey of DTP3 immunization coverage globally established that coverage 'has improved more gradually and not to the level suggested by countries' official reports', with reviewers concluding that increasing aid flows through GHIs such as GAVI need to be accompanied by the responsibility to ensure that resources are being utilized for intended purposes (Lim et al. 2008). The effect of all these governance difficulties is insufficient trust and transparency, a key ingredient for the success of development partnerships (Naimoli 2009).
One of the most important findings from this study is that, due to the social, health system and immunization programme developments outlined earlier, the strategic environment for immunization is shifting rapidly. Increasing pressures to decentralize, regulate and engage with civil society, private sector and public sector partners has created demand for wider participation in immunization policy making and implementation. The diversification and decentralization of modern societies in the SEARO Region is leading to a more complex decision-making and resource coordination environment. The cumulative effect of these social and system changes is that over-reliance on an ICC committee structure as a strategy for coordinating and guiding a large GHI investment serves to underestimate the complexity of the organizational task required in order to adjust to this more pluralistic management and social context. This being the case, it is now more relevant to refer to an immunization coordination system than solely to an immunization coordination committee (such as an ICC) as the overarching institutional context and governance mechanism for coordination of national and international immunization resources. This broader view of immunization programming and service delivery has a range of key entry points for coordination that includes civil society, sub-national health authorities, the academic and research community, regulators and health system planners (see Figure 3). Increasing social and institutional complexity, and the trend to health system strengthening, is likely to see the locus of authority and decision making on GHI coordination shift from national programme planners to central Ministry of Health agencies and coordination forums.
By coordinating and managing through an immunization coordination system (rather than over-emphasis on coordinating through an ICC), it will be possible to be more sensitive to country needs, trends in institutional development, stakeholder participation and hence national ownership of immunization and other public health programme agendas. It expands the opportunities for wider social participation in decision making about public health interventions and resource allocations. The potential effect of this widened participation is improvements in absorptive capacity of national governments for increasing levels of international health assistance, which has been identified as a key issue in one evaluation . Also, it increases the potential for improving the quality of national decision making through increased exchange of scientific, health system, private sector and civil society perspectives on immunization service improvement. This by no means reduces the significance of the ICC mechanism. On the contrary, it supports the ongoing strengthening of the ICC by distinguishing its prime and specialist function of international cooperation within this wider coordination context of the immunization coordination system. The main study finding that the ICC is being used for GAVI application and annual progress report purposes but is not being used well for strategic gap analysis and monitoring and evaluation highlights the need for strengthening ICCs to more effectively assist the Ministry of Health to coordinate national and international immunization resources.

From managing by programme to managing through systems
In all countries, there have been significant health system developments and reforms in recent years. In India, the National Rural Health Mission was launched in 2003, the main direction of which is to expand access to basic health services through a more integrated health management systems approach. Bangladesh has undertaken a period of health sector reforms in recent years (from 1998) that have included unifying service structures at the sub-district level (integration of family planning and health services) with the intention of unifying management and expanding access. An underlying intention of the Health Sector Reform Strategy (HSRS) in Nepal is to move towards a sector-wide approach (MOH Nepal 2004). The HSRS sets out three Programme Outputs (Prioritized Essential Health Care Services, Decentralized Management of Health Facilities, and Role for the Public Private Partnership). In Indonesia, recent decentralization of the health sector is presenting formidable management and policy challenges for central and district planners. In Sri Lanka, a Health Sector Master Plan 2006-2016 outlines the main health sector challenges, which include adjustments to demographic and epidemiological transition, the growth of the private sector, decentralization and human resource management change (MOH Sri Lanka 2007).
Notwithstanding the governance constraints alluded to earlier, the recent initiative by GAVI to open a health system strengthening window of funding is stimulating stronger partnerships between health system and immunization programme planners (Grundy and Moodie 2008), and the emergence of more defined health sector forums. Civil society and private sector service providers are all now well established in MCH service provision in most of the countries. These health system reforms and developments are taking place often within the context of ongoing civil instability (Sri Lanka and Nepal) or natural disasters (Indonesia and Bangladesh).
Amidst all these changes and developments, common themes in health system planning include strengthening health management in a decentralized context, sector-wide approaches, human resource motivation and planning, reaching out for the provision of 'essential packages of services' for hard to reach or vulnerable populations, and rural health infrastructure development. Given these trends towards decentralization and sector-wide approaches, there are now substantially increased policy pressures on national immunization programmes to integrate more effectively with health systems, and in particular, with MCH services in rural areas. This trend towards health system strengthening as a development paradigm is likely to continue, particularly given the findings from a review of 134 Global Fund grants that 'countries with stronger health systems and larger numbers of trained health workers are more likely to have successful programmes ' (Radelat and Siddiqi 2007) The governance of global health initiatives It was also observed during consultations that there is a spectrum of committee authority ranging from information sharing to recommendation and finally to decision-making authority. It was perceived by most that the ICC is currently on the far left-hand side of this spectrum (see Figure 4), but should shift further to the right, particularly in relation to strategic gap analysis and evaluation. There was no suggestion, however, that the ICC should shift further to 'decision making by committee'. There is some perception that the GFATM Country Coordinating Mechanism (CCM) model, for example, constitutes 'decision making by committee', and this is not the recommended path for ICCs. This finding is commensurate with a review of GHIs, which indicated that early country experiences of many GHIs have resulted in the establishment of parallel systems which are tending to reverse efforts toward coordinated development assistance (Brugha 2008, p.72). One other review indicated that the Global Fund has attracted more attention than other GHIs for poor harmonization of Global Fund and partner government institutional and financial planning processes (Buse and Harmer 2007, p.267).
On the other hand, one manager in this study reported that the GFATM CCM model is admired for its procedural effectiveness (governance procedures) and higher level representation. This reflects the main management challenge of GHIs of facilitating improved governance and sustainable programme results without developing parallel structures.
In this study, ICCs have invariably been viewed by respondents as a coordination body, and not as an executive agency. In fact, GAVI is highly valued in all countries for its capacity to work within government decision-making processes, rather than establishing parallel systems. In order to reduce duplication of separate coordination mechanisms for separate GHIs, there is probably a case for closer coordination of GHI forums with sector planning and monitoring and evaluation processes, particularly given this early success of GAVI in alignment to national plans and systems. At the same time, GAVI may

Recommendation by Committee Decision Making by Committee
Coordination & information sharing by Committee need to facilitate the development of more accountable governance procedures such as mandates, processes and membership of ICCs, including information on how constituencies are managed and how annual performance assessments are conducted (Buse and Harmer 2007, p. 265).

Global health initiatives and the absorptive capacity of implementers
A published review of Global Fund grants indicates that focusing resources on politically stable low-income countries will not run into difficulties of absorptive capacity as measured by grant implementation (Lu et al. 2006a). Level of grant implementation, however, does not equate with quality of grant implementation in all instances. In this study it was recognized that the ICCs were underperforming in relation to core functions of monitoring and evaluation and strategic gap analysis of GAVI grants, which are critical requirements for effective grant implementation. This lack of strategic gap analysis is further borne out by a global study of the impact of GAVI immunization service support grants on immunization coverage, which indicates that for countries with immunization coverage above 65%, the grants have had substantially reduced impact on coverage (Lu et al. 2006b).
This study confirms that GAVI globally has underestimated the institutional barriers to absorption (as defined by quality of implementation) and coordination of the investment at a country level. It has reinforced previous findings of lack of coherent and consistent systems for support for implementation (McKinsey & Co and Bill & Melinda Gates Foundation 2005). This in all likelihood stems from the organizational culture within GHIs of 'top down' models of governance (Muraskin 2004). ICCs have been viewed as a 'one size fits all' governance strategy for coordination of the GAVI international investment. Additionally, the significant variety in levels of health system development in GAVI and Global Fund eligible countries indicates that bodies such as ICCs and CCMs have ongoing but varying needs for oversight of programmes within countries (Radelet and Siddiqi 2007).
This study has established that there are significantly higher levels of institutional and contextual complexity at country level that require a more focused global response by GAVI on the governance challenges for institutions and partners implementing GHIs at the country level (see Box 1 for recommendations). It also confirms the findings of a global country consultation review of Phase 1 operations of GAVI, which indicated that 'major reform of ICCs was recommended by some partners to ensure a more active involvement of policy makers (including Ministry of Finance and Planning)' (HLSP 2005).

Conclusion
The evaluation has identified that there have been significant changes recently in the strategic environment for immunization, including developments in new vaccines and increased GAVI investment in Phase 2 of its operations (2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015). In tandem with developments in immunization, there are now well-established trends towards health system integration and decentralization, and institutional development of the nongovernment sector (academic, civil and private). In this context, the evaluation has found that, although ICCs are functioning well in relation to information sharing and GAVI application processes, they are performing less well in three main areas of monitoring and evaluation, strategic gap analysis, and coordination of technical co-operation.

Box 1 Recommendations for strengthening immunization coordination, SEARO ICC Evaluation 2007
1. Strategic directions: In order to facilitate closer linkages between immunization programmes and the health system, international health investments such as GAVI should be described, planned and evaluated within the system of country planning and monitoring systems such as Annual Programme Reviews and Health Sector Plans. 2. ICC structure and function: ICCs should review their country terms of reference (TOR) to describe essential coordination functions that include monitoring and evaluation, strategic gap analysis, expenditure tracking and resource mobilization. The TOR should clearly identify links to the broader immunization coordination system, as well as clearly identify that the coordination role is not limited to GAVI, but to other forms of international immunization health investment. 3. ICC structure and function: In some instances, particularly in countries where sector-wide approaches are more developed, it may be efficient to integrate ICC function into either the Global Fund CCM or a wider donor forum. This would be conditional on the effective functioning of an immunization operations technical working group and a scientific or policy National Committee of Immunization Practice (NCIP) that could recommend immunization policy development and resource mobilization strategies to the wider sector forum. 4. ICC support: GAVI should support countries to develop clear governance procedures for ICC function. Governance guidelines should incorporate role and function of the ICC and its membership, secretariat functions, rules and procedures governing chairmanship, membership renewal and participation, constituency representation, management of conflict of interest, proposal review and endorsement, technical cooperation strategy development, procedure for recommendation to MOH on GAVI applications or resource coordination, and links to the wider immunization coordination system and health system. 5. ICC support: In countries with a less developed immunization coordination system, investment should be to provide resource support for the development and operations of NCIPs (as regionally defined by WHO SEARO and adapted to country conditions).
On this basis, it is concluded that ICCs are no longer the sole point of reference for immunization coordination, but should act more in concert with government programme management, the health system, regulatory authorities, civil society and the private sector in matters relating to vaccine and programme decision-making and coordination. In this context, the ICCs should be maintained and strengthened as being one key international cooperation component of a wider immunization coordination system. Furthermore, the linking of immunization coordination to wider health system planning and delivery processes will strengthen key functions of strategic gap analysis and monitoring and evaluation, and reduce duplication of effort with other GHI initiatives such as the Global Fund. Specifically, integration of immunization with health system planning processes will promote efficiencies in implementation and, in doing so, expand the reach of immunization and maternal and child health care services in developing countries.