Strengthening systems for integrated early childhood development services: a cross-national analysis of governance

Authors


Abstract

While there has been substantial growth in early childhood development (ECD) services in low- and middle-income countries (LMICs), there is considerable inequity in their distribution and quality. Evidence-based governance strategies are necessary, but currently they are insufficient for widespread, quality implementation. In particular, there is a limited understanding of the use of systems approaches for the analysis of ECD services as they go to scale. The aim of this paper is to present findings from four countries, using a cross-national case study approach to explore governance mechanisms required to strengthen national systems of ECD services. While different sets of governance strategies and challenges were identified in each country, overarching themes also emerged with implications for systems strengthening. Study results focus on local, mid-level and central governance, with recommendations for effective coordination and the integration of ECD services in LMICs.

Introduction

In recent years, in spite of the expansion of services for young children and families, many of which have been accompanied by rigorous evaluations demonstrating effectiveness,[1] a majority of the world's youngest children still lack access to quality early childhood development (ECD) services.[2] This is due in part to poor systems-level coordination, and sometimes to the chaotic and unsystematic approaches used to scale up programs.[3] It is the capacity or ability of the system, therefore, that creates the opportunity to achieve desired outcomes.[4] Understanding the operations, strengths, and feasible entry points of ECD systems can help to ensure more effective service delivery, sustainability, and scalability.

Governance has been identified as a critical element of a system[5-7] because it provides its support and the operational infrastructure.[8-10] ECD governance processes enable the allocation of responsibility for services and multiple functions among specific stakeholders within and across levels of government and organizations of civil society and the private sector.[11] For integrated ECD services, understanding the governance of systems is important in order to ensure they are well planned, implemented, and coordinated. However, most work on improving ECD service equity, access, and quality has examined these features from a service-level perspective, that is, in the setting where the services are delivered.[12] Depending upon the country, relatively less attention has been given to the larger organizational and institutional structures within which the ECD services are situated. In comparison to other service systems, such as health and nutrition, there have been fewer inquiries into governance of ECD services per se.[9, 13]

To address issues of ECD governance, the authors designed and conducted a study to examine the governance functions of national ECD systems focusing on the planning, implementation, coordination, and financing of ECD services in a sample of three low- or low-middle-income countries (Cambodia, Lao People's Democratic Republic (PDR), and Kenya) and one upper-middle-income country (Peru). A mixed-methods phased design was employed with the use of semistructured interviews with key informants, focus group discussions with beneficiaries, and the examination of national policy documents. The main goal of the study was to understand the governance of national ECD systems primarily in the health, nutrition, sanitation, education, and protection sectors. Although a unique set of governance challenges and strategies was identified in each country, overarching themes also emerged with important implications for strengthening ECD systems. This article presents selected study results along with implications for scaling up ECD programs and especially for coordinating integrated ECD services.

Conceptual framework: theoretical and empirical underpinnings

In part, services can be improved by strengthening the larger systems within which they are implemented.[14] Historically, drawing on health systems as an example, targeted, sectoral approaches for service provision (e.g., immunizations, antiretroviral treatment for AIDS) were the dominant norm in the 1980s. However, efforts to improve the integration, local governance, and sustainability of services as well as the financing of national health strategies have led to using more systems-based approaches.[15] Similarly, ECD has transitioned from the evaluation of single interventions to exploring the scalability and sustainability of services across multiple sectors at national levels. Therefore, applying a systems-level approach allows the exploration of mechanisms to scale up services and incorporate them into the existing complex adaptive systems of service delivery for young children and families.[5]

Early childhood is a period from conception to 8 years of age when multiple domains of children's skills and competencies are intertwined in their growth, across physical, social, emotional, language, and cognitive domains.[16, 17] The sectors involved in ECD service provision typically include health, nutrition, sanitation, education, and protection.[18] In infancy, interventions tend to focus on health and nutrition services, parent and community support, and education and early childhood intervention for children with developmental delays and disabilities. After 3 years of age, services tend to shift to more formal education and learning programs of preprimary and primary schooling.[19]

A wide range of settings are used to deliver ECD services, and they vary within and across sectors and countries.[20] For example, programs can be home, center, clinic and/or community based. Services are also characterized by the generation of the beneficiary. For example, services can be for a single generation (serving either only the children or only their parents) or multigenerational (providing services to parents, children, and the extended family).[21, 22]

Finally, the auspices under which services are sponsored also differ, including governmental, nongovernmental (NGO), and private for-profit agencies. [23] These complex webs of services require good governance to enable their scaling up in an equitable, efficient, and effective manner.[4]

Governance processes allocating responsibility across levels and sectors of society are implemented directly through political processes, policies, plans, inter-agency agreements, and service support, and indirectly through registration, service and personnel standards, certification and recertification, regulations and protocols, and accountability and feedback mechanisms.[8]

Among the many governance functions, vertical coordination (across national, subnational, and local levels of government) and horizontal coordination (or intersectorality) are especially important for implementing integrated ECD services. Decentralization, a process especially affecting vertical governance through which decision-making and/or budgeting responsibilities are devolved or expanded from central to local levels of government, has greatly affected ECD in a range of countries.[24-27]

Ensuring the successful implementation and delivery of ECD services requires collaborative efforts among local government officials, practitioners, community residents, and families as important stakeholders.[28] Horizontal coordination is central to implementing and improving the quality of integrated and multisectoral services, for example nutrition-sensitive programs that include early childhood.[29] When a single sector is the primary lead in ECD services, the associated content area may dominate. More integrated services, in contrast, may include sectors of health, nutrition, sanitation, education, social welfare, and protection, each of which typically reside in different ministries or departments.[30] Some instances of combined ministries are found, such as health and protection ministries.[31] These tend to help enable greater multisectoral ECD coordination, but not always.

The governance of ECD services is situated within larger political, public policy, cultural, and economic contexts.[32, 33] These contexts are complex and too vast to cover in detail in this paper. Therefore, we highlight the key features of environmental conditions and capabilities so as to demonstrate that governance cannot be understood without taking into consideration the context within which it functions.[4] Environmental conditions typically include the distribution of political power, prevailing market structures, cultural and community norms, demographic transitions, and the availability of physical infrastructure (e.g., transportation). Implementation capability refers to the system's ability to design, implement, coordinate, and monitor the overall plan of action, and it can be further broken down into the capacity of knowledge and capacity of practice,[6] with the former being more relevant at the systems level and the latter to the service level.

Research questions

The study addressed the following questions:

  1. Who are the main actors in ECD and what role do they play in the governance of national systems?
  2. What are the horizontal (among the main actors (government, nongovernment and private for-profit)) and vertical (between national and subnational levels) coordination mechanisms for the governance functions of planning, developing, implementing, delivering, monitoring, evaluating, and financing ECD services?

Methods

Sample

A three-tiered approach was used to conduct sampling: (1) geographic sampling of countries and within countries including subnational levels; (2) sampling of policies, sectors, and services relating to ECD service provisions; and (3) key informant sampling. Each of these tiers is described further.

Geographic sampling

Two levels of geographic sampling were conducted. First, three regions were selected from a global sample (East Asia and the Pacific; Eastern and Southern Africa; and Latin America). For country selection within each region, the following criteria were used: (1) a combination of country indicators of ECD; (2) interest in the study; and (3) ability to support the implementation of the study in similar countries within a region. Based on these criteria, Cambodia and Lao PDR were selected from East Asia and the Pacific; Kenya and Uganda from Eastern and Southern Africa, and Peru from Latin America (see Fig. 1).1

Figure 1.

Sampling frame: global to local levels of geographic sampling.

Within-country sampling at subnational levels is displayed in Table 1. In each country, three regions were sampled from the most advantaged (represented by the capital city region in all countries), the median, and the least advantaged with respect to child well-being and access to services indicators.2

Table 1. In-country data collection overview
 CambodiaLao PDRKenyaPeru
Areas visited

Province 1: Kampong Thom

District: Kampong Svay district

Villages: Snor village and Traping Reusei village, both in Trapaing Reusei commune

Province 2: Oddar Meanchey

District: Chong Kal district

Villages: Kouk Pongro village, Cheung Tieng village, and Sre Prang village, all in Cheung Tien commune

Province 1: Vientiane

District 1: Xaysomboun

Villages: Hinhouaseaua; Muang Soum; Nam Thea

Province 2: Xienghouang

District 2: Pek

District 3: Mok

Villages: Thachock, Na Theur

Province 1: North Eastern

District: Garissa (provincial capital)

City/town: Garissa

Province 2: Eastern

District: Machakos

City/town: Machakos

Region 1:

Lima

Municipality:

Metropolitan

Municipality of Lima

City: Lima

Region 2: Ayacucho

Municipality:

Luricocha

Community: Chamana, Pichuirara

Region 3: Loreto

Municipality:

San Juan Bautista

Community: Belen, Primavera

Semistructured interviews and focus group discussion

National level:

n = 18

Provincial level:

n = 11

Village level:

Four focus groups per village (total n = 16 focus groups)

National level:

n = 18

Provincial level:

n = 7

District level: n = 5

Village level systems FGD: n = 10

Village level beneficiary FGD:

n = 6

National level:

n = 14

Provincial level:

n = 4

District level: n = 8 and 1 FDG, n = 6

NGOs: n = 1

Service level: n = 8

Beneficiary level: 1 FDG, n = 12

National level:

n = 44

Regional level: n = 57

Service level:

n = 27

Beneficiary level: 6 FDGs

Systems and services studied

Health: C-IMCI

Education: community preschool program

Home-based preschool program

State preschool program

Health: expanded immunization program

Education: public preschool and primary school

Health: essential health package, C-IMCI, EmOC

Education: public preschools

Protection: children's home

Health: national health program

Education: public preschools, community-based schools, early stimulation rooms

Protection: child protection units, conditional cash transfers

Integrated national programs: Targeted child health and development programs (e.g., Cuna Mas)

Policy, sector, and service sampling

In-country inventories of candidate sectors and services were generated to conduct research on ECD systems. To create the inventories using a set of policy-relevant coding criteria, the content of four categories of national documents were analyzed (i.e., ECD policies, social policies, legislation and laws, and national development policies) that influence ECD service provision in the country.[34] Second, a set of inclusion criteria were generated and applied to identify candidate services that (1) directly address children in the ECD age bracket as defined by the country;3 (2) are implemented at sufficient scale to enable an examination of access (typically across more than one region or province), which include government and NGO-sponsored services; (3) are considered to be feasible for observation and in-depth inquiry of operations; and (4) would grant permission to conduct interviews with service providers and systems-level staff associated with the program.

Informant sampling

There were three levels of informants for the study: system-level informants, service-level informants, and actual and potential service beneficiaries. These respondents cut across government and NGO agencies. System-level informants are actors who are part of the system that supports the provision of the service (e.g., the manager of a social development office at the regional level) but do not directly provide the services to children and families. These informants were sampled from national and subnational levels. Service-level informants are individuals who directly provide services or interact directly with children and families (e.g., teachers, nurses, midwives). They were typically sampled from the most local level (village/community/commune). Service beneficiaries are parents, caregivers, and children who receive the services. This paper focuses only on the first two categories of informants.

Procedure

This study was conducted in three phases in each of the countries selected.

Phase 1: desk review and ECD situation analysis

During the first phase, an ECD situation analysis was prepared with the following goals: (1) to create a country profile of ECD outcome and access indicators taking into account the political, economic, social, and cultural contexts; (2) to identify the specific policies, systems, and services to be studied; (3) to identify the regions, provinces, and municipalities to be studied; and (4) to determine the agenda for the country visit in Phase 2.

Phase 2: in-country data collection

The purpose of the second phase was to conduct an in-depth case study of ECD systems in the country (Fig. 1). All country visits were the result of highly collaborative planning between the study team and local partners (including UNICEF, the Bernard van Leer Foundation, and other relevant NGOs) and were conducted over an average of 2 weeks. The composition and size of the study teams varied slightly depending on the demands of the in-country work; however, each team was led by the one of the principal investigators in order to maintain consistency across the countries. Semistructured interviews with system- and service-level informants were conducted that lasted from 60 to 90 minutes. They were developed based on an extensive review of the current limited literature on governance and finance of ECD services in low- and middle-income countries (LMICs).[35] The protocols addressed the following areas or topics: (1) service goals and implementation (i.e., understanding the design of the service and how it is being implemented); (2) aspects of governance (i.e., planning, coordination, decision making, accountability, supervision, design, and monitoring and evaluation); (3) governance of the financing of ECD (i.e., generation of funds, decisions regarding the allocation and distribution of resources, and the costs and expenditures of programs); (4) quality of ECD services (i.e., training and certification of service providers, curricula, infrastructure, materials, and service provider and beneficiary interactions and relationships); (5) historical and ideological understanding of ECD and sociopolitical understandings of governance and finance; (6) understanding of the association between policies as guiding documents and service implementation.

The main topics of the study were adapted slightly to each country, depending on the context, actual content and provision of services, and respondent classification. Furthermore, not all questions could be asked of all respondents because of their designation or agency affiliation. No audio or video recording devices were used to protect the identity of informants and because many of the interviews were not conducted in English. Instead, team members took extensive field notes.

Phase 3: data analyses and reporting

The final phase of the study included an adapted version of concept mapping analyses[36, 37] of the interview data, and content analyses of the national policies and laws to extract trends and patterns in results. Each country study team first conducted this process independently; results were subsequently shared across study teams to extract overarching themes. Given that this study is one of the first systematic investigations into governance of ECD, an exploratory coding procedure was employed where data sources (e.g., field notes and analytic memos, document reviews, field observations, or interviews with local and central-level informants) were coded both manually and using Atlas Ti software. Coding results were subsequently shared between team members to ensure consistency and in the interpretation of the data and extraction of themes. Then, the analyses were shared across the country teams to draw comparisons and highlight unique characteristics of ECD systems in respective contexts. Cross-country comparisons included the identification of points of congruence and divergence and emergent themes. As the analysis proceeded, policy architecture and governance maps were drafted and revised iteratively with in-country partners.4

Results and discussion

This section focuses on the themes and patterns of findings relating to integrated ECD services that emerged across the countries included in the study.5 Despite wide variation across countries in ECD service provision and country socioeconomic and political structures, the analyses yielded interesting similarities. The patterns in the findings are presented in the themes outlined in succeeding sections. Results from individual country analyses are used to highlight findings presented in the following sections.

Main sectoral actors in ECD

ECD systems and services are complex because they cover a large age range (prenatal to 8 years old), often address the varying needs of different ethnic and population groups, and cut across several sectors, thereby involving a wide range of actors. A review of the national ECD policies revealed that while the age range covered was vast, the longstanding traditional sectors including health, nutrition, sanitation, education, protection, and social welfare were dominant in ECD service provision in specific age brackets, with health and related sectors focusing especially but not solely on the early phase of this age period and education focusing more on the preprimary and primary school years. However, differences emerged across countries in the composition of actors within each sector as well as in their roles and functions, some of which might provide some insights into ways to deal with the challenges of ECD leadership, multisectoral coordination, and integration.

First, a range of key partners usually exists within each of these sectors. For example, nutrition programs in Peru illustrate this finding well. The Programa Integral de Nutrición (comprehensive nutrition program) is executed by the Ministry of Women and Social Development, Ministry of Health, and Ministry of Education as well as local and regional governments, international cooperation agencies, and civil society organizations. Within the regional government of Lima, three units within the District's Management Office for Social Development are related to the delivery of nutritional programs, including the Submanagement Office of Health, the Division of Health Promotion, and the Division of Sanitary Prevention. In addition, the articulated nutritional program/Programa Articulado Nutricional is the official instrument that monitors the set of strategic programs executed by the Ministerio de Economía y Finanzas (Ministry of Economy and Finance) aimed at reducing chronic malnutrition of children under the age of five. Because each of these actors plays a role in the design, implementation, coordination, training, monitoring, and financing of nutrition services, a clear definition of governance roles is required.

Second, influence over budgeting and major policy decisions varied greatly among actors across the countries studied. In the relatively poorer countries (e.g., Cambodia and Lao PDR) in which donor spending outweighed government spending on ECD, international NGOs played a much more influential role in fundamental decisions regarding program models and policy than in the relatively richer countries in our sample (e.g., Kenya and Peru). Another pattern observed was the primacy of the Ministry of Finance (or in the Cambodian case, the Ministry of the Interior) among ministries for ECD spending. This was the case despite the fact that content expertise in ECD was more likely to reside in ministries of health, education, and protection, and despite the existence of departments of budgeting and planning within these other ministries. In some cases (e.g., Peru), major new initiatives in ECD were either established through national legislation or through particularly active roles conducted by senior staff members of the Ministry of Finance or their technical personnel with competency in children's issues.

Third, within each key sector there are multiple departments that need to work together to fulfill several essential ECD governance functions. In the ECD literature, an actor is often referenced in a general manner (e.g., Ministry of Health or Ministry of Education). The study provided a greater specification and identification of roles and responsibilities of each department within agencies in order for the agency to achieve desired outcomes. For example, within the Ministry of Education of Lao PDR, four different departments were involved in the different governance functions of the preprimary program: (1) the Department of Preschool and Primary Education was responsible for the design, development, and implementation of the service; (2) the Research Institute for Educational Sciences for the curriculum and content of educational materials; (3) the Department of Teacher Training for capacity building; and (4) the Department of Budget and Finance for providing funds from within the education budget for this program. This level of specification, obtained through preparing governance maps, enabled a clearer understanding of how system-level outcomes might be influenced. The definition of roles and responsibilities and an understanding of complementary functions within a sector have implications for improving access and quality.

Another key dimension of actors in ECD are the central-level bodies that lead ECD in the different national contexts. In Peru, the responsibility for leading major national programs for early childhood was transferred recently to the Ministry of Development and Social Inclusion (MIDIS). In Kenya, according to the National ECD Policy Framework, the Ministry of Education was designated as the lead agency for ECD coordination at the ministerial level. The implications of these leadership functions have yet to be determined. However, this study found a highly contextualized process for decision making regarding lead ministries included in this country sample.

Horizontal and vertical dimensions of governance: implications for integrated ECD

It is a well-established fact that ECD requires coordinated services because the holistic development of children must be taken into account. For example, at the service level in Jamaica, a program integrated nutrition supplementation with stimulation for development and achieved notable short- and long-term improvements in child development.[38] However, this is only one aspect of coordination, and it focused primarily on the service level without reference to the systems level. Other work on coordination in ECD services has taken a linear approach, one that spans initial single-sector approaches through cross- and multisectoral approaches into comprehensive approaches.[39]

Two dimensions of ECD coordination need to be examined simultaneously to identify additional ways to strengthen ECD systems for improving service delivery. Horizontal coordination occurs at a particular level of the administrative hierarchy and includes all the main actors who are involved in ECD services. Vertical coordination, on the other hand, includes all the relevant participants within the system, from national to local levels of government. Although vertical coordination has been well studied through models of decentralization, devolution, delegation, and deconcentration,[40] and calls for horizontal coordination are frequent, simultaneous analyses of vertical and horizontal governance are generally lacking in the ECD literature. An exception is a study of Latin American ECD systems, which revealed that large-scale, sustainable, and successful ECD programs included highly articulated horizontal and vertical coordination systems.[41]

In Kenya, the study revealed that the greatest level of coordination occurs across sectors at the national level; in other countries such a high degree of coordination is also noted at the community level. For instance, the government established a national coordinating body for children's services with strong representation of major ECD stakeholders. In some countries at more local levels of government, a similar articulation of coordination was noted. Similarly, in Peru and Cambodia, horizontal coordination between governmental and civil society leaders, and within government across ministries, led to national legislation and action planning in ECD in the last decade. Horizontal coordination was also observed to varying degrees in these countries—at the village level depending on communication between direct service providers in health and education, for example.

Simultaneous vertical and horizontal analyses showed that the middle levels of government (namely, provincial/regional and district) demonstrated the weakest horizontal coordination. Typically, at those levels, staff in health and education sectors did not work across sectors in the governance and implementation of ECD services. In some contexts, their offices were often physically separated, and neither representative bodies nor joint action planning occurred to the degree observed at local or national levels. In each of the country case studies, the interview data revealed that middle levels of government typically see their role as one of supporting information flow and communication within the sector and conducting monitoring and supervision to a lesser extent. The compliance and supervisory/monitoring roles of these intermediate levels were almost entirely vertically focused (that is, compliance oriented upward to higher administrative levels; supervisory/monitoring roles oriented downward toward lower administrative levels). Few incentives existed at these levels for staff to ensure horizontal coordination. Mid-level ECD administrators were found to work mainly within their sectoral silos. Furthermore, there was little to no mention of cross-sectoral collaboration in their job descriptions, tasks, or in their implementation of ECD services. It should be noted that in Kenya, the presence of Provincial Children's Officers, District Children's Officers, District Centers for Early Childhood Education Program Officers, and district and location area advisory councils (AACs) signaled the emergence of both vertical and horizontal coordination mechanisms at the middle levels of government. However, the interviews also revealed that engaging village networks (which are local-level forms of organized representation) was challenging for district AACs. In Peru, some efforts were beginning in national ministries to define the unique sets of competencies for mid-level administrators. This was being done, in part, to address a chronic issue across all countries in this study—the concentration of technical knowledge and management skills in the capital cities (or national level), to the detriment of subnational levels.

Study results suggest that mechanisms for horizontal and vertical coordination are needed at national, middle, and local levels of government. Although the sectors and levels will vary in each national context, guidelines and spaces for coordinated programming are required throughout the governance continuum. These guidelines should establish the roles and responsibilities as well as outline key processes not only for policy makers at national levels and service providers, but also for mid-level actors.

Local governance structures as key levers for improving integrated ECD service provision

Some of the work on ECD at the policy level has focused mainly on national and service levels. Several ECD policies have addressed all levels. Through this systematic study of all levels using the lens of governance, this study traced the especially significant role of local level of government in governance decisions that affect the integration of services.

National processes of decentralization provided the context for local involvement in ECD governance (including budgeting) in some of the sample countries. In Cambodia, decentralization occurred in the context of rebuilding local governance capacity, which was entirely destroyed during the Khmer Rouge regime and after. Commune councils, the chief mechanism of decentralization in governance, are responsible for the budgets of community preschools, which can represent a substantial proportion of local social service budgets. Decentralized decision making is reflected in the direct funding of Commune Councils by the Ministry of the Interior through apportionment of community-level (Sangkat) funds. Commune-level investment and commitment to ECD services are thus integrated into local budgeting priorities. While ECE is one of the four priorities for commune council spending guidelines being developed by the Ministry of Planning and Investment for the expenditure of UNDP social service funds (US$1,000 per commune), the ultimate decision on spending rests with the Commune Council and its representatives from health, education, and women's affairs, as well as a commune clerk charged with financial accounting functions. A somewhat similar process of local governance was observed in Lao PDR.

In Peru, as a result of decentralization, regional and local councils (which consist of a regional director, local mayors, and popularly elected representatives of civil society) generate annual local participatory budgets (Presupuesto Participativo Local (PPL)). With the creation of the incentive plan for municipal management improvement (Plan de Incentivos a la Mejora de la Gestión Municipal (PIM)), municipalities are able to access funds conditional upon attaining municipal goals that are set in a participatory manner at the beginning of each year. Local budgeting processes and incentives provide a context within which local investments in ECD services may be increased, although this does not always occur. For example, the incentive plan included as one of the prioritized goals for 2012 the reduction of malnutrition; however, municipalities tended to invest substantially more funds in infrastructure than in service provision. In some instances, ECD programs successful in achieving stated outcomes, such as the Proyecto de Atención Integral a Niños y Niñas Menores de Seis Años de la Sierra Rural in Peru ended when the Ministry of Education decentralized its funds for the project and they were diverted to building roads and municipal buildings.[41] Local spending was reported to focus more on local infrastructure than on social programming, including ECD services. Furthermore, differences were observed between two communities regarding the implementation of the PPL and PIM, with a higher level of implementation in less remote communities than in more remote ones.

Conclusion

This study represents one of the first attempts to map governance structures in ECD in LMICs. This research sought to understand ways that programs can be scaled up successfully through examining the roles and responsibilities of main actors engaged in the ECD enterprise. Several themes have emerged in this early analysis in the field.

First, the relative influence of actors in ECD varies substantially across countries. As nations increase their capacity to invest, it appears that NGO centrality may decline (compare Cambodia and Lao PDR in our examples to Kenya and Peru). Furthermore, successful national examples of policy agenda setting and national action planning, occur through coordination among government, NGO, and civil society sectors.[18] However, not all actors are equal in influence. The involvement of ministries of finance appears particularly critical to ECD prominence in national policy (as was the case of significant reduction in malnutrition in Peru across the last decade). This finding and those of another study imply that as national economies grow, the potential for the scale-up of ECD may rest mainly in government institutions leading partnerships with NGOs and civil society.[41]

Second, across all of the countries studied, the middle level of governance (e.g., district or provincial level) appears to suffer particularly in its overemphasis on vertical aspects of compliance, rather than horizontal aspects of coordination. This finding has clear implications for the scale-up of programs, in that the middle level may be the bottleneck that needs further exploration.[41] This study confirms earlier findings that building capacity at this level might be important for successful scale-up.

Third, innovative measures for improving and expanding local governance and control in budgeting and decision making affected ECD services in countries with recent initiatives in decentralization, which might offer potential lessons for scale-up of programs (e.g., Peru, Cambodia, Lao PDR) in yet other countries. Such innovations have the potential to create local buy-in regarding the importance of ECD when incentives to invest in service provision are in place, thereby growing the demand for services.

Finally, the study found that the integration of services across sectors was not commonly found in the four sample countries. It was primarily found in terms of multisectoral policy planning and coordination at the national level, where the most common approach was joint sponsorship of the national policy. The other form of integration was observed at the local level of governance, where decision making for a range of services rested with one decision-making body, for example, village counsels. The rhetoric of integration in the national policies did not appear to translate into integrated or even multisectoral governance functions at the middle levels of government in these countries.

Further research is required to explore additional areas of system strengthening such as leadership and finance. In the social sector, the finance ministry has an important and distinctive role with respect to sectoral allocation decisions, which in the case of ECD is further complicated given the multiple sectors engaged in providing services for young children and families. The findings suggest that combining approaches might be useful; however, the plans have to move beyond being articulated on paper to being implemented. For example Peru combines historical budgeting (e.g., some jurisdictions simply receive the same budget every year); needs-based budgeting (e.g., the budget is related to the varying needs of the different jurisdictions); and results-based budgeting (e.g., where budget lines are linked with specific population-based targets) as an approach to finance. However, the study also found that adjustments of budgets based on geography, staff incentives, and other contextual needs are not applied consistently. This suggests that more dynamic budgeting models are needed with an emphasis on implementation to achieve effective needs-based and results-based processes.

Within countries, due to limited time and resources, the researchers were unable to sample a wider range of geographic areas, provinces, and villages. Future work in this area would benefit from larger samples of subnational and local-level units. Such research would also benefit from including more middle-income countries, where not only governance and finance strategies differ but also priorities for ECD.

Finally with respect to integration, while clinical trials have demonstrated the positive effect of combining nutrition and stimulation interventions, that translation at a systems level needs further examination both with respect to the actors who should be included, for example, social protection, agriculture, and the range of services, for example, maternal mental health, community development.

Conflicts of interest

The authors declare no conflicts of interest.

  1. 1

    Uganda analyses and the process to select and conduct the study in the second Latin American country are underway.

  2. 2

    It should be noted that there was some variability in the application of the selection criteria of regions, districts, provinces, and villages because of variability in available data on child well-being, access to services, and ability to consult with local experts and the study in-country focal person. Another key consideration regarding the selection of regions was the feasibility and safety of travel within country to conduct interviews and observations.

  3. 3

    In the case of Peru, because of the emphasis on nutrition, this included prenatal age as well.

  4. 4

    Country reports provide the detailed analyses and results for each country study.

  5. 5

    Study results were examined for individual countries and also across countries. Contact authors for individual country results from Cambodia, Lao PDR, and Kenya.

Ancillary