Lessons learned from scaling up a community-based health program in the Upper East Region of northern Ghana

The original CHPS model deployed nurses to the community and engaged local leaders, reducing child mortality and fertility substantially. Key scaling-up lessons: (1) place nurses in home districts but not home villages, (2) adapt uniquely to each district, (3) mobilize local resources, (4) develop a shared project vision, and (5) conduct “exchanges” so that staff who are initiating operations can observe the model working in another setting, pilot the approach locally, and expand based on lessons learned.


INTRODUCTION
C ommunity-based health services programs are being launched and expanded throughout sub-Saharan Africa. Yet clinic-focused services remain the mainstay of most of these programs despite several convincing demonstrations that community-based operations can be more effective if static services are augmented with active provision of doorstep care. [1][2][3][4] Even the most successful of these small-scale projects often fail to be mainstreamed into large-scale operations, because experimental trials generally use many resources that are challenging to replicate in largescale operations. 5 Moreover, when doorstep service innovations are expanded beyond original target areas, a complex process of instituting system-wide adjustments to new supervisory structures, leadership dynamics, policies, resource allocation strategies, and plans are needed at each organizational level. 6 Such systemic changes are more complex to undertake than donors, researchers, and planners typically anticipate, 7 largely because leadership required to develop and test small-scale innovations sometimes contrasts with the type of managerial and political leadership required to change a large-scale system.
As early as the 1978 Alma Ata Conference, policies for achieving community-based primary health care became a pillar of Ghana's health policies. By the 1990s, however, mounting evidence that health development was not achieving national goals stimulated deliberations on feasible means of achieving health-sector reform. [8][9][10][11][12][13] Moreover, the specific means of improving program performance remained unclear. [14][15] Research identifying gaps in health outcomes called for national solutions, yet there was little concrete guidance to evidence-based policymaking and program development.
This paper describes the history of how an experimental study set the stage for a national program for promoting community-based primary health care-the Community-Based Health Planning and Services (CHPS) initiative. The paper also discusses factors of the initiative in the Upper East Region (UER) where CHPS was originally tested and scale up was most successful, despite being Ghana's poorest and most remote region. After a decade of implementation, the population of CHPS communities served by the CHPS program as a proportion of total district populations in the UER was 5 times the coverage achieved in the other 9 regions. To clarify factors that could explain the relative success of scale up, we consulted current and former Regional and District Directors of Health Services in the UER and compared their insights with those of leaders from an adjacent region where CHPS has been relatively slow to scale up.

WHAT IS CHPS?
CHPS is a national health policy to reorient primary health care services from subdistrict health centers to convenient community locations. Its goal is to transform the dynamics of rural health care service delivery from community health care providers who passively wait for patients into outreach workers who actively seek patients in communities and their homes, also known as doorstep services. The vision of CHPS is to accelerate progress toward Millennium Development Goals 4 and 5 (MDG 4 and 5) on child health and maternal health, respectively. The core strategy entails deploying trained and salaried nurses, known as community health officers (CHOs), to village locations where they provide basic preventive, curative, and promotional health services in homes or community clinics. The CHOs are supported by the health care program's community organizational activities, including the recruitment, training, and deployment of volunteer workers. Critically important to CHPS is the effective provision of family planning information and services, to include doorstep provision of oral, injectable, and barrier contraception, referral for IUDs and other long-acting methods, and promotional services that are targeted to the needs of men and organized mainly by male volunteers. [16][17]

PHASED PROGRAM DEVELOPMENT OF CHPS
Beginning in the early 1990s, Ghana instituted a partnership between applied health researchers and administrators to develop an action-oriented research agenda to guide health sector reform by resolving policy debates. 18 CHPS was informed significantly by national program development experience in Asia. 19 It was initially developed as a pilot project of the Navrongo Health Research Centre and progressed into a national policy over 4 overlapping phases ( Figure 1).

Phase 1: Navrongo Pilot to Develop Social Grounding for Service Strategies
In Navrongo, Kassena-Nankana District, UER, an 18-month pilot was initiated in 1994 by a team of social researchers, health scientists, and program implementers. The team consulted chiefs and elders, married women and men, and health care providers about appropriate strategies for implementing, managing, and sustaining community-engaged primary health care 14 in order to subordinate operation of the program to social institutions that shape reproductive preferences and health-seeking behavior. [20][21] Research scientists conducted qualitative research to determine the form of social interaction necessary for simplifying health communication and program mobilization processes and for facilitating volunteerism, consensus building, and ideational change. Qualitative research also identified Leadership to develop smallscale innovations is often different than the type of managerial and political leadership required to change systems at scale.
Lessons from scale up of a community-based health program in Ghana www.ghspjournal.org gender issues and possible strategies for addressing them, 22 including how to offset gate-keeping constraints in seeking health care, 23 how to engage the support and participation of men in reproductive health promotional activities that they might otherwise resist, 24 and how to sustain worker accountability for responsible service delivery. 17 Focus group discussions were conducted quarterly to understand health worker and community reactions and to revise and tailor program operations. Following 18 months of this participatory research and planning, an appropriate model was finalized. 25 Two sets of activities emerged from the Phase 1 model: 1. Existing clinical nurses were reoriented to community health care and assigned to village locations with the new designation of ''community health officers'' (CHOs). Nurses completed 18 months of training focused on basic curative health services, public health, immunization, and family planning.
2. Male volunteers from the communities were recruited and trained for 6 weeks to provide a limited set of services, such as oral rehydration and provision of condoms (Box 1). These zurugelu (''togetherness'') activities were based on existing traditional forms of governance, consensus building, and volunteerism, and they were designed to build male leadership and participation into reproductive health services, in addition to expanding women's participation in seeking reproductive and child health services. The project equipped volunteers with bicycles and startup kits of essential drugs, conducted training on service management, and set up revolving accounts so that the community financed and sustained the flow of supplies. 26 During Phase 1, the project clarified, documented, and translated operational details of mobilizing zurugelu and community health officer activities into practical and culturally appropriate implementation plans. 17  N Workers should be posted to their home district but not to their home village. Villages seek nurses who have an element of social distance and an ability to maintain family planning confidentiality. For this reason, the project constructed ''community health posts'' to provide a balance between integrating services into the local context and maintaining social distance between the service and social system.
for key lessons learned during the first phase.

Phase 2: Navrongo Experimental Trial to Test the Pilot Approach
Following the pilot, 37 communities of Kassena-Nankana District were grouped into experimental areas where zurugelu and CHO activities were deployed. Subdistricts were assigned to 1 of 4 experimental groups to evaluate the relative efficacy of the approaches ( Figure 2): 16 1. Zurugelu activities 2. CHO activities 3. Both zurugelu and CHO activities

No intervention
Relative efficacy was evaluated with data on child mortality and fertility trends collected by the Navrongo Demographic Surveillance System (NDSS). 16,27,28 Results indicated that comprehensive community-based health care provided by the CHOs with support from zurugelu volunteers was associated with statistically significant reductions in fertility rates and substantial improvements in child survival compared with other experimental groups. [29][30] In the first 3 years of the project, the total fertility rate declined by 1 birth in the combined CHO and zurugelu communities while it remained unchanged at nearly 5.5 in the comparison areas. In the communities where CHOs were deployedwhether with or without zurugelu volunteerschild mortality declined by one-half in only 3 years. Within 7 years, child mortality declined by two-thirds.
Communities in which CHOs were deployed but without support from community volunteers showed no changes in fertility rates. In addition, there was no fertility or mortality impact in the subdistricts where only zurugelu activities were implemented. In fact, volunteers posted without resident CHOs were less effective than providing no community care at all. 22,31 The non-intervention group was also ineffective, experiencing
Deploying nurses, with support from community volunteers, substantially reduced both fertility and child mortality rates. modest demographic changes unrelated to community-based care.
Because the CHO plus zurugelu approach produced improvement in both reproductive and child health, the combined approach was deemed to be the optimal strategy for national scale up. [29][30][32][33] See Box 3 for lessons learned during the second phase.

Phase 3. Replication Experiment in Other Districts to Validate Approaches
Although results of the Navrongo trial were impressive, district and regional health managers questioned relevance of the model to a national program. They debated whether replication of this model could be achieved in other parts of Ghana and whether it would improve demographic and health outcomes in areas outside the UER. Critics asserted that non-replicable institutional resources of the Navrongo Health Research Centre were responsible for the project's success, compromising relevance of the approach in other settings. Others argued that the Navrongo model was relevant only to the sociocultural circumstances of the UER.
To gain the necessary credibility for launching a national scale up of the model, the Navrongo experience was validated in another cultural and ecological zone of Ghana-the rural Nkwanta District of the Volta Regionwhereby only routinely available resources of the Ghana Health Service (GHS) were used. [14][15] To facilitate this process, the Volta Regional Health Administration sponsored a ''knowledge exchange,'' in which health workers from Nkwanta District visited the Navrongo project team in Kassena-Nankana to observe the Navrongo model of care and engage in candid discussions with counterparts about the feasibility of transferring the Navrongo model to Nkwanta. Although participants viewed the transfer of the model as requiring a daunting expansion of logistics capabilities, manpower, and supervisory workloads, direct dialogue with the Navrongo project team dispelled mystery about the CHPS development process and nurtured teamwork for developing pilot CHPS areas in Nkwanta. 34 To facilitate implementation, the Nkwanta management team adopted Navrongo's participatory planning process for adapting strategies to local and contextual circumstances as several distinct differences existed between the Nkwanta and Kassena-Nankana Districts, including their social organizational structures. Moreover, Navrongo had extensive research resources that the Nkwanta team lacked. Thus, the Nkwanta team developed training, logistics, and management information procedures that could be implemented at minimal cost. [34][35] The exchange program achieved operational success and contributed to increased health access and use. [35][36] In response, the GHS sponsored additional replication projects so that each region of Ghana would have learning localities where CHPS could be implemented using the Nkwanta scaling-up strategy. In each replication district, the community-based care model increased contraceptive prevalence, participation in antenatal and postnatal care, and childhood immunization coverage. 34,36 The success of the multidistrict replication process suggested that the Navrongo model could be implemented in similarly impoverished and health-deprived localities even with contrasting administrative, social, and cultural systems (Box 4). The exchange program also demonstrated that inter-district peer-exchange programs could be used to scale up CHPS. [40][41]14 Most importantly, monitoring evidence suggested that participatory team exchanges were a more successful strategy for fostering scale up N The combined CHO and zurugelu volunteers approach produced substantial improvements in both fertility and child mortality rates.
N Piecemeal approaches do not work: Volunteer-based outreach without a resident CHO had no impact on fertility or child mortality; CHO-based services without the support of volunteers had no family planning and fertility impact.
Exchange visits between newly implementing and advanced teams were more successful at scaling up the CHPS model than traditional workshops.
Lessons from scale up of a community-based health program in Ghana www.ghspjournal.org Global Health: Science and Practice 2013 | Volume 1 | Number 1 than workshops. Indeed, even today, most CHPS coverage is concentrated in districts that participated in exchanges with the Navrongo or Nkwanta teams. Much of the success with CHPS scale up in the UER is related to the fact that all workers in the region have had experience with the process of developing and implementing CHPS, even before they started the process in their home localities. 15

Phase 4. National Expansion of the CHPS Initiative
In 1999, the GHS reconvened district and regional health managers to assess the Nkwanta validation experience, and consensus emerged for promoting lessons from the Kassena-Nankana and Nkwanta Districts into a national program (Box 5). According to monitoring data from the GHS, 42 nearly all districts in Ghana had some degree of coverage of the CHPS program by 2008 ( Figure 3). Further observation and monitoring indicated that CHPS spread most rapidly in districts where pilots had been launched, suggesting that scale up followed patterns of change characteristic of diffusion processes. 15  N Redesigning training manuals to address technical inadequacy and improving documentation so that manuals are system implementation-focused rather than focused solely on health interventions was necessary.
N Decentralizing human resource development to the regional level ensured linguistic and cultural diversity were not constraints to scale up.

Box 5. Key Lessons from Phase 4 National Scale Up
Because core financing for CHPS is lacking, relying on other strategies helped to mobilize the necessary resources and accelerate scale up. These included: N Funding ''pilot CHPS'' zones within districts that give managers experience while politicians can observe operations and witness community enthusiasm for CHPS. Let the political gains of CHPS expansion become a political gain for local assemblymen and assemblywomen.
N Seeking local government and community support to fund start-up costs while health sector managers seek financing from the development sector for expansion costs N Prioritizing CHPS in regional staff meetings, budget discussions, and data review to catalyze organizational change. Follow through with regional technical staff visits to districts and CHPS zones, making CHPS expansion a key factor in performance reviews.

Methodology for Assessing Views of District Directors
To clarify stakeholders' perceptions of factors that either constrained or accelerated the scale up of CHPS in the UER, we interviewed all former and current District Directors of Health Services, 3 former and current Regional Directors of Health Services, and key staff and scientists of the Navrongo Community Health and Family Planning Project. For comparison, we interviewed the current Regional Director of the Northern Region and District Directors of Health Services in 4 northern regional health administrations. We also conducted a desk review of archival reports and documents of the GHS. We identified 2 sets of challenges to scaling up the CHPS program.
In general, issues related to manpower numbers, training, service capacity, and deployment summarized in Table 1 apply to the early CHPS implementation era. For example, CHOs in the early phases were unprepared to deliver essential health services, such as addressing maternal and neonatal complications. In addition, the initial training program for CHOs did not provide them with the necessary skills in how to engage with the community. Since 2009, a project of the GHS known as the Ghana Essential Health Intervention Programme (GEHIP) has focused on diagnosing barriers to CHPS expansion and developing interventions to address these problems. For example, the project trained CHOs in strategies for saving newborn lives and implemented a referral system to manage complicated cases. Table 2 summarizes constraints related to support systems for expanding CHPS, maintaining operations, and leading the program development process. For instance, stakeholders pointed to the complicated health management information system that overburdened  staff to the detriment of service delivery. Furthermore, lack of feedback on the collected data resulted in little use of the information to improve service delivery. In many districts, lack of leadership and political engagement, coupled with the absence of a budget line item for CHPS, resulted in inadequate resources and a lack of focus on and clarity about the program.

Strategies for Sustaining and Accelerating Scale Up in the UER
Interviews with the district and regional directors revealed 5 common strategies in the UER to cope with these challenges, which facilitated better scale up in the region compared with other regions.
Appropriate Manpower Recruitment, Training, and Deployment Strategies Due to the close proximity of the the Navrongo Centre to the Regional Health Administration in the UER, which deploys CHOs, the operational details of launching, sustaining, and scaling up a large cadre of community nurses, including developing appropriate strategies for selecting trainees, has always been well understood by regional health leaders. Before 2002, the UER lacked a community health nursing school, and most nurses came from outside the region. They often did not speak local languages, and it was challenging to find nurses for community relocation who had the cultural understanding that community work required. As one district director noted: It is all about commitment. In response, the UER opened a regional training school in Navrongo and pioneered a policy of recruiting nurse trainees from the districts to which they would be posted. The communities themselves nominate and financially support most nurse candidates to ensure

Evidence-Based Review and Modification of CHPS Strategies
Results of the Navrongo research project were shared with the District Health Management Teams (DHMTs) from the onset of the experiment, which provided district managers with the evidence needed to convince them to make the necessary changes in operations. Furthermore, a continuous review of research in management meetings and a general climate of openness to the role of research in guiding action prompted  29,44 The CHPS approach, as it was originally developed, lacked sufficient focus on ways to prevent maternal and neonatal death and disability. 45 To address this, UER regional and district health managers have invested in a program of in-service training designed to expand the range of services CHOs provide, with particular focus on improving and saving newborn lives, referral services, and other priority programs inadequately addressed by the initial Navrongo model.

Community Engagement and Grassroots Political Action to Mobilize Resources
The Navrongo project uncovered the need to develop health posts in which nurses could live and provide care to the community, which would establish a balance between community engagement and social distance. While the communities preferred nurses who were familiar with local languages and customs, they wanted the nurses to be socially removed from their own networks and extended families-someone the project termed as a ''trusted outsider,'' or a person who would keep secrets about family planning and avoid favoritism that might arise if the nurse was too closely linked to kindred groups.
To bypass GHS restrictions on spending resources for construction of new facilities, the UER Regional Health Administration encouraged communities to develop interim facilities with local labor, donated materials, and traditional construction methods. Then managers approached district assemblymen and development officers to replace functioning interim facilities with more permanent structures. A decade of systematic problem solving with community engagement has been the single N Female health volunteers are more committed to service activities than male volunteers, but male volunteers are critical to family planning promotion.

N Risk mitigation with field research:
Reliance on untested imported initiatives is risky.
N Partial IMCI does not work: Volunteer services can cause more harm than good unless volunteer deployment is coordinated with deployment of trained nurses and governed by rigorous supervision.

Shared and Consistent Vision and Commitment Among Leaders
Regional and district leaders in the UER have had a priority focus on and commitment to CHPS. From the onset, the UER Health Administration carried out strategies to mobilize resources, even cutting budgets of other activities and temporarily curtailing services to create budgetary space for CHPS. As UER district health managers noted when they were asked about scaling-up problems: Although launching CHPS is not expensive, its incremental costs are difficult to sustain. The estimated total expenditure for 1 fully functional CHPS zone covering a population of about 3,500 is US$33,345 (about US$9.50/capita) (including a solar panel but excluding CHO salaries, fuel for vehicles for monitoring activities, and training costs). The most costly components are facility construction ($20,240) and motorbike procurement ($5,300). However, if village volunteers conduct construction, using traditional material for building walls, the cost is reduced substantially (to about $3/capita). These modest costs are nonetheless a major challenge for district managers, who often lack adequate resources for implementing even the most basic health service agenda. The incremental costs of initiating the Navrongo model was the most dominant constraint to CHPS implementation identified CHPS expansion has been most successful in areas where GHS regional and district officials have persuaded local governments to add CHPS into their annual budgets. Simple-to-implement communication strategies have facilitated exchanges between local government authorities, politicians, and health leadership about CHPS. For example, GHS leaders have involved local officials in CHPS-sponsored community celebrations of implementation progress, linking popular support for CHPS to the political aspirations of elected officials.
But not all district health directors-the local focal points of health administration-consider CHPS as the most effective means to achieving health objectives, partly because most managers are unfamiliar with the Navrongo experiment but also because many are unconvinced that CHPS is cost-effective. To address these issues, the UER is testing a toolkit that will help managers compare different budgetary options and their implications on offsetting health and mortality risks. This tool has provided district managers with practical means of demonstrating the potential impact of investing in CHPS on health outcomes and may have contributed to accelerating the expansion of investment in CHPS between 2010 and 2012. 42 In addition to their commitment to CHPS, UER leaders also shared a coherent and consistent vision about CHPS, which contributed to the scaling-up process. According to GHS policy documents, CHPS is defined as ''the mobilization of community leadership, decisionmaking systems and resources in a defined catchment area (termed a 'zone'), the placement of reoriented frontline health staff, known as Community Health Officers, with logistics support and community volunteer systems to provide services according to the principles of primary health care (PHC-Plus).'' Regional and district leaders in the UER had a similar interpretation of this policy. For example, most UER directors considered CHPS to be functional even without a compound as long as CHOs provided services. In contrast, directors based elsewhere tended to disagree with this perspective. One such manager noted: The shared vision and commitment to the CHPS program among UER leaders facilitated the process of political and community CHPS expansion was most successful in areas where local politicians contributed funds to the program's budget.
Lessons from scale up of a community-based health program in Ghana www.ghspjournal.org Global Health: Science and Practice 2013 | Volume 1 | Number 1 engagement, which in turn facilitated resource mobilization. When health sector leadership catalyzes the process of community engagement and resource mobilization, CHPS scale up proceeds even if no health sector revenue exists for the program.

Peer Exchange Visits
Exchange visits between new and advanced districts helped to facilitate scale up by giving new CHPS program leaders an opportunity to directly observe how the model was working successfully. Dialogue between program leaders dispelled mystery about the CHPS development process and nurtured teamwork among the implementers. Peer exchanges were designed to achieve this by equipping visiting teams with the capability to implement the program in 1 or 2 zones.
To focus exchanges on implementation, visiting teams were encouraged to include participants who could represent the contrasting implementation responsibilities of each level of the system: the District Health Management Team, at least 1 subdistrict supervisory team, and 2 or more CHOs from the participating subdistrict. These individuals were then teamed with counterparts from an advanced implementation team to plan together how to implement a functioning CHPS zone in the participating team's home district.
The teams discussed the practical task of zone implementation, community engagement, replicating exchanges between communities, and building local political commitment to CHPS expansion. Each participating team also received seed funds to cover the cost of implementing a pilot zone, with training on how to cost and budget the program. Taken as a set of activities, exchanges equipped and financed pilot implementation of CHPS in ways that catalyzed scale up of operations once participants returned to their home districts.
Participants were also oriented to focus group methods that would enable managers to respond to community and worker opinion, adapt strategies to local conditions, decentralize planning, and take ownership of the program as a districtdirected, scaling-up initiative.
Although 38 districts throughout Ghana participated in exchanges, this process was far more intense in the UER than elsewhere. By 2004, all districts, subdistricts, and CHOs in the region had participated in 1 or more exchanges with the Navrongo team. National monitoring data have since revealed that peer-exchange visits have been more successful than workshops. In fact, by 2008 CHPS coverage was concentrated in the 38 districts that participated in exchange visits with the Navrongo or Nkwanta advanced CHPS districts. 15

CONCLUSION
Despite the challenges that have been identified, the CHPS initiative has begun to introduce health care reform in every region and district of Ghana. In the UER, where scale-up problems have been the focus of strategic review, trial, and dissemination, CHPS has established a sustainable service model, which ensures that progress is adapted to local realities and guided by evidence. Coverage of community-based care in the UER was 5 times the national average after a decade of scaling up, even though the region is the poorest in Ghana.
Nonetheless, problems with district health system leadership persist and merit further research, demonstration, and policy development. The new Ghana Essential Health Intervention Programme (GEHIP) has been launched to address the evidence gap on how to improve district leadership for continued CHPS expansion, 39 including the development of district planning tools, resource management systems, and intersectoral coordination strategies. 46 In addition, the new project is testing a simplified management information system to meet the information needs of CHOs and district leaders while minimizing the time and effort required for CHOs to capture, manage, and report information. 47 Although much remains to be accomplished, the CHPS experience in the UER attests to the practicality of scaling up CHPS throughout Ghana. Its core strategy is based on the principle of putting evidence into action through research, experimentation, multiple validations, and adaptation efforts-learning by doing.