PLOS

Introduction

In the past decade, the concept of planetary health has garnered global attention as the effects of climate change present grand challenges to human health and wellbeing worldwide. Planetary health is “the health of human civilization and the state of natural resources on which it depends” [1]. Describing the breadth of issues associated with the planetary health perspective, Whitmee et al. (2015) emphasize the need to closely consider “…the human systems—political, economic, and social—that shape the future of humanity” [2]. By highlighting impacts on our natural systems and humanity, the authors bring to the foreground the expansiveness of the planetary health concept. Both strategically and empirically, the planetary health approach aims to identify, understand, and quantify the human health impacts of global environmental change disruptions in order to develop solutions that will allow humanity and the natural systems we depend upon to thrive now and in the future [1]. To do so, planetary health frameworks urge collaborations between communities, governments, businesses, non-governmental organizations (NGOs), non-profits, and international agencies to improve management systems, so as to optimize health outcomes [2, 3].

Agroforestry, defined as the intentional integration of trees with crops and animal systems [4], is by definition a planetary health solution. Agroforestry addresses human and environmental health concerns as it rebuilds fertile soils [5–7], minimizes climate risk [8], diversifies farmer incomes [9–11], and often provides a source of food [8, 12], raw materials such as wood [13, 14], and other vital ecosystem benefits like water and temperature regulation [6, 15]. Agroforestry’s benefits in dimensions of human health and wellbeing and ecosystem health have been established in previous studies [10, 13, 15–18], underscoring the effectiveness of agroforestry as a planetary health solution [5, 18] when it is implemented where conditions (social, political, environmental ecological, etc.) are favorable. Human health and well being outcomes of agroforestry include livelihood resilience due to income diversification, food and nutrition security, and mental health [10, 13, 15–18]; as such, the outcomes contribute to the United Nations Sustainable Development Goals (SDGs) [19]. Furthermore, agroforestry’s potential to contribute to climate adaptation and mitigation has been demonstrated, for example, by sequestering carbon as trees grow [20–23]. Agroforestry can balance out the urgent economic needs of smallholder farmers while also addressing climate adaptation and mitigation [24]. For instance, trees can protect households from hot climate [10] and improve microclimates on farmland [18, 25].

Scaling up agroforestry programs for planetary health benefits

Although agroforestry has been identified as an effective adaptation and mitigation measure, the Intergovernmental Panel on Climate Change (IPCC) 2019 summary for policy-makers reports that gaps remain between current levels of agroforestry adoption and levels needed to reduce climate risks and impacts [26]. In short, despite successful programs, agroforestry as a planetary health solution has not yet been scaled up to achieve its full potential.

Scaling up is the process of spreading an intervention, practice, and its impact across geographical areas and to reach more people through a variety of mechanisms [24, 27–31]. Scaling from the farm level the to the landscape/community level provides an opportunity to take advantage of institutional dynamics that can foster adoption [32]. The presence of institutions across landscapes engaged in similar activities can be leveraged to support the scaling process and enhance the impact of agroforestry [33]. An illustration of this would be a company that manufactures low-cost composting toilets working in partnership with a local community-based organization. The partnership would both promote regenerative agriculture and enable the production sales of high-quality compost to the company [34]. Such a direct market linkage could incentivize adoption and sustainability of regenerative agricultural practices in degraded landscapes.

Current research advocates for the strategic scaling of agroforestry by way of Integrated Landscape Management (ILM) [32] in order to address the barriers to its adoption and spread, and secure its agricultural and environmental benefits to as many stakeholders [35]. Integrated landscapes, characterized as socio-ecological systems where nature conservation and livelihood goals are simultaneously pursued, reveal intersections and interactions across the agricultural and natural resources on which communities depend, as well as the sectors and stakeholders involved in their management [36–38].

Scaling up agroforestry via social networks

Situated within vast landscapes (e.g., geographic communities) are various social networks (e.g., villages, communities, affiliations) at different scales (e.g., communal, regional, provincial) that can help promote and spread natural resource management practices. Social network ties are the specific ways that people are related or connected (i.e., links), offering a mechanism for the spread of something tangible (e.g., training resources, tools) or intangible (e.g., behaviors, knowledge) [39, 40]. Social network ties play a critical role in translating small-scale interactions into large-scale patterns, which in turn influence local level dynamics [41]. Therefore, social network ties are essential in promoting the spread of new practices [42–44]. Prior studies recognize the vital role of social networks in promoting agroforestry [45, 46]. Additionally, Isaac & Matous (2017) [47] investigated agroforestry knowledge transfer through farmer advice networks in Ghana, and Gamboa et al. (2010) [48] studied the effects of social networks on the adoption of specific agroforestry tree species in Southern Ecuador. A question that remains in the literature is the extent to which institutions and individual actor networks are scaling and sustaining community-based agroforestry practices beyond their inception.

Current research suggests that while many development projects have substantially improved the health, well-being, and livelihoods of people, they have been restricted to small geographical areas where only a small number of people are likely to benefit from an intervention [49]. Therefore, studying the role of social networks in spreading and sustaining development interventions, particularly in low-income contexts, provides an opportunity to better understand how the structure of networks can function in natural resource situations, as recommended by Bodin & Crona (2009) [50]. Knowledge of how information and resources flow between farmers can help inform efforts to strategically scale up agroforestry throughout landscapes [51]. With global interest in scaling up natural climate solutions, including agroforestry [9, 52, 53], there is a critical need to understand the mechanisms that facilitate the adoption and spread of such solutions. To explore this, we sought to focus on a fragile state where planetary health impacts could bring extensive benefits to community wellbeing.

Study area: The fragility of rural Haiti

Haiti was the first sovereign black nation in the Americas to abolish slavery in 1804 [54]. Like other colonized nations it was long exploited for its natural resources and cleared of primary forests to support commercial agricultural exports [55–59]. Coupled with two centuries (1800s-2000s) of unjust reparations to the French colonizers [60], foreign invasion, natural disaster, social political conflict, and urban growth, the country experienced gradual deforestation and now is considered one of the most environmentally degraded countries in Latin America and the Caribbean [61]. Assessed on multiple dimensions (economic, environmental, political, security, societal, and human), Haiti has been characterized as one of the most extremely fragile contexts in the world; environmental, along with human and societal, all contribute significantly to the country’s overall fragility [62, 63].

In Haiti, about half of the country’s population (42%) lives in rural areas [64] which comprises smallholder farmer households, dependent on their land for both food and income. Deforestation, intensive agriculture, and climate change, therefore, create a demand for agricultural practices that are essential to the livelihood of farmers. Based on the 2021 Global Climate Risk Index, Haiti is also one of the most vulnerable countries to climate change [65], thus establishing the case for the need to build smallholder farmer resilience through regenerative practices that simultaneously provide ecosystem services and climate adaptation and resilience.

Agroforestry, when properly implemented and managed, can have a meaningful impact in places like Haiti [13] where, in addition to severe land degradation, the country is characterized by one of the highest rates of food insecurity in the world coupled with the most adverse health outcomes in the Americas [66]. Despite the planetary health benefits of agroforestry and its potential to address issues in Haiti’s rural landscape (e.g., slope stabilization, reduced soil erosion, natural disaster mitigation), the country is also amongst the most understudied regions, creating a significant gap in the literature. Most of our knowledge of agroforestry’s adoption drivers and effectiveness in Haiti is primarily limited to research dating back several decades (see for example, [59, 67–73]). Related to social networks, research studying networks within Haitian society has focused on populations outside of Haiti [74] and virtual connections [75], which although useful, do not contribute to the agroforestry and climate literature pertaining to Haiti.

The Haiti timber reintroduction program

Established in 2005, the Haiti Timber Reintroduction Program (HTRIP) is a community-based agroforestry program developed by a United States-based non-profit organization, Haiti Friends. Haiti Friends, in collaboration with a long-standing regional hospital, Hopital Albert Schweitzer (HAS) operates in the Lower Artibonite Valley [72]. Established in 1956, Hopital Albert Schweitzer regards itself as a provider of high quality clinical care for vulnerable patients in the region [76]. Ensuring the implementation of HAS services are dedicated office spaces, a fleet of private cars, a private fueling station, a water reservoir, and a solar power energy source. Haiti Friends is a Pittsburgh, USA-based non-profit that began its work in Haiti in 1985 by individuals with ties to HAS. Through this non-profit, the Haiti Timber Reintroduction Program (HTRIP) was founded in collaboration with HAS in 2005 as a means to strategically _target pressing issues of human health and deforestation in the vulnerable communities. HTRIP’s resources included agricultural inputs (e.g., seeds, tools), trained staff and experts in agriculture and agroforestry, and several sources of funding from dedicated donors to support their activities.

The HTRIP represents a proactive approach to community health by addressing the underlying issues of poverty and environmental degradation in the high elevation mountain communities [71]. Recognizing the health disparities (e.g., high rates of malnutrition and other diseases associated with extreme poverty) in the often remote upland mountain communities, the hospital requested HTRIP’s reforestation efforts, ultimately tackling some of the downstream human health issues caused by deforestation [72]. HTRIP seeks out communities where land is steep, and soil is eroded, and where residents show interest in participation [71]. The program’s approach is centered around a joint-venture effort between the local smallholder farmers who want to plant trees on their land and HTRIP.

HTRIP provides community-based agroforestry inputs and technical expertise [77]. The specific agroforestry practices implemented by HTRIP are agri-silviculture (i.e., forest farming) [78] which combines planting timber (e.g., Spanish cedar, Honduras mahogany) and fuelwood (e.g., Bloodwood, Siris) species, as well as fruit trees (e.g., mango, lime, orange, coconut, papaya soursop); all of which supplement farm crops (e.g., beans, corn, sorghum) on the plots. The species selected for each community are based on the needs and preferences of the residents and site suitability. Each year, new cohorts of 20–30 community members receive training, during which several agroforestry activities (e.g., nursery construction, seed collection and preparation, composting, and sapling care) are carried out by that year’s cohort and are overseen by the community leader who continues to lead each cohort yearly. At the end of the nine-month training the tree planting season begins. The trees obtained from participation in the program, once planted, become the property of HTRIP participants, and they decide what to do and how to use them when the trees reach maturity. This underscores HTRIP’s tree ownership concept.

Research objectives and questions

Consequently, to explore the mechanisms that facilitate the adoption and spread of agroforestry as a planetary health solution, the primary objective of our study was to investigate how institutional collaboration and social networks facilitated the adoption, scaling up, and sustainability of community-based agroforestry in the fragile state context of Haiti. The principal research question we examined was: how do institutional collaborations and social networks drive and sustain community-based agroforestry adoption and scaling? Through empirical investigation, the results of this question contribute to a larger objective designed to better understand the phenomena associated with adoption and scaling in a resource-limited context. For this study, we focused only on the HTRIP community-based agroforestry intervention. Additionally, we offer a theoretical framework for the scaling of innovations that researchers can utilize as they design and implement research for scaling planetary health solutions.

Materials and methods

To assess the spatial and temporal features of adoption and scaling of community-based agroforestry processes among communities in the study area, we used an exploratory sequential mixed methods approach [79]. Findings from qualitative data collection (e.g., focus groups, individual semi-structured interviews), which asked participants to recall their various network connections to residents of other HTRIP communities, were used to inform content of the social network analysis methods (network modeling and visualization). Subsequently, findings obtained from the social network analysis were compared to initial qualitative findings. Building on previous research that evaluated the HTRIP framework [72] and landscape factors associated with the program’s reforestation success [73], we conducted initial field visits between 2018 and 2019 that served as pilot trips to the study area to gain a better sense of the tree culture, tree cover, as well as HTRIP’s history operating in the study area. The field visits informed the development of interview instruments, which focused on participants’ motivations for practicing agroforestry with their community, as well as how they came to be involved in the HTRIP. A detailed account of the qualitative instruments can be found in supporting information (S1–S4 Texts).

Data collection

Following approval from research collaborators at HAS and HTRIP, participant recruitment and data collection took place between June and November 2019. Field observations took place in June and included a tour of the HAS campus and HTRIP offices, as well as a tour of various HTRIP communities. This observation data provided an understanding of the hospital’s operations and logistics, including how HTRIP was supported by them. To investigate adoption and social networks, 11 of the 66 HTRIP communities were purposively sampled [80] to represent variations associated with years of agroforestry adoption and community participation (e.g., less than five years, five to 10 years, more than 10 years) in the HTRIP program, and a total of 83 persons were interviewed (Table 1). The rationale for the selection of the 11 communities was based on capturing a wide range of years of involvement in the HTRIP network and providing insights into the temporal differences of network ties and trends in adoption mechanisms over time. To further strengthen our purposive selection of communities, HTRIP staff provided historical accounts of the communities and their inception into the program (names of 65 communities and total number of participants within each community). Across the 11 communities, at the time of data collection, we intentionally sampled five short-term adopters (less than five years of participation), three medium-term adopters (five to 10 years participation), and three long-term adopters (more than 10 years of participation). To understand the barriers to scaling, the short-term adopters also included communities that had practiced for a short period of time before concluding their active participation in HTRIP. Data points for the communities were obtained via data files from HTRIP staff who collected the points using handheld GPS devices. The community GPS points collected between 2005 and 2012 were provided in the form of KML files and were uploaded to the QGIS software. Polygon data for the study area, which served as the basemap for mapping the data, was obtained from a secondary dataset [81].

Expand

To participate in this study, individuals had to be residents of the community, had to have participated in HTRIP, and had to have been an adopter of community-based agroforestry. Following Cornell University Institutional Review Board (IRB) for Human Participant Research under protocol #1008001625, participants were made aware of the study’s purpose and that participation was voluntary. Verbal consent was obtained, and interviews audio recorded. The field team conducted in-person focus group interviews (55–70 min in length) and in-person individual semi-structured interviews (15–40 minutes in length). One remote individual semi-structured interview totaling 40 minutes with an HTRIP participant in Labadi was conducted in November of 2019 when he was later identified as a key informant. Participants were provided a monetary incentive for their time and involvement in the study during the workday. The focus group questions _targeted community-level observations, whereas the individual interview questions asked about people’s adoption of agroforestry, their social network ties in nearby communities, and personal observations about how the practice benefited their community (See S1–S4 Texts).

Fifty-eight individual semi-structured interviews took place throughout various spaces within communities such as home gardens and fields; HTRIP staff were interviewed at the HTRIP office. In addition, there were four focus group interviews with a total of 25 participants (with group sizes of six to seven). Three of the focus groups were held at community centers, and another one took place in a community forest. All interviews were administered verbally in Haitian Creole (Kreyòl) and responses were noted via the interviewers on a response sheet and audio recorded. Audio recordings of the interviews and focus groups were translated and transcribed to English in Microsoft Word for analysis by the first author who is a native Haitian Creole (Kreyòl) speaker. A total of 78 community members and five staff members were interviewed across the study area (Fig 1). Interviews with HTRIP staff, selection of which was based on their active role as program managers or field educators, complemented the data obtained from community residents and provided practitioner perspectives on the community adoption and scaling processes.

Expand

Analysis

Thematic data analysis was conducted using Dedoose qualitative data analysis software. An a priori coding framework (S5 Text) encompassing dimensions of community capitals (e.g., financial, human, social, natural, political) and agroforestry adoption (e.g., motivations, barriers, social network ties) was developed using themes from the literature [13, 67, 69, 82–85]. Coders utilized the a priori coding framework for analysis to ensure consistency in the coding process relative to the theoretical framework for the research. For community capitals the a priori parent codes derived from the literature were built capital, cultural capital, financial capital, human capital, natural capital, and social capital. For agroforestry adoption the a priori parent codes derived from the literature were motivation to adopt, factors affected ability to continue practicing agroforestry, and barriers to agroforestry adoption. To understand how institutional collaborations and social networks influenced adoption and scaling outcomes, two rounds of coding were conducted to identify drivers of adoption. The first round of coding focused on themes developed in the apriori coding framework [86] and was carried out by the lead author. Following the first round, the second author reviewed the coding application for confusion and discrepancies. We then reviewed coding problems and pinpointed areas of discrepancy. To enhance the reliability of the coding process, we omitted or combined unreliable codes to streamline the coding framework and clarify coding definitions. Moreover, to promote transparency and minimize bias, we ensured that at least one coder on the research team was external to the data collection process. This allowed them to approach the review of coding with a fresh perspective. In the second round of coding, emergent codes derived from round one that were not in the apriori coding framework, based on new dominant themes and patterns were applied [86]. Emergent codes are indicated in red font in S5 Text. Again, following the second round of coding, the second author reviewed the code applications for discrepancies. Additionally, both coders possessed expertise and prior experience in coding qualitative data, ensuring that the coding and theme development process were conducted rigorously [87]. The full coding framework with parent codes, sub-codes, and code descriptions are provided in S5 Text.

To visualize the social network ties between peers and family members within HTRIP communities and analyze their connectedness, each participant was asked to recall the names of the other HTRIP communities in which they had ties to family and peers. All connections obtained from the individual semi-structured interviews were recorded and summarized in Excel. Social Network Visualizer [88], an open-source social network analysis and visualization tool, was then used to analyze and visualize the network ties at the community level (i.e., an aggregation of individual network ties). A measure of centrality was conducted to quantitatively assess the influence of the sampled communities on the existing HTRIP network. Degree centrality is defined as the number of ties members of a community had within the HTRIP network. This metric is considered a measure of actor activity. In this case, the more connections a community had with other communities, the more active it was in the network [89, 90]. A total of 66 GPS points for HTRIP communities were collected to help map the network ties. To calculate relative distance, a map of HTRIP communities was created using QGIS (free and open-source) [91] as the spatial analysis platform.

Results

This research investigated how institutional collaboration and social networks drive the adoption, scaling up, and sustainability of community-based agroforestry in the Lower Artibonite Valley region of Haiti. We present results for our research question, how do institutional collaborations and social networks drive and sustain community-based agroforestry adoption and scaling?

The institutional collaboration

Field observations and interviews demonstrated how Hopital Albert Schweitzer provided multiple logistical advantages. Respondents noted that through the HAS-HTRIP collaboration, resources were pooled to address the health priorities of both institutions that converged through community-based agroforestry (Fig 2). Field observations found HAS provided fee-based access to the safe and secure office space that housed all HTRIP staff, materials and amenities, including meeting spaces, storage, internet, and other logistical necessities. This office space was located in the Integrated Community Health Department on the HAS campus. To address human nutrition, HAS provided a nutritional education program for the beneficiaries in order to enhance their knowledge of what foods can contribute to their health. As a complement, HTRIP managed an agroforestry garden on the hospital’s premises for educational and nutrition purposes, while also ensuring that their program’s agroforestry system considered the nutritional recommendations of the clinical team. Additionally, during moments of disease breakout, HAS provided assistance to HTRIP communities. This was illustrated by the cholera outbreak of 2010 as a result of mismanaged sewage at a United Nations peacekeepers camp [92, 93] which prompted HAS to support water and sanitation efforts in surrounding communities, including HTRIP communities. This vital contribution reduced communities’ risk of infection, allowing them to continue engaging in agroforestry activities.

Expand

Field observations also revealed how HTRIP operations were facilitated by the hospital’s long-standing infrastructure, which was made available for a reasonable overhead fee. HTRIP paid a flat 20% overhead fee to cover the administrative and financial services provided by HAS. Other costs associated with HTRIP activities were paid by HTRIP. Within this contract, HTRIP was supported by HAS’s departments that provided car maintenance and fueling, accounting services, administrative services, and a library to conduct research and hold large meetings. Lastly, due to the centralized nature of Haiti’s activities in the nation’s capital, the availability of office space in Port-au-Prince to conduct business (e.g., shipments, transfer and pickup points, document archive storage) was critical. HAS maintained an office space in the capital that was also at HTRIP’s disposal. Due to the convenient nature of these services, rather than establishing new administrative units, this collaboration brought HTRIP’s operating costs down, therefore allowing HTRIP to allocate more funds directly to its agroforestry activities.

Social networks, recruitment, and adoption

Social networks played a critical role in the recruitment and establishment of adopting communities. We found evidence that throughout phases of recruitment and implementation, both institutional and personal social network connections played a role in community-based agroforestry adoption and scaling throughout the various stages of the HTRIP. Relative to recruitment, respondents noted that in the first year of HTRIP’s implementation, HAS’s presence in eight communities helped establish HTRIP in those same communities. As a community leader in Labadi said,

“Even before HTRIP came, we were in the community development program that HAS had [here]. When that program ended, HTRIP came and told us about a new program that would teach us a lot of new things like soil conservation. They came to find us.”

This successful recruitment strategy was verified by the staff members when they described how the hospital had already established medical clinics in the communities where HTRIP could have an impact. Wherever the hospital was present became an opportunity to leverage existing trust between the hospital and communities in order to establish HTRIP. Of the 11 sampled communities, three were recruited through the HAS network, three through HTRIP staff, and five from peer referrals. Fig 3 illustrates the recruitment methods (via HAS, HTRIP, and peer referral) between the years 2005–2018 obtained from the interviews and secondary data, and the relative distance between the communities.

Expand

In addition to the institutional network recruitment methods, an investigation of the mechanisms at the individual level within communities demonstrated the influence of localized networks. Respondents noted how they were motivated to join their community’s HTRIP cohort. Community leader influence was the most prominent recruitment method at both the individual and community-level.

“The leader is the one that got me involved. Whatever he says is good I will get involved without a doubt. I stand with him.” ‐ resident of Kajil

“The leader really wanted us to be involved so we put our heads together.”

- resident of Koray

The community leaders represented the communities and maintained connections with all residents. It was through these leaders that those external to the community would gain access to all other residents. Despite the fact that some communities were divided by mountains that formed barriers to travel, it was common for community leaders to know residents in other isolated communities. Thus, agroforestry networks were a way to connect otherwise disconnected communities. Other interviewees described the ways that HTRIP staff and other members within their communities encouraged their involvement:

“So, the agronomist from HTRIP, a while back came across our leader and explained to him the program and then the leader informed us and asked us to come to the initial meeting to get more information.”–resident of Kaywit

“People here attract each other into the tree project here and other projects too. They even ask to come join.”–resident of Kaywit

Post-Pilot scaling

Through field observations and communication with program staff, the HTRIP pilot communities, established between 2005 and 2007, were exemplars of the long-term beneficial impact of the program. Respondents noted that these initial communities became the most influential in the expansion of HTRIP into other communities. Several referrals were observed to have stemmed from those long-standing communities. Some communities were referred by adjacent communities as well as communities farther away, which meant that people were not only getting others involved in their own communities, but also sharing information when traveling outside of their communities (Fig 3).

The HTRIP staff reported that most of the HTRIP communities had entered the program by way of peer-to-peer referral, thus demonstrating the critical role that peer networks play in the spread and scaling of an intervention. When community members were asked about their network connections to other HTRIP communities, 93 network ties were identified between 36 of the communities and most of the connections were to long-standing HTRIP communities (Dekouvet, Fonwo, Larok, Kaywit) with ten years or more experience in the program. Fig 4 illustrates the connectedness between the interviewed HTRIP community members.

Expand

Providing further support to the important role of peer networks, respondents also noted how program spread occurred naturally without active recruitment once people were either aware of the program’s existence or its benefits as illustrated in the following excerpts:

“They didn’t enter right away, but when they started seeing the little benefits that people were getting from being involved, it is then they started asking for themselves and when they could come join.” ‐ leader from Dekouvet

“Now I have trees, and one year I decide to cut them, and I make about four to five thousand gourdes. But this guy here, he doesn’t have, I can sit with him as usual since we grew up together. I tell him, there is HTRIP here, come in the program, use your time coming to the lessons, and later on you will have trees to plant on your land. And now you have trees that you can use in the future. Now, another person comes, we are each talking with one another, and we recruit more and more people into the program that way.” ‐ resident of Kajil

Social networks and the sustainability of community-based agroforestry practices

Beyond adoption, to secure this scalable form of community-based agroforestry, social networks also played a key role. At the institutional level, staff members noted that the HAS team would carry out HTRIP tasks or deliver messages to participants when the HTRIP staff could not make it to the field site. For example, there were long-standing communities that were no longer formally involved in the HTRIP program because all members of the community had been trained and became self-sufficient, therefore no longer needing the formal guidance of HTRIP. In order to maintain community-based agroforestry, these same communities, however, remained in touch with the HTRIP and HAS staff members. One HTRIP staff member described how they would often do check-ins with the community to assess if any assistance was needed with the practice:

“Even though we do not go there the same way we used to, we make sure to check on them and stay in touch with the leaders to see how things are being continued. There are times when they might see an HAS car pass by their area, they can let that staff know about something they need us to help them with. The message gets back to us. And if they ever need us urgently, they can always come to the hospital and find us. This is how it works.”

Social networks, within and between communities were said to be strengthened by participation in the HTRIP. Respondents noted that the social networks of community leaders and residents were multiplied as a result of the farmer exchange visits and cross community events. Two community leaders stated:

“As a leader, HTRIP multiplied my contacts. We would do community visits in which we would do community exchange and have conversations. Long ago, I didn’t know Anje or Koupwa. I only came to know these areas through HTRIP. I used to hear of these places but didn’t actually know them. We share our ideas and maintain contacts now.” ‐ leader from Larok

“The program developed a system of friends. Each person wanted to see each other after [the program].”- leader from Ottovin

Participation in HTRIP included a grand graduation ceremony, which further bolstered the social network ties of the region’s residents. Serving as one of the largest annual events in the Artibonite Valley, respondents noted that the graduation celebration facilitated interactions amongst the community members across the region. HTRIP newsletters described this as a grand cultural event where community members put together performances describing their HTRIP experience and illustrating the importance of trees in their respective communities. A 2011 newsletter described how 846 participants who were eligible for graduation along with about 1000 of their family and friends attended that year’s ceremony. One community leader stated that, “some people only joined HTRIP to get the chance to participate in the graduation” and described how the graduation was a significant driver of continued participation because it motivated people to want to be recognized for an accomplishment and connect with other people in the region.

The findings also identified cases of community-based agroforestry that were not sustained in HTRIP communities for various reasons. For example, members of Kapuis, a low-elevation community near a main road and the Artibonite river, chose to pursue small commerce activities (e.g., selling clothes, spices, packaged water) instead of prioritizing agroforestry. Similarly, in Labadi, another low-elevation location, community members wanted to maximize their land productivity and prioritized their cash crops, such as corn and rice, that were easily sold in nearby markets. The third community, Ottovin, experienced conflict with their leadership that caused internal issues. As a result of these conditions, community members’ attendance during training sessions and community-based agroforestry activities was reduced over time and ultimately led to the decline of their formal involvement in HTRIP. Nonetheless, HTRIP remained in touch with these communities in the event that their priorities or status changed. At the time of data collection, Ottovin had acquired a new leader and was planning to formally re-enter HTRIP.

Using data collected from qualitative interviews, an analysis of social network degree centrality which measured the number of ties members of a community had within the HTRIP network, provided further support for the positive role that HTRIP played on community ties over time. Respondents noted how their pre-existing ties were strengthened, while they also formed new ties with other members of the region by way of HTRIP. When asked about how many connections participants had in other communities, several connections were identified. Table 2 summarizes the 93 network ties and the percent degree centrality of the eight HTRIP communities that sustained their participation. The results show that long-term and medium-term communities represented 90% of the connections identified in the interviews, with long-term communities having slightly more connections (47.9%) than medium-term communities (41.5%). These same communities were also the most active in the sample’s network (89.1% degree centrality). Short-term communities had the lowest number of ties (10) and as a result, less influence in the network (10.6%), suggesting that longer participation in HTRIP had strengthened the networks of participants after some time.

Expand

Discussion

The results of our study demonstrate how a partnership between a community-based agroforestry program and a trusted institution such as a healthcare provider led to and increased the willingness of community members to be involved with the program, and ultimately adopting the practice. Operating as the only hospital within 610 square miles of the region, much of HAS’ services are in remote or mountainous communities that are sometimes only accessible when traveling by foot or animals such as donkeys and horses. The partnership between the Haiti Friends and Hopital Albert Schweitzer embodied a coordinated effort to create an enabling environment for rural upland communities. While the former provided trees with several benefits through a participatory learning and doing process, the latter provided clinical care and, in some instances, water and sanitation services, and nutritional education. Research in Dumont et. al (2019) [94] that documented how structural stakeholder engagement leads to more diverse and inclusive agroforestry systems in eastern Democratic Republic of Congo identified similar interventions to improve the enabling environment required to scale up agroforestry. The interventions included providing access to agricultural inputs, credits, information sharing, and road improvement, which provided support for these activities facilitating greater adoption of agroforestry practices in a fragile context.

When one institution could not make it to the field, the other took its place and handled the task (e.g., delivering equipment, delivering a message, transporting participants). This health and environment partnership model is in line with the recommendation of planetary health researchers who suggest promoting agroforestry through health centers as we move forward [18]. The results lend further support to the critical role that cross-sectoral networks among institutions that perform similar and different functions can play in achieving impact at broader scales [49]. This cross-sectoral approach also supports the important role of strategic partnerships in promoting the scaling up of agroforestry as demonstrated in the agroforestry literature [24, 32, 94].

When rural Haitian people have access to something that benefits their livelihood, they share it with others, highlighting the role of social capital and social network ties, and how together, they help spread beneficial innovations such as the Haiti Timber Reintroduction Program. In contrast to earlier research, which found negative effects on the adoption and scaling process due to remoteness [11], presence of mountains [95], distance from main roads and market centers [96], we did not find these factors to substantially impede HTRIP’s scaling in its _target region. This research supports prior research in Ghana (another fragile context) that found external institutional ties provided valuable information to facilitate the adoption of agroforestry, suggesting that external bridging social network ties can increase livelihood resilience [47]. Even in a limited-resourced context composed of isolated mountain communities, HTRIP was successfully able to establish itself in over 60 communities. Also in this context, a robust network between stakeholders facilitated the flow of information, resources, practices, and identities, and generated wider social acceptance of HTRIP; something researchers have suggested programs aim to achieve in rural areas [97].

The results of a social network analysis supported by qualitative accounts of how communities and their members joined HTRIP demonstrated that, in addition to institutional networks, peer-to-peer networks had the ability to facilitate the expansion and scaling of HTRIP through two mechanisms. The first mechanism was internal to a community, wherein the community leaders played an essential role in getting participants to commit to the program and attend class training sessions. This is consistent with previous studies describing community leaders as gatekeepers of information about programs [98], as well as transmitters of ideas and practices from influential or successful individuals in a society [99]. Our results on the important role of community leaders are in line with research that reviewed literature on other fragile contexts; in these studies authors found that central actors were public in nature and seen as providers of information and resources who helped strengthen trust between actors in Kenya and Ghana [100]. The second mechanism of scaling through social networks occurred between communities when the observable benefits of community-based agroforestry would motivate others to seek out the practice for their own communities. HTRIP participants would share the program’s information with those near and far (mountains away). The existence of long-standing HTRIP communities and experimental plots that could demonstrate the potential benefits of the practice, while simultaneously providing spaces for co-learning, played a critical role in influencing nearby and further communities to join HTRIP. This influential role of the long-standing HTRIP communities resembles what authors have described as demonstration effects of early adopters [101].

The social network analysis also demonstrated the positive role that community-based approaches to natural resource management could play in strengthening rural community ties. Due to their longstanding participation, select communities became influential in the HTRIP network, demonstrating a strong network of people engaged in similar activities, ultimately benefiting their livelihoods, as observed in earlier research [13]. These findings on the role of social networks in scaling community-based agroforestry in Haiti contribute supporting information to the argument that good networks are inclusive, facilitate collective learning, and allow sharing of success and generating wider social acceptance [97]. The HTRIP communities ultimately became communities of practice [102] that managed and innovated the community-based agroforestry system over time in collaboration with external institutions. Future work on the scaling of community-based agroforestry can apply an in-depth social network analysis to map community networks with the intent to facilitate innovation, as demonstrated in Balfour and Atler (2016) [39].

To scale and sustain natural climate solutions, we recommend that particularly in Haiti and similar contexts, that practitioners and planetary health professionals should leverage this collaborative human health and environmental approach to gain access to more communities. Similar to findings on adoption drivers of wildlife management initiatives in Mexico, trusted institutions with good reputations can also play an important role in supporting or promoting natural resource management programs [103]. Although situated in a fragmented landscape of 60 plus communities, the HTRIP presents a network of communities that spans the sub-region and that are actively working to change their communities. Combined, these communities can have meaningful impact on their landscape by harnessing the planetary health benefits of community-based agroforestry.

HTRIP’s ability to remain present within communities, even after program termination, was also identified as a mechanism for the practice’s sustainability. This mechanism was in line with natural resource management recommendations to continuously support the processes that help social groups form and help them mature [104]. HTRIP not only stayed in touch with community leaders to maintain contact, but also connected them with one another to strengthen network ties that could facilitate the sustainability of community-based agroforestry in the absence of program staff. This not only promoted accountability amongst participants but, importantly, also minimized feelings of abandonment by an external institution. In the case of the three communities that were not sustained in the formal HTRIP network, their termination of participation sheds light on the importance of aligning an intervention with local livelihoods. In agreeance with agroforestry adoption and natural resource management research in Sub-Saharan Africa [105, 106], agroforestry’s returns must be more favorable than alternative livelihood activities that exist within a community. Considering the heterogeneity of Haiti’s mountain and low-land regions, further research is needed to investigate how the social drivers of adoption (good social networks, livelihood activities) for scaling up of community-based agroforestry vary throughout the landscape.

Limitations

We encountered three limitations in this study. The first limitation was sample size; while the sampled communities represented various years of participation in HTRIP, a larger sample from HTRIP’s 60 communities could have revealed additional detail on the connectedness of HTRIP participants and whether there were other recruitment methods that were not identified during the study period. However, we were able to obtain supporting evidence from interviews with HTRIP staff that supported our finding that most of the communities became involved through peer-to-peer networks. The second limitation is the imbalance in gender representation of our interviewed sample. Although the HTRIP had many female participants, more men were willing and available to participate in the study during this time. Considering that women are typically responsible for making day trips to the market, taking care of the children, and preparing the day’s meal, this is the unfortunate reality behind the gender imbalance of the sample. Although the authors employed carefully formulated questions to obtain accurate and comprehensive information from respondents, a third limitation that remains is a potential risk of recall bias among participants in the social network analysis. While this risk is minimal given the saliency of relationships asked about in the social network analysis, it should be taken into account when interpreting the results. These limitations are associated with the difficulty of collecting data in a rural mountainous context with limited transportation capability under a limited timeframe and funding. As with fragile contexts, there were also delays due to ongoing socio-political protests that disrupted field work and at times, made it unsafe to travel far from the field office. Nevertheless, the results of this study do not detract from the importance of findings on community-based agroforestry adoption and scaling.

Conclusions

Our results provide initial insights in understanding how the scaling up of beneficial natural climate solutions such as community-based agroforestry can be further supported in limited resource contexts that remain the most vulnerable to climate change. Recent research suggests the adoption of the Scaling Innovations Theoretical Framework [107] which identifies several factors likely to influence the scaling of an innovation. Future research will build on the topic of this current paper and expand it to several other program characteristics likely to influence the scaling of natural resource management innovations in Haiti. Relative to those factors, in this paper, we identified the role that collaboration, partnerships, networks, and stakeholder engagement, among others played in the Haiti Timber Reintroduction Program’s success in achieving community-based agroforestry adoption, scaling, and sustainability. We invite other researchers to further explore and develop the concept of innovation scaling across disciplines as we work towards developing new cross-sector strategies and approaches to scaling up planetary health solutions in climate-vulnerable contexts.

Supporting information

S1 Text. Focus group interview guide with participating HTRIP communities.

doi:10.1371/journal.pclm.0000406.s001

(PDF)

S2 Text. Focus group guide for Kapuis community.

doi:10.1371/journal.pclm.0000406.s002

(PDF)

S3 Text. Individual interview guide for Kapuis Labadi Ottovin.

doi:10.1371/journal.pclm.0000406.s003

(PDF)

S4 Text. Individual interview guide for participating HTRIP individuals.

doi:10.1371/journal.pclm.0000406.s004

(PDF)

S5 Text. Coding framework for studying agroforestry in Haiti.

doi:10.1371/journal.pclm.0000406.s005

(PDF)

S1 Checklist. Inclusivity in global research.

doi:10.1371/journal.pclm.0000406.s006

(PDF)