Background
Micronutrient (vitamins and minerals) malnutrition is an important public health problem affecting populations in all regions of the world. According to the most recent estimate, 372 million preschool-aged children (aged 6–59 months) and 1.2 billion non-pregnant women of reproductive age (aged 15–49 years) globally are suffering from at least one micronutrient deficiency(Reference Stevens, Beal, Mbuya, Luo and Neufeld1). Additionally, 50–60% of the anaemia worldwide is due to iron deficiency(Reference Petry, Olofin and Hurrell2) affecting an estimated 27% of the world population in 2013, impacting developing nations at a disproportionately high rate(Reference Kassebaum and Collaborators3). Vitamin A and zinc deficiencies are also common and, coupled with undernutrition, have been attributed to approximately 45% of child deaths in 2011(Reference Black, Victora and Walker4). Micronutrient deficiencies have also been implicated in compromised immunity, retarded growth and lost developmental opportunity in children, and accelerated burden of chronic metabolic diseases with accumulated impact on health and productivity of humans. These deficiencies are due to lack of access to physiologically sufficient sources of food and are amplified by poor nutritional quality of crops. Therefore, increasing the consumption of essential micronutrients by the general population and especially of populations with high requirements remains a global priority.
Biofortification, the process of increasing the concentrations and bioavailability of essential nutrients in staple crops by conventional plant breeding and/or genetic engineering, is a promising approach to combatting micronutrient deficiencies(Reference Huey, Krisher and Bhargava5). Staple crops commonly targeted for biofortification are typically consumed in the daily diets of many populations and can include sweet potato, pearl millet, wheat, lentils, cassava, rice, beans and maize(6). Micronutrients commonly enhanced by biofortification techniques are provitamin A, zinc and iron. Ultimately, biofortification serves as a cost-effective (because once biofortified varieties are developed, small costs accrue for maintenance but their seeds are self-replicating, multiplying the potential benefits) and sustainable (farmers may receive the seeds initially but then retain seeds for future crop cycles, a self-sustaining agricultural solution) means to address the burden of micronutrient deficiency at a population level via already existing food systems, such as growing biofortified crops on smallholder farms or through purchasing biofortified food products in markets(6). To be fully realised as a sustainable solution to mitigate micronutrient deficiency, it is important to understand the factors that are associated with decisions to adopt and cultivate biofortified crops, including both facilitators and barriers to adoption.
Studies on adoption of agricultural innovation have often found that there is a large gap in time (ranging from 14 to 25 years) between availability of an innovation and its widespread adoption(Reference Walker and Alwang7). For example, the adoption level of vitamin A-biofortified cassava (VAC) variety was reported to be 35% in 2010 in sub-Saharan Africa(Reference Walker and Alwang7) and 39% in 2017 in Nigeria(Reference Ayinde, Adewumi, Ajewole and Ologunde8). Vitamin A-biofortified maize (VAM) yields have only a modest gain in sub-Saharan Africa compared with that in Brazil, China and the USA(Reference Atlin, Cairns and Das9), and varietal replacement has been extremely slow, with farmers still growing 20-year-old varieties(Reference Abate, Fisher and Abdoulaye10). High-iron beans (HIB) have seen a 29% adoption level following intensive delivery efforts(Reference Vaiknoras, Larochelle, Birol, Asare-Marfo and Herrington11) despite a demonstrated significant positive effect of HIB on health outcomes(Reference Haas, Luna and Lung’aho12). Indeed, orange sweet potato (OSP) is often cited because of multiple examples of its scaling-up exercises. Since 2009 when the level of OSP adoption was 6.9%(Reference Walker and Alwang7), 6.2 million households were reached by July 2019 in fifteen sub-Saharan African countries (with considerable variation between countries). The scaling process of the OSP innovation involved several distinct phases with key factors occurring during each of these phases. Low and Thiele(Reference Low and Thiele13) provided a detailed description of the initiatives aimed at scaling up the promotion and adoption of biofortified orange sweet potatoes. Experiences, findings, implementation challenges and lessons learned from these initiatives have been described previously(Reference Girard, Brouwer, Faerber, Grant and Low14). Across these initiatives, households significantly increased OSP production(Reference Low and Thiele15). Huey et al.(Reference Huey, Krisher and Bhargava5) in a systematic review described the impact pathway by which these initiatives increased OSP production leading to improved micronutrient status. Despite the many lessons learned, there remain unknowns in terms of implementation best practices and innovations needed to enhance delivery, engagement, uptake and impacts(Reference Low and Thiele15). Therefore, there is a need to understand the factors driving crops’ slow adoption. In fact, there has been a call by the US government food security strategy for the identification of pathways for faster scaling in the early phase of research and collaboration among scale-up partners(16). Understanding the factors that enable and inhibit the adoption of biofortified crops is crucial for designing future programmes focused on scaling.
Throughout the past decade, there has been a surge of studies reporting on biofortified crops. Several of them have explicitly examined the factors influencing adoption of OSP in Ghana(Reference Adekambi, Okello, Abidin and Carey17), Nigeria(Reference Adekambi, Okello, Abidin and Carey17,Reference Chah, Anugwa and Nwafor18) , Mozambique(Reference Jenkins, Shanks, Brouwer and Houghtaling19–Reference Jogo, Bocher and Grant21), Tanzania(Reference Shikuku, Okello, Sindi, Low and McEwan22) and Uganda(Reference de Brauw, Eozenou, Gilligan, Hotz, Kumar and Meenakshi20). A few other studies evaluated the determinants of adoption of VAC among farmers(Reference Ayinde, Adewumi, Ajewole and Ologunde8,Reference Ayodele, Fasina and Osundahunsi23,Reference Onyeneke, Emenekwe and Munonye24) . Vaiknoras et al.(Reference Vaiknoras, Larochelle, Birol, Asare-Marfo and Herrington11,Reference Vaiknoras, Larochelle, Birol, Asare-Marfo and Herrington25) evaluated the impact of different delivery approaches on household time to adoption of iron-biofortified beans in Rwanda. While several reviews based on these primary studies have been conducted in the recent past restricting their foci on one country/region and/or one biofortified crop/product(Reference Girard, Brouwer, Faerber, Grant and Low14,Reference Mulongo, Munyua, Mbabu and Maru26–Reference Okwuonu, Narayanan, Egesi and Taylor29) , these reviews did not use systematic review methods, which result in a synthesis of all available evidence across multiple databases and sources to limit bias in the overall conclusion. Systematic reviews are therefore more comprehensive and objective than reviews that are not systematic, as only some types of evidence may be included in the review. Two formal systematic reviews available with focus on implementation outcomes (e.g. acceptability and adoption) of a biofortified crop-based intervention were published in 2017(Reference Talsma, Melse-Boonstra and Brouwer30) and 2023(Reference Huey, Bhargava and Friesen31). The 2023 review summarised the results of testing sensory acceptability across ten conventionally biofortified food products. The review published in 2017 summarised the determinants of both acceptance and adoption(Reference Talsma, Melse-Boonstra and Brouwer30). However, in this 2017 review, studies included predicted the adoption of biofortified crops (i.e. acceptance and adoption remained mostly hypothetical or lab-in-the-field studies, based on choice experiments or experimental auctions) or participants were exposed to the biofortified crops for the first time (i.e. possibly biased findings from one-time exposure rather than actual experiences using of the biofortified crop), not based on real-world adoption or uptake data. In addition, large-scale effectiveness studies of biofortified crops are lacking, and reported adoption studies were mostly conducted under strict intervention conditions. Most of the studies in that review were not carried out among the groups most at risk of micronutrient deficiencies, such as young children and women of reproductive age. To harness the full potential of biofortified crops, it is imperative to understand the actual factors affecting all aspects of adoption and leverage on those that facilitate implementation and scale-up of biofortified crops. To the best of our knowledge, no review has been conducted that provides a comprehensive and consolidated overview of the barriers and enablers of adopting biofortified crops and their products at multiple levels including producers. Using the Consolidated Framework for Implementation Research (CFIR)(Reference Damschroder32,Reference Damschroder, Reardon, Widerquist and Lowery33) , a determinant framework from the implementation science discipline, the objective of this review was to summarise the evidence around the factors that positively or negatively impact adoption decisions on biofortified crops and its products among smallholder farmers and consumers across several domains (e.g. individuals delivering/receiving the innovation, the settings/context, implementation processes and the characteristics of the innovation itself). Understanding these factors could support the decision making of multiple actors (e.g. practitioners, researchers, policy makers) in the design, implementation and scale-up of programmes for biofortified crops and their products.
Methods
Inclusion and exclusion criteria
Participants
Any human population was considered eligible for inclusion, and as such, in terms of sampling design, we considered growing farm households who planted, cultivated or marketed biofortified seeds, roots or vines, either provided to them by programmes or those which had been retained from previous years’ crops. We did not limit by region, and households at either local or national scales were eligible.
Interventions
We included studies utilising biofortified foods and food products, including those that have undergone processing post-harvest and that have been delivered as crops only or in the form of food products (as defined by study authors). Crops included those biofortified by conventional plant breeding approaches. We included any type of conventionally biofortified crop.
We did not include studies using agronomic biofortification methods, genetic engineering-based biofortification methods in order to focus the review on conventionally biofortified crops. We also excluded animal-based biofortified foods, such as dairy products or meat from animals that consumed biofortified feed, and plant-based protein-biofortified crops, such as quality protein maize, to focus on micronutrient biofortification.
Comparators
Comparators included either (A) a non-biofortified (i.e. conventional) version of the same crop or food product made using non-biofortified crops, or (B) food products industrially fortified with the same micronutrient.
Primary outcomes
The primary outcome was adoption determinants (i.e. any barrier or facilitator empirically shown or believed to are associated with the adoption of biofortified crops, foods and their products). We defined adoption as households’, farmers’ or other stakeholders’ engagement in planting, cultivating, retaining for future cultivation, or marketing seeds, roots or vines that produce biofortified crops and/or their consumption of such biofortified crops and products made thereof. A barrier was defined as any variable reported be associated with impeding or obstructing the adoption of biofortification interventions. In contrast, a facilitator was considered as any of those variables that were associated with promotion of the adoption of biofortification interventions. Studies that provided usable data for either of these two factors were included. Studies that only reported (with no unit of text relevant to adoption determinants) the effect of adopting biofortification intervention on health/nutrition were not included (e.g. Ref. (Reference Kairiza, Kembo, Pallegedara and Macheka34)).
Study designs
As barriers and facilitators could be identified either from intervention studies assessing the effectiveness of biofortification interventions and/or associated implementation strategies or observational studies analysing factors of adoption of biofortification interventions, studies using any of the research designs ranging from qualitative to quantitative to mixed methods were eligible. Therefore, any relevant primary study conducted in the T2-2 phase (effectiveness research) and beyond along the translational research continuum (as described by Refs. (Reference Brown, Curran and Palinkas35–Reference Fort, Herr, Shaw, Gutzman and Starren37)) were included. We included primary studies conducted in the T2-1 phase (efficacy research) only if they reported any determinants of adoption in addition to health outcomes.
Literature searches
We performed a search of relevant literature databases including MEDLINE (PubMed), AGRICOLA, AgEcon, CABI Abstracts (Web of Science) and organisational websites (e.g. Harvest Plus, CGIAR and partners) as described previously(Reference Huey, Krisher, Friesen, Mbuya, Monterrosa and Mehta38).
As a preliminary assessment of the literature on biofortification, we conducted a broad search in MEDLINE (PubMed) on 29 March 2021, using the following key terms: Biofortification[MeSH] OR biofortif*[tiab] OR “bio-fortif*“[tiab]. This resulted in 1434 results. After screening these results and ascertaining key words to use for increasing the sensitivity of the search, we conducted searches in additional databases, using broad or narrower searching depending on the topic focus of the database. These, including the original MEDLINE search, are summarised in Table 1. We also hand-searched websites of organisations that play roles in growing or distributing conventionally biofortified crops. The results are included in Table 2.
Table 1. Search strategy across included databases
Table 2. Results from hand-searching organisation websites
We also identified 1147 potential citations outside of the original search during the screening process. These included studies that were cited in review papers but did not include variations of the term ‘biofortification’ in their abstracts; not indexed in any of the literature databases described above and were thus missed by the original search; or published recently in 2021, which we identified from journals’ table of contents alert feeds. Some of the latter included full-text versions of conference abstracts that were found and included in the original screening pool.
Data screening and extraction
We screened all records for eligibility, first at the title/abstract level and subsequently at the full-text screening level, and used a subset of articles to improve consistency among review authors. We then extracted data for each identified study, according to the following: study level details including authors or research group, study year and location; study design, population targeted, sample size and how crops were biofortified; and outcomes of the study.
We extracted core data relevant to implementation determinants of biofortified crops, foods and their products. Data extracted at this level included, in addition to barriers and facilitators of adoption, measures (definition) of adoption used, methods employed for identifying factors of adoption, and implementation strategies reported (if any) in each study. An inductive approach was taken to extract data on barriers and facilitators (which were later categorised using the CFIR codebook (see https://cfirguide.org/ for more information)(Reference Damschroder, Aron, Keith, Kirsh, Alexander and Lowery39) to allow capturing data segments on implementation determinants that did not fit the current CFIR constructs with potential to modify the CFIR matrix (i.e. testing applicability of CFIR in biofortification interventions). The decision to follow an inductive approach to data extraction was based on the anticipation that the CFIR might not fully reflect the implementation context of biofortification interventions.
Data synthesis and analysis
We conducted a directed (deductive) qualitative content analysis (DQCA)(Reference Assarroudi, Heshmati Nabavi, Armat, Ebadi and Vaismoradi40) to identify constructs related to implementation determinants extracted from the included studies. DQCA involves identifying a theoretical framework based on what the synthesis will be executed. This identification involved two steps and resulted in the selection of the CFIR as the theoretical framework(Reference Damschroder, Aron, Keith, Kirsh, Alexander and Lowery39). Initially, a list of candidate frameworks was developed by referring to three sources: (a) the website dissemination-implementation.org that lists theories, models and frameworks identified by Tabak and colleagues(Reference Tabak, Khoong, Chambers and Brownson41), (b) Grol et al.’s ‘Planning and studying improvement in patient care: the use of theoretical perspectives’(Reference Grol, Bosch, Hulscher, Eccles and Wensing42), and (c) Nilsen’s ‘Making sense of implementation theories, models and frameworks’(Reference Nilsen43). Then, candidate frameworks were rated against three criteria (familiarity, applicability and usability) online using the Theory Comparison and Selection Tool (T-CaST) (available at https://impsci.tracs.unc.edu/tcast/), and the CFIR scored the highest, indicating good fit to be used in this review. Although it is recommended to assess the CFIR constructs for their salience and adapt them for a particular context(Reference Damschroder, Aron, Keith, Kirsh, Alexander and Lowery39), this study assessed for each of the forty-eight constructs across five domains of the updated version of CFIR (i.e. CFIR 2.0)(Reference Damschroder, Reardon, Widerquist and Lowery33) as it was not possible to specify the most relevant constructs without having prior knowledge. Whether or not studies contributed data to the CFIR constructs was displayed through a matrix checklist, and the nature of the data contributed (i.e. barrier or facilitator or both) was indicated by differential colour-coding. The number of studies reporting on barriers and facilitators was aligned with each of the a priori constructs, and domains were counted and presented in tabular form. In addition to the frequency table, a summary table of barriers and facilitators was presented in tabular form for each construct.
Results
Study selection
For the four review topics, we found a total of 5142 records (Figure 1). Ultimately, we found overall 308 eligible records across the four review topics outlined previously.
Figure 1. PRISMA diagram(Reference Page, McKenzie and Bossuyt44).
Absence of data on implementation determinants was the major reason for excluding studies at the full-text screening phase: out of 308 articles assessed for eligibility, 267 studies were excluded as they did not report on implementation barriers and/or facilitators. As shown in Figure 1, forty-one studies met the eligibility criteria and were included in DQCA.
Characteristics of included studies
All the reports/articles included in this review were published within the past 15 years (2007 and later), with the majority (n = 37, 90%) being in or after 2016, the year the World Food Prize was awarded to four researchers who pioneered the development and dissemination of a biofortified crops(45). The included studies reported on four conventionally biofortified crops and/or their products: OSP, HIB, VAC and VAM. All studies specified the approach of biofortification used. Studies in this review predominantly represent OSP (n = 28, 68%), followed by VAC and HIB (n = 8, 20% each). While only one study reported VAM, three studies examined multiple biofortified crops. Geographically, East Africa (n = 29, 71%) and West Africa (n = 10, 24%) were most represented in this review, with a few studies from other regions including South America (n = 1, 2%) and South Asia (n = 1, 2%). This review presents evidence from nine different countries from Africa, with more studies coming from Kenya (n = 9, 22%), Nigeria (n = 9, 22%), Mozambique (n = 7, 17%), Rwanda (n = 6, 15%) and Uganda (n = 6, 15%) than from Tanzania (n = 3, 7%), Ghana (n = 2, 5%) and Malawi (n = 1, 2%) in Africa, India (n = 1, 2%) in South Asia, and Brazil (n = 1, 2%) in South America.
All but two studies included farming households in the rural setting: one study focused on purchase for consumption of OSP bread among urban consumers, while the other studied consumption and knowledge of OSPs and IBs in both rural and peri-urban areas. In terms of research phase, eight studies were a pre-implementation assessment to inform choice and development of implementation strategy and two other studies explicitly focused on external validity utilising effectiveness trial methodology (i.e. T2-2 phase). Five studies primarily had joint aims to test an implementation strategy (primary) and to assess health outcomes (secondary) (i.e. hybrid type 3 design). The majority of the studies (n = 19) only focused on implementation research, assessing the effect of implementation strategy(ies) on adoption status (n = 11) or post-implementation impact evaluation (n = 8). Only four studies assessed adoption and its drivers at scale (e.g. T4-1 phase in the translational research continuum). In terms of study type, this review included thirty-one (76%) quantitative studies compared with a few purely qualitative studies (n = 4, 10%). Six (15%) were mixed-methods studies.
Characteristics of adoption measures
Definitions and measurements used for the adoption outcome varied greatly across the studies. Eleven out of forty-one studies used both categorical and continuous measures of adoption, while thirty studies (73%) used only a categorical outcome measure. The definitions of adoption were not consistent across studies and ranged from one to five per study (Table 3). Twenty-five articles (61%) used a single definition, with the remaining using multiple definitions. The most frequently applied adoption measure was whether a household was growing a biofortified crop or not. Other measures included but were not limited to the decision to grow or consume in future, retention/preservation of planting materials, proportion of dedicated cultivation area, total cultivation area, and previous or ever consumption/production/purchase of biofortified crops and related products. The included studies differed in terms of their focus on the stages of adoption. Most studies (n = 36, 88%) conceptualised adoption in terms of agricultural and/or economic aspects (i.e. cultivation, multiplication, retention for future cultivation, and selling/purchasing of biofortified crops and/or their products). Nine studies (22%), however, had explicit focus on consumption (including three of these studies with sole focus on consumption). Studies also differed regarding the continuity aspect of the adoption measure. Thirty-one studies (76%) captured cross-sectional (first time) adoption, while ten studies (24%) focused on continuous adoption (e.g. voucher redemption for seeds/planting materials, storage or retention of seeds/planting materials, and cultivation of biofortified crops for multiple seasons/years).
Table 3. Mapping of barriers and facilitators to the Updated Consolidated Framework for Implementation Research (CFIR 2.0) constructs
Notes: CFIR 2.0 (Reference Damschroder, Reardon, Widerquist and Lowery33). HIB, high-iron beans; OSP, orange sweet potato; VAC, vitamin A cassava; VAM, vitamin A maize.
Blue, consistent facilitator of adoption; 
pink, consistent barrier; 
yellow, facilitator and barrier.
Barriers and facilitators to adoption of biofortified crops
Barriers to and facilitators of adoption of biofortified crops and products were spread across all the five domains of the CFIR 2.0: innovation, outer setting, inner setting, individuals and processes. The included studies reported factors of adoption that were captured by twenty-four out of the forty-eight constructs under these domains. Table 3 provides a mapping of each of the included studies based on which construct(s) and domain(s) it contributed data (i.e. factors of adoption) to align with. This matrix checklist also provides a visualisation of studies reporting only facilitators, only barriers or both facilitators and barriers. Some constructs of the CFIR 2.0 manifested as either facilitators (cells highlighted as blue) or barriers (cells highlighted as pink) to adoption. Other constructs captured factors that were reported to simultaneously be associated with adoption as barriers and facilitators (cells highlighted as yellow) in a single study.
Orange sweet potato (OSP)
The specific factors affecting adoption of OSP under the reported CFIR constructs are provided in Tables 4–8. Anecdotally, in one study, researchers found that, in 2011, conditions for maintaining vines were difficult and many households lost their vines; however, households who had received agricultural extension (the treatment group) were more likely to have been able to keep vines through the particularly dry year and continued to grow OSP in 2012(Reference de Brauw, Moursi and Munhaua85).
Table 4. Factors affecting adoption of orange sweet potato (OSP) and its products
Table 5. Factors affecting adoption of orange sweet potato (OSP) and its products
Table 6. Factors affecting adoption of orange sweet potato (OSP) and its products
Table 7. Factors affecting adoption of orange sweet potato (OSP) and its products
Table 8. Factors affecting adoption of orange sweet potato (OSP) and its products
Innovation characteristics
Characteristics of OSP that offered relative advantage and facilitated its adoption included early maturity(Reference Adekambi, Okello, Abidin and Carey17,Reference Jogo, Bocher and Grant21,Reference Behrman51,Reference Cole, Levin and Loechl54,Reference Adekambi, Okello and Rajendran57,Reference Chah, Anugwa and Ifeanyi64) , superior yield(Reference Adekambi, Okello, Abidin and Carey17,Reference Jogo, Bocher and Grant21,Reference Low, Arimond, Osman, Cunguara, Zano and Tschirley49,Reference Behrman51,Reference Cole, Levin and Loechl54,Reference Adekambi, Okello and Rajendran57,Reference Chah, Anugwa and Ifeanyi64) and sensory attributes(Reference Jenkins, Shanks, Brouwer and Houghtaling19,Reference Okello, Shiundu and Mwende48,Reference Low, Arimond, Osman, Cunguara, Zano and Tschirley49) , and high sugar (dry matter) content(Reference Adekambi, Okello, Abidin and Carey17,Reference Jogo, Bocher and Grant21,Reference Chah, Anugwa and Ifeanyi64) . Popularity of OSP as a staple food crop enhances its local adaptability to be adopted by farmers(Reference Arimond, Ball and Bechoff71). In addition, resistance to disease (e.g. virus(Reference Adekambi, Okello and Rajendran57,Reference Arimond, Ball and Bechoff71) ) and drought(Reference Jogo, Bocher and Grant21,Reference Adekambi, Okello and Rajendran57) and shortened time for cooking(Reference Adekambi, Okello and Rajendran57) were reported to favour OSP adoption. Ease of storage of OSP in the ground for longer durations may also favour adoption(Reference Adekambi, Okello and Rajendran57), but pest infestation remains a barrier(Reference Phorbee, Mulongo, Njoku, Maru and Munyua67). However, a few studies reported that OSP was sensitive to diverse weather conditions limiting its adoption; these include dying of planting materials in dry growing conditions(Reference Behrman51,Reference Mudege, Mwanga, Mdege, Chevo and Abidin70,Reference Arimond, Ball and Bechoff71) , lower tolerance to drought in white-fleshed sweet potato(Reference Jenkins, Shanks, Brouwer and Houghtaling19), and perishability of vines when it rains(Reference Behrman51,Reference Cole, Levin and Loechl54) . Some studies also reported relative disadvantages of OSP including shorter shelf life and cracking of tubers(Reference Phorbee, Mulongo, Njoku, Maru and Munyua67), lower dry matter content(Reference Adekambi, Okello and Rajendran57,Reference Phorbee, Mulongo, Njoku, Maru and Munyua67) and inferior taste(Reference Adekambi, Okello and Rajendran57,Reference Chah, Anugwa and Ifeanyi64) . Furthermore, farmers perceive OSP as more demanding in terms of field preparation and weed maintenance(Reference Jenkins, Shanks, Brouwer and Houghtaling19), requiring intensive knowledge and labour(Reference Mudege, Mwanga, Mdege, Chevo and Abidin70). Therefore, they find recommended OSP production practices difficult to integrate into their existing production systems(Reference Chah, Anugwa and Ifeanyi64). All factors are summarised in Table 4.
Characteristics of individuals
Reported capability-related factors in the literature can be grouped as physical (age(Reference Adekambi, Okello, Abidin and Carey17,Reference Jogo, Bocher and Grant21,Reference Shikuku, Okello, Sindi, McEwan and Low58,Reference Chah, Anugwa and Ifeanyi64,Reference Babatunde, Omoniwa and Adeniyi65) , skill(Reference Cole, Levin and Loechl54,Reference Babatunde, Omoniwa and Adeniyi65,Reference Arimond, Ball and Bechoff71) and experience(Reference Arimond, Ball and Bechoff71)) and psychological (education(Reference Jogo, Bocher and Grant21,Reference Chah, Anugwa and Ifeanyi64,Reference Phorbee, Mulongo, Njoku, Maru and Munyua67) and knowledge(Reference Jogo, Bocher and Grant21,Reference Petry, Wirth and Friesen47,Reference Okello, Shiundu and Mwende48,Reference Behrman51,Reference Gilligan, Kumar, McNiven, Meenakshi and Quisumbing55,Reference Gilligan, Kumar, McNiven, Meenakshi and Quisumbing56,Reference Shikuku, Okello, Sindi, McEwan and Low58,Reference Prakash, Kishore, Roy, Behura and Immanuel62,Reference Arimond, Ball and Bechoff71) ). Older age was reported as both barrier(Reference Shikuku, Okello, Sindi, McEwan and Low58,Reference Babatunde, Omoniwa and Adeniyi65) and facilitator(Reference Jogo, Bocher and Grant21,Reference Chah, Anugwa and Ifeanyi64) among farmers. One study reported mixed views (i.e. both enabler and barrier) on the association of getting older and OSP adoption(Reference Adekambi, Okello, Abidin and Carey17). Education level of farmers or household head was directly related with OSP adoption(Reference Jogo, Bocher and Grant21,Reference Chah, Anugwa and Ifeanyi64,Reference Phorbee, Mulongo, Njoku, Maru and Munyua67) – with low education serving as a barrier and more education serving as a facilitator – as was their knowledge on vitamin A(Reference Jogo, Bocher and Grant21,Reference Gilligan, Kumar, McNiven, Meenakshi and Quisumbing55,Reference Gilligan, Kumar, McNiven, Meenakshi and Quisumbing56,Reference Arimond, Ball and Bechoff71) , source for OSP(Reference Petry, Wirth and Friesen47,Reference Behrman51,Reference Shikuku, Okello, Sindi, McEwan and Low58,Reference Prakash, Kishore, Roy, Behura and Immanuel62) and nutritional benefits of OSP(Reference Petry, Wirth and Friesen47,Reference Behrman51,Reference Shikuku, Okello, Sindi, McEwan and Low58,Reference Prakash, Kishore, Roy, Behura and Immanuel62) or its product (e.g. OSP bread)(Reference Okello, Shiundu and Mwende48). Implementers’ skill to serve clients well(Reference Cole, Levin and Loechl54) and farmers’ ability to practice intercropping(Reference Babatunde, Omoniwa and Adeniyi65) and maintain OSP vines between seasons(Reference Arimond, Ball and Bechoff71) were facilitators. Having previous experience of growing OSP was positively associated with adoption in contrast to being new to OSP(Reference Arimond, Ball and Bechoff71).
Opportunity factors influencing adoption of OSP were related mainly to the physical opportunity (environment) reflected in access to agricultural inputs including vines(Reference Jenkins, Shanks, Brouwer and Houghtaling19,Reference Jogo, Bocher and Grant21,Reference Behrman51) , land(Reference Jogo, Bocher and Grant21,Reference Behrman51,Reference Gilligan, Kumar, McNiven, Meenakshi and Quisumbing55,Reference Gilligan, Kumar, McNiven, Meenakshi and Quisumbing56,Reference Okello, Muoki, Kwikiriza, Wambaya and Heck59,Reference Larochelle, Labarta and Katungi63,Reference Mudege, Mwanga, Mdege, Chevo and Abidin70,Reference Arimond, Ball and Bechoff71) , labour(Reference Jenkins, Shanks, Brouwer and Houghtaling19,Reference Arimond, Ball and Bechoff71) and water(Reference Mudege, Mwanga, Mdege, Chevo and Abidin70). Having access to or ability to purchase vines was reported as a facilitating factor(Reference Jogo, Bocher and Grant21,Reference Behrman51) , while lack of timely access to sufficient, good-quality vines impeded adoption(Reference Jenkins, Shanks, Brouwer and Houghtaling19,Reference Behrman51) . The size of land area to which farmers had access has been found to have conflicting associations with different measures of adoption. We found mixed associations of land both within(Reference Gilligan, Kumar, McNiven, Meenakshi and Quisumbing55,Reference Gilligan, Kumar, McNiven, Meenakshi and Quisumbing56,Reference Arimond, Ball and Bechoff71) and across studies(Reference Behrman51,Reference Gilligan, Kumar, McNiven, Meenakshi and Quisumbing55,Reference Gilligan, Kumar, McNiven, Meenakshi and Quisumbing56,Reference Larochelle, Labarta and Katungi63–Reference Babatunde, Omoniwa and Adeniyi65,Reference Phorbee, Mulongo, Njoku, Maru and Munyua67,Reference Arimond, Ball and Bechoff71) . However, having larger land area cultivated with sweet potato was consistently reported as a facilitator(Reference Gilligan, Kumar, McNiven, Meenakshi and Quisumbing55,Reference Gilligan, Kumar, McNiven, Meenakshi and Quisumbing56,Reference Arimond, Ball and Bechoff71) . Labour constraints(Reference Jenkins, Shanks, Brouwer and Houghtaling19,Reference Arimond, Ball and Bechoff71) including heavy domestic responsibilities of women(Reference Mudege, Mwanga, Mdege, Chevo and Abidin70) constrained adoption. Consistent to the relationship of labour constraint and adoption, greater household size(Reference Arimond, Ball and Bechoff71) and higher dependency ratio(Reference Okello, Bocher and Low66) were found as facilitators of adoption. Lack of water sources or access to wetland(Reference Mudege, Mwanga, Mdege, Chevo and Abidin70) and lack of purchasing power(Reference Low, Arimond, Osman, Cunguara, Zano and Tschirley49) or capital(Reference Jenkins, Shanks, Brouwer and Houghtaling19,Reference Mudege, Mwanga, Mdege, Chevo and Abidin70) to get access to inputs such as land and irrigation limited adoption. Participation in salaried employment, however, remained as a barrier to adoption. Crop diversity – reflected in growing and selling of higher number of crop varieties – was associated with higher adoption(Reference Jogo, Bocher and Grant21,Reference Arimond, Ball and Bechoff71) . Similarly, two studies reported positive association between being wealthier and adoption status. However, owning livestock assets did not have unidirectional associations with adoption across studies(Reference Adekambi, Okello, Abidin and Carey17,Reference Okello, Muoki, Kwikiriza, Wambaya and Heck59) . Such lack of unidirectionality was also observed for the sex of household head controlling resources including land(Reference Jogo, Bocher and Grant21,Reference Behrman51,Reference Gilligan, Kumar, McNiven, Meenakshi and Quisumbing55,Reference Gilligan, Kumar, McNiven, Meenakshi and Quisumbing56,Reference Chah, Anugwa and Ifeanyi64) . Regarding social opportunity, we almost exclusively found positive associations of better social network and support and adoption. Social networking including participation in community farmer organisations or community groups(Reference Adekambi, Okello and Rajendran57,Reference Mudege, Mwanga, Mdege, Chevo and Abidin70) and being the leader of farmer groups(Reference Gilligan, Kumar, McNiven, Meenakshi and Quisumbing55,Reference Gilligan, Kumar, McNiven, Meenakshi and Quisumbing56,Reference Arimond, Ball and Bechoff71) facilitated adoption. Likewise, adoption was positively associated with a support system reflected in better access to extension services(Reference Adekambi, Okello and Rajendran57) and higher frequency of interaction between extension and farmers(Reference Adekambi, Okello, Abidin and Carey17). One study, however, reported unlikeliness of cooperative society members to cultivate a large proportion of their farmland to OSP.
Motivation factors affecting adoption were both reflective and automatic in nature. Farmers’ aversion of risk(Reference Behrman51), unwillingness to try innovative crops and/or recommended practices(Reference Behrman51,Reference Cole, Levin and Loechl54) , and having poor yield in the first growing season(Reference Behrman51) acted as barriers that fall under the reflective motivation category. We also found four facilitators that align with this sub-component of motivation; these include presence of child or pregnant women in the household(Reference Adekambi, Okello, Abidin and Carey17,Reference Lagerkvist, Mutiso, Okello, Muoki, Oluoch-Kosura and Heck50) ; being cosmopolite (having had a cosmopolitan outlook in the last year) to seek information on OSP(Reference Chah, Anugwa and Ifeanyi64); the satisfaction of serving one’s own communities(Reference Cole, Levin and Loechl54); and women’s goal setting concerning when and how to provide children with proper complementary foods(Reference Lagerkvist, Mutiso, Okello, Muoki, Oluoch-Kosura and Heck50). All the reported factors that fall under the automatic motivation category were cited as facilitators. These include positive attitude towards OSP(Reference Shikuku, Okello, Sindi, Low and McEwan22,Reference Shikuku, Okello, Sindi, McEwan and Low58) , having belief that children enjoy eating OSP tubers(Reference Shikuku, Okello, Sindi, Low and McEwan22); having belief to make adequate returns from OSP/vine sales(Reference Shikuku, Okello, Sindi, Low and McEwan22,Reference Behrman51,Reference Arimond, Ball and Bechoff71) ; feeling in control of using OSP-based complementary foods(Reference Lagerkvist, Mutiso, Okello, Muoki, Oluoch-Kosura and Heck50); perception that vine multiplication provides employment opportunity(Reference Okello, Jogo, Kwikiriza and Muoki60) and brings good health(Reference Shikuku, Okello, Sindi, Low and McEwan22,Reference Behrman51,Reference Okello, Jogo, Kwikiriza and Muoki60) ; and desire to earn ‘more income’(Reference Okello, Jogo, Kwikiriza and Muoki60). All factors are summarised in Table 5.
Outer setting
Factors within the outer setting domain concerned critical incidents, local attitudes and conditions, partnerships and connections, existing policies and laws, and external pressure. Critical incidents acted as barriers and included extended hot dry season(Reference Arimond, Ball and Bechoff71), high virus pressure(Reference Arimond, Ball and Bechoff71), flooding(Reference Larochelle, Labarta and Katungi63), pest damage(Reference Larochelle, Labarta and Katungi63), theft(Reference Behrman51) and grazing animals(Reference Behrman51). Community recognition of implementers’ work(Reference Cole, Levin and Loechl54) and approval of OSP use in child foods(Reference Lagerkvist, Mutiso, Okello, Muoki, Oluoch-Kosura and Heck50) helped facilitate adoption. In contrast, prevailing gender ideology limiting women’s mobility and bias of extension services towards male heads of households constrained adoption(Reference Mudege, Mwanga, Mdege, Chevo and Abidin70). Also, perceptions of OSP as ‘a women’s crop’ and ‘a poor person’s food’(Reference Cole, Levin and Loechl54) were cited as barriers. Local conditions that favoured adoption include proximity and good access to markets for roots (vines) and means of transport(Reference Arimond, Ball and Bechoff71). Areas with two growing seasons a year also facilitated adoption(Reference Arimond, Ball and Bechoff71). In contrast, drought-prone areas characterised by high temperatures and little moisture posed hindrance to adoption(Reference Low, Arimond, Osman, Cunguara, Zano and Tschirley49,Reference Larochelle, Labarta and Katungi63,Reference Mudege, Mwanga, Mdege, Chevo and Abidin70) as did areas with soil of low fertility and instances of soil erosion(Reference Larochelle, Labarta and Katungi63). High costs associated with transport, delivery of OSP vines, or agricultural inputs (labour, fertilisers, herbicides or vines) were cited by four studies(Reference Larochelle, Labarta and Katungi63,Reference Chah, Anugwa and Ifeanyi64,Reference Phorbee, Mulongo, Njoku, Maru and Munyua67,Reference Arimond, Ball and Bechoff71) . Greater distance to key places including agricultural field office, main road, vine multipliers’ locations and health facility were inversely associated with adoption(Reference Adekambi, Okello, Abidin and Carey17,Reference Cole, Levin and Loechl54,Reference Adekambi, Okello and Rajendran57,Reference Okello, Bocher and Low66,Reference Arimond, Ball and Bechoff71) . Other local infrastructural barriers included lack of access to markets (clients) for selling vines(Reference Mudege, Mwanga, Mdege, Chevo and Abidin70), inefficient transport system(Reference Chah, Anugwa and Ifeanyi64) and poor storage conditions(Reference Phorbee, Mulongo, Njoku, Maru and Munyua67).
Market pressure (i.e. competition of OSP (or vines) with peer crops) was associated with farmers’ adoption of OSP. Availability of markets and traders for roots and vines(Reference Arimond, Ball and Bechoff71), profit from selling roots/vines(Reference Chah, Anugwa and Ifeanyi64), and presence of market price incentives for OSP and its products(Reference Low, Arimond, Osman, Cunguara, Zano and Tschirley49) helped facilitate adoption. In contrast, inadequate market for the sale of OSP(Reference Behrman51,Reference Chah, Anugwa and Ifeanyi64,Reference Babatunde, Omoniwa and Adeniyi65,Reference Phorbee, Mulongo, Njoku, Maru and Munyua67) , unstable markets(Reference Jenkins, Shanks, Brouwer and Houghtaling19), lack of demand(Reference Prakash, Kishore, Roy, Behura and Immanuel62) and lower price premium(Reference Prakash, Kishore, Roy, Behura and Immanuel62) hindered adoption.
Partnerships and connections with entities external to the implementing organisation were associated with implementation outcome. Adoption was reported to be facilitated by the greater numbers of nongovernmental organization (NGO) engaged in agricultural activities with smallholder farmers and the presence of strong public sector extension services or public institutions. However, protraction and difficulty in contracting suitable local partners constrained the adoption pathway(Reference Arimond, Ball and Bechoff71). In addition to NGO and extension services, availability of socially cohesive pre-existing groups and the existence of strong farmer organisations/associations at the local level were cited as facilitators(Reference Arimond, Ball and Bechoff71), but limited sharing of information and planting material across farmer networks limited adoption(Reference Jenkins, Shanks, Brouwer and Houghtaling19). Weak linkages among producers, processors, manufacturers and marketers were also mentioned as barriers to the pathway to adoption(Reference Phorbee, Mulongo, Njoku, Maru and Munyua67). Although the existence of government policies, programmes and strategies for promotion and consumption of OSP helps facilitate adoption(Reference Arimond, Ball and Bechoff71), lack of collaboration at the national level about the promotion of OSP hindered adoption(Reference Phorbee, Mulongo, Njoku, Maru and Munyua67,Reference Arimond, Ball and Bechoff71) . Similar hindrance effects were also reported for under-financed, under-staffed and under-recognised nutrition programmes and policies(Reference Cole, Levin and Loechl54). All factors are summarised in Table 6.
Inner setting (implementing organisation)
The characteristics of inner setting (i.e. implementing organisations) including institutional capacity, available resources, relational connections, and access to knowledge and information were associated with implementation of programmes aimed at increasing adoption. Failure to supply healthy planting materials of the desired varieties timely in sufficient quantities deterred adoption(Reference Low, Arimond, Osman, Cunguara, Zano and Tschirley49,Reference Larochelle, Labarta and Katungi63,Reference Arimond, Ball and Bechoff71) . Human resource inadequacy as barriers to programme delivery was cited by two studies that characterised extension staff as insufficient (resulting in high client loads), overcommitted (multitasking), underpaid and inadequately qualified(Reference Cole, Levin and Loechl54,Reference Arimond, Ball and Bechoff71) . While high-quality training of extension personnel and farmer multipliers was reported to be critical to successful adoption(Reference Arimond, Ball and Bechoff71), lack of training and training materials was the most frequently mentioned constraint by implementers(Reference Cole, Levin and Loechl54). Lack of adequate office space and transportation for implementers was also reported as barriers(Reference Cole, Levin and Loechl54). However, a working environment with effective supervision and cooperative teamwork helped implementers stay motivated to effectively carry out their work to promote adoption(Reference Cole, Levin and Loechl54). All factors are summarised in Table 7.
Implementation processes
The aspects of planning, teaming and assessing context that facilitated implementation of programmes aimed at increasing adoption were cited by two studies(Reference Cole, Levin and Loechl54,Reference Arimond, Ball and Bechoff71) . Specific activities within these three sub-domains were having sufficient preparatory period for improvements in intervention design (i.e. planning), diagnostics analyses to identify implementation challenges (i.e. assessing context) and having local implementation partners with established systems and procedures to support coordination, networking, organisation of implementers, and ongoing monitoring and evaluation (i.e. teaming). These two studies also identified the positive associations between reflecting and evaluating as in timely detection of problems in implementation and their resolution using mixed-methods implementation evaluation(Reference Cole, Levin and Loechl54) and operations research(Reference Arimond, Ball and Bechoff71). For example, following a dramatic drop of voucher redemptions for vines in the dry season, vine multipliers were provided with pumps for irrigation to ensure a steady flow of healthy, mature vines while farmers were allowed to redeem vouchers for an extended period as they could not keep vines alive during the dry season(Reference Cole, Levin and Loechl54). Two separate studies identified the positive association between adoption and modifying the innovation (i.e. biofortified OSP crop) without altering its any core feature (i.e. adapting). These included altering the length of exposure to the OSP intervention(Reference Jogo, Bocher and Grant21) and frequency of exposure to the OSP intervention(Reference Maredia, Porter, Jin and Farris69). Tailoring of strategies to increase adoption was cited by twelve studies and included use of vouchers redeemable for vines(Reference Shikuku, Okello, Wambugu, Sindi, Low and McEwan46,Reference Cole, Levin and Loechl54,Reference Arimond, Ball and Bechoff71) , agronomic training(Reference Adekambi, Okello, Abidin and Carey17,Reference Shikuku, Okello, Wambugu, Sindi, Low and McEwan46,Reference de Brauw, Eozenou, Gilligan, Hotz, Kumar and Meenakshi52,Reference de Brauw, Eozenou, Gilligan, Hotz, Kumar and Meenakshi53,Reference Adekambi, Okello and Rajendran57,Reference Okello, Muoki, Kwikiriza, Wambaya and Heck59,Reference Mudege, Mwanga, Mdege, Chevo and Abidin70) , market development(Reference Arimond, Ball and Bechoff71), nutrition education through multiple channels(Reference Ojwang, Otieno, Nyikal, Muoki and Okello61) and articulation of agriculture, nutrition and market sectors(Reference Low, Arimond, Osman, Cunguara, Zano and Tschirley68). Engaging innovation recipients to attract and encourage them to adopt OSP was mainly through nutrition education and awareness campaigns emphasising the benefits of growing and consuming OSP(Reference Adekambi, Okello, Abidin and Carey17,Reference Shikuku, Okello, Wambugu, Sindi, Low and McEwan46,Reference Behrman51–Reference Cole, Levin and Loechl54,Reference Adekambi, Okello and Rajendran57,Reference Okello, Muoki, Kwikiriza, Wambaya and Heck59,Reference Maredia, Porter, Jin and Farris69) . However, de-incentivising implementers (i.e. community health workers) was associated with reduced probability of growing and consuming OSP(Reference Okello, Bocher and Low66). All factors are summarised in Table 8.
High-iron beans (HIB)
The specific factors affecting adoption of HIB under the reported CFIR constructs are provided in Tables 9–13.
Table 9. Factors affecting adoption of high-iron bean (HIB) and its products
Table 10. Factors affecting adoption of high-iron bean (HIB) and its products
Table 11. Factors affecting adoption of high-iron bean (HIB) and its products
Table 12. Factors affecting adoption of high-iron bean (HIB) and its products
Table 13. Factors affecting adoption of high-iron bean (HIB) and its products
Innovation characteristics
Characteristics of HIB that offered relative advantage and facilitated its adoption included better taste, better market return potential and shortened time for cooking(Reference Asare-Marfo, Herrington and Birachi84). However, its poor yield was cited as a barrier(Reference Asare-Marfo, Herrington and Birachi84). Nevertheless, tolerance to drought and resistance to pests and disease were reported to favour HIB adoption(Reference Asare-Marfo, Herrington and Birachi84). All factors are summarised in Table 9.
Characteristics of individuals
Reported capability-related factors in the literature can be grouped as physical age(Reference Muthini80,Reference Muthini, Nzuma and Nyikal81) and experience(Reference Vaiknoras, Larochelle, Birol, Asare-Marfo and Herrington11,Reference Vaiknoras, Larochelle, Birol, Asare-Marfo and Herrington25,Reference Vaiknoras and Larochelle77,Reference Vaiknoras and Larochelle78) and psychological (education(Reference Vaiknoras, Larochelle, Birol, Asare-Marfo and Herrington11,Reference Vaiknoras, Larochelle, Birol, Asare-Marfo and Herrington25,Reference Vaiknoras and Larochelle77,Reference Vaiknoras and Larochelle78,Reference Muthini80,Reference Muthini, Nzuma and Nyikal81,Reference Asare-Marfo, Herrington and Birachi84) ) and knowledge(Reference Petry, Wirth and Friesen47,Reference Muthini80,Reference Muthini, Nzuma and Nyikal81) . Education was found to have mixed effect on HIB adoption(Reference Vaiknoras, Larochelle, Birol, Asare-Marfo and Herrington11,Reference Vaiknoras, Larochelle, Birol, Asare-Marfo and Herrington25,Reference Vaiknoras and Larochelle77,Reference Vaiknoras and Larochelle78,Reference Muthini80,Reference Muthini, Nzuma and Nyikal81,Reference Asare-Marfo, Herrington and Birachi84) . Lack of awareness of or knowledge about the source and benefits of HIB were cited as barriers to HIB adoption(Reference Petry, Wirth and Friesen47,Reference Muthini80,Reference Muthini, Nzuma and Nyikal81) . Older age of farmers was reported as barrier(Reference Muthini80,Reference Muthini, Nzuma and Nyikal81) . However, more years of experience of growing HIB was associated with lower dis-adoption of HIB(Reference Vaiknoras, Larochelle, Birol, Asare-Marfo and Herrington11,Reference Vaiknoras, Larochelle, Birol, Asare-Marfo and Herrington25) . Growers of traditional climbing beans were found less likely to adopt the improved bush variety(Reference Vaiknoras and Larochelle77,Reference Vaiknoras and Larochelle78) .
Two physical opportunity factors – high price of improved seeds and lack of land – were barriers(Reference Larochelle, Labarta and Katungi63). Other physical opportunity factors including wealth(Reference Sellitti, Vaiknoras and Smale79,Reference Asare-Marfo, Herrington and Birachi84) , land area(Reference Sellitti, Vaiknoras and Smale79–Reference Muthini, Nzuma and Nyikal81,Reference Asare-Marfo, Herrington and Birachi84) , crop diversity(Reference Muthini80,Reference Muthini, Nzuma and Nyikal81) , household size(Reference Vaiknoras and Larochelle77–Reference Sellitti, Vaiknoras and Smale79) and number of agricultural equipment owned(Reference Vaiknoras and Larochelle77,Reference Vaiknoras and Larochelle78) were positively related with HIB adoption. Households with women as the main decision maker for bean production was associated with better adoption(Reference Vaiknoras, Larochelle, Birol, Asare-Marfo and Herrington11,Reference Vaiknoras, Larochelle, Birol, Asare-Marfo and Herrington25) . This was also true for households with better access to information, markets and financial resources(Reference Asare-Marfo, Herrington and Birachi84). Regarding social opportunity, we exclusively found that a better social network and support system were positively associated with adoption. Social networking reflected in membership to highly trusted social groups(Reference Asare-Marfo, Herrington and Birachi84) and a higher number of fellow farmers with whom to discuss agriculture and nutrition(Reference Vaiknoras, Larochelle, Birol, Asare-Marfo and Herrington11,Reference Vaiknoras, Larochelle, Birol, Asare-Marfo and Herrington25,Reference Muthini80,Reference Muthini, Nzuma and Nyikal81) facilitated adoption. Likewise, adoption was positively associated with support systems such as having access to agricultural extension agents(Reference Vaiknoras, Larochelle, Birol, Asare-Marfo and Herrington11,Reference Vaiknoras, Larochelle, Birol, Asare-Marfo and Herrington25,Reference Vaiknoras and Larochelle77,Reference Vaiknoras and Larochelle78) . All factors are summarised in Table 10.
Outer setting
Factors within the outer setting domain concerned critical incidents, local conditions and external pressure. All factors cited under these three constructs posed barriers to HIB adoption and included drought(Reference Larochelle, Labarta and Katungi63), remoteness (distance to produce markets)(Reference Muthini80,Reference Muthini, Nzuma and Nyikal81) and low market price for HIB(Reference Larochelle, Labarta and Katungi63). No facilitators were reported. All factors are summarised in Table 11.
Inner setting (implementing organisation)
The characteristics of inner setting (i.e. implementing organisations) that were associated with adoption of HIB were related to availability of resource materials. Failure to supply HIB planting material (seed) timely in sufficient quantities deterred adoption and was reported by two studies(Reference Larochelle, Labarta and Katungi63,Reference Asare-Marfo, Herrington and Birachi84) . No facilitators were reported. All factors are summarised in Table 12.
Implementation process
Two aspects of implementation process (adapting and tailoring strategies) were found to facilitate the implementation of programmes aimed at increasing adoption. Adapting included modifying the size of HIB seed packets. This helped farmers get access to small packets without travelling long distances to formal delivery channels while providing small-scale producers with large packets(Reference Vaiknoras, Larochelle, Birol, Asare-Marfo and Herrington11,Reference Vaiknoras, Larochelle, Birol, Asare-Marfo and Herrington25) . Tailoring of strategies to increase adoption was cited by three studies and included use of seed delivery payback scheme(Reference Vaiknoras, Larochelle, Birol, Asare-Marfo and Herrington11,Reference Vaiknoras, Larochelle, Birol, Asare-Marfo and Herrington25) , increasing visits from extension officers(Reference Muthini80,Reference Muthini, Nzuma and Nyikal81) , and provision of nutrition training in addition to intensive agricultural training by extension officers(Reference Ogutu, Fongar and Gödecke82,Reference Ogutu, Fongar and Gödecke83) . No barriers were reported. All factors are summarised in Table 13.
Vitamin A cassava (VAC)
The specific factors affecting adoption of VAC under the reported CFIR constructs are provided in Tables 19–21.
Innovation characteristics
Characteristics of VAC that offered relative advantage and facilitated its adoption included early maturity and superior yield(Reference Esuma, Nanyonjo, Miiro, Angudubo and Kawuki76). Several studies, however, reported disadvantages of VAC limiting its adoption including bitter taste of fresh roots(Reference Esuma, Nanyonjo, Miiro, Angudubo and Kawuki76), poor yield(Reference Onyeneke, Emenekwe and Munonye24) and lower dry matter content (high moisture)(Reference Onyeneke, Emenekwe and Munonye24,Reference Phorbee, Mulongo, Njoku, Maru and Munyua67,Reference Esuma, Nanyonjo, Miiro, Angudubo and Kawuki76) . Resistance to drought(Reference Esuma, Nanyonjo, Miiro, Angudubo and Kawuki76) was reported to favour VAC adoption, but time-consuming root peeling of VAC and frying of gari(Reference Olaosebikan, Abdulrazaq and Owoade72), difficulty in sprouting(Reference Onyeneke, Emenekwe and Munonye24) and perishability(Reference Onyeneke, Emenekwe and Munonye24,Reference Phorbee, Mulongo, Njoku, Maru and Munyua67) are complexities associated with VAC constraining its adoption. All factors are summarised in Table 14.
Table 14. Factors affecting adoption of vitamin A cassava (VAC) and its products
Characteristics of individuals
Among the psychological capability factors affecting VAC adoption, level of education was positively associated with adoption(Reference Onyeneke, Emenekwe and Munonye24,Reference Gonzalez, Perez, Estevao Cardoso, Andrade and Johnson73,Reference Esuma, Nanyonjo, Miiro, Angudubo and Kawuki76) . Similarly, having increased awareness of the benefits of VAC facilitated adoption(Reference Ayodele, Fasina and Osundahunsi23,Reference Phorbee, Mulongo, Njoku, Maru and Munyua67,Reference Gonzalez, Perez, Estevao Cardoso, Andrade and Johnson73–Reference Ayinde75) . Physical capability factors including older age and higher experience in farming activities were negatively associated with VAC adoption(Reference Ayinde, Adewumi, Ajewole and Ologunde8,Reference Ayodele, Fasina and Osundahunsi23,Reference Ayinde74,Reference Ayinde75) .
Lack of physical opportunity factors including capital(Reference Ayinde74,Reference Ayinde75) , planting materials(Reference Phorbee, Mulongo, Njoku, Maru and Munyua67,Reference Ayinde74,Reference Ayinde75) and labour(Reference Olaosebikan, Abdulrazaq and Owoade72) constrained adoption. Still, larger household size acted as a barrier to VAC adoption(Reference Onyeneke, Emenekwe and Munonye24). Being a female youth was associated with more engagement in cassava production activities(Reference Esuma, Nanyonjo, Miiro, Angudubo and Kawuki76), and two studies reported a positive association between getting married and VAC adoption(Reference Ayodele, Fasina and Osundahunsi23,Reference Esuma, Nanyonjo, Miiro, Angudubo and Kawuki76) . Further physical opportunity factors supporting higher adoption included access to credit(Reference Onyeneke, Emenekwe and Munonye24,Reference Gonzalez, Perez, Estevao Cardoso, Andrade and Johnson73) , VAC stem(Reference Ayinde, Adewumi, Ajewole and Ologunde8) and media including phone and internet(Reference Ayinde, Adewumi, Ajewole and Ologunde8,Reference Onyeneke, Emenekwe and Munonye24,Reference Gonzalez, Perez, Estevao Cardoso, Andrade and Johnson73) . Social opportunity facilitating adoption came in the forms of contact with extension agents to receive training on improved agronomic practices(Reference Ayinde, Adewumi, Ajewole and Ologunde8,Reference Onyeneke, Emenekwe and Munonye24,Reference Olaosebikan, Abdulrazaq and Owoade72,Reference Gonzalez, Perez, Estevao Cardoso, Andrade and Johnson73) and membership of cooperative society(Reference Onyeneke, Emenekwe and Munonye24).
Motivation factors affecting VAC adoption were both reflective and automatic in nature. Within the reflective motivation category, income from non-farm activities(Reference Ayinde74,Reference Ayinde75) and higher number of under five children(Reference Onyeneke, Emenekwe and Munonye24) acted as barriers. However, increased income from VAC further motivated farmers to adopt VAC(Reference Ayodele, Fasina and Osundahunsi23). The only automatic motivation factor supporting adoption of VAC was having a positive perception about it(Reference Ayodele, Fasina and Osundahunsi23). All factors are summarised in Table 15.
Table 15. Factors affecting adoption of vitamin A cassava (VAC) and its products
Outer setting
Factors within the outer setting domain concerned critical incidents, local attitudes and conditions, partnerships and connections, and external pressure. Critical incidents acted as barriers and included activities of Fulani herders destroying first planting of VAC(Reference Phorbee, Mulongo, Njoku, Maru and Munyua67), pest damage and disease infestation(Reference Ayinde74,Reference Ayinde75) , theft(Reference Ayinde74,Reference Ayinde75) and farm invasion of grazing animals(Reference Ayinde74,Reference Ayinde75) . A prevailing general perception that the yellow colour of VAC indicates it is used for sick people also constrained adoption(Reference Phorbee, Mulongo, Njoku, Maru and Munyua67).
Local conditions that favoured VAC adoption include ease of transporting products to designated markets(Reference Olaosebikan, Abdulrazaq and Owoade72). In contrast, inadequate local capacity to supply freshly harvested VAC to large-scale buyers and processors posed hindrance to adoption(Reference Phorbee, Mulongo, Njoku, Maru and Munyua67). Expensive processing and storage facilities(Reference Olaosebikan, Abdulrazaq and Owoade72,Reference Ayinde74,Reference Ayinde75) coupled with poor road networks(Reference Ayinde74,Reference Ayinde75) particularly discouraged farmers’ adoption of VAC. These problems were further compounded by poor market infrastructure and middlemen exploitation(Reference Olaosebikan, Abdulrazaq and Owoade72) with associated fluctuations in VAC price(Reference Ayinde74,Reference Ayinde75) . We also found one study citing poor soil fertility as a barrier to adoption(Reference Onyeneke, Emenekwe and Munonye24).
Competition of cassava with peer crops was associated with farmers’ adoption of VAC. Non-availability of ready-made markets for VAC(Reference Ayinde74,Reference Ayinde75) , low price(Reference Chah, Anugwa and Ifeanyi64) and oversupply of cassava in local markets(Reference Olaosebikan, Abdulrazaq and Owoade72) deterred VAC adoption. In contrast, higher consumer demand facilitated adoption(Reference Olaosebikan, Abdulrazaq and Owoade72).
Partnerships and connections with entities external to the implementing organisation were associated with implementation outcome. Absence of collaboration among agricultural stakeholders was cited by one study as the main issue behind poor adoption of VAC(Reference Phorbee, Mulongo, Njoku, Maru and Munyua67). Lack of extension services and poor extension supervision were also found to limit adoption(Reference Phorbee, Mulongo, Njoku, Maru and Munyua67,Reference Olaosebikan, Abdulrazaq and Owoade72) . All factors are summarised in Table 16.
Table 16. Factors affecting adoption of vitamin A cassava (VAC) and its products
Inner setting (implementing organisation)
The characteristics of inner setting (i.e. implementing organisations) that were associated with the adoption of VAC were related to availability of resource materials. Failure to supply VAC planting material (seed) even when farmers were willing to pay for them constrained adoption and was reported by two studies(Reference Phorbee, Mulongo, Njoku, Maru and Munyua67,Reference Gonzalez, Perez, Estevao Cardoso, Andrade and Johnson73) . No facilitators were reported. All factors are summarised in Table 17.
Table 17. Factors affecting adoption of vitamin A cassava and its products
Implementation process
An aspect of implementation process found to be associated with VAC adoption was engaging innovation recipients. Low-awareness campaigns on VAC and its products were cited as a limiting factor(Reference Phorbee, Mulongo, Njoku, Maru and Munyua67), while involving farmers in participatory research process facilitated adoption(Reference Gonzalez, Perez, Estevao Cardoso, Andrade and Johnson73). All factors are summarised in Table 18.
Table 18. Factors affecting adoption of vitamin A cassava and its products
Table 19. Barriers affecting adoption of vitamin A maize (VAM) and its products
Table 20. Barriers affecting adoption of vitamin A maize (VAM) and its products
Table 21. Barriers affecting adoption of vitamin A maize (VAM) and its products
Vitamin A maize (VAM):
The specific factors affecting adoption of VAM under the reported CFIR constructs are provided in Table 19. As a note, we found only one study, and this study only reported barriers of VAM adoption.
Innovation characteristics
Characteristics of VAM that offered relative disadvantage and hindered its adoption included lower yield and harder texture(Reference Phorbee, Mulongo, Njoku, Maru and Munyua67). Aflatoxin contamination and damage by storage pests were complexities associated with maize constraining its adoption(Reference Phorbee, Mulongo, Njoku, Maru and Munyua67). All barriers are summarised in Tables 19–21; no facilitators were reported in our included studies.
Characteristics of individuals
Among the psychological capability factors, we found low level of education and lack of awareness of VAM to be associated with lower adoption(Reference Phorbee, Mulongo, Njoku, Maru and Munyua67). Similarly, limited agronomic insights was negatively associated with VAM adoption(Reference Phorbee, Mulongo, Njoku, Maru and Munyua67). All barriers are summarised in Table 20.
Outer setting
Factors within the outer setting domain concerned local conditions and external pressure, extracted from one study(Reference Phorbee, Mulongo, Njoku, Maru and Munyua67). All factors cited under these two constructs posed barriers to VAM adoption and included inefficiencies in the credit market, consumers’ high preference for conventional maize with associated low demand of VAM, and lack of markets for selling VAM(Reference Phorbee, Mulongo, Njoku, Maru and Munyua67). All barriers are summarised in Table 21.
Summary: OSP, HIB, VAC, VAM
The frequency of studies reporting barriers and facilitators aligned with each of the constructs and specific to each of the four crops and/or their products (OSP, HIB, VAC, VAM) is presented in Table S1.
Among the characteristics of a biofortified crop, relative advantage and complexity were common across all the crops as positively associated with adoption. More studies (n = 11, 26.8%) identified relative advantage as facilitators rather than barriers (n = 7, 17.1%), while complexity was reported more commonly as barriers (n = 8, 19.5%) than facilitators (n = 5, 12.2%). Within the outer setting domain, local conditions and external pressure were common factors behind adoption. External pressure, particularly those related to market, emerged as hindrance to adoption in nine studies (22.0%). Aspects of local conditions were more common as barriers being cited by approximately one-third studies (n = 14, 34.1%). Within the context of implementing organisation, the only construct common to more than one crop was available resources being cited as barriers by several studies (n = 6, 14.6%). Constructs under the domain of implementation process were virtually absent in the literature reporting studies on VAC, HIB and MAC. While at least two studies on OSP reported each of the six constructs (teaming, engaging, assessing context, planning, adapting, tailoring strategies, and reflecting and evaluating) included in Table 22, they were almost exclusively identified as facilitators of adoption. Engaging innovation recipients (n = 8, 19.5%) and tailoring strategies to improve implementation (n = 13, 31.7%) were most frequently reported implementation processes facilitating adoption. In contrast to implementing organisation and implementation process domains, “characteristics of individuals” domains were heavily coded. The opportunity construct was most frequently cited – both as barriers (n = 15, 36.6%) and as facilitators (n = 23, 56.1%) – among all twenty-four constructs reported in this review. We also found capability features of individuals to have mixed associations with adoption with thirteen (31.7%) and sixteen (39.0%) studies reporting negative and positive associations, respectively, with adoption. Unlike capability and opportunity, motivation as a factor of adoption was not common across all the crops. Adoption of only OSP and VAC was reported in the literature to be affected by aspects of motivation among individuals.
Table 22. Geographies of include crops and studies
Regions based on World Health Organization regions(101). HIB, high-iron beans; OSP, orange sweet potato; VAC, vitamin A cassava; VAM, vitamin A maize; AFR, African Region; AMR, Region of the Americas; SEAR, South-East Asian Region; EUR, European Region; EMR, Eastern Mediterranean Region; WPR, Western Pacific Region.
Discussion and conclusions
In this review, we synthesised forty-one studies to elucidate the barriers to and facilitators of adoption of conventionally biofortified crops. OSP represented nearly 70% of studies, followed by HIB and VAC (20% each), and one study in VAM, which is expected given the earlier roll-out of OSP in target regions. This is the first review to apply a comprehensive determinants framework (i.e. CFIR) from the implementation science field to identify factors associated with adoption of a biofortification intervention. We identified a wide array of barriers and enablers/facilitators to the adoption of biofortified crops and foods. For provitamin A-biofortified OSP, the most reported factors that promoted adoption were farmers having knowledge about importance, relative advantage, and efficient production and management practices of OSP, and the main barriers were lack of timely access or inability to maintain good-quality vines especially in drought-prone areas and remoteness to key places including markets. For HIB, the main factors that promoted adoption were farmers’ networking and high farming experience and the main barriers were lack of knowledge in the form of education and experience about HIB. For VAC, the main factors that promoted adoption were having awareness of the benefits of VAC and access to information, while the main barriers were poor road networks and inadequate processing and storage facilities. Finally, for VAM, factors identified were only barriers mainly related to low awareness about VAM among farmers and consumers in addition to concerns regarding yield, texture and aflatoxin contamination. Overall, agronomic and nutritional benefits of the biofortified crops, especially when farmers are aware of these relative advantages, were found to facilitate adoption. Conversely, production, storage and processing-related constraints arising from local environmental and market conditions coupled with inadequate agronomic insights of farmers limit their adoption.
A comprehensive understanding of adoption determinants is required and should act as an evidence base to inform the development and/or selection of breeding and scaling strategies. Arguably, the Consolidated Framework for Implementation Research (CFIR) that has been developed by Damschroder et al. is the most well developed and comprehensive among the determinant frameworks(Reference Damschroder, Aron, Keith, Kirsh, Alexander and Lowery86). By using a meta-theoretical determinant framework (i.e. CFIR) to capture adoption factors acting at multiple levels, we found adoption determinants of biofortified crops and their products to occur within the existing domains of the CFIR. This highlights that studies contributing to the evidence of this review explored a diverse array of aspects influencing adoption of biofortified crops. Kirk et al. studied how researchers utilise the CFIR in implementation studies and found that it is mostly employed to analyse what facilitates or hinders implementation (aligned with our goal)(Reference Kirk, Kelley, Yankey, Birken, Abadie and Damschroder87). Commonly cited determinant frameworks other than the CFIR are limiting in one way or other (e.g. do not explicitly address the relevance of end-users (farmers)) to be used as the synthesising architecture. As biofortified crops are an agricultural innovation with the goal of scaling up nationally rather than replicating a clinical innovation across healthcare facilities (i.e. the basis of developing the CFIR), it was difficult to delineate the CFIR’s implementing organisation and outer setting domains within the agriculture context. CFIR, being an organisation-centric framework, is often criticised for downplaying the central role of end users. However, CFIR has been revised on the basis of user feedback, and the version (i.e. CFIR 2.0)(Reference Damschroder, Reardon, Widerquist and Lowery33) we have employed in this review incorporates constructs based on Michie et al.’s(Reference Michie, van Stralen and West88) COM-B system that shape behaviour.
This comprehensive review, being the first of its kind to assess drivers of adoption among producers, to our knowledge, limited us in comparing the adoption factors with previous literature. However, the constructs and domains identified agree with factors previously reported to affect the implementation of agriculture-nutrition interventions in general(Reference Fiorella, Erin, Lia and Rona89–Reference Di Prima, Wright, Sharma, Syurina and Broerse91). Although factors of adoption identified in this review were summarised by using only 50% (twenty-four out of the forty-eight) of the CFIR constructs, this is not a perceived limitation of CFIR. Rather, this is possibly a reflection of the nascent state of implementation research and authors not reporting or incorporating many of the implementation aspects in their research. Other studies utilising the CFIR framework have also not utilised all the domains/constructs(Reference Warren, Ndwiga and Sripad92,Reference Huang, Nakigudde and Rhule93) .
We found a relatively uneven distribution of factors across domains with relatively low representation from the inner setting (i.e. characteristics of the implementing organisation) and implementation processes domains. That we included in our review only those factors drawn directly from empirical evidence explains this discrepancy in the frequency of factors across domains. Another related reason behind the relatively infrequent reporting of inner setting factors and implementation processes is the low number of studies reporting process evaluations or implementation science research (i.e. those testing different implementation strategies with exclusive focus on improving implementation outcomes (e.g. adoption)). Most of the studies included in this review covered the initial stages of implementation that included support from research teams. Once this support is withdrawn and implementation progresses to advanced stages, e.g. real-world conditions, determinants from the inner setting (e.g. resources available to the implementing organisation) and implementation processes (i.e. activities and strategies used to implement the intervention) domains may start acting upon the adoption process(Reference Aarons, Sklar, Mustanski, Benbow and Brown94).
For example, successful implementation of the biofortified OSP in Kenya has been largely attributed by Cole et al.(Reference Cole, Levin and Loechl54) to the additional time, funding and other core resources available to the project team from project partners and donors during the implementation. Although the majority of studies lacked exclusive focus on these two domains, authors frequently mentioned potential association of these domains as recommendations to be considered before future development of biofortified crop project. However, our sole reliance on data-derived factors instead of authors’ subjective project experiences adds strength to the evidence presented in this review. In contrast to inner setting and implementation processes domains, we found a relatively common association between adoption and characteristics of individuals as reflected in the frequency with which studies reported factors that fell under capability, opportunity and motivation. This finding reflects the importance of not overlooking the individual characteristics when developing implementation strategies to promote adoption of biofortified crops and supports Powell et al.(Reference Powell, Beidas and Lewis95), who emphasised assessing the characteristics and preferences of receivers and deliverers of interventions before their delivery.
Although biofortified crops have been shown to have superior characteristics to the many non-biofortified varieties currently cultivated by farmers, factors within the intervention characteristics domain were not exclusively identified as facilitators. Empirical findings from the studies in this review confirm that adoption is contingent upon a comprehensive list of agronomic and sensory attributes of the biofortified crops and their products. For VAM, HIB and VAC, economic demands of higher yield should also be addressed in addition to palatability standards. For OSP, more attention should be paid especially to sensory attributes including dry matter content (sugar and starch composition), skin colour and root shape. There were far more barriers than facilitators identified in the inner setting and outer setting domains. For example, barriers including failure to timely supply adequate planting material and human resource inadequacy were reported across all the fortified crops in the inner setting domain. This is not surprising given the underfinanced public sector extension services in countries where the studies included in this review took place. Similarly, commonly reported barriers in the outer setting including physical remoteness, seasonality (especially dry season) and lack of markets and agro-products storage and processing facilities mostly fell beyond the control of implementing organisations. However, implementation processes that recognised local conditions during planning(Reference de Brauw, Eozenou, Gilligan, Hotz, Kumar and Meenakshi52–Reference Cole, Levin and Loechl54), made choices that fit with existing conditions(Reference Vaiknoras, Larochelle, Birol, Asare-Marfo and Herrington11,Reference Vaiknoras, Larochelle, Birol, Asare-Marfo and Herrington25,Reference Gonzalez, Perez, Estevao Cardoso, Andrade and Johnson73) , allowed flexibility in project implementation from planned interventions in response to obstacles identified through ongoing monitoring and evaluation(Reference Shikuku, Okello, Wambugu, Sindi, Low and McEwan46,Reference Cole, Levin and Loechl54,Reference Arimond, Ball and Bechoff71,Reference Muthini80,Reference Muthini, Nzuma and Nyikal81) , and provided combined nutrition-agricultural training to generate demand and compensate for limited agronomic know-how(Reference Adekambi, Okello, Abidin and Carey17,Reference Shikuku, Okello, Wambugu, Sindi, Low and McEwan46,Reference Behrman51–Reference Cole, Levin and Loechl54,Reference Adekambi, Okello and Rajendran57,Reference Okello, Muoki, Kwikiriza, Wambaya and Heck59,Reference Maredia, Porter, Jin and Farris69,Reference Ogutu, Fongar and Gödecke82,Reference Ogutu, Fongar and Gödecke83) allowed to overcome these barriers. This explains why the implementation process domain (especially the engaging construct) was predominantly coded as facilitator in sharp contrast to the inner setting and outer setting domains.
The determinants identified in this review should be interpreted not as the ones with the strongest association with adoption of biofortified crops, but the ones which were studied and reported by the authors of the included studies. This review only included determinants that were directly supported by the quantitative data or qualitative statement (self-report) describing the real experience of the study participants (i.e. no ‘hypothetical’ barriers, such as authors’ extrapolation of probable subsequent events acting as a barrier or facilitator, were included). Moreover ‘adoption’ was measured differently across studies. It is likely that adoption happens in stages, rather than being a discrete, one-off event. For example, Rogers(Reference Rogers96) describes at least four stages of adoption (i.e. intention, initial decision, action to try or use, continued use). A major limitation of this review is that we are unable to explain how the identified factors are causally linked to adoption (or the stage of adoption) and their degree of influence.
Future studies on adoption of biofortified crops should consider the stages of adoption and identify the determinants across these stages(Reference Aarons, Hurlburt and Horwitz97). Researchers should also synthesise and recommend various implementation strategies based on identified barriers using the CFIR-Expert Recommendations for Implementing Change (ERIC) implementation strategy matching tool(Reference Waltz, Powell, Fernández, Abadie and Damschroder98). It is imperative that the future studies adequately report the different characteristics of the implementation strategies (e.g. action, dose and duration) employed to address adoption following published recommendations (e.g. Proctor et al.(Reference Proctor, Powell and McMillen99)). Future research on adoption determinants should also focus on the understudied crops, such as VAM, for which we found only one study, and understudied geographies (most included studies were done in the African Region as defined by the World Health Organization (Table 22)), mainly focused on OSP, due to its being rolled out there first as discussed previously(Reference Huey, Krisher and Bhargava5). As biofortified crops have been rolled out in thirty out of over 120 countries tested as of 2020(100), across all the WHO regions (Table 22), excluding the European Region, research is needed to inform adoption determinants in these varied settings, accounting for not only differences in adoption but also in delivery and diffusion, considering the diversity in crop type: vegetatively propagated crops like cassava and sweet potato, hybrid grain crops like maize and open pollinated crops like beans and pearl millet. Additionally, as non-biofortified wheat and rice are commonly consumed, it will be useful in a future study to compare the adoption determinants between these and biofortified crops. Finally, understanding the facilitators and barriers to adoption of crops biofortified through other means, including either agronomic or genetic engineering – which was beyond the scope of the current review – remains a research gap necessary to assess in a future review, given the varying appropriateness and applicability of different biofortification methods across different settings and geographies(Reference Talsma, Melse-Boonstra and Brouwer30).
Conclusion
This review analysed forty-one studies reporting on aspects of adopting biofortified crops and their products and provides a set of factors that enables or constrains their across twenty-four constructs and five domains of a meta-theoretical determinant framework from implementation science. The barriers identified in this review are a useful starting point for implementation researchers to move beyond the currently dominant effectiveness research towards testing implementation strategies for scaling up biofortified crops and continuous optimisation of ongoing projects promoting adoption of biofortified crops. This review has revealed that scaling up research would be well served by studies that compare different delivery models (e.g. free versus subsidised distribution of planting materials), delivery platforms (e.g. commercial versus public to address the underfinanced inner setting) and delivery channels (e.g. facility-based versus community-based to address physical remoteness) and implementation processes to adoption of biofortified crops. Also, in-depth studies should be conducted in different contexts to delineate the interactive influences of different CFIR constructs including biological and cultural characteristics of individuals (e.g. age and sex) and economic factors (e.g. wealth and income) on adoption. Finally, as the current literature is heavily biased towards OSP, one of the first crops to be developed and rolled out, future works should involve examining adoption of other types of biofortified crops as their dissemination increases, such as VAM, iron and zinc pearl millet, zinc wheat and iron lentils, as well as VAC and HIB.
Financial support
The authors thank BMZ and the Netherlands Ministry of Foreign Affairs for the Commercialisation of Biofortified Crops programme co-led by the Global Alliance for Improved Nutrition (GAIN) and HarvestPlus. SLH was supported by the NIH under award 5T32HD087137.
Competing interests
S.M. holds equity in a diagnostic start-up focused on developing assays for low-cost and point-of-care measurement of certain nutrients from a drop of blood using results from his research as a faculty member at Cornell University. GAIN is a not-for-profit organisation supporting and promoting biofortification programmes; V.M.F., M.N.N.M., E.M. and A.M.N. are employees of GAIN. All other authors declare that they have no known conflicts of interest.
Author contributions
V.M.F., M.N.N.M., E.M. and A.M.N. from GAIN were involved in conceptualisation, interpretation of the results, and reviews and edits of the manuscript, but were not directly involved in conducting the search, deciding on study/report eligibility, data extraction and statistical analyses. S.L.H., V.M.F., M.N.N.M., E.M., A.M.N. and S.M. conceptualised the review. S.L.H. conducted the searches. S.L.H., S.I., N.H.M., E.M.K. and A.B. screened the records for eligibility and extracted data. S.I. and S.L.H. synthesised the results. All authors were involved in interpretation of the results. S.L.H. and S.I. wrote the first draft of this manuscript. All authors critically reviewed and revised the manuscript. The German Federal Ministry of Economic Cooperation and Development (BMZ) and the Netherlands Ministry of Foreign Affairs as the funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript.
Open access
