A scoping review of feed interventions and livelihoods of small-scale livestock keepers

In low- and middle-income countries (LMICs), one billion people derive part of their livelihoods from livestock, which also provide key nutrients in contexts with high child stunting. Demand for animal-source foods is growing due to urbanization and rising incomes, yet livestock productivity in LMICs remains far below that of high-income regions (for example, average cow milk yields in Western Europe are about 20 times higher than in Eastern Africa). Improving per-animal productivity is also critical for environmental sustainability, as low yields drive high greenhouse gas emissions intensity in sub-Saharan Africa. Multiple constraints limit productivity, including genetics, veterinary access, and input scarcity, but insufficient access to adequate, quality feed is widely recognized as the main constraint. Numerous initiatives have promoted improved feed options (such as improved grasses and legumes, multipurpose trees, enhancing the intake and nutritive value of crop residues, and preserving fresh feed), yet there is limited rigorous evidence on their uptake by smallholders, related productivity gains, and livelihood impacts. This scoping review aims to assess the availability and quality of evidence on adoption of improved ruminant feed options by small-scale producers, the effects on livestock productivity, and the extent of consequent improvements in smallholder livelihoods, to identify promising strategies for improving feeding and livelihoods.

The field historically emphasized technical feeding trials and agronomic performance, with limited assessment of farmer uptake and livelihood impacts. Despite decades of research and development to deliver high-quality feed options, evidence of effectiveness in smallholder systems is sparse and often anecdotal. Earlier literature is dominated by trials rather than adoption studies. There is also a mismatch between research focus and on-farm importance: forages and agroforestry receive substantial attention, whereas crop residues—ubiquitous and often the largest feed component in tropical smallholder systems—are comparatively under-studied and lack a clear disciplinary home.

The scoping review followed PRISMA-SCR guidance with a preregistered protocol (Open Science Framework, 5 June 2019, https://osf.io/6ywht/). A comprehensive search strategy (developed and peer-reviewed by librarians) covered key concepts: feed improvement/conservation, small-scale producers/pastoralists, LMICs in Africa/Asia/Latin America, and innovation/adoption indicators. On 5 June 2019, searches were conducted in CAB Abstracts, Web of Science Core Collection, Scopus, and ProQuest Dissertations and Theses Global, and 20 grey literature sources (including AGRIS, AgEcon Search, CEE, Campbell, Cochrane, IFAD, IIED, World Bank, WHO, UNEP, WFP, and others). Results were deduplicated; 22,981 unique records remained. Screening proceeded in three blinded stages with two reviewers and a third as tie-breaker: (a) an ML-assisted title screening using metadata excluded 12,195, leaving 10,786; (b) title/abstract screening excluded 10,243, leaving 543; (c) full-text screening excluded 470 for reasons including lack of adoption or effects analysis (n=257), no improved feed options (94), wrong study type (74), inaccessible (30), wrong language (9), only industrial byproducts/concentrates (4), wrong country (2). Seventy-three studies met inclusion criteria: primary empirical research on small-scale or agro-pastoral ruminant keepers in LMICs, analyzing adoption of improved feed options and/or effects on productivity and livelihoods, in English/French/Spanish/German. Data extraction captured study characteristics; intervention types; duration; methods; outcomes on adoption, productivity, and livelihoods; and adoption drivers/constraints. Quality assessment rated (1) study methodology, (2) justification of methodology, and (3) overall quality on ordinal scales; two assessors scored each paper, with scores averaged. A subset of 25 studies that explicitly analyzed adoption drivers/constraints was re-examined to categorize reported factors. Descriptive synthesis summarized evidence across intervention types (planted forages, agroforestry, crop residues) and along the impact pathway (adoption → productivity → livelihoods).

- Evidence base: From 22,981 records, 73 studies met criteria; only 6 examined the full pathway from adoption through productivity to livelihoods. Of the 73, 58 reported adoption, 19 productivity outcomes, and 22 livelihood outcomes. - Geographic and systems coverage: Nearly half the studies were from East Africa/Horn of Africa; Southeast Asia was second. Mixed crop–livestock systems dominated (53 studies); pastoral/agro-pastoral systems were less represented (13 studies). Many studies were quantitative; some used mixed methods; few were qualitative, and none of the productivity studies were purely qualitative. - Intervention types: 53 studies on planted forages, 26 on agroforestry (multipurpose trees), and 7 on crop residues. Most studies assessed a single intervention type. The research focus is skewed toward forages/trees despite crop residues being central to diets in smallholder systems. - Adoption ranges (studies with usable data): forages 0–90% (32 studies), agroforestry 8–87% (11 studies), crop residues 20–86% (3 studies). Adoption was more frequently reported than productivity. - Productivity effects: Only nine papers across interventions provided sufficient data; reported changes ranged from 7% to 61% (milk yield, weight gain, body condition, herd/flock growth). Evidence linking crop residue improvements to quantified productivity change was scarce. - Livelihood effects: 22 papers reported livelihood outcomes; 14 quantified impacts. Reported ranges: household income change 6–285% (5 forage, 3 agroforestry papers), gross margin increases 58–519% (3 forage, 1 agroforestry), and labour/workload reductions −24% to −70% (5 forage papers). Effects varied widely by context and study design. - Adoption drivers (most frequently cited among 25 studies): farmer experience/education (10 papers); expected productivity or income gains (8); access to extension/training (8); labour availability (7); good market access (6); also noted: access to credit/off-farm income, market orientation, group membership/social pressure, land scarcity; soil improvement was infrequently cited as a motive. - Adoption constraints: increased labour requirements (6 papers); low perceived net benefit (4); limited access to technology/inputs (e.g., seed/planting material) and weak forage seed systems (4); technology complexity (4); competition with other land uses (4).

The review reveals a paucity of rigorous studies that trace the full impact pathway from adoption of feed interventions to animal productivity and household livelihoods in smallholder LMIC settings. The evidence base is biased toward technical feeding trials and toward interventions in planted forages and agroforestry, with comparatively little on crop residues despite their central role in diets. Large variability in reported outcomes complicates comparisons across interventions, driven by heterogeneity in the specific technologies assessed, study designs (e.g., project-linked versus independent evaluations), time horizons (with likely S-curve dynamics), and strong site-specific biophysical and socioeconomic conditions. Adoption and impact are influenced by technology management demands (skills, extension, input access), alignment with farmer objectives (perceived and realized benefits), and availability and opportunity costs of land and labour. These findings suggest that successful scaling of feed interventions must consider not only technical performance but also enabling services, market access, and whole-farm trade-offs.

This scoping review highlights a limited and uneven evidence base on the adoption of improved ruminant feed options by small-scale livestock keepers in LMICs and on their effects on productivity and livelihoods. Only six studies followed the full adoption–productivity–livelihood pathway, and overall study quality was average to low. Reported adoption and impact ranges are wide and context-dependent. To improve outcomes, future efforts should: (1) invest in rigorous longitudinal and quasi-experimental evaluations that measure adoption, animal performance, and household-level impacts; (2) expand research on crop residue improvements and food–feed crop breeding; (3) address adoption barriers by strengthening extension/training for knowledge-intensive technologies and by developing reliable forage seed and planting material systems; (4) target semi-commercial smallholders with resources and incentives to invest, and consider whole-farm land and labour trade-offs; and (5) tailor interventions to local biophysical and market contexts. These priorities, along with the recommendations in Box 3, can enhance the uptake and effectiveness of feed technologies and their contributions to smallholder livelihoods.

- Limited number of eligible studies (73 out of 22,981 records), with only six covering the full adoption–productivity–livelihood pathway. - Average to low methodological quality across many studies; heterogeneity in designs, indicators, and time horizons precluded meta-analysis and complicates cross-study comparisons. - Potential bias toward project-linked studies with short lifespans, limiting observation of longer-term productivity and herd dynamics. - Geographic and systems skew, with underrepresentation of pastoral systems and some regions (e.g., West Africa) and a dominance of mixed systems. - Intervention skew toward forages and trees; sparse quantified evidence on crop residue improvements. - Site-specific contexts and differing measurement protocols for productivity (e.g., species-specific metrics, seasonal vs annual measures) limit generalizability.