We conducted a community-based cross-sectional study using mixed methods in the Butezi health district, former Ruyigi Province, eastern Burundi. Among the four districts in Ruyigi, Butezi has the highest rates of chronic malnutrition among children. The district, which comprises two communes, Butezi and Nyabitsinda, is predominantly rural and relies mainly on subsistence and small-scale livestock farming for income. The map below shows the district’s location within Ruyigi Province (Figure 1).

The source population included all children aged 24 to 59 months residing in the Butezi health district. For the quantitative part of the study, we selected children within this age group whose mothers or primary caregivers lived in the designated hills and agreed to take part. Children whose caregivers could not be contacted despite several attempts or who were ill during the interview were excluded.

We determined the sample size using the proportion formula n = p (1 − p)Zα²/i², where Z= 1.96 (95% confidence level), with an estimated prevalence of iron-fortified bean consumption (p) of 50% and a margin of error (i) of 0.03. This yielded a total of 544 children. Adding 10% to account for potential non-response

Figure 1. Butezi health district. From the Burundian sanitary health Card, 2016.

increased this to 598, which we rounded up to a final target of 602.

This total was then allocated proportionally across the hills in both communes, based on the estimated number of eligible children in each hill, ensuring that no hill would be under- or overrepresented relative to its population.

We randomly selected 14 of the 54 hills in the 2 communes of the Butezi health district: 4 from Butezi commune and 10 from Butaganzwa commune. Within each selected hill, eligible households were identified using household lists compiled by community health workers, and children were selected through simple random sampling until the hill’s quota was reached. In households with more than 1 child under 5, we randomly selected one child.

This two-stage approach, consisting of proportional allocation across the hills followed by random selection of children within each hill, ensured that the sample was representative of the study area and made the selection process transparent and easy to replicate.

We designed a structured questionnaire, informed by a literature review of similar studies, covering sociodemographic characteristics, household and child information, and knowledge, attitudes, and practices regarding iron-fortified beans. A supplementary semi-structured interview guide was developed for the qualitative component to explore perceptions of anemia as a health problem, the perceived role of iron-fortified beans in its prevention, and the availability and accessibility of these beans in the community.

Nutrition experts validated both tools and pilot-tested them in the Ruyigi Health District, adjacent to Butezi but outside the study area, to ensure that their wording was clearly understood before fieldwork began and to avoid influencing the population that would later be recruited for the main survey.

The interviewers were trained in data collection tools, the study’s objectives, and procedures to ensure data quality; daily supervision sessions were held throughout the fieldwork. Before each interview, the study’s objective was explained to participants, and those who agreed to participate signed a consent form. Quantitative interviews were conducted with mothers or primary caregivers, randomly selected from the list of eligible children in the study areas; each interview lasted approximately 20 minutes and was conducted in Kirundi, the local language.

For the qualitative component, key informants were purposively selected from seven categories of actors involved at different levels in the production, distribution, promotion, or consumption of iron-biofortified beans in the Butezi health district: caregivers of children aged 24 - 59 months (n = 28), community health workers (n = 14), agricultural extension officers (n = 14), health facility staff, including health center heads and hospital-based providers (n = 6), the district’s Chief Medical Officer (n = 1), the provincial Director of Agriculture and Livestock (n = 1), and bean seed multipliers (n = 2). This resulted in 66 semi-structured interviews, one per participant.

This targeted, multi-category sampling approach was designed to gather insights from all key stakeholders involved in adopting iron-biofortified beans. It included provincial agricultural authorities and seed multipliers at the upstream level, along with community health workers, agricultural extension agents, and caregivers as end users. The goal was to capture diverse viewpoints across and within these groups, rather than focusing on a single perspective.

Within each category, interviews continued until no new themes emerged, indicating saturation. This was most evident among the groups of mothers, community health workers, and agricultural extension officers, which had the highest respondent counts. Single-informant groups were included to capture unique institutional perspectives, as only one individual holds each role in the study area. The two seed multipliers interviewed represent nearly all seed multipliers active in the study communes. Interviews were conducted in Kirundi.

Data were collected over 10 days: quantitative data from November 23 to 28, 2023, and qualitative data from November 30 to December 5, 2023.

For the “knowledge, attitudes, and practices” (KAP) aspects of our study, we used the following variables: knowledge of the type of iron-fortified beans, source of information about iron-fortified beans, perceived importance of consuming iron-fortified beans, types of iron-fortified beans consumed, frequency of bean consumption, taste of iron-fortified beans, and side effects of iron-fortified beans. The type of iron-biofortified beans reported as known and consumed was caregiver-reported; biofortification status was not independently verified.

For the qualitative component, the questions and topics were designed to understand the factors limiting the consumption of iron-fortified beans.

The completeness and accuracy of the data were verified daily. In Stata, quantitative data were recorded and analyzed; categorical variables were summarized as proportions, and continuous variables were described using measures of central tendency and spread.

The interviews were transcribed in Kirundi, translated into English for analysis, and a manual thematic analysis was conducted in Excel to identify recurring sub-themes across the informant categories. The interviews were coded by five encoders using a hybrid thematic approach that combined deductive and inductive elements. For each main theme derived from the interview guide, sub-themes were identified through a thorough reading of the transcripts. For each sub-theme, recurring keywords were extracted from the corpus to build a hierarchical coding grid (theme, sub-theme, keywords). The five encoders worked independently before comparing their respective grids to assess coding consistency. Disagreements were resolved through collective discussion until consensus was reached. The final categories were validated by the entire team before the qualitative results were integrated with the quantitative results.

The abbreviations are defined as follows: ASC for community health worker; MCD for head of the medical district; DPAE for provincial director of agriculture and livestock; and ISABU for the Institute of Agricultural Sciences in Burundi. In the results, citations are followed by an anonymized participant code that indicates the informant’s category, interview number, and source hill (e.g., mother, community health worker, agricultural extension officer). These codes were assigned during data encoding to protect participant privacy.

The National Ethics Committee of Burundi (CNE/14/2023) approved the study protocol. Before fieldwork began, we met with local administrative and health authorities in the study area to present the project and obtain their support.

Before any data were collected, participants received a consent form and an explanation of the research’s benefits. Only participants who consented to the study were invited to participate in data collection. In data management, we assigned a code to each participant to ensure anonymity and protect personal data.

Written informed consent was obtained from the mother or primary caregiver for each participating child before data collection. Participants were informed of the study’s purpose, procedures, and their right to withdraw at any time without penalty. For the qualitative component, verbal consent to record interviews was also obtained from key informants.

3.1.1. Description of the Sample

The study population consisted of 602 children aged 24 to 59 months. Table 1 below describes the sociodemographic characteristics:

Table 1. Distribution of respondents by sociodemographic and economic characteristics (N = 602).

In our sample of 602 children, 251 (41.69%) were between 24 and 35 months old, and more than half were male (50.50%). The majority (60.80%) were the third child in their family.

Based on the characteristics of the children’s families, 343 children came from families with two or more children under 5, representing 56.98% of our sample.

Most mothers were farmers (98.50%), had a primary school education (62.46%), and were married (96.01%). The 20 - 34 age group was the largest among mothers, accounting for 61.79% of cases. The majority of the fathers of the surveyed children were farmers (94.35% of cases), and the head of the household was male in 94.19% of cases.

3.1.2. Knowledge of Iron-Biofortified Beans

The following Table 2 shows the results regarding knowledge of iron-fortified beans:

Table 2. Distribution of respondents by knowledge of iron biofortification in beans (N = 602).

Based on knowledge of bean biofortification, only 17 of 602 respondents reported hearing of biofortified beans; all 17 (100%) cited at least one type of biofortified bean they knew. The most frequently cited sources of information on bean biofortification were health care facilities, agriculture monitors, seed multipliers, and community health workers.

3.1.3. Consumption of Iron-Biofortified Beans

In terms of practices, they are explained in more detail in Table 3 below.

Table 3. Distribution of respondents according to practices related to beans and biofortified beans (N = 602).

More than half of the children in our sample (51.66%) had eaten beans the day before the survey. Only 2.82% of caregivers reported that their children had eaten biofortified beans. Nearly half of the children reported eating beans three times a week. Mothers interviewed reported that their children generally like beans (96%). Most bean seeds came from markets.

There is a contrast between common and non-biofortified iron beans: beans, as a food category, are accepted and consumed by children, with 96.01% of respondents saying they like to eat them and more than half having eaten beans the day before; however, biofortified varieties are still rarely seen in diets (2.82%).

3.1.4. Attitudes Related to Iron-Biofortified Beans

The following Table 4 shows the participants’ perceptions and attitudes toward iron-biofortified beans:

Table 4. Distribution of respondents by attitudes about biofortification of beans with Iron (N = 602).

Based on respondents’ attitudes toward biofortification, the vast majority (97.18%) are unaware of both its taste and its side effects.

3.2.1. Factors Limiting the Consumption of Iron-Fortified Beans

1)Perceptions of the prevalence of childhood anemia, its causes, and ways to address it

Most participants agreed that anemia among children under 5 is a reality in their area, as one of the interviewed mothers attested:

“Anemia exists in our community. Even my child has had to deal with it. It really does exist here.” BMZA1

The causes cited included poverty and inadequate nutrition, as highlighted by a community health worker who took part in the study:

“In my opinion, the main cause is a lack of proper nutrition, which leads to a gradual decrease in the child’s blood levels until full-blown anemia sets in.” ANBI1

This perception was echoed at the health system level, as follows:

“Infantile anemia is a fact in our health district. So, when children suffer from malaria, their hemoglobin levels decrease, and sometimes we transfuse them. This means that even with this malaria, they had a low rate of hemoglobin, and malaria highlights the already pre-existing low rate.” MDB1

For most participants, the foods that help combat anemia are the three major food groups: carbohydrates, fats, and proteins. Only a few specifically mentioned beans as an iron-rich food, as in the following quote:

“I could mention beans, vegetables, and oils as foods that can help reduce anemia.” AMRA1

These accounts reveal a generally correct understanding of the link between diet and the occurrence of anemia. Beans were not spontaneously identified as a full-fledged source of iron, but were instead subsumed into a broad and undifferentiated category of foods that help.

This gap between general awareness of anemia related to malnutrition and a lack of specific knowledge about iron-rich foods mirrors the patterns already described in the complementary quantitative survey conducted in the same district, where iron intake from dense foods of animal origin, such as meat and eggs, is low.

This suggests that nutrition education messages in this context may need to go beyond generic dietary advice and adopt more targeted, diet-specific communication about iron sources.

2)Awareness of iron-biofortified beans among key informants

Most participants were unfamiliar with this type of bean, including community health workers. One of them put it this way:

“No, I don’t know these beans, but I hope that among the beans available in our area, they are included as well.” ARTA1.

This mother also emphasized: “These beans don’t exist where we live.” BMGE2

Taken together, these testimonies illustrate a form of “invisibility” of biofortified iron beans in the community’s food landscape: even health professionals and community health workers have described these varieties as absent, unknown, or hypothetical rather than as a concrete option for production or consumption, suggesting that the seed distribution chain has not yet reached the region in a structured or continuous manner.

This is consistent with the low proportion of respondents able to spontaneously name a biofortified variety in the quantitative component (2.82%), and confirms that the barrier at this stage is exposure rather than non-acceptability.

3)Structural and economic barriers to seed access and cultivation of iron-biofortified beans

The main reasons cited for the low consumption of biofortified beans were a lack of awareness and limited availability of seeds and manure. As one participant noted: “People aren’t well-informed about this type of bean; the cost of seeds and manure also hinders bean cultivation.” MUNBIC66.

Beyond the lack of information, participants highlighted concrete structural constraints: the cost of quality seed and manure needed to support their cultivation has repeatedly discouraged farmers from adopting new bean varieties, even when interest exists.

This finding aligns with the overall results of the quantitative survey, in which 88.87% of household bean seeds were purchased from regular markets rather than from specialized propagation or distribution channels (1.16%). This indicates that most families lack a readily accessible supply route for biofortified seeds.

The seed multipliers have also reported this point of view: « One of the challenges we face is the cost related to these beans, and also, once the beans are available, we have difficulties with manure. » MSN2

3.2.2. Stakeholders’ Suggestions and Perspectives on Increasing Consumption

The main suggestions from the participants were to increase awareness of iron-biofortified beans and their availability. As one participant said: “What I can ask for is greater availability of seeds and more awareness sessions to encourage the practice of consuming these beans.” AKGAC1

The provincial director of agriculture and livestock spoke about the need for more awareness among all stakeholders and greater accessibility to these bean varieties: “Awareness must be raised about this type of bean, not only among farmers, but also among all stakeholders involved in child nutrition. We have platforms where we meet; we need to spread these messages on these beans so that we can speak the same language in the field and raise awareness. And on our side, if we have a partner who wants to get involved in agriculture, we will tell them to mobilize to make available these types of beans that help combat anemia.” DPR1

This call for greater seed availability, coupled with awareness raising, reflects a community that is not resistant to innovation but rather constrained by a lack of practical opportunities to adopt it.

This suggests that demand responsiveness already exists and that the programmatic effort would be more effectively directed toward providing seeds and secondary communications, expanding local access to certified biofortified seeds, and equipping community health workers and agricultural agents with simple, visual, and repetitive messages.

Our results suggest that a single intervention will not be sufficient to increase the consumption of biofortified iron beans in the Butezi district. Given that rural farm households in the region are heterogeneous in terms of food and agricultural decision-making structures and modes [42], these interventions should be designed to reach not only mothers but also other household members who influence food and seed choices.

A successful approach would involve: 1) connecting with the community through existing channels, such as community health workers, agricultural extension agents, and prenatal care programs that already reach most mothers, while also expanding mass communication; 2) improving local seed multiplication and market distribution to ensure biofortified varieties are as accessible as traditional beans; 3) providing support to reduce the cost of agricultural inputs for these varieties; and 4) incorporating specific messages about iron and food into current nutrition education sessions, moving from general ‘good nutrition’ advice to explicit guidance on iron-rich foods. Leveraging these existing platforms is likely more cost-effective than establishing new communication strategies.

The use of a mixed-methods approach enabled the integration of both quantitative and qualitative data, enhancing our understanding of the factors affecting biofortified bean consumption. Participants might have inaccurately reported not consuming biofortified beans due to limited knowledge. Additionally, because the study was conducted in only one health district, the findings may not apply to other regions within the country.

This mixed-methods design is a particular strength of the study: the consistency of findings across markedly different participant groups (caregivers, community health workers, agricultural agents, and health professionals) reinforces confidence in the conclusion that limited awareness and access, rather than poor acceptability, are the principal barriers to adopting iron-biofortified beans. In addition, the cross-sectional design of the quantitative component limits our ability to establish causal relationships between awareness and consumption, and self-reported information on bean consumption may have been subject to recall or social desirability bias.

Furthermore, because knowledge of biofortified beans was low in our sample, the taste and side effect estimates are based on a very small number of respondents (n = 17). They should be interpreted as indicative rather than representative. Finally, in the qualitative survey, although opinions were received from various key informants (mothers, health workers, and agricultural agents), social desirability may have influenced certain responses, particularly regarding the perceived causes of anemia.

Further studies are needed to assess the effect of awareness-raising interventions on household adoption of biofortified beans. Longitudinal or experimental studies would also help examine their impact on iron status and the prevalence of anemia among children under 5 years of age. Research on local names for iron-bean varieties biofortified in the region, also comparing them with the names of those promoted by the Institute of Agronomic Sciences of Burundi, could help create a cartography of the beans available there and contribute to awareness of their consumption. Finally, research into the production, distribution, and marketing systems for biofortified seeds could help identify the constraints that limit their availability and use in rural communities.

Finally, given the low level of knowledge about biofortified iron beans found in our research, a major short-term research priority would be a formative evaluation of community outreach strategies: for example, assessing the effectiveness of messages delivered by community health workers and agricultural monitors, and comparing their effectiveness with that of media campaigns in this specific rural context, before large-scale investments are made.