Determinants of knowledge about aflatoxin and fumonisin contamination in sorghum and postharvest practices among caregivers of children aged 6–59 months in Kerio Valley, Kenya

Abstract Stunting among children under five years old is still a problem in many developing countries including Kenya. However, there is little information linking stunting with mycotoxin contamination of complementary foods. The aim of this study was to assess knowledge about aflatoxin and fumonisin contamination in sorghum alongside postharvest handling and storage practices among caregivers of children under five years old in Kerio Valley, Kenya. A cross‐sectional study was conducted to obtain data from 353 randomly selected caregivers of children aged 6–59 months. Qualitative data were obtained through Focus Group Discussions and Key Informant Interviews. Overall, majority of the caregivers of young children had poor knowledge (61.8%) about mycotoxin contamination of food, and poor postharvest handling and storage practices (74.5%). The caregiver's knowledge about mycotoxins was significantly associated with age [(AOR=4.629, (95% Cl: 2.530–8.472), p < .001], education level [(AOR=0.275, (95% Cl: 0.088–0.434), p = .001], marital status [(AOR=15.187, (95% Cl: 1.830–126.007), p = .012], and household monthly income [(AOR=2.623, (95% Cl: 1.550–4.439), p < 0,001]. Furthermore, the caregiver's age [(AOR=3.845, (95% Cl: 1.558–9.490), p = .003], education level [(AOR=0.196, (95% Cl: 0.088–0.434), p < .001], monthly income [(AOR=3.291, (95% Cl: 1.550–6.986), p = .002], and knowledge on mycotoxin contamination of sorghum [AOR, 5.428 (95% Cl: 2.855–10.319), p < .001] were significantly associated with postharvest handling and storage practices except for marital status [AOR, 3.579 (95% Cl: 0.403–31.775), p = .252]. In conclusion, caregivers of young children had poor knowledge about mycotoxin contamination of complementary foods and suboptimal postharvest handling and storage practices of sorghum. This increases the risk of mycotoxin exposure to young children and necessitates mitigation measures including sensitization campaigns and social behavior change communication.


| INTRODUC TI ON
Mycotoxins are harmful substances produced by some species of fungi that contaminate food material when stored in conditions that favor fungal growth (Obonyo & Salano, 2018). The most significant mycotoxins found in cereals are aflatoxins and fumonisins due to their ubiquitous nature as well as their acute and chronic effects causing death (Probst et al., 2007) or stunting (Knipstein et al., 2015), respectively. There are studies that reported contamination of cereal-based complementary foods by diverse mycotoxins (Kimanya et al., 2014;Ojuri et al., 2018). This could explain why dietary exposure to mycotoxins is identified as a major cause of undernutrition, morbidity and mortality among children (Paudyal et al., 2017). Children are more sensitive to mycotoxin exposure than the rest of the population due to their underdeveloped immune system, lower body weight, and less acidic stomachs among other factors (Magoha et al., 2014).
In Kenya, sorghum is an important staple food crop for many households especially in the arid and semi-arid lands (ASALs) (Orr et al., 2016) due to its adaptability to harsh environmental conditions. In this regard, the Kenyan government has continued to develop and promote the production of improved sorghum varieties as a means to improving income and food security (Government of Kenya (GoK), 2018). The Kerio Valley, a semi-arid area, with unreliable rainfall, chronically experiences food and nutrition insecurity (Elgeyo Marakwet County, 2018) and is currently leading or ranks poorly in the country as far as child nutrition is concerned. A recent study (Kipyego & Mugalavai, 2019) (Kiome et al., 2016).
It was anticipated that the child nutrition outcomes would improve after the interventions. Therefore, the persistent undernutrition rates necessitated the assessment of underlying factors that contribute to stunting. This is especially important because caregivers of young children from Kerio Valley use sorghum as a sole or major component of complementary foods (in form of porridge) as a common practice (Kiome et al., 2016).
In addition, adequate evidence from neighboring counties (e.g., Nandi County, Kenya) indicates that sorghum is susceptible to aflatoxin and fumonisin contamination at levels above regulatory limits (Sirma et al., 2015). It is thought that co-occurrence of aflatoxins and fumonisins in complementary foods contributes to stunting in an additive manner (World Health Organization (WHO), 2018) possibly through the toxin-induced liver and intestinal injury (Knipstein et al., 2015), or impaired immune system function (Leroy et al., 2015). Therefore, this study investigated the factors that determine knowledge about aflatoxin and fumonisin contamination in sorghum and postharvest practices such as handling and storage practices of sorghum among children's caregivers as contributors to possible contamination of complementary foods.
There is limited information on the contribution of occurrence and exposure to aflatoxin and fumonisin in sorghum-based complementary foods to the persistent undernutrition among children under five years in Kerio Valley, which is considered an ultimate goal of this study.

| Study area
This study was undertaken in major sorghum growing areas (Emsoo, Endo, and Arror Wards) of Kerio Valley in Elgeyo Marakwet County.

| Survey design, population, and sampling
A cross-sectional study was conducted in Kerio Valley, three months (June-July, 2020) after government lockdown due to the COVID-19 global pandemic. To obtain the study participants, multistage sampling procedures were adopted. First, Elgeyo Marakwet County was purposively selected being one of the counties where production and utilization of sorghum has been promoted in the AVCD program.
Kerio valley was further purposively selected due to the high malnutrition rate reported among children under five years and being the major sorghum-growing zone in the County, especially in Emsoo ward in Keiyo North, Endo (Marakwet East), and Arror (Marakwet West). The list of the villages growing sorghum in the three wards was obtained from the Ward Agriculture Office, and thirty villages were selected randomly based on probability proportional to population size (PPS) technique. A second sampling frame was generated from the randomly selected villages by enlisting all sorghum growing households with children aged 6-59 months, with the help of the Ward Agriculture Officer, village elders, and lead farmers. Lastly, using Microsoft Excel, simple random sampling was used to select 353 households (Emsoo, = 118; Endo, = 117 and Arror, =118) from the sampling frame generated. The study respondents were the mothers/caregivers of children 6-59 months of age.

| Structured Household Questionnaires
Structured household questionnaires were used to obtain information about respondents' socio-demographic information, knowledge about aflatoxin and fumonisin contamination in foods, the health effects of ingesting these toxins in food, postharvest handling, and storage practices of sorghum. guides were used to collect qualitative data. One FGD was conducted from each ward among mothers/caregivers of children 6-23 months of age to understand how they handled sorghum and if mycotoxin contamination was of concern when handling children's food. Children aged 6-23 months are nutritionally vulnerable and the introduction of complementary foods that are contaminated with mycotoxins could further impair their growth and development (Alamu et al., 2020). Therefore, it was of importance to have a focus group discussion with the mothers/caregivers of children 6-23 months of age. The FGD participants were selected from villages that were excluded from the household survey. For the key informant interviews, 14 participants including chairpersons for aggregation centers, Ward Agricultural Officers (WAOs), lead farmers, and community health volunteers (CHVs) were selected since they are the frontline agents of change at the community level. They act as intermediaries between researchers and farmers by training and motivating farmers to adapt the good agricultural practices as well as providing free health education training on essential healthcare for mothers and children. The thematic areas of discussion during the in-depth interviews included food safety issues, aflatoxin and fumonisin awareness, and their management strategies.

| Data analyses
All quantitative data were analyzed using Statistical Package for Social Software (SPSS) version 26. Data on caregivers' sociodemographic characteristics, knowledge about mycotoxins, and sorghum postharvest handling and storage practices were analyzed using the chi-square test. To assess knowledge and practices, every correct answer to the questions was assigned a score of 1 while a score of 0 was given to a wrong answer and where the respondent did not know. Scores on knowledge and practices for each respondent were calculated by summing up the scores attained on each question and the overall score ranked as good or poor. Twelve knowledge questions were asked and thus the scores ranged from 0 to 12. Individuals scoring 7 and below (score below 60%) were categorized as having poor knowledge and scores above 7 as good knowledge (ul Haq et al., 2012;Wang et al., 2015). Regarding the caregivers' postharvest practice, 18 questions were asked and a total score per individual of less than or equal to 10 (score below 60%) was categorized as poor practices and scores above 10 as good practices (Papagiannis et al., 2020;Wang et al., 2015). Binary logistic regression was used to determine the association between variables of interest. p <.05 was considered statistically significant. The qualitative data were analyzed using the thematic content analysis method by identifying similarities, differences, and trends between the individuals and group responses. Common themes that emerged from each individual and throughout the interviews were identified as well as any differences in responses based on the demographics. The findings from the FDGs and KII were then triangulated onto the household survey data to provide in-depth understanding of the concepts studied.

| Sociodemographic characteristics of the study population
The sociodemographic characteristics of the study participants are shown in Table 1. The household heads for majority of the farmer household were male (93%). More than half (52.1%) of the caregivers were aged 31-45 years and were in monogamous marriage (87.3%). About 45% of farmers had at least primary school education, indicating some level of literacy while more than a third (38%) had a monthly income of less than Ksh. 5,000. All sociodemographic characteristics (sex of the household head (p =.034), age of the respondent (p =.025), education level (p =.001), and monthly income (p =.001)) were significantly different in the three administrative wards except for marital status (p =.232).

| Knowledge about mycotoxin contamination in foods
The average score for knowledge about mycotoxins among the study respondents was 6.48 ± 2.94, out of a possible 12 score, indicating low level of knowledge among caregivers ( Figure 1). Of the 353 respondents, 218 (61.8%) were within the poor knowledge score, with more caregivers from Emsoo (71.2%) having poor knowledge scores compared to those from Arror (57.6%), and Endo (56.4%) ( Table 2). Although more than half (60.9%) of the respondents had heard about aflatoxins, only 25.2% and 19.8% were familiar with the terms fumonisin and mycotoxins, respectively. Furthermore, out of the caregivers who heard about aflatoxin and fumonisin, 49(13.9%) and 27(33.8%) knew that sorghum grain can be contaminated by aflatoxin and fumonisin respectively. Majority (67.4%) of them knew that consumption of mycotoxin-contaminated food could lead to adverse health effects such as abdominal pain (55.5%) and diarrhea (22.0%), while 40.8% and 45.9% thought it could also lead to childhood stunting and immune suppression, respectively. The caregivers from Arror (81.4%) were more knowledgeable (χ 2 =28.939, p =.001) of the health effects of mycotoxins consumption in foods was as compared to Endo (63.2%) and Emsoo (57.6%) ward.

| Postharvest practices favoring mycotoxin contamination in foods
Overall, score of the average practices among caregivers was 9.15 ± 1.90, out of 18 total scores ( Figure 1). The postharvest practices for majority (75.4%) of the caregivers were poor and diverse (χ 2 =12.237, p =.002) across the three wards. Caregivers from Emsoo ward recorded the highest number of respondents (86.4%) with poor postharvest practices followed by Endo (71.8%) and Arror (70.8%), respectively (Table 3). Overall, less than half of the respondents (35.1% and 31.2%) knew that inadequate drying and poor storage of grains contribute to mold growth resulting in food spoilage ( Figure 2). The caregivers from Emsoo (37.8%) and Arror (36.3%) considered poor drying as the main cause of spoilage of sorghum while those from Endo ward (34.1%) pointed out storage as the critical point of contamination. Majority of caregivers dried sorghum after harvest; however, 12.2% left their crops to dry while standing in the field. Among those who dried their grains after harvest (87.8%), duration of drying (χ 2 =20.807, p =.001) and materials used for drying (χ 2 =15.157, p =.004) were significantly different across the wards. Drying sorghum for more than a week was popular in Emsoo while drying grains on bare ground was mainly practiced in Arror ward. To confirm grains dryness, none of the caregivers used a moisture meter but instead, they used traditional methods such as biting (51.0%), looking (24.6%), and touching/squeezing (24.1%).
Nearly, all the caregivers (96.3%) used manual threshing methods to shell sorghum, and before shelling, they separated spoilt sorghum panicles from the clean ones. Although more than a third (32.3%) of the caregivers threw away the spoilt grains, majority (51.6%) used them as animal feed, prepared local brews (13.6%) while some from Emsoo ward (5.1%) reported using such grains for human consumption. More than half (52.4%) of the caregivers stored their grains in traditional granaries and before grains storage, less than half sorted (41%) their grains, cleaned (21.8%), or fumigated storage area (7%).
Majority of the caregivers from Arror (81.4%) used hermetic bags to store sorghum grains compared with those from Emsoo and Endo (χ 2 =40.828, p =.001). Additionally, a large proportion of the caregivers from Arror (84.7%) and Endo (84.6%) wards used plastic buckets to store sorghum flour for preparing children's porridge with majority of them storing the flour for less than a week (χ 2 =11.393, p =.003).

| D ISCUSS I ON
The findings of the current study demonstrated that majority of children's caregivers had poor knowledge about aflatoxin and fumonisin contamination of sorghum-based foods. Poor knowledge may have contributed to poor postharvest handling and storage practices.
During focus group discussions (FGDs), caregivers revealed that mycotoxin contamination was not their concern when preparing sorghum-based complementary foods. Majority of them thought that aflatoxin and fumonisin solely affect maize, another major staple crop in Kenya. These findings corroborate several other studies that have reported that large rural populations in Kenya are only knowledgeable about aflatoxins contamination of maize-based foods (Njeru et al., 2019;Obonyo & Salano, 2018;Walker & Davies, 2013).
Aflatoxin in maize may have received considerable attention in Kenya because of severe aflatoxin contamination of maize that has caused human deaths repeatedly over the years (Probst et al., 2011).
Comparable with the findings reported by Njeru et al., (2019), majority of caregivers in this study were not familiar with fumonisins.
This could be explained by the current national efforts that are solely focused on aflatoxins, a situation that is similar in most African countries (Matumba et al., 2014). while still having high residual moisture content (Hell et al., 2008).
In the current study, nearly all caregivers threshed/shelled sorghum using manual methods (beating sorghum panicles in a sack or using a traditional pestle (stone) and mortar (hollow tree trunk), with some of them threshing sorghum on bare ground. Manual threshing exposes grains to fungi through kernel breakage and eventually, mycotoxin production (Taye et al., 2018).
Storing properly dried grains in a clean and dry environment as well as controlling storage pests by use of insecticides and hermetic bags can reduce the build-up of mycotoxins in the store (Abass et al., 2018). In the present study, more than half of the caregivers from Endo and Arror ward used hermetic bags to store sorghum grains attributable to the intervention by different community development projects. Some caregivers, however, highlighted placing sorghum bundles directly on the floor. Worse still, a good number of them placed the storage bags directly on the floor, which may increase fungal growth in stored food commodities due to moisture absorption (Suleiman & Rosentrater, 2015 humidity (Maina et al., 2016). In this study, over two-thirds of caregivers stored sorghum grains in traditional granaries and in dwelling houses. Additionally, the material for keeping flour is of critical importance to maintain the quality of the flour (Hemery et al., 2020).
Furthermore, the flour for preparing children's porridge was mainly stored in plastic buckets, which are known to retain heat and moisture and easily promote fungal growth and mycotoxin contamination (Mutegi et al., 2013). Sacco et al., (2020) and Selemani (2018) suggested that paper bags could be a suitable packaging material fewer trainings as compared to their counterparts from Tanzania (Katengesya, 2018). Inadequate awareness on issues related to mycotoxins poses a risk of producing and feeding young children with contaminated complementary foods (Katengesya, 2018). For instance, mixing sorghum flour for preparing child's food with more than two other flours such as maize, millet, groundnuts, or cassava was a common practice in the study area. Apart from this practice interfering with the absorption of important nutrients, it could also increase children's exposure to several mycotoxins (Ojuri et al., 2018).
Furthermore, it was revealed during FGDs that traditional tools that were used for grains threshing and grinding were not cleaned between each batch of grains. One participant explained that the reason for not washing some of these tools, for example, the hollow tree trunk (wero konee) for threshing, was that it would crack. This indicates that caregivers of young children do not have adequate knowledge on the risks of cross-contamination of children's food.
A remarkable finding from the study is that the older caregivers (≥30 years) were four times more likely to be aware of aflatoxin to understand (Logrieco et al., 2018). Hence, the findings suggest the need to possibly include topics on mycotoxin and food safety in the available infant and young child feeding guidelines as well as to learn about or pay for aflatoxin or fumonisin control. This implies that caregivers with less monthly income are less likely to purchase nonfood items such as drying materials, hermetic bags, or radio that will enhance good postharvest practices and mycotoxins' awareness.
Additionally, people with high incomes are likely to be more careful about food and are more willing to pay for food safety than those with lower incomes (Redzwan et al., 2012). This suggests that poverty and lack of sufficient income might contribute to low levels of awareness and knowledge about mycotoxin, leading to high mycotoxin exposure (Leroy et al., 2015) through poor food handling and preparation.

| CON CLUS ION
The study demonstrated that caregivers of young children imple-

CO N FLI C T O F I NTE R E S T
The authors declare that they do not have any conflict of interest.

E TH I C A L A PPROVA L
This study protocols and procedures were reviewed and approved and measures on the prevention of COVID-19 were followed during interviews.

DATA AVA I L A B I L I T Y S TAT E M E N T
Due to privacy/ethical restrictions, the data that support the findings of this study are available on request from corresponding author.