Farming Systems, Food Security, Dietary Intakes, and Nutrition Status Among Young Children in Rural Tanzania Before and After Harvest.

Introduction

Undernutrition during early life is a major global concern. Undernutrition is common in subsistence‐farming households were farming practices and seasons impact nutrition status through effects on food consumption. The UNICEF conceptual framework describes the immediate, underlying, and basic causes of malnutrition (Moursi et al.2008). There are several examples of nutrition‐specific interventions that target immediate causes of undernutrition such as maternal and child food intake, promoting appropriate infant feeding, and treating infectious diseases (Arsenault et al.2014). However, focusing on the immediate causes of undernutrition is not enough. Nutrition‐sensitive programmes addressing underlying determinants such as household food insecurity, inadequate prenatal care for women, and unhealthy living environments are essential for sustainable improvements in nutrition (Hotz and Gibson2007). The Food and Agriculture Organisation (FAO) defines food insecurity as a situation that exists when people lack secure access to sufficient amounts of safe and nutritious food for normal growth and development and active and healthy life (Masset et al.2011). Studies show that children living in food insecure households have low diet diversity and are at higher risk of undernutrition (Ali et al.2019; Chandrasekhar et al.2017; Haddad2013). Rural food insecure households are often smallholder farming households that rely on their own production as their main source of food. Therefore, subsistence‐level farming is a key target for improving food security and potentially impacting nutrition status (Darapheak et al.2013; Faber et al.2016; Melese2013). Several high‐level discussion documents have noted the potential for agriculture interventions to improve food security and nutrition outcomes in rural low‐income settings (Dewey2001; Disha et al.2012; Rodríguez et al.2007). While interventions targeting farming practices have positive changes in infant feeding practices, dietary diversity, and food security (Dutta and Sharma2020) very few report significant effects on nutrition status outcomes like stunting, wasting and underweight (Leonard et al.2000; Michaelsen et al.2017).

In farming households, time of year with respect to the growing season will impact food availability and access to food and therefore food security status, diet diversity and consequently nutrition status. However, very few studies have reported changes in food security status and food intake for farming households before and after harvest. Better knowledge of these relationships could help shape tailored nutrition‐sensitive interventions for households within different farming systems and at different times of the year. The current study examined the relationship between farming systems, food security, dietary intakes and nutrition status among infants and young children pre and post‐harvest seasons.

Methods

Study Design, Population and Sampling

This was a longitudinal survey design such that data were collected pre‐harvest and post‐ harvest. The study was conducted in the Mvomero district in the Morogoro region and the Kishapu district in the Shinyanga region, Tanzania Mainland. The regions were selected based on household farming practices representative of Tanzanian agriculture. The main crops in both regions include maize, rice, sorghum, sweet potatoes, green gram and groundnuts. Agriculture in both regions is dependent on season and rainfall, which follows a unimodal or bimodal rainfall pattern. Districts (within regions) are divided into divisions, and divisions are further sub‐divided into wards and villages. One division was randomly selected from each district, followed by the random selection of two wards per division, and one village per ward. The study villages were Makuyu and Milama in Mvomero district, and Lubaga and Mwakipoya in Kishapu district. The study population comprised of farming households with mothers and children < 2 years of age (nutrition status survey) randomly invited to the study from a larger study cohort, referred to here as the ‘general household survey’. Sample size calculations were performed using method described by the United Nations, powered at 90% with a type 1 error rate of 0.05. A sample of 277 households was recruited to the general household survey from each district (n= 554 households in total) based on estimates of food insecurity prevalence. A sample of 110 households from each district was recruited to the nutrition status survey based on estimates of wasting prevalence, including 10% attrition rate to mitigate potential losses to follow‐up post‐harvest. Pre‐harvest, considered a period of food shortage, was February‐March, 2014 and post‐harvest, considered a period of food surplus, was August, 2014.

Ethical Approval

This study was conducted according to the guidelines laid down in the Declaration of Helsinki, and all procedures involving human subjects/patients were approved by the National Institute of Medical Research (NIMR), Tanzania (Reference: 1679). Verbal informed consent was obtained from all subjects/patients. Verbal consent was witnessed and formally recorded. All data was collected by a team of trained researchers following standard protocols.

Demographic Information

A pre‐tested structured questionnaire was used to record information on age, sex and marital status of the head of the household, family size, headship of household, education, employment status, income, infant birth weight (reported on clinic card) and illness over the past 2 weeks (Muhimbula et al.2019).

Identifying Farming Systems Within Farm Households

Farming systems were defined by researchers based on the types of crops cultivated, the degree of dependence and market orientation for crops, and the number of livestock units owned by households (Wordofa and Sassi2020). The degree of household dependence on a particular crop was calculated based on the percentage of total land area cultivated that was allocated to that crop. The household was considered dependent on a particular crop if it had at least 70% of the total cultivated land area allocated to that particular crop. The placement of households into a specific farming system was based on the 2013 farming year data reported from the pre‐harvest survey (Wordofa and Sassi2020). The farming systems were Single Food Crop (SFC) which grows a SFC on ≥ 70% of the total cultivated land; Mixed Food Crop (MFC) which grows several food crops, (i) none occupy ≥ 70% of the total cultivated land, and (ii) the total proportion of land dedicated to food crops is ≥ 70%; Cash Crop (CC) which grows a cash crop on ≥ 70% of the total cultivated land; and Mixed Crop Livestock (MCL) which grows crops and has more than 7.3 Tropical Livestock Units (TLU) (Wordofa and Sassi2020).

The Household Food Insecurity Access Scale (HFIAS)

Household food insecurity was assessed using the Household Food Insecurity Access Scale (HFIAS) developed by the Food and Nutrition Technical Assistance (FANTA) project (Coates et al.2007). The HFIAS tool consists of nine questions that capture all three core domains that reflect a household's inadequate access to food in the preceding 30 days. The response was recorded as 0 if food insecurity did not happen, 1 if it rarely happened (1–2 times), 2 if it sometimes happened (3–10 times) and 3 if it often happened (> 10 times) in the previous 30 days. Answers to questions were combined, and the household was allocated to a food security category as follows: food secure, mild food insecure, moderate food insecure and severely food insecure.

Anthropometry

Weight was measured using a seca weighing scales model 874 (seca Hamburg, Germany). Length and height were measured using the UNICEF portable length and height measuring system SET‐2 (S0114540). All anthropometric equipment was calibrated prior to use, and measurements followed standard protocols (NBS and ICF Macro2015–16). Length‐for‐age z‐score (LAZ), weight‐for‐age z‐score (WAZ), and weight for‐length z‐score (WLZ) were calculated using WHO ANTHRO Software version 3.2.2 (WHO, Geneva, Switzerland) and WHO Child Growth Standards (De Onis2006). Stunting, underweight and wasting were defined as z‐scores below −2 standard deviations (SD) of the median values of the reference data. Mid upper arm circumference (MUAC) was recorded using a flexible colour‐coded measuring tape (UNICEF, MUAC Child 11.5 Red/PAC‐50). The MUAC tape was wrapped around the mid‐upper left arm to measure its circumference, and measurements were taken to the nearest 0.1 cm. Based on the WHO/UNICEF recommended cut‐off points, children with MUAC ≤ 11.5 cm were classified as severely malnourished 11.5–12.5 cm as moderately malnourished and ≥ 12.5 cm as normal (WHO and UNICEF2009).

Dietary Intake and Food Quantification

Individual semi‐quantitative 24 h dietary recalls were conducted pre and post‐harvest to determine the amount and types of foods consumed by children. Mothers recalled and described every item of food and drink the child had consumed over the past 24 h for 1 day only. Detailed information was obtained through systematic repetition of open‐ended questions. Amounts were estimated using standard household measures e.g. bowls, plates, cups, spoons. Where possible a measuring cylinder was used to estimate volumes or weights of foods consumed (Ferguson et al.1994). All mixed dishes were probed for the type and amounts of ingredients used. All the information was recorded on the 24 h recall sheet. The 24 h recall was conducted on weekdays. Atypical days (religious feast or celebrations) were not included. The 24 h recall data was entered into NUTRITICS software research edition version 4.2 (Nutritics, Dublin Ireland), which included data from the Tanzanian Food Composition Database (Lukmanji et al.2008). All foods consumed were categorised into 18 food groups defined using the nutrient composition of the foods and the Tanzania Food Composition database categories TableS1. Individual Dietary Diversity Scores for infants (IDDS) were also calculated based on WHO guidelines (WHO2008). All foods that were consumed in the past 24 h were grouped into 7 food groups. Each group scored 1 point, otherwise 0. The IDDS scores were calculated by summing the number of food groups consumed by the individual children over the 24 h recall period. A score of ≥ 5 indicated minimum dietary diversity. Outliers (over‐or under‐reporters) for energy intake (EI) were identified prior to data analysis using a defined protocol as follows: (1) we calculated age dependent energy intake (EI)/energy requirements (ER) from complementary foods based on WHO criteria for breastfed children (Dewey and Brown2003); (2) the ratio of EI/ER from complementary foods was determined for each child; (3) infants and young children with less than 50% of EI/ER or more than 150% EI/ER were identified as ‘at risk of misreporting’; (4) WAZ was examined for those ‘at risk of misreporting’; (5) food security status, breastfeeding frequency and infections for infants and young children identified as ‘at risk of misreporting’ were also examined. Over‐reporters were defined as those with WAZ < −2 and EI/ER > 150% coming from insecure households with no infections/illness in the months preceding the survey. Under‐reporters were defined as those with EI/ER < 50% from food‐secure households, with no reported illnesses. Based on the evaluation of data described in this protocol, food intake data from 3 children pre‐harvest classified as ‘at risk of underreporting’ were excluded from the analysis.

Statistical Analysis

A study database was compiled and imported into IBM SPSS Statistics for Windows for analysis version 20.0 (IBM Corp, NY, USA). Descriptive statistics were generated for all data, and normal distribution of continuous variables was checked graphically using box plot, Q‐Q plots andp‐values from the Shapiro–Wilk test. Non‐parametric tests were used when the data were not normally distributed. The findings were summarised using frequencies and proportions for categorical variables and mean and SD or median with interquartile ranges (IQR) for continuous variables. Where necessary, adjusted means and standard error of the mean (SEM) are presented. Baseline characteristics were compared for each of the farming systems using Chi‐square test for proportions and one‐way ANOVA with post‐hoc test to compare the means for continuous variables. Kruskal‐Wallis's test with post‐hoc ranking was used to compare the median amount of food consumed from food groups between the farming systems; the Mann‐Whitney Test was used to compare the amount of food consumed from food groups between food security status groups, and the Friedman test was used to examine the significant changes in the amount of food consumed from food groups pre‐ to post‐harvest. Univariate GLM ANCOVA was used to examine differences in WAZ, LAZ and WHZ between farming systems and food security status groups when pre‐ and post‐harvest data were examined separately. A two‐way mixed ANCOVA was used to test the effect of season on the changes in WAZ, LAZ and WHZ from pre‐ to post‐harvest and to examine interaction effects of farming system and food security on changes. Where the interaction term was significant, post hoc analysis involved separating groups to examine intra‐group changes between pre and post‐harvest. For univariate analysis all data was included except for individual missing observations for certain variables, for example if a measurement could not be taken for any reason. For two‐way analysis participant data was included if measurements were available for both time points. Mother's age and BMI, and infant and young child age, sex and birthweight were considered as potential covariates. In all cases, covariate selection followed the methods proposed by Greenland (1989). Infant and young child age and sex were significant and were included as the covariates in the final model.p‐values < 0.05 were considered statistically significant.

Results

Farming Systems, Demographics, and Food Security Status

Each household (n= 209) was allocated to one of four farming systems; SFC (23%), Mixed Food Crops (28%), Cash Crops (22%) and Mixed Crop‐Livestock (27%). Thirteen households were lost to follow‐up leaving a cohort of 196 post‐harvest. Characteristics of households by farming systems pre‐harvest are presented in Table1. Household size was significantly different between farming systems (p< 0.001). MCL households had the highest number of household members (n= 10) and the highest proportion achieving formal education (> 40%,p< 0.001). Mothers from SFC households had the highest BMI in both seasons. There was no difference in infant birth weight across farming systems. About 40% of households experienced severe food insecurity pre‐harvest, but this decreased to 4% post‐harvest. Household food security status was different between farming systems pre (p= 0.01) and post‐harvest (p= 0.03). Food insecurity was lowest for SFC households (56%) and highest for MCL households (82%) pre‐harvest. Post‐harvest, food insecurity prevalence was lowest for MCL households (21%) and highest for SFC households (49%) (Table2).

Table: Household characteristics by farming system groups.

Table: Food security status of each farming system group pre and post‐harvest.

Farming Systems and Breastfeeding Practices

Some key breastfeeding indicators are presented in TableS1. There was no significant difference in the proportion of mothers who initiated breastfeeding within 1 h of birth across farming systems. Approximately 30%–50% of mothers offered extra fluids other than breast milk within 3 days after delivery; however, the proportion of mothers offering fluids other than breast milk was not significantly different across farming systems. Plain water was the most frequent fluid offered in the first 3 days after birth. Approximately 80%–85% of infants were not exclusively breastfed from 2 to 5 months, but there was no difference in the proportion of infants exclusively breastfed across farming systems.

Farming Systems, Food Security, and Nutrition Status

Table3presents mean infant nutrition status measurements grouped by farming system and food security status pre‐ and post‐harvest. During pre‐harvest, infant WAZ (p= 0.01) and LAZ (p= 0.03) were different across farming systems, however the differences were not significant post‐harvest. Post‐hoc analysis revealed lower WAZ and LAZ among infants and young children from MFC households and higher WAZ and LAZ from CC households pre‐harvest after controlling for age and sex. WAZ, LAZ, WLZ, and MUAC also differed by food security status (Table3). Infants in food‐insecure households had significantly lower WAZ (p= 0.01), WLZ (p= 0.01), and MUAC (p= 0.03) pre‐harvest. Two‐way ANCOVA revealed a significant interaction between food security status and the farming system on LAZ in the post‐harvest season (p= 0.01). GLM profile plots indicated that infants and young children from food‐insecure MCL had significantly higher LAZ than infants and young children from food‐secure MCL, whereas the opposite was the case for all other farming systems. FigureS1There was no significant interaction between food security status and farming systems for WAZ, WLZ, and MUAC pre and post‐harvest (data not shown). The proportion of infants classified as stunted, underweight, wasted, and malnourished with low MUAC stratified by farming systems are presented TableS2. The proportion of underweight infants was significantly higher in the post‐harvest within CC farming households compared to other farming households (p= 0.04). Stunting rate increased in all farming systems during post‐harvest, with a significant increase for infant's CC farming households (p< 0.001). Similarly, the proportion of malnourished infants within SFC, MFC and MCL was significantly lower post‐harvest (p> 0.001).

Table: Nutrition status indicators for farming system groups and food security status groups.

The second stage of GLM analysis focused on changes from pre‐ to post‐harvest TableS3. When the cohort was considered as a whole, there was a significant change in MUAC pre‐ to post‐harvest (p< 0.001). When each farming system was considered separately, a significant increase in MUAC was noted for children within the MFC (p< 0.001) and MCL (p= 0.01) systems; however, there were no significant changes in any other parameters. When food secure and food insecure groups were considered separately, there was a significant change in LAZ (p= 0.01) and MUAC (p= 0.01): LAZ decreased significantly in food insecure households, while MUAC increased significantly in food secure and insecure households. Two‐way repeated measures ANCOVA showed no interaction between food security and the farming system on nutrition status indicators. However, when food security status was included as a factor in repeated measures analysis for each farming system, there was a significant difference in the change of WAZ for food secure versus food insecure children from MFC households (p= 0.04). GLM profile plots revealed a greater decrease in WAZ for the food insecure group in the MFC farming system compared to the food secure group. No other differences were noted for food‐secure versus food‐insecure groups within the different farming systems.

Food Security and Infant and Young Child Food Intake

Based on one 24 h recall, all children in the study consumed ‘Cereals dishes’, ‘Vegetables dishes’ was the next most commonly consumed food group (41% pre, 51% post), followed by ‘Pulses dishes’ (19% pre, 39% post). ‘Meat, poultry, eggs, and fish’ were consumed by 10% pre‐harvest and 29% post‐harvest, and ‘Milk and milk products’ were consumed by 11% pre‐harvest and 13% post‐harvest. No child consumed foods from the ‘Vegetables’ or ‘Sugars’ groups pre‐harvest, only 4 children consumed the sugars group post‐harvest. There were no significant differences in the proportion of consumers from food‐secure and food‐insecure households; however, there was a significant difference in the amounts of foods consumed pre‐harvest (Table4). Infants and young children from food‐insecure households had higher intakes of ‘Vegetable dishes’ (p= 0.04) and lower intakes of ‘Fruits and fruit juices’ compared to those from food‐secure households (p= 0.02). The % consumers of food groups from food insecure and food secure households increased post‐harvest except for ‘Pulses dishes and Milk and milk products’. While acknowledging the fact that children were 6 months older post‐harvest and they would consume greater amounts of foods, the median amount of ‘Cereal dishes’ consumed by 12–18 month old children increased from 242 g pre‐harvest to 297 g post‐harvest. Changes in the median amount of food consumed from each food group pre to post‐harvest were also compared for food‐secure and food‐insecure households separately depending on the available data at both time points. There was no significant difference in the amount of food consumed pre‐ to post‐harvest for any food group within food‐secure or food‐insecure groups.

Table: Percentage of consumers and the amount (g) of food consumed by children aged 6–24 months by food security status groups pre‐ and post‐harvest.

Farming Systems and Infant and Young Child Food Intake

The three most commonly consumed food groups for SFC and MFC farming systems were ‘Cereals dishes’, ‘Pulses dishes’, and ‘Vegetable dishes’. ‘Cereals dishes’ also ranked highest for CC, followed by ‘Vegetable dishes’ and ‘Local broths’. Again, ‘Cereals dishes’ and ‘Vegetables dishes’ were highest for MCL, followed by ‘Milk and milk products’. Food group intake by farming systems during pre‐harvest is presented in Table5. There was a significant difference in the % consumers of ‘Vegetables dishes’ (p= 0.02) and ‘Pulses dishes’ (p= 0.02) across farming systems. More infants and young children in MFC and MCL farming systems consumed ‘Vegetables dishes’ compared to SFC and CC. SFC were the largest consumers of ‘Pulses dishes’. When comparing the amounts of food from food groups consumed, MCL had higher intakes of ‘Pulses dishes’ (p= 0.01) compared to all other farming systems.

Table: Percentage consumers and the amount (g) of food consumed by children aged 6–24 months by farming systems groups pre‐harvest.

During post‐harvest, there was no change in the 3 most commonly consumed food groups for SFC and MFC households Table6. The proportion of children consuming ‘Local broths’ in the CC households decreased, so the 3rd most commonly consumed food group was ‘Pulses dishes’. The proportion of MCL children consuming Milk and milk products also decreased so that ‘Pulses dishes’ were the 3rd most commonly consumed food group in this farming system. There were significant differences in the %consumers of 5 food groups post‐harvest, with MCL having the highest %consumers of ‘Vegetables dishes’ (p= 0.01) and ‘Pulses dishes’ (p= 0.02), but the lowest %consumers of ‘Pulses’ (p< 0.001). In contrast, SFC had the highest %consumers of ‘Pulse dishes’ but the lowest %consumers of ‘Pulses dishes’ and ‘Vegetables dishes’. MFC and CC had somewhat similar patterns of food intake post‐harvest, with slightly higher % consumers of Meat poultry eggs fish food group in the CC farming system. When amounts of food consumed by groups were compared, MFC and MCL had higher intakes of ‘Vegetable dishes’ (p= 0.04), ‘Pulses dishes’ (p= 0.003), and Meat poultry eggs (p= 0.04) and lower intakes of ‘Fruit and fruit juices’ (p= 0.02) compared to SFC and CC. There was a significant difference in the amount of ‘Pulses dishes’ (p= 0.01) consumed with the highest intakes in MFC and lowest in MCL households.

Table: Percentage consumers and the amount (g) of food consumed by children aged 6–24 months by farming systems post‐harvest.

Changes in median amount of food consumed from food group's pre to post‐harvest were compared by splitting the file by farming systems. The amounts of ‘Cereals dishes’ consumed increased significantly from pre to post‐harvest for MCL (p= 0.003) and SFC (p= 0.04) households. Children's intakes of ‘Vegetables dishes’ also increased significantly in SFC farming households (p= 0.04). Intakes of other food groups could not be compared given the number of valid cases pre and post.

Discussion

This study reports differences in food security and nutrition status outcomes for infants and young children from different farming systems pre and post‐harvest. Infants and young children from MFC households had lower mean WAZ and LAZ compared to other farming systems, while infants and young children from CC households had higher mean WAZ and LAZ. Infants and young children from food‐insecure households had lower WAZ, WLZ, and MUAC pre‐harvest, whereas there was no difference post‐harvest. Farming systems with a low prevalence of food security pre‐harvest improved post‐harvest; for example, the proportion of MFC and MCL households classified as food secure increased from 20% to 80%. Changes in food security coincided with changes in food intake and MUAC for MFC and MCL households. The SFC farming system was different from others in that the proportion of food‐secure households did not change much from pre‐ to post‐harvest. Overall, CC farming households recorded better WAZ, LAZ, WLZ and MUAC compared to other farming systems. While acknowledging the fact lower LAZ observed post‐harvest may not be directly related to higher nutritional risk observed pre‐harvest season due to influences on linear growth as infants were 6 months older post‐harvest; Improvements in nutrition status post‐harvest were linked to increased food security and food group consumption. When comparing diets across farming systems, there were no major differences in the proportions consumed or the amounts of foods consumed pre‐harvest. However, farming system‐related dietary patterns were more apparent post‐harvest.

Among all farming systems, MCL and MFC households were the most sensitive to seasonal changes in food accessibility, which was reflected in improved food security, improved food consumption and higher MUAC post‐harvest. Food group diversity was poor pre‐harvest for MCL and MFC households; both % consumers and the amount of food consumed increased post‐harvest, particularly for ‘Pulses, nuts and seeds’, ‘Pulses nuts and seeds dishes’ and Meat poultry eggs fish dishes. The difference between these farming systems is that MCL includes livestock. Research suggests that farming systems that mix crops and livestock can contribute to both farm income and food security with potential implications for improved nutrition status (Nabarro and Wannous2014). In support of this finding, MCL households had better nutrition status compared to MFC who had lower WAZ, LAZ, and WLZ in both seasons. However, the high levels of food insecurity in MCL households pre‐harvest highlights the significant impact of season on food security and nutrition status regardless of farming practices, which is well documented in the literature (Ferro‐Luzzi et al.2002; Harding2008; Schalekamp2009). From the general household survey (described in the methods), MCL households had high levels of food insecurity during pre‐harvest, had the biggest households, and had the highest age‐dependent ratio which is consistent with our findings (Massawe2016). Using data collected from the households reported in this manuscript, as well as other households in both regions with children > 24 months, Massawe (2016) reported that MCL household income was generated from off‐farm activities mainly post‐harvest, suggesting poor access to cash and reserves for these households in the pre‐harvest season. Additionally, Massawe and his colleagues argued that although MCL households had the highest overall income, such income was seasonal, and it is possible that it was depleted in just a few months after the harvest period (Massawe2016).

Similar high rates of food insecurity in pre‐harvest among livestock keepers were reported previously in Uganda (Mayanja et al.2015). Although one might expect livestock owners to sell animals in times of food shortage for income and so that children can remain in school, this is typically not the case for households with livestock (Hedges et al.2016). Research suggests that pastoralists prioritise livestock as they view livestock as a financial asset (Coppock et al.2018), and therefore, children often leave school early to take care of livestock (Hedges et al.2016). Therefore, income is not directly related to years of schooling in certain instances. Research suggests that this is not happening, possibly due to market access and the perceived impact on socioeconomic status (Ruhangawebare2010; Stroebel et al.2011). Additionally, from these findings, MCL households had the largest household sizes, while a large household size in MCL was important for providing labour for both livestock keeping and cropping activities, it could be associated with high food insecurity levels depending on their contribution to household income.

An unexpectedly significant interaction was observed between food insecurity and LAZ for infants and young children in MCL households. Infants and young children from food‐insecure MCL households had higher LAZ than food‐secure MCL post‐harvest. In contrast, all other systems showed better nutrition status in food‐secure than in food‐insecure households. It is difficult to know what mechanism might be at play here; however, the data shows that during the pre‐harvest season, a larger proportion of children from MCL households consumed ‘Milk and milk products’ compared to other households. At the same time, a small proportion of children were exclusively breastfed (~15%) and so a majority were receiving fluids and foods other than breastmilk. It is possible that a larger proportion of children in MCL households received milk as a substitute fluid during the pre‐harvest, food‐insecure period and that this led to somewhat better outcomes relative to other farming households. Whilst this is an extrapolation of the available data, we know from previous research that it is common to introduce animal milk to infants and young children in households with livestock in Tanzania (Hanselman et al.2018; Kibona and Mwanri2020). Several studies have shown a significant positive relationship between milk intake and linear growth (Ånestrand2013; Hoddinott et al.2015; Mosites et al.2017; Shreenath et al.2011; Wagah et al.2015) & Haile and Headey (2023). Despite this observation and discussion around the potential that cow's milk was used as a substitute fluid for infants and young children in this period, it should be noted that the WHO/UNICEF recommend exclusive breastfeeding until 6 months, when suitable complementary foods are introduced alongside breastfeeding, with further advice to delay cow's milk until after 1 year of age (Victora et al.2016).

Within the SFC system, the proportions of food secure and food insecure households did not vary much pre‐ and post‐harvest. Although food production diversity has potential to improve food security (Melese2013; Sunderland2011), mono‐cropping has the advantage of increasing yield (Weitzman2000). However, in subsistence farming systems mono‐cropping is vulnerable to crop failure leading to higher food insecurity, poor diet quality and nutrition (Sunderland2011). This might explain the observation that SFC households had higher food security in pre‐harvest compared to other farming systems, but very little change post‐harvest. Indeed, ~50% of SFC households were food insecure post‐harvest compared to other farming systems where the proportions of food insecure households dropped significantly. Food insecurity within SFC was related to poor nutrition status which recorded lower WAZ, LAZ and WLZ second to MFC for both pre and post‐harvest. Yet again, diet diversity was poor, reflected in the high intakes of ‘Cereal dishes’ and low intakes of ‘Pulses nuts and seeds’ ‘Milk and milk products and Meat, poultry, eggs, fish dishes’. SFC households also had low income relative to other farming systems and smaller land areas which probably explains their dependence on maize mono‐cropping for home consumption (Wordofa and Sassi2020).

Agriculture generates income for food and non‐food expenditures, which can translate to expenditure on nutrition‐enhancing goods and services including health care and education and consequently impact child nutrition status (Gillespie and van den Bold2017). Studies indicate that children in higher‐income households tend to eat a higher‐quality diet than those in poorer households (Arimond and Ruel2004; Wiegers et al.2011). However, other studies suggest that introducing cash crops in poorer households does not translate to improved nutrition status (Ecker et al.2011). The current study showed that infants and young children from CC households had higher WAZ and LAZ compared to other farming systems. CC households had a relatively small proportion who had no schooling, were the farming system with the highest number of school years (although not statistically significant) and had higher income relative to SFC and MFC. The combination of these factors would suggest that CC households had some advantage over other households, which might result in better access to food, and health services, hence better nutrition status.

Food intakes were quantified based on groups defined in the Tanzanian Food Composition Database (Lukmanji et al.2008). The amount and % consumers of food groups increased post‐harvest, and there were differences between farming systems, suggesting that dietary patterns differ when food is available. Several studies have reported similar findings in relation to production diversity and diet diversity (Ferguson et al.2015; Hirvonen and Hoddinott2016; Melese2013; Organisation2023). However, despite improved intakes post‐harvest, very few infants and young children consumed animal source foods, particularly meat, poultry, fish and eggs. This is concerning as low intakes of animal source foods are associated with nutrient deficiencies and increasing intakes are associated with improved growth and lower rates of stunting (Hetherington et al.2017; Iannotti et al.2017). The changes noted here when food is available suggests that an intervention that improves food production could have an impact on nutrition status. However, improving access to foods pre‐harvest remains a challenge. Multicomponent interventions that target food production, processing and preservation while also educating households and providing economic support systems could provide a more sustainable change and alleviate food shortages in the pre‐harvest season.

This study collected data from households in two regions of Tanzania selected as the farming practices broadly represented Tanzanian agriculture. The main strength of this study was the longitudinal study design which provided the means to link farming practices, food security, food consumption, and infant nutrition status while accounting for seasonal changes. However, the fact that children were 6 months older post‐harvest must be acknowledged when comparing food intakes pre‐ and post‐harvest as older children will consume greater amounts of foods. In this study, food records were quantified, and intakes were analysed from a food group perspective, which eases translation to practical recommendations. However, food intake was assessed based on one 24 h recall which has the potential to introduce recall bias and will not account for day‐to‐day variation and therefore is not a true reflection of habitual diet. Finally, the sample size in this survey was calculated to assess nutrition status; however, when split by farming system and given the low intakes of certain food groups it was not possible to calculatep‐values for certain comparisons which limited the analysis of food intake data.

Conclusion

This study revealed high food insecurity, poor diet diversity, and nutrition status pre‐harvest in four farming systems. While seasonal food insecurity affected all farming systems, children from CC households recorded better WAZ and LAZ than children from MFC and MCL households pre‐harvest. Food security and food intake increased post‐harvest, which was reflected in improved nutrition status, particularly for MFC and MCL households. This suggests that CC farming systems provide income that can support households during food shortages, while MFC and MCL systems improve dietary diversity post‐harvest. There were no major differences in dietary intake between the farming systems, particularly in pre‐harvest, when food availability was low for all. Dietary diversity only improved post‐harvest, particularly through increased consumption of legumes, pulses, nuts, and animal‐source foods.

Generally, our findings highlighted two important pathways: the income‐generating pathway, which was stronger in the CC farming system, and food group diversity, which was observed among MCL and MFC households post‐harvest. Interventions to improve nutrition status should therefore focus on strengthening both pathways while addressing seasonal gaps through improved food storage and preservation, enhancing market access for nutrient‐rich foods, and promoting nutrition education for better child nutrition.

Author Contributions

H.M., J.K. and A.M.O.S. designed the study and carried out the study. H.M. and A.M.O. analysed the data and interpreted the findings. All authors contributed to writing the article.

Conflicts of Interest

The authors declare no conflicts of interest.

Acknowledgements

We would like to acknowledge the contribution of the AgriDiet Tanzanian team (Prof Jim Kinsella, Dr Deirdre O'Connor, Prof Amon Mattee, Prof Thadeus Mkamwa, Dr Goodluck Massawe and Achilana Mtingele), and all project partners (University College Cork, Institute of Development Studies (UK), Ethiopian Development Research Institute, Mekelle University, and Haramaya University, Ethiopia). We also thank Sokoine University of Agriculture and Saint Augustine University, Tanzania for their support with transport and accommodation during the data collection phase. Finally, we would like to thank the villages and families who participated in this research. Funded by Irish Aid, Department of Foreign Affairs, Ireland and the Higher Education Authority of Ireland under the Programme of Strategic Cooperation.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.

Associated Data

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.

References

1. Ali, N. B. , Tahsina T., Hoque D. M. E., et al. 2019. “Association of Food Security and Other Socio‐Economic Factors With Dietary Diversity and Nutritional Statuses of Children Aged 6–59 Months in Rural Bangladesh.” PLoS One 14, no. 8: e0221929. doi.org/10.1371/journal.pone.0221929
2. Ånestrand, G. 2013. Milk Consumption and Growth of Children in the Kilimanjaro Region in Rural Tanzania: An Ethnographic Research Done Through Quantitative and Qualitative Methodes. In.
3. Arimond, M. , and Ruel M. T.. 2004. “Dietary Diversity Is Associated With Child Nutritional Status: Evidence From 11 Demographic and Health Surveys.” Journal of Nutrition 134, no. 10: 2579–2585. doi.org/10.1093/jn/134.10.2579
4. Arsenault, J. E. , Nikiema L., Allemand P., et al. 2014. “Seasonal Differences in Food and Nutrient Intakes Among Young Children and Their Mothers in Rural Burkina Faso.” Journal of Nutritional Science 3: e55. doi.org/10.1017/jns.2014.53
5. Chandrasekhar, S. , Aguayo V. M., Krishna V., and Nair R.. 2017. “Household Food Insecurity and Children's Dietary Diversity and Nutrition in India. Evidence From the Comprehensive Nutrition Survey in Maharashtra.” Maternal & Child Nutrition 13: e12447. doi.org/10.1111/mcn.12447
6. Coates, J. , Swindale A., and Bilinsky P.. 2007. Household Food Insecurity Access Scale (HFIAS) for Measurement of Household Food Access: Indicator Guide (Version 3). Food and Nutrition Technical Assistance Project, Academy for Educational Development.
7. Coppock, D. L. , Bailey D., Ibrahim M., and Tezera S.. 2018. “Diversified Investments of Wealthy Ethiopian Pastoralists Include Livestock and Urban Assets That Better Manage Risk.” Rangeland Ecology & Management 71, no. 1: 138–148.
8. Darapheak, C. , Takano T., Kizuki M., Nakamura K., and Seino K.. 2013. “Consumption of Animal Source Foods and Dietary Diversity Reduce Stunting in Children in Cambodia.” International Archives of Medicine 6, no. 1: 29. doi.org/10.1186/1755-7682-6-29
9. Dewey, K. G. 2001. “The Challenges of Promoting Optimal Infant Growth.” Journal of Nutrition 131, no. 7: 1879–1880. doi.org/10.1093/jn/131.7.1879
10. Dewey, K. G. , and Brown K. H.. 2003. “Update on Technical Issues Concerning Complementary Feeding of Young Children in Developing Countries and Implications for Intervention Programs.” Food and Nutrition Bulletin 24, no. 1: 5–28. doi.org/10.1177/156482650302400102
11. Disha, A. , Rawat R., Subandoro A., and Menon P.. 2012. “Infant and Young Child Feeding (IYCF) Practices in Ethiopia and Zambia and Their Association With Child Nutrition: Analysis of Demographic and Health Survey Data.” African Journal of Food, Agriculture, Nutrition and Development 12, no. 2: 5895–5914.
12. Dutta, M. , and Sharma M.. 2020. “Complementary Foods: A Review on Types, Techniques and Nutritional Content.” International Journal of Home Science 6: 90–96.
13. Ecker, O. , Breisinger C., and Pauw K.. 2011. Growth is Good, But Is Not Enough to Improve Nutrition.
14. Faber, M. , Laubscher R., and Berti C.. 2016. “Poor Dietary Diversity and Low Nutrient Density of the Complementary Diet for 6‐to 24‐Month‐Old Children in Urban and Rural Kwazulu‐Natal, South Africa.” Maternal & Child Nutrition 12, no. 3: 528–545. doi.org/10.1111/mcn.12146
15. Ferguson, E. , Chege P., Kimiywe J., Wiesmann D., and Hotz C.. 2015. “Zinc, Iron and Calcium Are Major Limiting Nutrients in the Complementary Diets of Rural Kenyan Children.” Supplement, Maternal & Child Nutrition 11 Suppl 3, no. S3: 6–20. doi.org/10.1111/mcn.12243
16. Ferguson, E. L. , Gibson R. S., and Opare‐Obisaw C.. 1994. “The Relative Validity of the Repeated 24 h Recall for Estimating Energy and Selected Nutrient Intakes of Rural Ghanaian Children.” European Journal of Clinical Nutrition 48, no. 4: 241–252.
17. Ferro‐Luzzi, A. , Morris S. S., Taffesse S., Demissie T., and D'Amato M.. 2002. “Seasonal Undernutrition in Rural Ethiopia: Magnitude, Correlates, and Functional Significance.” Food and Nutrition Bulletin 23, no. 2: 227–228. doi.org/10.1177/156482650202300211
18. Gillespie, S. , and van den Bold M.. 2017. “Agriculture, Food Systems, and Nutrition: Meeting the Challenge.” Global Challenges 1, no. 2: 1600002. doi.org/10.1002/gch2.201600002
19. Greenland, S. 1989. “Modeling and Variable Selection in Epidemiologic Analysis.” American Journal of Public Health 79, no. 3: 340–349. doi.org/10.2105/ajph.79.3.340
20. Haddad, L. 2013. “From Nutrition Plus to Nutrition Driven: How to Realize the Elusive Potential of Agriculture for Nutrition?” Food and Nutrition Bulletin 34, no. 1: 39–44. doi.org/10.1177/156482651303400105
21. Haile, B. , and Headey D. D.. 2023. “Growth in Milk Consumption and Reductions in Child Stunting: Historical Evidence From Cross‐Country Panel Data.” Food Policy 118: 102485. doi.org/10.1016/j.foodpol.2023.102485
22. Hanselman, B. , Ambikapathi R., Mduma E., Svensen E., Caulfield L. E., and Patil C. L.. 2018. “Associations of Land, Cattle and Food Security With Infant Feeding Practices Among a Rural Population Living in Manyara, Tanzania.” BMC Public Health 18, no. 1: 159. doi.org/10.1186/s12889-018-5074-9
23. Harding, K. B. 2008. “Dietary Intakes and Nutritional Status of Rural Ghanaian Children: Are Season and Attending Daycare Important Determinants?” Master's thesis, McGill University.
24. Hedges, S. , Borgerhoff Mulder M., James S., and Lawson D. W.. 2016. “Sending Children to School: Rural Livelihoods and Parental Investment in Education in Northern Tanzania.” Evolution and Human Behaviour 37, no. 2: 142–151.
25. Hetherington, J. B. , Wiethoelter A. K., Negin J., and Mor S. M.. 2017. “Livestock Ownership, Animal Source Foods and Child Nutritional Outcomes in Seven Rural Village Clusters in Sub‐Saharan Africa.” Agriculture & Food Security 6, no. 1: 9.
26. Hirvonen, K. , and Hoddinott J.. 2016. “Agricultural Production and Children's Diets: Evidence From Rural Ethiopia.” Agricultural Economics 47, no. 3: 261–270.
27. Hoddinott, J. , Headey D., and Dereje M.. 2015. “Cows, Missing Milk Markets, and Nutrition in Rural Ethiopia.” Journal of Development Studies 51, no. 8: 958–975.
28. Hotz, C. , and Gibson R. S.. 2007. “Traditional Food‐Processing and Preparation Practices to Enhance the Bioavailability of Micronutrients in Plant‐Based Diets.” Journal of Nutrition 137, no. 4: 1097–1100. doi.org/10.1093/jn/137.4.1097
29. Iannotti, L. L. , Lutter C. K., Stewart C. P., et al. 2017. “Eggs in Early Complementary Feeding and Child Growth: A Randomized Controlled Trial.” Pediatrics 140, no. 1: e20163459. doi.org/10.1542/peds.2016-3459
30. Kibona, M. G. , and Mwanri A. W.. 2020. “Infant and Young Child Feeding Practices Among Pastoralist and Crop Farming Communities in Mvomero District, Tanzania.” Tanzania Journal of Agricultural Sciences 19, no. 1: 22–33.
31. Leonard, W. R. , DeWalt K. M., Stansbury J. P., and McCaston M. K.. 2000. “Influence of Dietary Quality on the Growth of Highland and Coastal Ecuadorian Children.” American Journal of Human Biology 12, no. 6: 825–837. doi.org/10.1002/1520-6300(200011/12)12:6<825::AID-AJHB10>3.0.CO;2-E
32. Lukmanji, Z. , Hertzmark E., Mlingi N., Assey V., Ndossi G., and Fawzi W.. 2008. Tanzania Food Composition Tables. Muhimbili University of Health and Allied Sciences (MUHAS) & Tanzania Food and Nutrition Centre (TFNC).
33. Massawe, G. D. 2016. Farming Systems and Household Food Security in Tanzania: The Case of Mvomero and Kishapu districts, Published PhD thesis, University College Dublin, Dublin: Ireland.
34. Masset, E. , Haddad L., Cornelius A., and Isaza‐Castro J.. 2011. A Systematic Review of Agricultural Interventions That Aim to Improve Nutritional Status of Children. In: London: EPPI‐Centre, Social Science Research Unit, Institute of Education, University of London.
35. Mayanja, M. N. , Rubaire‐Akiiki C., Greiner T., and Morton J. F.. 2015. “Characterising Food Insecurity in Pastoral and Agro‐Pastoral Communities in Uganda Using a Consumption Coping Strategy Index.” Pastoralism 5, no. 1: 11.
36. Melese, T. 2013. Nutritional Status of Ethiopian Weaning and Complementary Foods: A Review. [Institutional report]. Addis Ababa University.
37. Michaelsen, K. F. , Grummer‐Strawn L., and Bégin F.. 2017. “Emerging Issues in Complementary Feeding: Global Aspects.” Maternal & Child Nutrition 13: e12444. doi.org/10.1111/mcn.12444
38. Mosites, E. , Aol G., Otiang E., et al. 2017. “Child Height Gain Is Associated With Consumption of Animal‐Source Foods in Livestock‐Owning Households in Western Kenya.” Public Health Nutrition 20, no. 2: 336–345. doi.org/10.1017/S136898001600210X
39. Moursi, M. M. , Arimond M., Dewey K. G., Trèche S., Ruel M. T., and Delpeuch F.. 2008. “Dietary Diversity Is a Good Predictor of the Micronutrient Density of the Diet of 6‐to 23‐month‐old Children in Madagascar.” Journal of Nutrition 138, no. 12: 2448–2453. doi.org/10.3945/jn.108.093971
40. Muhimbula, H. S. , Kinabo J., and O'Sullivan A.. 2019. “Determinants of Infant Nutritional Status in Rural Farming Households Before and After Harvest.” Maternal & Child Nutrition 15, no. 3: e12811. doi.org/10.1111/mcn.12811
41. Nabarro, D. , and Wannous C.. 2014. “The Potential Contribution of Iivestock to Food and Nutrition Security: The Application of the One Health Approach in Livestock Policy and Practice.” Revue scientifique et technique (International Office of Epizootics) 33, no. 2: 475–485. doi.org/10.20506/rst.33.2.2292
42. NBS and ICF Macro . 2015. ‐16. Tanzania Demographic and Health Survey and Malaria Indicator Survey (TDHS‐MIS) 2015‐16.
43. De Onis, M. 2006. WHO Child Growth Standards: Length/Height‐For‐Age, Weight‐for‐Age, Weight‐for‐Length, Weight‐for‐Height and Body Mass Index‐for‐Age. WHO.
44. Organization, W. H. 2023. Levels and Trends in Child Malnutrition: UNICEF/WHO/World Bank Group Joint Child Malnutrition Estimates: Key Findings of the 2023 Edition. In Levels and Trends in Child Malnutrition: UNICEF/WHO/World Bank Group Joint Child Malnutrition Estimates: Key Findings of the 2023 Edition.
45. Rodríguez, S. C. , Hotz C., and Rivera J. A.. 2007. “Bioavailable Dietary Iron Is Associated With Hemoglobin Concentration in Mexican Preschool Children.” Journal of Nutrition 137, no. 10: 2304–2310. doi.org/10.1093/jn/137.10.2304
46. Ruhangawebare, G. K. 2010. “Factors Affecting the Level of Commercialization Among Cattle Keepers in the Pastoral Areas of Uganda.” Master's thesis, Makerere University.
47. Schalekamp, C. V. Z. 2009. “Seasons of Hunger: Fighting Cycles of Quiet Starvation Among the World's Rural Poor, Stephen Devereaux, Bapu Vaitla and Samuel Hauenstein Swan: Book Review.” African Safety Promotion 7, no. 2: 55–56.
48. Shreenath, S. , Watkins A., Wyatt A., et al. 2011. “Exploratory Assessment of the Relationship Between Dairy Intensification, Gender and Child Nutrition Among Smallholder Farmers in Buret and Kipkelion Districts, Kenya.” International Livestock Research Institute (ILRI).
49. Stroebel, A. , Swanepoel F. J. C., and Pell A. N.. 2011. “Sustainable Smallholder Livestock Systems: A Case Study of Limpopo Province, South Africa.” Livestock Science 139, no. 1: 186–190.
50. Sunderland, T. C. H. 2011. “Food Security: Why Is Biodiversity Important?” International Forestry Review 13, no. 3: 265–274.
51. Victora, C. G. , Bahl R., Barros A. J. D., et al. 2016. “Breastfeeding in the 21st Century: Epidemiology, Mechanisms, and Lifelong Effect.” Lancet 387, no. 10017: 475–490. doi.org/10.1016/S0140-6736(15)01024-7
52. Wagah, M. , Hodge J., and Lewis A.. 2015. Leveraging Agriculture for Nutrition in East Africa (LANEA). In.
53. Weitzman, M. L. 2000. “Economic Profitability Versus Ecological Entropy.” Quarterly Journal of Economics 115, no. 1: 237–263.
54. World Health Organization , United Nations Children's Fund, Food and Agriculture Organization of the United Nations, International Food Policy Research Institute, and United States Agency for International Development. 2008. Indicators for Assessing Infant and Young Child Feeding Practices: Part 1: Definitions. World Health Organization.
55. WHO and UNICEF . 2009. WHO Child Growth Standards and the Identification of Severe Acute Malnutrition in Infants and Children: A Joint Statement by the World Health Organisation and the United Nations Children's Fund.
56. Wiegers, E. , Dorp M. van, and Torgerson S.. 2011. Improving Nutrition Through Agriculture: Viewing Agriculture‐Nutrition Linkages Along the Amallholder Value Chain.
57. Wordofa, M. G. , and Sassi M.. 2020. “Impact of Agricultural Interventions on Food and Nutrition Security in Ethiopia: Uncovering Pathways Linking Agriculture to Improved Nutrition.” Cogent Food & Agriculture 6, no. 1: 1724386.

---

Republished from the open web under CC-BY. Authors: Muhimbula H, Kinabo J, O'Sullivan AM. Read the original.