Running head: ANIMAL SOURCE FOOD (ASF) CONSUMPTION 1 Animal Source Food (ASF) Consumption in Rural Niger: What are the Drivers? Krinna Patel University of Florida
ANIMAL SOURCE FOOD (ASF) CONSUMPTION 2 Abstract This paper identif ied drivers of animal source food (ASF) among rural populations in Niger. This region suffer s from high rates of chronic malnutrition and its people are vulnerable to food insecurity the consequence of depleted agricultural lands, severe droughts and floods, and extreme heat, all of which is being exacerbated by climate change. P opulations in thi s area, including agricultural, agropastoral, and pastoral populations, typically have livestock and have access to their byproducts, including meat milk, and sometimes eggs y et, consumption of ASF nutrient dense foods remains low. This paper examine d the role that food security, livelihood, agroecology, and demographic context play in determining ASF consumption in a population of rural Nigerien agropastoral and pastoral households This paper described ASF consumption using data collected in 2011 in eastern Ni ger. Key findings suggest ed that household (HH) ethnicity, head of household (HHH) education level, HH agroecology, and food security play a significant role in determining consumption of ASF within the household U nderstanding how these characteristics influence ASF consumption may assist policy makers and global health practitioners in developing, targeting, and implementing nutritional intervention s and programs that improve nutritional outcomes through increase d ASF consumption in low income settings Keywords: animal source food (ASF), food security, Niger, pastoralism, livelihood
ANIMAL SOURCE FOOD (ASF) CONSUMPTION 3 Animal Source Food (ASF) Consumption in Rural Niger: What are the Drivers? Background Niger, a country located in the African Sahel, is one of the least developed countries in the world (United Nations Human Development Program [UNDP] 2016). Due to complex, widespread structural determinants of poverty, life expectancy is 61.9 years and 45.7% of the population liv es below the international poverty line of $1.90 a day ( UNDP 2016). The prevalence rate of child stunting, an indicator of chronic malnutrition, is 43.0%, and child wasting, an indicator of acute malnutrition, is 19.0%, both of which are significantly greater than the global threshold s [UNICEF] 2016). These persistent rates of malnutrition reflect a complex set of underlying determinants, including poverty, po or hygiene and sanitation, and limited access to health rapidly growing population, land degradation, climate change, reduced income generating opportunities, and gender disparities combine to position it consistently among the poorest an d most malnourished countri es in the world (Baro & Duebel, 2006 and UNDP, 2016). Further aggravating malnutrition in Niger is the arid, dry climate that makes land use for agricultural production tenuous and has resulted in extensive food insecurity and s everal food crises over the last century (Gambo Boukary, Diaw, & Wnscher, 2016) The World Food availability and adequate access at all times to sufficient, (World Food Programme [WFP] 2017). On an annual basis, 22% of Niger's population suffers from chronic food insecurity (United States Agency for International Development, 2011). These food insecure households suffer from weak livelihood bases and spend u p to 72% of their income on food (World Health Organization [WHO] 2006). An assessment of the daily food intake of rural Nigerien women indicate s that starchy foods, including grains, are the most commonly consumed items, while animal source foods ( ASF ) a re the least commonly consumed food items (Cisse Egbuonye, Ishdorj, McKyer, & Mkuu, 2017).
ANIMAL SOURCE FOOD (ASF) CONSUMPTION 4 Although the arid and semi agroecolog y makes crop production extremely tenuous, the environment has fostered livestock production in various forms for hundreds of years. Pastoralism in the Sahel has been described as the specific relationship that has evolved between livestock, people, and the environment in resource limited, arid areas of the world ( McKune & Silva, 2013). Pastoralism a nd livestock production are practiced by nearly all households throughout rural Niger. The value of livestock to households, not onl y pastoral households, but agro pastoral and a gricultural households as well, is well documented. Agricu lture represents 40 % o f national GDP, with livestock constituting 33% percent of agriculture related GDP (International Trade Commission, Department of Commerce, United States of America, 2017 ). Livestock production provides a financial liquidity, allowing households to meet short and long term family needs, especially in times of drought and food crises ( Ayantunde, De Leeuw, Turner, & Said, 2011). Livestock also provide draft power and inputs, through manure, that improve soil fertility and increase crop yields among farmers ( Powell, Pearson, & Hiernaux, 2004). However, the contributions of livestock in meeting the nutritional needs of the household is often overlooked. Particularly among poorer families, self consumption of ASF produced by household livestock may not be prio ritized ; families may focus more on maintaining their livestock herds, selling livestock or byproducts when available, and ensuring that there is any food (most often, grain) available to be consumed every day. ASF can provide a variety of macro and micronutrients that are difficult to obtain in adequate quantities from plant source foods alone and are essential to proper development of infant and young children (Murphy & Allen, 2003) ASF are associated with increased cognitive function, resistance to infection, improved pregnancy outcomes, and improved school performance of children ( Neumann, Harris, & Rogers, 2002). A randomized control trial conducted in 2015 showed that consumption of one egg a day among children aged 6 9 months can reduce stunti ng in children by 47% (Iannotti et al., 2017). The same study found a 74% reduction in the number of underweight children by the
ANIMAL SOURCE FOOD (ASF) CONSUMPTION 5 conclusion of the trial (Iannotti et al., 2017). Supplemental ASF consumption in the diets of school children in developing nat ions has been shown to improve Vitamin B 12 concentrations and limit the incidence of malnutrition ( McLean et al., 2007) ASF supplementation among school children in Kenya showed increase cognitive performance, as indicated by RAVEN scores at school ( Neuma nn et al., 2007 ) Though there is evidence supporting ASF consumption for improved full child growth and development, research also indicates significant socioeconomic disparities in ASF consumption. A recent study in urban Ethiopia found households of greater socioeconomic status consumed significantly more ASF th an households with lower socioe conomic status ( Workicho et al., 2016). Studies conducted within Niger show limited variation in the consumption of ASF between rural populations i n the Zinder and Maradi regions with most household diets containing non ASF (C isse Egbuonye, Ishdorj, McKyer, & Mkuu, 2017). Thus, a full understanding of the barriers and facilitators of ASF consumption is needed Aims and Hypotheses This analysis of ASF consumption among h ouseholds surveyed in 2011 examine s a number of demographic factors historically associated with nutritional outcomes These include sex of the head of household (HHH), education level of the HHH, and household (HH) ethnicity. In addition to these demographic factors this analysis explore s the relationship between HH agroecology (pastoral, agropastoral or agricultural), livelihood (pastoral, agropastoral or agric ultural ) and food security (secure, moderate, or insecure) with ASF consumption with in these communities. The aim of this study is to identify the factors associated with increased ASF consumption among agricultural, agropastoral, and pastoral households Niger. Using secondary dat a collected in 2011 in the Zinder region of eastern Niger, t his study provides important baseline information on the nutritional intake of ASF in Niger and delin eates the specific types of ASF consumed in these communities This information serve s as a fou ndation from which
ANIMAL SOURCE FOOD (ASF) CONSUMPTION 6 researchers and development practitioners may operate to increase ASF consumption among high risk populations specifically in rural, low income livestock rich environments, such as Niger, where malnutrition remain s endemic. Methods Study D esign, S etting, P articipants This analysis was based on data collected in 2010 2011 through a cross sectional study of rural communities in eastern Niger, whose methods included collection through household interviews, focus group discussions, key i nformant interviews, and anthropometric measurements (McKune, 2012). The original study sought to understand household vulnerability, food insecurity, nutrition, adaptation, and resilience to climate change. Surveys of 133 households provided information o n dietary intake, including 24 hour food recall for men, women, and children on specific foods. This data included information on meat, poultry, fish, dairy (milk or yogurt), and egg consumption. Re analysis of the dietary diversity data allow for formatio n of the principal measurement of this study ASF consumption. Statistical P rocedures and S tatistical M easures SPSS (Statistic al Package for the Social Sciences) software was used to execute descriptive statistics and qualitative data analysis on all the variables included in this study. Binary logistic regression and bivariate association (Chi computed to investigate the effects of independent variables on ASF consumption. Independent variables included in this analysis include an agroecology zone, livelihood continuum index, and a food security in order to examine their role in household ASF consumption (McKune, 2012). The agroecology zones were defined by the 2012 Famine Early Warning System livelihood/food zones of Niger (Famine Early Warning Systems Network [ FEWSNET], 2012). Households were a ssigned to one of three zones based on their location: agricultural, agropastoral, or pastoral, which represented their location in 2011. The livelihood continuum is a multi item index developed to reflect a range of
ANIMAL SOURCE FOOD (ASF) CONSUMPTION 7 livelihoods found in these areas. The i ndex range is from 1 livelihood is represented at the other (7). The index was created based on both objective and subj ective indicators of livelihood, ranging from self identification, ethnicity, and ancestral livelihood patterns (all indicated through household surveys), to herd size, and mobility (see ( McKune, 2012 ) for more detail). The index captures livelihoods at a specific moment in time, allowing for relative comparison across groups as well as the potential to compare the same population over time. The food security index, based on (Maxwell et al., 1999), was calculated based on the frequency and severity of copin g mechanisms used by households in times of crisis, including decreased portions at meals, decreased nutritional quality at meals, changed eating habits, decreased number of meals, collection of wild famine foods, day without eating, and sale o f personal b elongings (Maxwell, 1999 & McKune, 2012). The food security index ranges from 1 15. Using SPSS, data were then binned into three categories at natural breaks and households were assigned to one of the following categories: most secure, moderately secure, a nd most insecure. It is important to note that this is a relative description from within the sample population, thus all households may be food insecure (or secure), but the categories denote food security relative to the whole sampled population. The inc lusion of this measure will assist in determining if ASF consumpt ion differed among households based on their different experiences of food insecurity in 2011. Using the 24 hour food recall data, a variable for overall ASF consumption was created which in cluded information on the intake of meat, poultry, fish, and dairy. Initial analysis was conducted for men, women, and children within each household. No intra household differences were determined between men, women, and children. Given that reported data on ASF were indicator for household ASF consumption. The ASF indicator used is a dichotomous variable describing if the household consumed any ASF or not within the p revious 24 hours.
ANIMAL SOURCE FOOD (ASF) CONSUMPTION 8 In addition to the ASF variable describing consumption of any dairy, meat, fish, poultry, or eggs within the household, ASF consumption was broken down into more detailed indicators. The Meat ASF variable was created using data reported d uring 24 hour recall of consumption of meat, poultry, and fish for men, women, and children, separately. Similar to the ASF indicator, no intra household differences were identified (men, women, and children), so a dichotomous variable of the overall house consumption. Likewise, the Dairy (milk and yogurt) and Eggs variables were analyzed separately from Meat to determine differences in the consumption of these products at the household level. The Dairy variable was created using the same methodology for the ASF and Meat variables. Egg consumption was zero across all households, thus it has bee n excluded from further analysis Distinction of these variables ( Meat Dairy and Eggs ), in addition to the more general ASF variable, is important in these populations, as literature indicates that milk consumption is typically higher among more mobile pastoral populations and is historically credited with keeping nutritional status of mob ile pastoral children better than children of nearby settled families ( Fratkin E. & Roth, E.A. (Eds.) 2005 ). Thus, independent variables (agroecology, food security, and livelihood) were assessed against not only ASF, but each of these outcome variables as well. D emographic indicators that were explored as confounders include HHH sex, HH ethnicity, and HHH education level as reported in 2011. Results Variable Descriptions A total of 133 household surveys were analyzed. Table 1 1 depicts the characterist ics of the participants and households in the survey. Although most of the survey participants were male (79.7%), female HHHs were present, making up 20.3% of the total survey respondents. Based on the (FEWSNET, 2012) map of agroecology zones, 14.3% of households in this sample were located within the agricultural zone, 28.6% were in the agropastoral zone, and
ANIMAL SOURCE FOOD (ASF) CONSUMPTION 9 57.1% were in the pastoral zone. Households within the agricultural zone were used as the reference population in later analysis. The livelihood continuum index showed households across a range of livelihoods, ranging from those highly dependent upon crop based agriculture (25 households, or 18.8%) to those dependent on highly mobile pastoralism (5 households, or 3 .8%). Table 1 1 presents the distribution of households across the livelihood continuum, rounding to the nearest whole number between 1 7. Notably, no household existed at the highly agricultural pole (1), but 5 households were determined to exist at the highly pastoral pole (7) during the time of this survey. Likewise, descriptive analysis of the food security index showed 35.3% households as most food secure, 17.3% as moderately food secure, and 45.1% as most food insecure. F ood insecure households wer e utilized as the reference population in later analysis. ASF consumption data, the key outcomes in this study, were categorized into four distinct categories: Meat, Dairy, Eggs, and any ASF Among households reporting 24 hour food recall data, 14.3% consu med meat, 36.8% consumed dairy, and 0% of households consumed eggs. These combine for a total of 38.3% of households reporting any ASF consumption in the past 24 hours, or 51 of the 133 households sampled. During the survey, demographic data were collecte d on HHH education, HHH sex, and household ethnicity. Education data indicate that 78.9% of HHHs had received no formal education, 2.3% attended primary school but did not complete their studies, 4.5% completed primary school, 1.5% attended some secondary school, and 12% attended only Koranic school. For analysis purposes, the no education population was used as a reference group to determine statistically significant associations between education levels and ASF Of the 133 households, 24.8% identified as Hausa, 33.8% Tuareg, 23.3% Peul, and the remaining 18.0% were Beri Beri. The Djerma ethnicity was not represented in the sample population. The Tuareg ethnicity,
ANIMAL SOURCE FOOD (ASF) CONSUMPTION 10 being the largest ethnic group repr esented in the data set, were used as the reference populat ion in later analysis of ASF consumption between ethnicities. Table 1 1. Characteristics of the population 2010 2011 Eastern Niger Data Set Overview Frequency (Percentage) of Sample Population Frequency (Percentage) Consum ing any ASF Total 133 (100) 51 (38.3) HHH Sex Male Female 106 (79.7) 27 (20.3) 43 (40.6) 8 (29.6) Household Ethnicity Hausa Djerma Tuareg Peul Beri Beri 33 (24.8) 0 (0.0) 45 (33.8) 31 (23.3) 24 (18.0) 13 (39.4) 0 (0.0) 26 (57.8) 5 (16.1) 7 (29.2) HHH Education None Primary Incomplete Primary Complete Secondary Koranic School only Missing 105 (78.9) 3 (2.3) 6 (4.5) 2 (1.5) 16 (12.0) 1 (0.8) 37 (35.2) 0 (0.0) 4 (66.7) 0 (0.0) 9 (56.3) 1 (0.0) Household Agroecology Zone Agricultural Hausa Tuareg Agropastoral Hausa Beri Beri Pastoral Hausa Tuareg Peul Beri Beri 19 (14.3) 1 (0.8) 18 (13.5) 38 (28.6) 19 (14.3) 19 (14.3) 76 (57.1) 13 (9.8) 27 (20.3 ) 31 (23.3) 5 (3.8) 11 (57.9) 1 (100.0) 10 (55.6) 7 (18.4) 5 (26.3) 2 (10.5) 33 (43.4) 7 (53.8) 16 (35.6) 5 (16.1) 5 (100.0) Household Livelihood Continuum 1.00 2.00 3.00 4.00 5.00 6.00 7.00 Missing 0 (0.0) 25 (18.8) 47 (35.3) 22 (16.5) 14 (10.5) 15 (11.3) 5 (3.8) 5 (3.8) 0 (0.0) 13 (52.0) 14 (29.8) 9 (40.9) 6 (42.9) 7 (46.7) 0 (0.0) 2 (40.0) Household Food Security Most Secure 47 (35.3) 23 (48.9)
ANIMAL SOURCE FOOD (ASF) CONSUMPTION 11 Moderately Secure Most Insecure Missing 23 (17.3) 60 (45.1) 3 (2.3) 11 (47.8) 16 (26.7) 1 (33.3) Animal Source Food (ASF) Consumption Any ASF (meat, poultry, fish, dairy) Missing Consumed no ASF Meat (meat, poultry, fish) Missing Dairy (milk and yogurt) Missing Eggs 51 (38.3) 2 (1.5) 80 (60.2) 19 (14.3) 1 (0.8) 49 (36.8) 1 (0.8) 0 (0.0) Statistical Tests and Outcomes Demographic associations with ASF Ethnicity. Binary regression tests were calculated to distinguish associations between HH ethnicity and ASF consumption. Statistical tests illustrated that ethnicity is a predictor for not only ASF, but meat and dairy consumption, respectively, as well. Table 2 1. ASF Consumption by ethnicity Household Ethnicity Unstandardized Coefficient Tuareg (reference population) --Hausa .856* Peul 2.074** Beri Beri 1 .312** Constant .425 Model Fit Adjusted R Square + .163 ** N (# of participants) 131 Notes: *p<0.1; **p<0.05; --reference population; + Nagelkerke R Square Table 2 1 emphasizes the role that ethnicity plays as a determinant of ASF consumption. Examination into the relationship of ethnicity and ASF consumption showed that when
ANIMAL SOURCE FOOD (ASF) CONSUMPTION 12 compared to the Tuareg, each other ethnic group within the sample pop ulation was sig nificantly less likely to consume any ASF during the time period. Table 2 2. Meat Consumption by ethnicity. Household Ethnicity Unstandardized Coefficient Tuareg (reference population) --Hausa .526 Peul 19.538 Beri Beri .056 Constant 1.665 Model Fit Adjusted R Square + .155 ** N (# of participants) 132 Notes: *p<0.1; **p<0.05; --reference population; + Nagelkerke R Square For ethnicity and Meat consumption using binary logistic regression tests, no significant differences were found when comparing each of the ethnic groups to the Tuareg, ind icating that ethnicity was not predictive of meat consumption for this particular sample population (Table 2 2). Table 2 3. Dairy Consumption by ethnicity. Household Ethnicity Unstandardized Coefficient Tuareg (reference population) --Hausa 1.061** Peul 2.016** Beri Beri 1.255** Constant .368 Model Fit Adjusted R Square + .158 ** N (# of participants) 132 Notes: *p<0.1; **p<0.05; --reference population; + Nagelkerke R Square
ANIMAL SOURCE FOOD (ASF) CONSUMPTION 13 Likewise, the use of regression tests outlined major associations for ethnicity and Dairy consumption (Table 2 3). Compared to the Tuareg, each ethnicity was less likely to consume dairy ASF Historically, the Peul are known for their high rates of dairy c onsumption. T hese data indicate that they were less likely to consume dairy than the Tuareg population, a finding that is further explored in the discussion section. Addi tionally, this analysis provided spe cific evidence that ethnicity is a significant pre dictor of dairy consumption. HHH sex. Chi associations between HHH sex and ASF consumption. When analyzed to determine if HHH sex was associated with ASF consumption, the results were not si gnificant (p=.499). Correspondingly, Chi associated with Meat consumption, the results of which were not significant (p=.120), and for Dairy consumption the results were not significant (p=.824 ). The lack of associations between HHH sex and indicators of ASF consumption indicate d that living in a mal e or female headed household was not predictive of ASF consumption in these populations. This variable was removed from any further statistical proc edures and analysis. HHH Education. Similarly to HH ethnicity, binary logistic regression tests were run to identify any associations between the educational status of the HHH and ASF consumption. A significant association between HHH education and Meat consumption was identified. Specifically, completion of primary school by the HHH was associated with increased consumption of meat ASF (p=.011) when compared to the uneducated group. Those educated through Koranic school also showed significance for impr oved meat consumption (p=.022). Likewise, HHHs educated through Koranic school were also significantly positively associated with Dairy consumption (p=.088) compared to the uneducated population These findings suggest ed that higher education levels of HH Hs are associated with greater consumption of meat and dairy products No significant association between HH education levels and the any ASF variable were identified.
ANIMAL SOURCE FOOD (ASF) CONSUMPTION 14 HH Agroecology. Binary Logistic regression tests were applied to identify associations between HH agroecological zones and ASF, Meat, and Dairy consumption. Findings indicate d that HH s located within the agropastoral agroecolog ical zone are associated with greater ASF consumption overall (p=.002) and Dairy consumption (p=.004) Agropastoral zoned HHs were significantly less likely to consume any ASF or dairy products than the agricultural group. Agroecolog ical zone did not prove to be a significant predictor of Meat consumption. Food security. Similarl y regression tests were calculated to determine associations between food security and ASF, Meat, and Dairy consumption. Findings suggest ed that most secure HHs and ASF (p=.011) and moderately secure HHs and ASF (p=.070) were positively associated with increased intake of any ASF Findings for food security and Meat consumption found no significant association. However, the most food secure HHs were more likely to consume Dairy ASF (p=.026) than the most food insecure population. These developments suggest ed that improved food security is associated wit h higher consumption of ASF and specifically, dairy ASF Livelihood continuum. Chi associations between livelihood and ASF, Meat, and Dairy consumption. Results showed no association between the livelihood continuum and ASF consumption (p=.207). This finding indicated that there is no association between livelihood and the intake of ASF within this sample population. Additionally, no significant association was found between livelihood and Meat associations between livelihood and Dairy consumption (p=.255 ). Cofounder Models of Independent Variable A ssociations with ASF Multiple binary logistic regression models were generated to examine the relationship between the independent variables of agroecology, food security, and livelihood on ASF consumption. Potential cofounders were included in each consecutive model. The se included HH ethnicity and HHH education level.
ANIMAL SOURCE FOOD (ASF) CONSUMPTION 15 Table 3 1. Binary Logistic Regression Models Investigating the Effects of HH Ethnicit y, HHH Education, HH Agroecology, Food Se curity, and Livelihood on ASF consumption. Independent Variables Model 1 Model 2 Model 3 Model 4 Model 5 HH Ethnicity Tuareg (reference population) ----------Hausa .865* .512 .273 1.081 1.020 Peul 1.927** 1.863** 1.891** 1.659** 1.673** Beri Beri 1.579** .202 .266 .632 .863 HHH Education No education (reference populati on) ----------Primary incomplete 20.087 19. 515 19.561 19.524 19.612 Primary complete 1.558 1.115 .955 .951 .787 Secondary 20.294 21 .036 20.960 21.493 21.405 Koranic School .995 .998 .919 .954 858 HH Agroecology Zone Agricultural (reference popul ation) --------Agropastoral 2.418** 2.170** 2.942** 2.774** Pastoral .171 .078 .226 .248 Food Security Index Insecure (reference population) ----Moderate .409 .476 Secure .756 .868* Livelihood Continuum .047 .017 Constant .274 .381 .022 .516 .168 Model Fit Adjusted R Square + .235** .331** .354** .352** .380** N (# of participants ) 130 130 127 125 122 Notes: *p<0.1; **p<0.05; --reference population; + Nagelkerke R Square Table 3 1 presents the binary logistic regression associations for the independent variables and ASF consumption. Each of the predictor and confounding variables were included in the models to estimate their influence on ASF intake. Among cofounders, HH ethnicity, specifically being of Peul ethnicity, was significantly negatively associated with consumption of ASF compared to the Tuareg. Results show ed that there is a significant relationship between
ANIMAL SOURCE FOOD (ASF) CONSUMPTION 16 being in an agropastoral zone and the intake of ASF In other words, households within the agropastoral agroecology zone were significantly less likely to consume ASF than those within the agricultural zone (reference population) across all the models. Findings also showed that when agroecological zone food security, and livelihood were included in a model together, HHs within the agropastoral zone (p=.062) were still less likely to consume ASF than those in the agricultural zone and food secure HHs were more likely to consume ASF than food ins ecure HHs (p=.089). Regression analysis found that the livelihood index is unassociated with ASF consumption for this population. Table 3 2. Binary Logistic Regression Models Investigating the Effects of HH Ethnicit y, HHH Education, HH Agroecology, Food Security, and Livelihood on Meat consumption. Independent Variables Model 1 Model 2 Model 3 Model 4 Model 5 HH Ethnicity Tuareg (reference population) ----------Hausa .599 1.778** 1.677** 1.427 1.360 Peul 19.167 18.698 18.685 18.535 18.547 Beri Beri .238 1.269 1.265 .795 .807 HHH Education No education (reference population) ----------Primary incomplete 18.850 18.102 18.122 18.217 18.223 Primary complete 1.982** 1.839* 1.811* 1.607 1.600 Secondary 19.383 19.734 19.967 19.869 20.025 Koranic School 1.242* 1.329* 1.327* 1.234* 1.247* HH Agroecology Zone Agricultural (reference population) --------Agropastoral 2.574** 2.389** 2.420** 2.286** Pastoral .996 .935 .572 .541 Food Security Index Insecure (refe rence population) ----Moderate .505 .327 Secure .129 .063 Livelihood Continuum .362 .335 Constant 2.060 1.543 1.72 0 .320 .509 Model Fit
ANIMAL SOURCE FOOD (ASF) CONSUMPTION 17 Adjusted R Square + .258** .332** .339** .357** .359** N (# of participants ) 131 131 128 126 123 Notes: *p<0.1; **p<0.05; --reference population; + Nagelkerke R Square Table 3 2 presents the binary logistic regression model for agroecology, food security, and livelihood on Meat In the presence of the food security indicator we found that HHH education was significantly associated with increased meat consumption. Specif ically, among HHs where the HHH completed primary school, more meat is consumed and among HHs where the HHH atten ded Koranic school more meat was consumed. Findings also indicated that being Hausa was significantly positively associated with meat consumption compared to the Tuareg, when food secu rity and agroecological zone were included in the model. Additionally, the models found that food securi ty on any level and livelihood, were not predictive of meat consumption beyond the effects o f household education and HH ethnicity. However, the models consistently found that households within the agropastoral zone of agroecology were significantly associ ated with less meat consumption compared to those within the agricultural zone. These findings remained significant (p=.048) in the final model, where all three independent variables (agroecology, food security, and livelihood) were considered, suggesting that households within the agropastoral zone were less likely to consume meat ASF than the agricultural zoned group. Table 3 3. Binary Logistic Regression Models Investigating the Effects of HH Ethnicity, HH H Education, HH Agroecology, Food Security, and Livelihood on Dairy consumption. Independent Variables Model 1 Model 2 Model 3 Model 4 Model 5 HH Ethnicity Tuareg (reference population) ---------Hausa 1.127** .026 .113 .642 .631 Peul 1.868** 1.903** 1.935** 1.761** 1.823** Beri Beri 1.592** .116 .183 .579 .823
ANIMAL SOURCE FOOD (ASF) CONSUMPTION 18 HHH Education No educa tion (reference population) ----------Primary incomplete 19.911 19.339 19.365 19.400 19.456 Primary complete 1.720* 1.295 1.165 1.184 1.065 Secondary 20. 066 20.708 20.578 21.052 20.909 Koranic School 1 .154* 1.164* 1.126 1.105 1.050 HH Agroecology Zone Agricultural (reference population) --------Agropastoral 2. 213** 2.013** 2.661** 2.534** Pastoral .075 .154 .058 .079 Food Security Index Insecure (reference population) ----Moderate .1 86 .244 Secure .587 .704 Livelihood Continuum .037 .116 Constant .205 .166 .126 .011 .616 Model Fit Adjusted R Square + .237** .358** .346** .340** .362** N (# of participants ) 130 131 128 126 123 Notes: *p<0.1; **p<0.05; --reference population; + Nagelkerke R Square Similar to ASF and Meat, regression analysis was conducted to determine associations between the independent variables and Dairy consumption. Regression analysis showed that households within the agropastoral zone were less likely to consume dairy ASF compared to those within the agricultural zone, across all the models, including the final model (p=.012) when agroecology, food security, and livelihood were all included. Being of the Peul ethnic group was significantly negatively associated with dairy consumption, compared to the T uareg, across all models. These findings remaine d significant in the final model (p=.011) when agroecology, f ood security, and livelihood were consider ed indicating that the Peul were less likely to consume dairy than the Tuareg p opulation. The food security and livelihood indicators did not show association with dairy consumption for this sample population. Discussio n
ANIMAL SOURCE FOOD (ASF) CONSUMPTION 19 Animal source food has emerged as a critical source of necessary vitamins, minerals, and amino acids, demonstrating significant impact on child nutritional outcomes. The purpose of this study is to identify factors associated with ASF consumption in a samp le of households across an agricultural pastoral continuum in rural Niger where poverty and malnutrition are endemic. In the sample of 133 households, 38.3% of the households consumed any ASF in 24 hours, 14.3% of the population consumed a meat product, and 36.8% of the sample population consumed a dairy product. As indicated in the results section, most of the ASF consumed within the study population were dairy products, with significantly less meat being consumed, and no egg consumption whatsoever The analysis yielded a number of interesting results, some of which corroborate existing literature (e.g. a significant positive association between food security and ASF consumption), while other results ran counter to what literature and general knowledge o f the region and its people would have suggested (e.g. a significant negative association between being of Peul ethnicity and milk consumption). Significant as sociations between HH ethnicity and ASF, and HH agroecology and ASF, arose across multiple models Additionally, a significant positive association between HHH educational level and m eat consumption was identified. The implication and interpretation of these findings are discussed below. The association between HH ethnicity and ASF may be considered the most interesting finding of the study. While strong associations between HH ethnicity and ASF consumption were anticipated, the discovery that the Peul, a historically pastoral ethnic group in the region, consumed significantly less ASF, particularly milk than oth er groups is difficult to explain. Despite extensive understanding and knowledge that many of the Peul practice sedentary agricultural and agropastoral livelihoods in the region, they are well known for their production and sale of milk. Thus further research will be required to explain this finding. Some possible explanations that should be explored include the vulnerable position many Peul HHs may have been forced into, following the 2010 food crisis, when the data utilized in this study wa s
ANIMAL SOURCE FOOD (ASF) CONSUMPTION 20 collect ed in 2011 During the crisis, entire livestock groups decimated including loss of donkeys leaving the Peul in a moment of vulnerability, which may have been reflected in the study survey. Alternately, as Fratkin and others have documented, when pastoralists settle, they become extremely vulnerable, as it tends to be the most vulnerable who are forced out of mobile pastoralism, and their ability to produce crops and engage in sedentary livelihoods is limited by a lack of traditional knowledge of farming practices. The lack of milk consumption amongst the Peul, an ethnic group that has historically been associated with high rates of dairy consumption, may be explained using either rationale. Another perplexing finding was the significant associa tion of agroecology and ASF specifically that households located within the agricul tural zone are more likely than other groups to consume ASF. However, important to interpreting these data is the fact that nearly all households within the agricultural zon e were identified Tuareg, another historically pastoral e thnic group within the region. More research is needed to understand who these Tuareg households are. They may be mobile agropastoralist or pastoralists still practicing pastoralism but were captured in the s urvey during a time they were residing in an agricultural zone. Another interpretation is that they are in fact settled Tuareg populations who consume high levels of milk consumption. Finally, the significant positive association between HHH edu cat ion level and meat was not unexpected. Results showed there is an association between completing primary school and increased meat consumption, which corroborates widespread understanding that increases in education levels are asso ciated with decrease d prevalence of childhood health conditions and disabilities and lower rates of malnutrition. Results also suggested a more consistently positive association between attending Koranic school and meat consumption. In the context of Niger, the finding of an as sociation between Koranic schooling and meat consumption is excludes individuals who are competent in reading and writing in Arabic (which is learned in
ANIMAL SOURCE FOOD (ASF) CONSUMPTION 21 Koran ic school for purposes of reading and writing the Koran). More research is needed to understand more about those HHHs who attended Koranic school in order to perceive its relationship to HH meat consumption. If future research is to improve the intake of ASF, it is important that we begin to better understand how independent factors, such as ethnicity and livelihood interact with food security to improve nutritional outcomes of the population. Though this rese arch indicates that ethnicity, education, and agroecology are all significantly related to ASF consumption, it has also generated a number of hypotheses for future research. Limitations There are significant limitations to this research. First, there is a small sample population utilized in the survey and the participants are from the same geographical region, restricting the representativeness of the data. The data used in this study represent a subset of communities from a larger study conducted in 2005 thus were not randomly selected. T hese results may not be generally representative of the ethnicities included in this sample. For an extensive and comprehensive understanding of the drivers of ASF consumption in this regi on, a significantly larger sample size should be gathered, and surveys should be administered at various times of the year to account for seasonal differences (temporality) in agroecology, food security, and livelihood. Another important component to con sider is that the reported low rates of ASF consumption and the limited types consumed leave room for uncertainty in determining the drivers for the consumption of ASF The data collected for this analysis is entirely dependent on statistical software presents room for sampling and measurement errors. Additionally, it is critical to note that the households in this research are all food insecure, but they were categorized into most insecure, moderately secure, and most secure to simplify the data calculations and to create a baseline grouping of food insecurity levels for this population. All the
ANIMAL SOURCE FOOD (ASF) CONSUMPTION 22 households represent ed in this sample are highly vulnerable to food insecurity. Furthermore, this study did not take into consideration cultural and direct financial/economic influences as predictors of ASF consumption. Specific traditions, religious/ethnic beliefs, and custo ms may influence the consumption of ASF as well as financial insecurity. The thriving market for livestock and its efficient liquidity value was not considered in this study; livestock production and sales have shown to be significant predictors of ASF, es pecially in poverty stricken populations. Future Research A deeper analysis of pastoralist lifestyles, agroecology, ethnicity, and food intake practices may prove beneficial to developing sustainable nutritional practices for all livelihoods in the region More research on the teachings and education al practices of Koranic school will provide a better understanding of its function in improving ASF consumption specifically meat intake within these populations Targeted nutritional and educational programs for increasing ASF consumption among high risk populations can lead to improvements in multiple facets of health especially in the case of Niger, where dairy consumption is high among multiple ethnic groups, including the Peul. Understanding if and how milk consumption is evolving within these communities specifically the Peul, can provide further insight to improving nutritional outcomes in the region. An examination of their dietary habits, particularly, the consumption of meat and dairy products, may be used to determine if there is a scope for using these strategies to increase ASF consumption across the communities. Conclusion In this study, HH education level HH ethnicity, and HH agroecology are identified as significant drivers of consuming ASF The ( FEWSNET 2012) agroecological zones and food security index proved to be significantly associated ASF consumpt ion. The livelihood index proved to be an insignificant indicator of consuming ASF To p rocure a broader understanding of ASF in the African Sahel, strategic and targeted interventional programs considering a
ANIMAL SOURCE FOOD (ASF) CONSUMPTION 23 multitude of variables, including accessibility to improve dietary diversity, and affordability should be considered, and policy should be implemented to diminish food insecurity, increase livelihoods, and better the health outcomes of malnourished high risk populations, such as these communities in Niger.
ANIMAL SOURCE FOOD (ASF) CONSUMPTION 24 References Ayantunde, A. A., De Leeuw, J., Turner, M. D., & Said, M. (2011). Challenges of assessing the sustainability of (agro) pastoral systems. Livestock Science 139 (1), 30 43. Baro, M., & Deubel, T. F. (2006). Persistent hunger: Perspectives on vulnerability, famine, and food security in sub Saharan Africa. Annu. Rev. Anthropol. 35 521 538. Cisse Egbuonye, N., Ishdorj, A., McKyer, E. L. J., & Mkuu, R. (2017). Examining n utritional a dequacy and d ietary d iversity a mong w omen in Niger. Maternal and Child Health Journal 1 9. Famine Early Warning Systems Network (FEWSNET). (2012). Niger Livelihood Zones (2011). Retrieved from http://www.fews.net/fews data/335. Fratkin E. & Roth, E.A. (Eds.). (2005). As Pastoralists Settle. Springer US: Springer Verlag US. resilience to food insecurity in Niger. Sustainability 8 (3), 181. Iannotti, L. L., Lutter, C. K., Stewart, C. P., Riofro, C. A. G., Malo, C., Reinhart, G., & Waters, W. F. (2017). Eggs in e arly c omplementary f eeding and c hild g rowth: a r andomized c ontrolled t rial. Pediatrics e20163459. International Trade Administration, Department of Commerce, United States of America. (2017). Niger Agricultural Sector. Retrieved from https://atlas.media.mit.edu/en/profile/country/ner/. Maxwell, D., Ahiadeke, C., Levin, C., Armar Klemesu, M., Zakariah, S., & Lamptey, G. M. (1999). Alternative food security indicators: revisiting the frequency and severity of coping strategies'. Food Policy 24 (4), 411 429. McKune, S.L., (2012). Climate Change, l ivelihood and h ousehold v ulnerability in e astern Niger. (Unpublished doctoral dissertation). University of Florida, Gainesville, Florida.
ANIMAL SOURCE FOOD (ASF) CONSUMPTION 25 McKune, S. L., & Silva, J. A. (2013). Pastoralists under pressure: double exposure to economic and environmental change in Niger. Journal of Development Studies 49 (12), 1711 1727. McLean, E. D., Allen, L. H., Neumann, C. G., Peerson, J. M., Siekmann, J. H., Murphy, S. P. & Demment, M. W. (2007). Low plasma vitamin B 12 in Kenyan school children is highly prevalent and improved by supplemental animal source foods. The Journal of Nutrition 137 (3), 676 682. Murphy, S. P., & Allen, L. H. (2003). Nutritio nal importance of animal source foods. Th e Journal of Nutrition 133 (11), 3932S 3935S. Neumann, C., Harris, D. M., & Rogers, L. M. (2002). Contribution of animal source foods in improving diet quality and function in children in the developing world. Nutrition Research 22 (1), 193 220. Neumann, C. G., Murphy, S. P., Gewa, C., Grillenberger, M., & Bwibo, N. O. (2007). Mea t supplementation improves growth, cognitive, and behavioral outcomes in Kenyan children. The Journal of Nutrition, 137 (4), 1119 1123. Powell, J. M., Pearson, R. A., & Hiernaux, P. H. (2004). Crop livestock interactions in the West African drylands. Agronomy J ournal 96 (2), 469 483. Workicho, A., Belachew, T., Feyissa, G. T., Wondafrash, B., Lachat, C., Verstraeten, R., & Kolsteren, P. (2016). Household dietary diversity and a nimal s ource f ood consumption in Ethiopia: evidence from the 2011 Welfare Monitoring Survey. BMC Public Healt h 16 (1), 1192. UNICEF ) (2016). The State of the New York: UNICEF. United Nations Development Program (UNDP). (2016). Human Development Report 2016 New York: UNDP.
ANIMAL SOURCE FOOD (ASF) CONSUMPTION 26 United State Agency for International Development. (2011). USAID Office of Food for P eace Niger Bellmon Estimation. Retrieved from http://pdf.usaid.gov/pdf_docs/PBAAB942.pdf. World Food Programme [WFP] (2017). What is food security? Retrieved from https://www. wfp.org/node/359289. World Health Organization (WHO) (2006). Food s ecurity in Niger. Retrieved from http://www.who.int/hac/crises/ner/appeal/Niger_advocacy_Oct06.pdf.