Disponibilités alimentaires, niveaux de risque apparent et état nutritionnel en milieu rural sahélien burkinabé

Since the food crises of the 70's and 80's, sahelian countries have explored several strategies to deal with the food insecurity risk. However, the different approaches used often gave most importance to the apparent order of things but failed to take into account the complex causality of the food vulnerability and of its nutritional consequences. Our research was undertaken in a province in the East Burkina Faso, to investigate similarities or disparities between food vulnerability as assessed on a territorial basis through apparent risk levels, food vulnerability as measured by in-depth information on its determinants, and the nutritional status of the inhabitants. Indicators on climatic or geographical constraints and on environment management were gathered in each of the 280 villages of the province. They allowed us to define four areas according to the severity and repeatability of food risks. Repeated cross-sectional surveys were carried out on a randomly selected sample of 180 “collective dwelling units” (CDUs or compounds) in 30 villages, including 600 households and 5600 people. The data collected encompassed social, demographical and economic characteristics, management of the last seasonal food shortage, assessment of cereal availabilities, and standardized anthropometric measures. The nutritional indices used were the body mass index (BMI) for adults, the age-adjusted BMI for adolescents (10-20 y), and the weight-for-age Z-score (WAZ) for 0-9 year old children. As compare to the other areas, the malnutrition rates were lower in the one where the apparent food risk was the highest: in this area, the mean WAZ was higher among 0-4 y as well as among 5-9 y children (p=0.005 and p<0.0001, respectively); the mean BMI was higher among 10-20 y boys (p=0.012) and girls (p=0.0003), and among adults, males (p=0.01) and female as well (p<0.0001). In the same area of highest apparent risk, the seasonal food shortage was earlier anticipated (p=0.01), before this season the mean household income was higher, and the final cereal availabilities were better (p=0.03). When adjusting on the above characteristics, the relationship between apparent risk areas and malnutrition rates is only slightly modified. But when adjusting also on social and demographical characteristics, this relationship remained significant only among 5-9 y children (p=0.01), 10-20 y girls (p=0.03) and female adults (p=0.008). The results showed that apparent risk levels, as assessed through environmental constraints, and malnutrition rates are far to be superimposable. Our analysis suggested that households living where food production is regularly difficult have developed on the long run adapted strategies to reduce the consequences of seasonal hazards.Depuis les crises alimentaires des années 70 et 80, les pays sahéliens ont exploré diverses stratégies pour gérer le risque d'insécurité alimentaire. Cependant, privilégiant souvent l'ordre apparent des choses, ces approches intègrent mal les déterminants complexes de la vulnérabilité alimentaire et de ses conséquences nutritionnelles. La recherche que nous avons menée, dans une province de l'Est du Burkina Faso, avait pour objectif d'explorer les concordances ou disparités entre un diagnostic territorial de la vulnérabilité alimentaire, hiérarchisant des niveaux de risque apparent, une étude approfondie de ses déterminants et l'état nutritionnel des individus. Des indicateurs des contraintes climatiques, géographiques et de gestion du milieu, recueillis dans les 280 villages de la province, ont permis de hiérarchiser quatre zones selon l'intensité et la récurrence des risques alimentaires. Des enquêtes ont été menées dans un échantillon aléatoire de 180 cours collectives (pour 30 villages, 600 ménages et 5600 individus). Elles ont recueilli des informations socio-démographiques, économiques, sur la gestion de la soudure alimentaire (anticipation, durée, vente de bétail, aide), une estimation des disponibilités céréalières et des mesures anthropométriques des individus. Seuls les indices nutritionnels de déficit énergétique sont analysés ici : indice poids-taille (P-T) de 0 à 9 ans, indice de masse corporelle (IMC) chez les adultes et IMC-pour-âge de 10 à 20 ans. Un niveau de malnutrition plus faible est observé dans la zone de risque apparent maximal : meilleur indice P-T chez les moins de 5 ans (p=0.005) comme chez les 5-9 ans (p<0.0001), IMC plus élevé chez les adolescents garçons (p=0.012) et filles (p=0.0003), et chez les adultes hommes (p=0.01) comme femmes (p