Objective For measurement of progress towards the Millennium Development Goal (MDG) 1, reliable data on nutrition indicators of specific countries are essential. Malnutrition is also the main determinant for childhood mortality, which is addressed in MDG 4.
Methods In the health and demographic surveillance area of Kossi Province in north-western Burkina Faso, nutritional parameters were compared in two cohorts of young children of the same age range from eight villages. Surveys took place in June and December of the year 1999 and 2009. A multivariate model was used to control for confounding variables.
Results For the 1999 study, data were analysed for 179 and 197 children who took part in the June and December survey respectively. In 2009, corresponding data were analysed for 460 and 409 children. Prevalence of underweight was highest in December 1999 (42.6%) and lowest in December 2009 (34.1%). After adjustment for age, sex and village, there was a slight but not always significant improvement in the z-scores of weight-for-age, weight-for-length, length-for-age, and mid-arm circumference over time.
Conclusions The findings from this study confirm the still unacceptable high prevalence of malnutrition in young children of rural sub-Saharan Africa (SSA). Progress in the reduction of malnutrition remains slow on this continent making it rather unlikely that the corresponding MDGs will be achieved. Large-scale multi-sectoral community-based interventions are urgently needed for a sustainable improvement of child health in SSA.
Objectif: Pour la mesure des progrès réalisés vers l’Objectif du Millénaire pour le Développement (OMD) 1, des données fiables sur les indicateurs de nutrition spécifiques des pays sont essentielles. La malnutrition est également le principal déterminant de la mortalité infantile, qui est abordé dans l’OMD 4.
Méthodes: Dans la zone de surveillance démographique et de santé de la province de Kossi, dans le nord-ouest du Burkina-Faso, les paramètres nutritionnels ont été comparés dans deux cohortes de jeunes enfants de la même tranche d’âge provenant de huit villages. Les enquêtes ont eu lieu en juin et décembre 1999 et 2009. Un modèle multivarié a été utilisé pour contrôler pour les variables confusionnelles.
Résultats: Pour l’étude de 1999, les données ont été analysées pour 179 et 197 enfants qui ont participé respectivement à l’enquête de juin et décembre. En 2009, les données correspondantes ont été analysées pour 460 et 409 enfants. La prévalence de l’insuffisance pondérale était la plus élevée en décembre 1999 (42,6%) et la plus faible en décembre 2009 (34,1%). Après ajustement pour l’âge, le sexe et le village, il y avait une légère amélioration, mais pas toujours significative dans les z-scores du poids pour l’âge, le poids pour la taille, la taille pour l’âge et le périmètre brachial, au cours le temps.
Conclusions: Les résultats de cette étude confirment la prévalence toujours élevée et inacceptable de la malnutrition chez les jeunes enfants des régions rurales de l’Afrique subsaharienne. Les progrès réalisés dans la réduction de la malnutrition restent lents sur ce continent, rendant plutôt improbable l’atteinte des OMD correspondants. Des interventions multisectorielles à grande échelle basées sur la communauté sont urgemment nécessaires pour une amélioration durable de la santé de l’enfance en Afrique subsaharienne.
Objetivo: Para medir el progreso hacia el Objetivo de Desarrollo del Milenio (ODM) 1, es esencial contar con datos fiables sobre los indicadores nutricionales para países específicos. La desnutrición es también el principal determinante de mortalidad infantil, del cual se ocupa el ODM 4.
Métodos: En el área de seguimiento sanitario y demográfico de la provincia de Kossi, al noroeste de Burkina Faso, se compararon los parámetros nutricionales para dos cohortes de niños pequeños, con el mismo rango de edad, pertenecientes a ocho poblados. Las encuestas se realizaron en Junio y Diciembre de los años 1999 y 2009. Se utilizó un modelo multivariado para controlar variables de confusión.
Resultados: Se analizaron datos del estudio de 1999 de 179 y 197 niños, que participaron en las encuestas de Junio y Diciembre respectivamente. Para el 2009 se analizaron datos correspondientes a 460 y 409 niños. La prevalencia de bajo peso era mayor en Diciembre de 1999 (42.6%) y menor en Diciembre del 2009 (34.1%). Después de ajustar para edad, sexo y población, había una pequeña pero no siempre significativa mejora en los z-scores de peso-por-edad, peso-por-altura, altura-por-edad, y circunferencia del brazo a lo largo del tiempo.
Conclusiones: Los hallazgos de este estudio confirman que la desnutrición continúa teniendo una prevalencia inaceptablemente alta entre niños pequeños de zonas rurales del África subsahariana. El progreso de reducción de la desnutrición continúa siendo lento en este continente, lo cual hace que sea improbable que los ODM correspondientes puedan ser alcanzados. Se requiere urgentemente realizar intervenciones a gran escala, multisectoriales y basadas en la comunidad, para conseguir una mejora sostenible en la salud infantil en África subsahariana.
The target of the first Millennium Development Goal (MDG) of reducing by half the proportion of people who suffer from hunger between 1990 and 2015 is measured by the percentage of children under 5 years who are underweight (United Nations 2011). The prevalence of under-five children being underweight in developing countries has only decreased from 30% in 1990 to 23% in 2009 with the highest prevalence and burden still being in southern Asia (United Nations 2011). In sub-Saharan Africa (SSA), the prevalence of underweight in under-five children has also only marginally declined over the last 20 years, from 27% in 1990 to 22% in 2009 (United Nations 2011).
In under-five children, malnutrition is responsible for globally about one-fifth of disability-adjusted life years (DALYs) and about one-third of deaths (Black et al. 2008; Bhutta et al. 2010). Malnutrition, thus, remains the leading risk factor for morbidity and mortality in young children of poor countries (Müller & Krawinkel 2005; Black et al. 2008, United Nations 2011). The target of the 4th MDG is to reduce the rate of under-five mortality by two-thirds between 1990 and 2015. Recent publications show a global decline from 11.9 million deaths in 1990 to 7.7 million deaths in 2010 (Bhutta et al. 2010; Rajaratnam et al. 2010). Although rates of decline accelerated in the whole of SSA during the last 10 years, West Africa remains among the regions with the highest under-five mortality (Rajaratnam et al. 2010).
This article reports on the development of nutritional indicators in one district of rural Burkina Faso over a period of 10 years (from 1999 until 2009).
Materials and methods
Burkina Faso is a landlocked low-income country in the Sahel zone of West Africa. The study took place in the Nouna Health District (NHD) in north-western Burkina Faso, which is holoendemic for malaria (Müller et al. 2001). In 2009, the population of NHD was about 320 000 living in some 300 villages (Sie et al. 2010). The area is a dry orchard savannah, inhabited mainly by subsistence farmers and cattle keepers of different ethnic groups. The short rainy season usually lasts from June to October (Müller et al. 2001). The dry season includes two parts: a dry, cold and dusty period (November to February) and a dry and very hot period (March–May). The formal health services in the NHD comprise the hospital in Nouna town and 29 village-based health facilities (Sie et al. 2010).
The fieldwork for this study was conducted in collaboration with the Nouna Health Research Center (CRSN), which is situated in Nouna town. A health and demographic surveillance system (HDSS) has been operating in the area since 1992 (Sie et al. 2010).
Study design and procedures
The main aim of the study was to assess the development of malnutrition in young children of the Nouna study area over a 10-year period, that is, to compare anthropometric data collected in 1999 with corresponding anthropometric data collected in 2009. In 1999, anthropometric data were collected within the frame of a large placebo-controlled community-based zinc supplementation trial (Müller et al. 2001, 2003). During the trial, cross-sectional surveys took place in the study area in June and December 1999. Study participants were a cohort of 708 children aged 6–31 months from 18 villages, randomly chosen from the HDSS register of the CRSN. For this study, only children from the eight largest villages of the 1999 trial (Bourasso, Cisse, Kodougou, Koro, Nokui, Seriba, Sikoro and Solimana) were included, and only the data from the placebo arm were used (i.e. 30 children per village). In 2009, a comparable cohort of children was enrolled with data collected from two cross-sectional surveys in June and December. Children from the eight villages and within the same age range as those recruited for the 1999 study were included through random selection from the HDSS register.
The study teams for the cross-sectional surveys consisted of study physicians, study nurses, laboratory technicians and specifically trained field workers, collecting data on anthropometric characteristics with similar methods in 1999 and 2009 (Müller et al. 2001, 2003). In brief, weight was measured to the nearest 0.1 kg with a hanging scale (UNICEF Salter Model 235 6S), which was calibrated and controlled daily before use. Recumbent length was measured to the nearest 0.1 cm with a length board (infantometer model no. BCE05, Global Products Corporation, http://www.gpcmedical.com) with an upright base and a moveable headpiece. Mid-upper arm circumference (MUAC) was taken to the nearest 0.1 cm with a standard colour-coded, stretch-resistant but flexible plasticised paper measuring tape.
From the anthropometric measurements, z-scores were derived for weight-for-age (WFA), length-for-age (LFA), weight-for-length (WFL) and mid-upper arm circumference-for-age (MUAC/A). A z-score <−2 standard deviations (SD) is an indicator for malnutrition: a WFA z-score <−2 SD indicates underweight, a LFA z-score <−2SD indicates stunting and a WFL z-score <−2 SD indicates wasting. If the z-score is <−3SD, the child is considered severely malnourished.
Sample size of the two surveys
The sample size calculation for the 2009 surveys was based on the primary outcome WFL. To demonstrate an overall 50% difference (from 25% in 1999 to 12.5% in 2009) and applying a design factor of 2 for clustering, the required sample size was estimated as 480 children (approximately 60 children from each of the eight villages). Comparing this sample to the data of the 240 children of the 1999 study allows for detection of a 50% difference with 80% power (P < 0.05).
Data management and statistical analysis
The data from 1999 were available as an cleaned SAS data file. In 2009, all data collected were systematically screened for completeness, plausibility and consistency and prepared for data entry. Entered data were cleaned, and differences were resolved by checking against the original field data forms. Data analysis was performed with SAS version 9.2 (SAS Corporation, Cary, USA).
Z-scores were derived using the WHO Child Growth Standards and the SAS macro available at the WHO website. Extreme (i.e. biologically implausible) z-scores for each indicator were flagged according to WHO guidelines (WHO 2006a,b, WHO SAS macro: http://www.who.int/childgrowth/software/en/). To compare the levels of malnutrition (regarding mean WFA, LFA, WFL and MUAC/A z-scores) across study years for the June and December surveys, a multivariate analysis was performed adjusting for age, sex and village.
Ethical approval was obtained from the Ethical Committee of the Heidelberg University Medical School (1999 and 2009) and the local authorities (Burkina Faso Ministry of Health in 1999 and the local Ethical Committee of Nouna in 2009). Before the start, the trial was explained in detail to and discussed with all relevant district authorities and the concerned communities. Community consent was sought. During the surveys, individual oral (in 1999) and written (in 2009) informed consent was sought from the respective caretakers of study children. All study children found ill during the survey were treated free of charge. In case of significant illness including severe malnutrition diagnosed by the study physicians, they were referred to the Nouna hospital. However, the final decision whether the children were taken to hospital remained with the parents/caretakers of the children.
For the 1999 study, data were available for analysis from 179 and 197 children who took part in the June and December survey, respectively. In 2009, data were available from 460 and 409 children who took part in the June and December survey respectively.
Table 1 compares the demographic and anthropometric data from the 1999 and 2009 samples. Children recruited in June 2009 were slightly older compared with the children recruited in June 1999 (mean age 20.5 vs. 17.6 months, P ≤ 0.0001). Furthermore, there were significant differences between 1999 and 2009 in sex distribution and in the proportion of children recruited from the eight study villages.
Table 1. Baseline demographic and anthropometric characteristics of young children in Nouna Health District in Burkina Faso, 1999 and 2009
June 1999 (N = 179)
June 2009 (N = 460)
December 1999 (N = 197)
December 2009 (N = 409)
SD, standard deviation; MUAC, mid-upper arm circumference.
Table 2 and Figure 1 show the prevalence of malnutrition among the 1999 and 2009 study children. Malnutrition rates were high in all surveys, but showed a tendency to decrease from 1999 to 2009. Prevalence of underweight was highest in December 1999 (42.6%) and lowest in December 2009 (34.1%). Prevalence of stunting was highest in December 1999 (46.9%) and lowest in June 2009 (29.6%). Prevalence of wasting was highest in June 2009 (25.9%) and lowest in December 2009 (15.6%). Prevalence of MUAC<−2SD was highest in December 1999 (36.7%) and lowest in December 2009 (8.8%).
Table 2. Prevalence of nutritional parameters in young children of Nouna Health District in Burkina Faso in 1999 and 2009
Prevalence in % (95% CI)
Prevalence in % (95% CI)
SD, standard deviation; WFA, weight-for-age; LFA, length-for-age; WFL, weight-for-length; MUAC, mid-upper arm circumference.
16.5 (10.7, 22.2)
39.8 (32.3, 47.3)
12.4 (9.3, 15.5)
35.4 (31.0, 39.9)
17.9 (12.3, 23.6)
42.6 (35.4, 49.8)
13.5 (10.0, 16.9)
34.1 (29.3, 38.8)
23.0 (16.4, 29.5)
45.4 (37.7, 53.1)
10.9 (7.9, 13.9)
29.6 (25.3, 33.9)
22.4 (16.2, 28.6)
46.9 (39.6, 54.2)
18.9 (15.0, 22.8)
44.7 (39.8, 49.7)
10.8 (5.9, 15.7)
25.6 (18.8, 32.3)
7.0 (4.5, 9.4)
25.9 (21.8, 30.0)
8.7 (4.5, 12.9)
22.1 (16.0, 28.1)
4.4 (2.3, 6.5)
15.6 (12.0, 19.3)
6.8 (2.8, 10.8)
21.6 (15.2, 28.0)
2.0 (0.6, 3.3)
12.0 (8.9, 15.1)
14.9 (9.5, 20.2)
36.7 (29.5, 43.9)
2.0 (0.5, 3.4)
8.8 (5.9, 11.7)
Table 3 shows the differences in mean z-scores for the malnutrition indicators adjusted for age, sex and village. Comparing 1999–2009, WFA and WFL z-scores improved significantly in December, while LFA z-scores improved significantly in June. MUAC/A z-scores improved significantly both in June and in December.
Table 3. Comparison of the mean z-scores of nutrition parameters in young children of Nouna Health District in Burkina Faso between 1999 and 2009 (adjusted for sex, age and village)
In this study, survey data on malnutrition prevalence in a rural area of West Africa were compared in two cohorts of young children recruited in 1999 and 2009, respectively. Overall, the study showed slight improvements in the anthropometric indicators between 1999 and 2009. Looking at the age-, sex- and village-adjusted mean z-scores of the nutrition parameters, significant improvements can be observed for WFA (December surveys), LFA (June surveys), WFL (December surveys) and MUAC/A. These improvements are in line with national survey data from Burkina Faso on children under 5 years, collected every 3–5 years and summarised in the WHO Global Database on Child Growth and Malnutrition, based on the WHO child growth standards (http://www.who.int/nutgrowthdb/database/countries/who_standards/bfa.pdf). In the Western greater region (Boucle du Mouhoun) encompassing the study area over the period 1992–2009, the prevalence of underweight decreased from 34.0% to 23.4%, the prevalence of stunting decreased from 46.2% to 34.3%, and the prevalence of wasting decreased from 13.7% to 9.6%. Compared with these regional survey data, our study demonstrated an overall higher prevalence of malnutrition and a not so marked decline over time. This can partly be explained by differences in the methods used. Most importantly, the national surveys include all under-five children, while our study included only children under 3 years, an age-group where malnutrition is more prevalent. Further methodological differences concern the determination of date of birth (which is more precise in HDSS-based surveys), different sample sizes (which is slightly higher in the national survey) and different months of surveys. The national surveys had collected data during the months December–March (1992/1993), November–March (1998/1999), June–December (2003), March–June (2006) and August–September (2009).
Looking specifically at underweight prevalence in the Nouna study area, changes during the last 10 years are small. According to our data, underweight prevalence went down from 40% in June 1999 to 35% in June 2009 and from 43% in December 1999 to 34% in December 2009. This is far too slow to reach the MDG 1 target of reducing by half the percentage of children under 5 years of age who are underweight (United Nations 2011).
Stunting is a measure of chronic malnutrition. The high prevalence of stunting in all our surveys, thus, supports the conclusion that the malnutrition prevalence in young children has not changed much over time. The observed increase in stunting prevalence between the June and December surveys can be explained by the ageing of the cohort children.
In contrast, wasting is an indicator for acute malnutrition. Wasting and severe wasting prevalence remained rather high in under-five children of this study, both in the June and December surveys and decreased significantly only between December 1999 and December 2009. The high prevalence of wasting in Nouna District is in line with data published for the West African region in general (Black et al. 2008). Short-term changes in this indicator can be explained by various factors such as the quality of harvest, epidemics or other events that impact on the purchasing power of households and the morbidity of individuals in respective years (Müller & Krawinkel 2005).
This study has some strengths and some limitations. Firstly, anthropometric data were compared between 2 years only. We do not know how these indicators developed in the years between 1999 and 2009, and the study years may also not be representative for the time period. Secondly, although care was taken to use in principle the same methods and comparable tools, the surveys in 1999 and 2009 were performed by different teams who may have used slightly different procedures. Thirdly, for ethical reasons, study children found ill during the surveys were treated appropriately by the study physicians. Data collected on the study children in the December surveys could, thus, have been influenced by treatment of sick children in the June surveys. Fourthly, as the study has only followed the development in one district of one country, the data are not fully representative for Burkina Faso and clearly not representative for the whole of SSA. One of the main strengths of the study is the fact that children were sampled from the HDSS data bank. Dates of birth have therefore been known with a high level of accuracy, which is usually not the case with national surveys such as those conducted in Burkina Faso (Institut National de la Statistique et de la Démographie, et Macro International Inc 2000, Institut National de la Statistique et de la Démographie (INSD) et ORC Macro 2004, Institut National de la Statistique et de la Démographie et UNICEF 2008, Ministry of Health Burkina Faso 2009). Finally, the information from this study is based on carefully implemented research projects which guarantee a high quality of the data.
The study has been conducted in an HDSS area where a number of research projects have been implemented between 1999 and 2009, and this could also have biased the comparison. Although most of these studies were disease-specific interventions mainly in the field of malaria control (Müller et al. 2006, 2008), they could have contributed to improvements in the nutrition status of children. However, the lack of improvement found in this study clearly points to the conclusion that more comprehensive interventions are needed to reduce malnutrition in children of SSA (Müller & Krawinkel 2005).
Today, most of the funding for international health concentrates on disease-specific (vertical) interventions instead of addressing more broadly (horizontal) the underlying causes of disease such as poverty, low education and prevailing malnutrition (Egilman et al. 2011). Although some more attention has recently been devoted to strengthening health systems, the main focus is still on supporting the development and implementation of non-sustainable and often even counter-productive selected technical interventions such as drugs, vaccines or micronutrient supplementations (Müller & Krawinkel 2005; Swanson et al. 2010; Egilman et al. 2011). A combination of large-scale implementation of nutrition interventions, strengthening of the health and education systems, investment in environmentally sustainable agricultural techniques and improvements in hygiene and sanitation in the frame of sustainable community development would likely have more sustained long-term effects on morbidity and mortality in populations of low-income countries (Bryce et al. 2008, Egilman et al. 2011).
In conclusion, the unacceptably high prevalence of malnutrition in young children of rural SSA has not changed much over the last decades, which makes the achievement of the corresponding MDGs very unlikely. Large-scale multi-sectoral community-based interventions are urgently needed for a sustainable improvement of child health in SSA.
Acknowledgements and funding
We thank the staff of the CRSN for their continuous support during data collection. The study was funded by a grant from the German Research Foundation (Graduiertenkolleg 793) at the Ruprecht-Karls-University Heidelberg.