Menu
Menu
Menu
Menu
Background: Maternal malnutrition, which has been a problem in Madagascar for several years, has been rising despite interventions to improve the situation. This study aims to identify the socioeconomic determinants of malnutrition among mothers who are one of the most vulnerable groups. Methods: A cross sectional study was carried out among 670 mothers aged 18 to 45 living in the Amoron’i Mania region of Madagascar. The study was conducted during the post-harvest period. The nutritional status of mothers was assessed by anthropometry. A Body Mass Index (BMI) lower than 18.5 kg/m2 or an arm circumference lower than 220 mm were used to define malnutrition. Data on the characteristics of the mothers and their households were also collected. Multiple logistic regression was used to identify factors associated with maternal malnutrition. Results: The prevalence of maternal undernutrition is estimated at 17% (95% CI: 14–20) according to BMI and 9% (95% CI: 7–11) for Mid Upper Arm Circumference (MUAC). In the multivariate analysis, using BMI, the factors significantly associated with malnutrition were: the household size equal to or greater than 6 (AOR = 1.59 [1.04–3. 42], p = 0.029) and use of unsafe water source (AOR = 1.99 [1.02–3.85], p = 0.030). For the MUAC, the factors associated are: use of unsafe water source (AOR = 2.82 [1.01–7.97], p = 0.041) and increased number of children under five years old (AOR = 1.38 [1.02–1.89], p = 0.025). Conclusion: This study confirmed the importance of mothers’ malnutrition in the study area. Fight against maternal malnutrition needs interventions to improve access to safe drinking water and to promote family planning.
The study was conducted in Amoron’i Mania, one of Madagascar’s 22 regions. The region comprises 4 health districts, 53 communes, of which 52 are located in rural areas, and about 580,000 inhabitants. This region has the highest undernutrition (BMI < 18.5) prevalence among women of reproductive age in Madagascar. It was estimated at 41.6% in the Demographic Health Survey (DHS) results in 2008–2009 [11]. A cross-sectional study was carried out. The study population included non-pregnant mothers between 18 and 45 years of age who had given birth more than 6 months earlier. Mothers under 18 years old were not included because of the difficulty in getting the guardian’s consent. Mothers over 45 years old were not included so as to only take into account mothers in their reproductive period, knowing that the fertility rate is very low among woman aged between 45 and 49 years old [10]. To ensure the validity of the weight measurements, we also excluded women who had given birth within the last 6 months. Indeed, there is a gradual reduction in pregnancy weight gain and stabilization in mothers’ weight around sixth months after delivery [14, 15]. A two-stage cluster sampling was used. The first stage aimed at selecting 30 “fokontany” (smallest administrative structure) out of the 760 in the region. It was done by systematic random sampling. The second stage was used to select, for each “fokontany”, eligible mothers from a list established by community workers. This was done by simple random sampling. The sample size was calculated on the basis of maternal undernutrition national prevalence (27%), 5% margin of error, 95% confidence level and a design effect of 2 [16]. The sample size was estimated at 606. Twenty-one subjects per cluster therefore had to be included. During data collection, 670 women were actually interviewed. Data collection was conducted in July and August 2015, during the post-harvest period in the region. Nutritional status was assessed by anthropometric measurements such as weight, height and Mid Upper Arm Circumference (MUAC). Women were weighed with 100 g accuracy with a SECA electronic scale. Height was measured with 1 cm accuracy with a SECA wall mounted height rod. MUAC was measured on the left arm, midway between the acromion and the olecranon, with 1 mm accuracy with an adult-specific measuring tape. Information about dietary, socioeconomic, health and reproductive characteristics were collected by interviewers. Regarding the mothers’ dietary practice, a 24-h recall was used [17, 18]. Interviewers asked and established the list of all foods taken by the mother the day before the survey. To minimize omissions, they focused on food intake based on the pre-established list to assess women’s dietary diversity. Information about the place and the people with whom the mother took the meal as well as the occurrence of an unusual event the day before the survey was collected to detect unusual food consumption [17] . The recall took into account any day of the week except for days with unusual consumption such as festive days or stays away from the household. Mother’s age, education level (last school year taken into account, primary education takes six years in Madagascar), occupation, marital status and husband’s occupation were collected. Information about gravidity, parity, number of children aged less than 5 years old, breastfeeding status and birth interval were collected. Interviewers were recruited locally. Data collectors received training according to their mission. The investigator, a technician from the Nutrition Department of the Ministry of Public Health, and the regional nutrition manager supervised the data collection. The study was approved by the Malagasy Ministry of Health’s Ethics Committee. The Body Mass Index (BMI) was calculated by dividing weight in kilograms by height in square meters. WHO defined standards were used to identify undernourished women, i.e. BMI below 18.5 kg/m2, height below 145 cm, and MUAC below 220 mm [14]. The dietary diversity score was calculated using the 10 categorized food groups according to the latest recommended Women’s Dietary Diversity Score [18]. The consumption of one or more foods in one group was worth 1 point and the maximum score was 10. The birth interval was calculated for the last two deliveries within the last five years. Afterwards, that interval was grouped using a 24-month threshold. Mothers who did not have two childbirths within the last five years and three primiparous women were classified in the 24 months or more group. Household-related variables were studied, i.e. the household head sex, the quality of the water source used for meal preparation, the toilet type, the fuel used and the house type. The water source is considered safe if it comes from standpipes, public faucets, covered or protected wells, wells or boreholes with pumps and faucets inside or outside the dwelling. Three indicators of economic profile were created considering the possession of household goods. We used the DHS Madagascar list to establish our list of goods. The first indicator refers to possession of movable property (furniture, radio, TV, telephone, bicycle, etc.), the second refers to possession of farming equipment and the third to possession of farm animals. The corresponding scores for these properties were established by principal component analysis (PCA). The scores were categorized into three groups (high, medium and low) based on possible values close to the tertiles. The period (number of months) in which a household consumes its annual rice production was also collected. Rice production is considered as an important element of food security in the study area and can reflect the economic level of households. Rice is a Malagasy staple food and rice cultivation remains the main activity of most farming households in the region studied [19]. Stata / IC 13.1 (StataCorp LP, College Station, USA) software was used to analyze the data. In bivariate analysis using logistic regression, the association between maternal undernutrition and other variables was estimated by the Odds Ratio (OR) with its confidence interval. The chi-square or chi-square for trend tests were used. Variables with p-value < 0.20 in bivariate analysis were considered for inclusion in a logistic regression model. A stepwise backward method was used for selection of statistically significant covariates. Categorical variables with more than two categories were transformed into indicators. The backward procedure used to select variables in the final model was based on the likelihood ratio. The adequacy of the final model was checked using Hosmer Lemeshow test. The adjusted ORs and their 95% confidence intervals were computed from the final multivariate logistic model. The significance level (p-value) was set at 0.05.
N/A
