Situation and determinants of the infant and young child feeding (IYCF) indicators in Madagascar: analysis of the 2009 Demographic and Health Survey

A total of 40.7% of the children aged 6–23 months were stunted (Table 1). More than half of the mothers (53.9%) were younger than 19 when they first gave birth. A substantial proportion of the mothers had no formal education (28.6%) and almost half of them only attended primary school (49.7%). Exposure to media was very low because most mothers never read newspapers nor magazines (84.9%), listened to radio (46.6%) nor watched television (85.9%). The great majority of the households lived in rural areas (82.2%). Nearly half (49.6%) of the children lived in households within the two lowest quintiles of the wealth index. Additionally, more than half of the households (57.6%) did not have toilets and only 35.4% utilized improved sources of drinking water such as piped water, protected wells and rainwater.

A substantial number of infants were put to the breast within the first hour after birth (77.2%) and continued breastfeeding at one year was almost universal (99.6%) as seen in Fig. 1. However, exclusive breastfeeding was relatively low as less than half of the mothers (48.8%) exclusively breastfed their infants during the first 6 months. A total of 88.3% of the infants received complementary foods at 6–8 months. Only 22.2% of the infants aged 6–23 months ate from four or more food groups. A small proportion (19.6%) reported eating an iron-rich food the previous day.

Because of the high compliance with continued breastfeeding at one year (only one observation did not comply), the indicator was not tested in regression models. Infants whose mothers gave birth for the first time at an age younger than 19 years were more likely not to have breastfed within the first hour. Also, the households with more resources (third and fourth quintiles) were more likely to have appropriate early initiation of breastfeeding practices (OR = 0.63 and OR = 0.58, respectively) compared to the poorest households (Table 2).

Maternal age at first birth was the only factor significantly associated with exclusive breastfeeding under six months. Infants of mothers who first gave birth at an age younger than 19 years had 59% higher risk (OR = 1.59) of not being exclusively breastfed during the first 6 months.

For the complementary feeding indicators, none of the factors tested were associated with the time of introduction of solid, semi-solid or soft foods. Infants whose mothers attended primary school or higher had lower odds of having inadequate dietary diversity. Maternal exposure to media was also associated with adequate dietary diversity in infants. For instance, infants whose mothers did not listen to radio at all were more likely to have inadequate dietary diversity compared to those whose mothers listened to radio almost everyday. Likewise, infants were more likely to have inadequate dietary diversity if their mothers did not read newspaper or did not watch TV at all. Additionally, infants born to younger mothers had higher risks of being fed with inappropriate dietary diversity. Infants living in rural areas were also more likely to not have reached the minimum dietary diversity (OR = 4.00) compared to infants in urban areas. Infants of mothers who worked outside the home were at higher risk of having inadequate dietary diversity. The odds of having inappropriate dietary diversity decreased with increased wealth of the family.

Girls were more likely to not have been given iron-rich foods (OR = 1.34). Children whose mothers attended secondary school or higher education had lower chances of inadequate consumption of iron-rich foods. Infants born to mothers who do not read newspaper at all were more likely to not be given iron-rich foods. Infants whose mothers did not work outside the home and those living in a household with higher wealth were more likely to consume iron-rich foods. Finally, children living in rural areas had higher odds of having inadequate iron-rich food consumption.

For the breastfeeding indicators, our results suggest that although the rate of exclusive breastfeeding under 6 months was low, the indicators for two other breastfeeding practices, i.e. early initiation and continued breastfeeding, were high in Madagascar. A total of 77.2% of the children were put to the breast within one hour after birth and continued breastfeeding at one year was very high (99.6%). Compared to the data previously published by WHO [8], all of the breastfeeding indicators showed improvement except for the exclusive breastfeeding rate (67% in 2004 and 48.8% in 2009). In 2004, the WHO report did not communicate data on mothers working outside the home, but the difference in the reported rates of exclusive breastfeeding between 2004 and 2009 might be due to the increase in the number of working mothers since 2004. In 2009, a total of 87.5% of the mothers of infants aged 0–6 months were working outside the home, of which 51.6% did not practice exclusive breastfeeding. Because of time constraints and other logistic challenges, they were more likely to have practiced mixed feeding.

For the complementary feeding indicators, the rate of timely introduction of complementary foods from 2004 to 2009 also improved from 78% to 88.9%. A decrease from 31% to 22.2%was noticed in the percentage of infants who received the minimum dietary diversity between 2004 and 2009. For the DHS 2009, data collection was conducted from November 2008 to August 2009, which coincided with a political crisis in the country. Protestations against the then-government led to the resignation of the President and the rise of a transitional government. The transitional government was unrecognized by the international community, including donors and foreign governments, which sanctioned Madagascar by suspending their assistance programs [23]. As most of the public investment relevant to health and education relied on external funding, basic social services were diminished leaving the population very vulnerable, especially women and children [24]. In fact, the level of poverty was increased from 74.1% in 2005 to 81.8% in 2010 [25]. Because dietary diversity is closely linked to food access and availability [26], this might have contributed to the drop in the percentage of children achieving the minimum dietary diversity. Additionally, political instability is one of the basic causes of inadequate nutritional status in children [24]. In 2010, there were even fewer infants in rural areas who had received the minimum dietary diversity (14.1%) according to the CFSVA + N analysis [16].

Only the minimum dietary diversity variable was positively associated with LAZ as children who had four or more food groups were more likely to have higher z scores, and thus would be less likely to be stunted. This outcome is concordant with the results found in the literature in different countries with high prevalence of stunting. Using pooled data from 14 low-income countries, Marriott et al. reported that attaining the minimum dietary diversity was associated with lower odds of stunting [17]. Also, studies using nationwide survey data from Bangladesh [27] and India [6] concluded that higher dietary diversity was associated with higher LAZ/HAZ. These findings highlight the importance of the quality of the diet for child growth especially during the first two years of life.

No significant association was found between the other three IYCF indicators (for which we had data) and stunting. Similar results were found in Cambodia, Haiti, Kenya and Uganda while using DHS data [4]. The only significant associations were between stunting and the introduction of solid, semi-solid or soft foods in Bangladesh and the minimum acceptable diet in Zimbabwe. The authors argued that the 24 h-recall nature of the indicators does not allow the investigator to capture the whole situation of feeding practices. Because the questions only consider the feeding practices of the previous day (the last 24 h), no matter how the infant had been fed in the past, as long as the caregivers had appropriate feeding practices the previous day, they will still be compliant with the indicator. The researchers suggested that the lack of sensitivity of the indicators might explain the lack of association between some IYCF indicators and child anthropometric measurements [4].

However, besides IYCF indicators there are multiple determinants of stunting, as it is a multifactorial problem. Rakotomanana et al. [28] found that increased age, more severe anemia level, being a male and decreased maternal height were associated with higher risk of stunting in infants aged 0–23 months in Madagascar. Also, infants currently being breastfed, infants living in households using iodized salt and in urban areas had lower odds of stunting. Region of residence was also a factor affecting stunting risk in the study.

Among the IYCF practices, sex was only a determinant for the consumption of iron-rich foods. No difference was seen in the consumption of flesh foods between sexes. However, there were more boys than girls being given organ meats even though the overall consumption meats was only 3.4% at 6–23 months.

Being born to a mother who gave birth at 19 or older or living in a family with somewhat more resources (in the third and fourth wealth quintile) was associated with an improved chance of initiating breastfeeding within the first hour after birth. A study using the Nigeria DHS reported that mothers with higher wealth were more likely to comply with the early initiation of breastfeeding [29]. More wealthy households may have more access to health care and may have been in contact with health care professionals who encouraged better breastfeeding practices.

Lower odds of inadequate diet diversity in infants were associated with higher maternal education, greater maternal exposure to media and household wealth. Similar results were found in Bangladesh [10], Uganda [9] and Ethiopia [13] as children of mothers with no formal education had higher odds of not meeting the minimum dietary diversity criteria. Regarding maternal exposure to media, a nationwide study in India [6] also reported that mothers who were less exposed to media might have less nutrition knowledge, as it is an important channel to spread nutrition information. In Southern Ethiopia, children had higher dietary diversity score when their mothers had received IYCF information on mass media in the last one month [13]. Additionally, studies across countries confirmed that increased family wealth was associated with better chances of meeting minimum dietary diversity [9, 15]. Higher household wealth generally means having access to more diverse food and more having more resources to be allocated to childcare and nutrition. Our results also showed that infants of working mothers had higher odds of inadequate dietary diversity. There is no clear explanation about the association between the maternal working status and the dietary diversity score except that working mothers generally spend time away from their children, which may lead to poor complementary feeding practices. Caretakers may be less knowledgeable about infant and young children nutrition and good feeding practices.

The consumption of iron-rich foods was associated with secondary or higher maternal education and higher wealth. Children living in rural areas were less likely to eat flesh foods. The overall poor dietary diversity and poor consumption of iron-rich food may be linked to access to food and to the general food insecurity in the country. Analyses from the CFSVA + N concluded that the seasonality of earnings due to a non-diversified income source in Malagasy households affects the access to food throughout the year [16]. Thus, 84% of the households experienced a time in the year when they did not have sufficient food. Yet, when food was available, other factors may play a role in inappropriate complementary feeding such as maternal or caretaker nutrition knowledge or attitudes [30].

Achieving feeding indicators appropriately has been associated with children’s anthropometric measurements in several countries [4]. Despite a few improvements in some IYCF indicators between 2004 and 2009 in Madagascar, efforts are still needed to improve breastfeeding and complementary feeding practices.

This study showed that several maternal and household characteristics are associated with inappropriate feeding practices in Madagascar. Younger maternal age (less than 19 y) at first birth was associated with sub-optimal breastfeeding, failure to achieve minimum dietary diversity and inadequate consumption of iron-rich foods. Sex education programs among youth and a specific focus on delaying early onset of first pregnancy may be needed as 53.3% of the Malagasy mothers in this sample were younger than 19 when they delivered their first child. Additionally, pregnant adolescents are more likely to give birth to low birthweight babies and are at risk of adverse birth outcomes [31].

Higher maternal education also has been linked to improved complementary feeding practices in other developing countries [9, 32]. Additionally, women’s empowerment interventions may lead to better nutritional outcomes. For instance, a study done across sub-Sahara African countries demonstrated that women who made decisions on large purchases and who had control over the household income were more likely to have adequate complementary feeding practices [33]. Thus, there is a need for nutrition-sensitive interventions in educational programs for women and girls to improve feeding practices in Madagascar. With the increasing number of mothers working outside home, these interventions should target other caretakers in the household as well such as grandmothers or elder siblings. Grandmothers’ knowledge regarding IYCF practices has been associated with better child feeding in Nepal [34].

Integrated nutrition interventions aiming to improve overall nutrition outcomes also have been shown to be efficient in ameliorating inadequate IYCF practices. A pilot study implementing the Essential Nutrition Actions (ENA) framework, was done in two districts of Madagascar in 2000, reported a significant increase in early breastfeeding, in exclusive breastfeeding under 6 months, and in meeting the minimum meal frequency for young children [35]. Through technical support and various trainings, as well as micronutrient supplementation, the ENA promotes key messages about maternal and child nutrition including feeding practices. The authors reported that involving a wide array of partners and creating a favorable policy environment through policy analysis and development were the key factors for the success of the pilot study. Thus, implementing such programs on a larger scale should be helpful in an effort to improve IYCF practices in Madagascar.

Despite being one of the few nationwide analyses of the IYCF indicators and associated factors in Madagascar, our study had limitations. Because of the lack of precision in the instruments during data collection, children’s weights were not reported for the 2009 DHS for Madagascar preventing us from looking at the association between wasting and underweight with the IYCF indicators. Additionally, although our study used the most recent nationwide data that includes nutrition information, the situation might have changed since 2009. This emphasizes the need for another nationwide representative survey to monitor the changes in the nutrition situation of the country.