The national exclusive breastfeeding rate was 23.5% according to the 5th MICS (2016) in Cote d’Ivoire (15). The descriptive analysis involved 980 children under six months of age. The results indicate that maternal age at delivery, exposure to medical discourse on breastfeeding (a composite variable combining the number of antenatal cares and assistance at delivery by skilled personnel), and the child’s sex were significantly associated with the duration of exclusive breastfeeding. Regarding the mother’s age at delivery, after one month of life, the EBF prevalence curves differ significantly from each other until the end of the 6th month (Fig. 2). Moreover, the differences increased over time. The curve representing mothers aged 35 years or more remains above the others throughout the observation period and that of mothers under 20 years of age remains below them (except for the 6th month). This means that children born to older mothers experienced cessation of exclusive breastfeeding later than children born to younger mothers.

Prevalence of EBF 0–6 months by age of mother at delivery

Regarding exposure to counseling on breastfeeding or the number of antenatal cares and the qualification of the person who assisted the woman at delivery, we note that the proportion of exclusively breastfed children remains higher among children whose mothers have achieved more than three antenatal cares and had benefit from a skilled attendant at delivery (Fig. 3).

Prevalence of EBF 0–6 by exposure to medical discourse on breastfeeding

Otherwise, the proportion of children who were fed other foods remained lowest among mothers with high exposure to counseling session on breastfeeding. Thus, children whose mothers had achieved more than three antenatal cares and were assisted by a skilled attendant at delivery were experienced cessation of exclusive breastfeeding less quickly over the 6-month period than other children.

Virtually over the entire duration of exposure to exclusive breastfeeding cessation, the prevalence curve for male children is above that for female children. This means that male children are less at risk of stopping EBF than female children (Fig. 4).

Prevalence of EBF 0–6 by sex of child

The Multiple Correspondence Analysis (MCA) was used to profile the children. Factor analysis seeks to reduce a large amount of information (in the form of values on variables) to a few large dimensions (factors). It attempts to decompose the patterns of correlations to explain them by a limited number of factors. Regarding the profile of children (Fig. 5), the MCA identified two groups of children. The target group was composed exclusively of non-exclusively breastfed children. They lived in Abidjan and in urban areas in general. They live in households with high economic well-being. Their mothers had a secondary education or higher and were highly exposed to the media. They had achieved more than three antenatal cares and had been assisted by a skilled attendant at delivery and they have no more than three children.

Profile of children aged 0-6 months by EBF status

The second group consists of exclusively breastfed children whose mothers lived in the northern region and in rural areas in general. These women are largely of the Gur ethnic group and from neighboring countries. They belonged to households with low economic well-being with more than four children. They were uneducated, with little or no exposure to breastfeeding counseling and usually live with their partners.

The results of the discrete-time logistic models are presented in Tables 2 and 4. The values provided by these tables are odds ratios but the Table 4 present crude OR, adjusted OR and the 95% CI. The regression models consider 980 children under six months. Child age was a control variable for each model. The Table 3 highlights the explanatory factors of the EBF discontinuation, ranked in decreasing order of their explanatory power: region of residence, child’s health status, household standard of living, child’s sex, mother’s exposure to breastfeeding counseling and mother’s level of education (Tables 3 and 4).

Compared to children in the North, children in the East and West regions were 0.54 (95% CI 0.31, 0.97 and 0.65 (95% CI 0.47, 0.89) times less likely to experienced cessation of EBF before six months, respectively. In contrast, children in Abidjan were about two times more likely to be deprived of exclusive breastfeeding than children in the North. The difference in risk of discontinuing EBF was not significant between children in the North and those in the Central and Southern regions without Abidjan. Thus, exclusive breastfeeding is practiced over a relatively long period of time in the East and West regions than in the Center, North and South regions (with Abidjan).

In models M1 to M14, region of residence remained significant, whereas it was not significant in the crude effects model. The coefficients associated with its modalities changed with the introduction of variables such as religion, household standard of living, cohabitation of spouses, level of education, age at delivery, and degree of exposure to medical discourse on breastfeeding and the media. Therefore, region of residence primarily affects the duration of exclusive breastfeeding indirectly through the other variables. In other words, some of the differences in risk of EBF discontinuation associated with the mother’s region of residence are explained by differences in standard of living, education, and exposure to medical discourse on breastfeeding. In addition, the city of Abidjan became significant with the inclusion of standard of living in the M5 model and decreased from 5 to 1% from the M11 model onwards with the introduction of mothers’ media exposure, then remained stable until the final model. The influence of the city of Abidjan is mediated by household standard of living and women’s media exposure.

The risk of being deprived of exclusive breastfeeding among children who were not sick in the two weeks prior to the survey was greater than among children who had been sick (adjusted OR 1.8; 95% CI 1.452, 2.234). Among children who showed signs of cough, diarrhea, or fever, the duration of EBF tended to be longer than if they were healthy. In both the crude and final models, the health status of the child had a significant influence on the duration of the exclusive breastfeeding of children before their 6th month.

The duration of exclusive breastfeeding is significantly influenced by the household’s level of economic well-being. According to the Table 4, children from households classified in the very low and low standard of living quintile more likely (adjusted OR 1.59; 95% CI 0.95, 2.65 and adjusted OR 2.15; 95% CI 1.33, 3.5, respectively) to be deprived of EBF than children from households classified in the very high standard of living quintile. Children from households classified in the middle and high standard of living quintile are 1.85 (95% CI 1.19, 2.87) and 2.53 times (95% CI 1.69, 3.8) more likely to be deprived of exclusive breastfeeding before six months, respectively.

While the interaction between EBF and this variable was insignificant in the crude model, its significance appeared in the model (M5). Thus, its influence on exclusive breastfeeding cessation is indirect and reinforced by other variables. Moreover, the effect of the “Very weak” modality fades with the introduction of the “Exposure to medical discourse” variables. This could be because mothers in households with very low economic power do not have access to health services to the same degree as mothers living in households with very high economic power. This could result in some inequality in access to breastfeeding information.

The sex of the child significantly influences the duration of the exclusive breastfeeding. Indeed, the risk of being deprived of EBF before six months is higher for girls (adjusted OR 1.48; 95% CI 1.22, 1.80) than for boys. Thus, girls are less likely to be exclusively breastfed than boys.

Children of mothers with high exposure to medical discourse were less likely (adjusted OR 0.74; 95% CI 0.60, 0.91) to be denied exclusive breastfeeding before six months than those whose mothers had low exposure. Otherwise, children of mothers who had at least four antenatal cares and were attended by skilled health personnel at delivery were more likely to receive EBF than other children. The adjusted OR associated with unexposed (adjusted OR 1.36; 95% CI 0.87, 2.13) and highly exposed (adjusted OR 0.74; 95% CI 0.59, 0.91) mothers show that increasing the degree of exposure to medical speech increases the odds of exclusive breastfeeding.

The mother’s education level is relevant to explain the duration of exclusive breastfeeding. Children of uneducated mothers and those of mothers with primary education are about 0.73 (95% CI 0.52, 1.02) and 0.64 (95% CI 0.45, 0.91) times less likely to be deprived of exclusive breastfeeding, respectively. Women with no education and those with primary education practice exclusive breastfeeding longer than other women.

In models M7 to M12, the mother’s education level had no influence on the dependent variable. The effect of this variable on the continuation or cessation of exclusive breastfeeding became significant with the introduction of the child’s health status in model M13. The influence of this variable is therefore boosted by the child’s health status.