Promising trends and influencing factors of complementary feeding practices in Côte d’Ivoire: An analysis of nationally representative survey data between 1994 and 2016

Poor complementary feeding (CF) challenges early childhood growth. We examined the trends and influencing factors of CF practices among children aged 6–23 months in Côte d’Ivoire. Using data from Demographic and Health Surveys (DHS, 1994–2011) and Multiple Indicator Cluster Surveys (MICS, 2000–2016), the trends and predictors of World Health Organization-United Nations International Children’s Emergency Fund CF indicators including the timely introduction of complementary foods (INTRO), minimum meal frequency (MMF), minimum dietary diversity (MDD) and minimum acceptable diet (MAD) were determined. Using 2016 MICS data, we applied multivariate logistic regression models to identify factors associated with CF indicators. Between 1994 and 2016, the mean proportion of children aged 6–8 months achieving INTRO was 56.9% and increased by about 25% points since 2006. Over 2011–2016, the proportion of children aged 6–23 months meeting MMF, MDD and MAD increased from 40.2% to 47.7%, 11.3% to 26.0% and 4.6% to 12.5%, respectively. Older children and those from urban households had higher odds of meeting MDD and MAD. Maternal TV watching was associated with higher odds of meeting MDD. The secondary or higher education levels of mothers significantly predicted higher odds of meeting INTRO and MDD. Currently, breastfeeding was also positively associated with odds of meeting MMF and MAD. Children from poorer households had lower odds of meeting MMF, MDD and MAD. Despite the improvements, CF practices remain suboptimal in Côte d’Ivoire. Influencing factors associated with CF were distributed across individual, household and community levels, calling for future programmes and policies to implement multi-level strategies to improve young children’s diet in Côte d’Ivoire. To understand the CF trends (aim 1), we extracted data on the key CF indicators from five nationally representative survey reports: the 1994 and 2011 Côte d’Ivoire Demographic and Health Surveys (DHS) reports and 2000, 2006 and 2016 Côte d’Ivoire Multi‐Indicator Cluster Survey (MICS) reports. Four independent researchers conducted data extraction in pairs. Any discrepancy (i.e., inaccurate data extracted from wrong tables in reports) between the researchers were resolved through group discussion until consensus was reached. To explore the current influencing factors of CF (aim 2), we analyzed the most recent 2016 MICS data. Information on the 2016 MICS survey methodology, sampling procedure and questionnaires has been published previously (Institut National de la Statistique et al., 2016). Briefly, eligible women and children were included based on a two‐stage stratified sampling procedure. At the first stage, a total of 512 census enumeration areas were selected with probability as the primary sampling units (PSUs). At the second stage, 25 households were selected by systematic sampling within each PSU. Based on prior studies looking at CF practices in low‐ and middle‐income countries (Na, Aguayo, Arimond, Dahal, et al., 2018; Na, Aguayo, Arimond, Mustaphi, et al., 2018; Na et al., 2017), the inclusion criteria of mother–child pairs to be included in our analysis were: (1) mothers were between 15 and 49 years of age; (2) the youngest singleton child was aged 6–23 months; (3) children were alive at the time of the survey; and (4) children lived with their mothers. We included mothers aged 15–49 years old to decrease the possibility of enroling children with potential health problems born from teenage (<15 years) or older (>49 years) mothers. The study further defined the youngest singleton children aged 6–23 months to avoid potential recall bias and to prevent enroling more than one child from each household. In addition, only alive children living with their mothers were included, so the surveys were able to collect their CF practices from mother–child pairs. Four CF indicators defined by WHO in 2010 were analyzed in this study, including the introduction of solid, semisolid, or soft foods (INTRO), minimum meal frequency (MMF), minimum dietary diversity (MDD) and MAD (World Health Organization, 2010). Data for INTRO were available in 1994 and 2011 DHS and 2000, 2006 and 2016 MICS reports. Data for MMF, MDD and MAD were only available in the 2011 DHS and 2016 MICS reports. The CF indicators were defined per WHO definitions as follows (WHO, 2010): INTRO: The proportion of infants 6–8 months of age who received solid, semisolid and soft foods in the previous day or night. MMF: The proportion of breastfed and nonbreastfed children 6–23 months of age, who received solid, semisolid or soft foods (including milk for nonbreastfed children) the minimum recommended number of times or more in the previous day or night. For breastfed children, MMF is met if at least two solid/semisolid feeds occurred for children aged 6–8 months and at least three feeds occurred for children aged 9–23 months. For nonbreastfed children, MMF is met if at least four feeds of complementary food or milk for children aged 6–23 months occurred. MDD: The proportion of children 6–23 months of age who received foods from four or more food groups in the previous day or night out of the following seven food groups: (1) grains, roots and tubers, (2) legumes and nuts, (3) dairy products, (4) flesh foods, (5) eggs, (6) vitamin‐A‐rich fruits and vegetables and (7) other fruits and vegetables. MAD: The proportion of children 6–23 months of age who received the minimum recommended dietary diversity and the minimum recommended meal frequency in the previous day or night. For breastfed children, they are classified as having a MAD when they meet the MMF and MDD standards. For nonbreastfed children, they are classified as having a MAD when they meet the MMF standards and receive at least two milk feedings along with at least four food groups other than milk products. The selection of the influencing factors at the individual, household and community levels was based on the conceptual framework developed by Stewart et al. (2013) and our previous work in South Asia (Na, Aguayo, Arimond, Dahal, et al., 2018; Na, Aguayo, Arimond, Mustaphi, et al., 2018; Na, Aguayo, Arimond, Narayan, et al., 2018; Na et al., 2017). Individual‐level factors included child, maternal and paternal characteristics. For children, the following variables were included: sex, age, birth order, birth interval, measured birthweight, perceived birthweight and child morbidity including diarrhoea, fever and cough. Maternal characteristics included age, smoking status, education, marital status, occupation, nutritional status (height and body mass index), breastfeeding practices, utilization of reproductive health care, exposure to media and women’s attitude towards domestic violence. Paternal characteristics included age and education. Household‐level factors included household structures and socioeconomic status. The household characteristics included the place of residence, sex of household head, number of household members, number of children under five years, types of cooking fuel, water characteristics (source and location of drinking water, time to get to water sources) and quintiles of overall household wealth index (higher quintiles indicate poorer households). The community‐level factor described access to health care within the community where the selected subjects lived. Based on the utilization of maternal and child nutrition and health care services among all respondents, the rank score for community access to health care was generated first and then categorized into quintiles. Higher rank scores or quintiles indicate poorer community access to health care. A detailed description of the factors is available elsewhere (Na et al., 2020). All data analysis was performed using STATA/SE 15.1 (StataCorp). The prevalence of four CF indicators was extracted from the national DHS and MICS reports. Multivariable models were applied to determine the associations between influencing factors and CF indicators among children aged 6–23 months: (1) the bivariate associations between influencing factors and CF indicators were examined first to select the significant risk factors at p= 0.1, and (2) the selected variables from the bivariate analysis were included in the multivariate risk factor analysis.