1. Background of the study

Complementary feeding, as defined by the World Health Organization (WHO), is the process of introducing additional foods and liquids alongside continued breastfeeding, typically beginning at six months of age  (WHO, 2023). According to UNICEF, complementary foods refer to any solid or liquid food—whether locally prepared or manufactured (WHO, 2019)—that supplements or replaces breast milk. Introducing diverse foods during infancy is essential for optimal growth and development, as this period represents a critical window for establishing healthy dietary habits (UNICEF, 2022).

Globally, the prevalence of poor complementary feeding practices varies significantly. For instance, rates range from 40% in Latin America and the Caribbean (Ulisses et al., 2020) to as high as 89% in Southern Asia, such as in India(NFHS-5, 2019). Other regions report prevalence rates of approximately 48% in Guatemala (Movement & Progress, 2020) and 41.5% in China(J. Liu et al., 2021). In high-income countries like the United States and those in Europe, challenges often relate to dietary quality—particularly the overconsumption of sugar and insufficient intake of fruits and vegetables during early childhood (X. Liu et al., 2021).

In Africa, the prevalence of poor complementary feeding is also alarmingly high. Rates vary from 64.8% in Ghana  (Saaka et al., 2022) to 91.4% in Liberia (M. et al., 2022). Uganda reports a prevalence of 68.7%(Scarpa et al., 2022), while Gambia stands at 78.2%(M. et al., 2022), Ethiopia at 81.8%(Assefa & Belachew, 2022), and the Eastern and Southern Africa region collectively reports 76.0%(Gatica-Domínguez et al., 2021).

Exclusive breastfeeding and timely introduction of complementary foods during the first 1,000 days of life are crucial for normal child growth and development (WHO, 2019). Adequate complementary feeding significantly contributes to improved cognitive development and linear growth (Panjwani, A., & Heidkamp, R., 2017).  However, developed regions continue to face challenges, including poor dietary diversity and reliance on nutrient-poor foods high in sugar like in the USA and Europe (X. Liu et al., 2021). Complementary feeding practices also vary across culture, religion, beliefs and socioeconomic classes (Sajini Varghese et al., 2017).

Inappropriate and low-quality complementary feeding practices remain the greatest challenges to child survival, growth, and development (UNICEF, 2022). ). Inadequate quality and quantity of complementary foods, along with unsuitable complementary feeding practices, pose a hazard to the health and nourishment of children (White, J. M.. et al, 2017). Current data from various global child nutrition reports show that more than 48% of children below 2 years are fed with the minimum acceptable meal and 71% do not have a minimally diversified diet  (UNICEF, 2022).

In Rwanda, only less than 22% of children aged from 6 to 23 months are fed with a minimum acceptable diet as complementary feeding packages (RDHS, 2019), and this has contributed to 18% of children below 2 years becoming stunted (UNICEF, 2022). In the context of Nyamasheke district, more than 40% of children below 5 years are stunted, and the level of complementary feeding practices remains critically low, with only less than 10% of children below 2 years receiving an acceptable diet (RDHS, 2019), (UNICEF Rwanda, 2019). However, limited maternal nutrition, knowledge, household economic status, and social norms have been suggested to influence appropriate complementary feeding practices in the western province  (RDHS, 2019).

The practice of complementary feeding in Rwanda is influenced by a number of factors such as maternal health, nutrition education, broader health challenges, and socioeconomic challenges (Uwiringiyimana, V. et al., 2019). However, RDHS reported that only less than 22% of children aged 6 to 23 months receive a minimum acceptable diet in complementary feeding packages (RDHS, 2019). . Furthermore, UNICEF Rwanda reported that less than 18% of children aged 6 to 8 months are stunted, and this condition has been linked to delayed or poor and inappropriate complementary feeding practices  (UNICEF, 2022).

However, there is a scarcity of data about complementary nutritional practices in Nyamasheke district, and there is a need to determine the association of complementary nutritional feeding practices and the increased level of stunting, especially among children below 2 years but above 6 months old in Nyamasheke district. Therefore, this study aims to explore factors influencing complementary feeding among lactating mothers in Nyamasheke catchment area, and the study outcome was used to guide in developing measures to improve infant feeding practices needed in the reduction of stunting in the western province, Rwanda.

2. Methods

The authors followed various research approaches to conduct this study as described below.

2.1 Study design

This was a cross-sectional study that collected quantitative data from September 2 to 16, 2024.  

2.2 Setting

This study was conducted in Nyamasheke district, which comprises two distinct regions: one near Nyungwe National Park and the other bordering Lake Kivu. The district is situated in the Western Province of Rwanda, in Eastern Africa. Figure 1 provides further information about the geographical location of Nyamasheke district.

2.3 Study Population, Sample Size, and Selection of Participants

This study targeted a population of 2,490 lactating mothers and caretakers of children aged between 6-24 months in the communities of Nyamasheke district (Health Management Information System).

1. Sampling design and sample size

Data was collected from 385 mothers or guardians of children aged 6 to 24 months. The sample size was determined using the Cochran formula(Cochran, 2002).

2. Sample size

Given the known study population, the sample size was calculated using the statistical formula developed by Cochran (1963). The mathematical illustration of the Cochran model is provided below.

n= z2*p*(1-p)e2                            Where:

no =sample size  ; Z2 =confidence interval: 1.96; e2⁼ Degree of accuracy: 0.05;

Q= Estimated population: 1-P ; P= Population assumed portion: 50%

    n=1.962 0.5(1-0.5)0.052

no   = 385

Therefore, current study considered a sample size of 385.

3. Selection of participants

Study participants were conveniently enrolled in the study until the desired sample size was achieved.

2.4 Data collection

Following approval from the Ethical Review Committee of Mount Kenya University and administrative clearance from Nyamasheke District, the researcher proceeded with formal introductions and preparations for data collection. Authorization letters were presented to the Executive Secretaries of the selected sectors and to the Heads of the respective health centers involved in the study.

At this stage, preliminary data were retrieved from the Antenatal Care (ANC) units of the selected sectors. This included information such as the number of eligible respondents per village and Isibo, the age of the children, and the details of mothers or caregivers with children aged between 6 and 24 months.

To select study participants, a simple random sampling technique was applied to the ANC data. This ensured that every eligible individual had an equal probability of being selected, thus enhancing the representativeness of the study population.

Once participants were identified, the researcher mapped and located them within their respective communities, based on the local administrative structure of villages and Isibo. Before initiating data collection, the Research Assistants introduced themselves to all potential participants, explained the purpose, methods, and expected outcomes of the study, and guided them through the informed consent process.

Participants who met the inclusion criteria were then interviewed through face-to-face interactions. The administration of the structured questionnaire took approximately 15 to 30 minutes per participant. The data collection process followed a pre-developed research protocol, which included:

1. Obtaining ethical approvals from Mount Kenya University and Nyamasheke District;
2. Mapping and identifying eligible participants;
3. Conducting pre-testing of the data collection tools;
4. Introducing the study to participants and obtaining informed consent;
5. Collecting data using structured, closed-ended questionnaires and face-to-face interviews.

2.5 Data collection tools

This study utilized a structured questionnaire to gather data. The questionnaire included closed-ended questions regarding the demographic information of the children and the feeding practices of their mothers or guardians. Additionally, the English version was proofread by two additional individuals in the field of nutrition to identify any potential errors. Interviews were conducted using the questionnaire in the native language of the study participants and then translated into English after proofreading to correct any errors.

2.6 Variables

Figure 2 shows an overview of the independent and outcome variables for this study. Independent variables include those related to social-demographic factors, health-economic factors, and child feeding knowledge. The dependent variable is based on complementary feeding practices. Complementary feeding is defined as any food that is cooked and chopped for the purpose of feeding a child aged between six to twenty-four months. Based on Minimum Dietary Diversity (MDD), we estimated the prevalence of poor complementary feeding. MDD is defined as receiving at least 5 out of 8 recommended food groups in the day preceding data collection. MDD estimates the probabilities that a baby consumes (1) grains/roots/tubers, (2) legumes/nuts, (3) dairy products (milk, yogurt, and cheese), (4) flesh foods (meat, fish, poultry, and liver/organ meat), (5) eggs, (6) vitamin A-rich fruits and vegetables, and (7) other fruits and (8) vegetables(RDHS, 2021). Of the 385 respondents, about 34.8% were classified as having poor complementary feeding practices.

2.7 Bias

To reduce bias and ensure the accuracy of the collected data to answer current research questions, this questionnaire was proofread by three different people to check the clarity of the questions. To ensure easy reproducibility of answers to the questions, the questionnaire was translated into English. Lastly, the questionnaire was administered to 20 breastfeeding mothers. The questionnaire was newly developed by adopting questions on complementary feeding from similar studies globally and in Africa (Gatica-Domínguez et al., 2021; Mulugeta et al., 2024; Nurokhmah et al., 2022; Saaka et al., 2022; Scarpa et al., 2022; Ulisses et al., 2020).

2.8 Quantitative variables

During the analysis, we transformed continuous variables into categorical variables to facilitate the analysis of trends and relationships for interpretation. The following are the baseline groupings:

The income of the participants or their husbands was grouped as:

• 0–9,999: No or very low income (possibly unemployed or in informal jobs).
• 10,000–29,999: Low to moderate income.
• 30,000–900,000: Middle to high income.

We collected and analyzed husbands' income because they often contribute significantly to household income. Categorizing this variable allows for exploring how household financial strength impacts child nutrition and caregiving.

Baby age in months is also grouped to reflect developmental stages. These categories help assess age-appropriate feeding practices and identify when complementary feeding challenges are most likely to arise. The groupings were as follows:

• 6–12 months: Period when complementary feeding starts, alongside breastfeeding.
• 13–18 months: Transitional feeding stage; increased reliance on family foods.
• 19–24 months: Approaching weaning age; expected to be eating more like the rest of the family.

2.9 Statistical methods

The data analysis was conducted in three main steps. Firstly, descriptive statistics were performed using STATA version 17 to determine the prevalence of poor complementary feeding practices and describe the sociodemographic characteristics of the study participants. Secondly, inferential analysis using Pearson’s correlation was carried out to assess the associations and relationships between independent and outcome variables. Finally, multivariate logistic regression analysis was employed to identify and evaluate the independent effects of factors on poor complementary feeding practices among mothers and caregivers of children aged 6 to 24 months, while adjusting for potential confounders.

3. Results

This section covers the results of the study, including the provision of socio-demographic characteristics of the respondents, the prevalence of poor complementary feeding (PCF), and factors associated with poor complementary feeding practices (PCFPs).

3.1 Demographic characteristics of the study participants

This study collected data from 385 babies aged between 6 to 24 months with a mean age of 15 months, with nearly equal representation of both sexes: 51.9% males and 48.1% females, and 99% born in a health facility. Respondents were biological mothers or grandmothers of the babies and caregivers. They were reached from 9 sectors of Nyamasheke district, including Macuba, Nyabitekeri, Shangi, Kagano, and Kanjongo, considered to be surrounding Lake Kivu, and Rangiro, Bushekeri, Cyato, and Karambi sectors located nearby Nyungwe National Park, which is the biggest forest in Rwanda. They aged between 19 to 62 years old, with nearly half (46%) of them aged 30 to 39 years old, and residing in sectors around Lake Kivu (47.5%). The majority had primary education as their highest attained schooling level (66.5%) and were engaged in agriculture and subsistence farming as their occupation (63.1%). Farming activities among families of the participants mainly involved tea and coffee farming/cropping activities, which took up most of their time from 5:00 AM to 6:00 PM in farmlands.

Regarding financial status, based on self-report, respondents were earning amounts of money in Rwandan francs ranging from 0 to 900,000 per month. About 71.7% of them were earning less than 30,000 frw per month. Similarly, the majority of their husbands were earning below 30,000 frw per month. Detailed information about the study participants' characteristics can be found in Table 1.

Table 1: Demographic characteristics of the study participants

3.2 Prevalence of poor complementary feeding of children aged 6 to 24 months in Nyamasheke district, Rwanda

Based Minimum dietary diversity (MDD), we estimated the prevalence poor complementary feeding. MDD was defined as receiving at least 5 out of 8 recommended food groups in the last day preceding data collection.  MDD estimates probabilities that a baby (1) grains/roots/ tubers, (2) legumes/nuts, (3) dairy products (milk, yogurt, and cheese), (4) flesh foods (meat, fish, poultry, and liver/organ meat), (5) eggs, (6) vitamin A-rich fruits and vegetables, and (7) other fruits and (8) vegetables(RDHS, 2021). Of the 385 respondents, about 34.8% were classified as having Poor Complementary feeding practices.  Figure 3 provides further information.

Based on Minimum Dietary Diversity (MDD), we estimated the prevalence of poor complementary feeding. MDD is defined as receiving at least 5 out of 8 recommended food groups in the day preceding data collection. The recommended food groups include (1) grains/roots/tubers, (2) legumes/nuts, (3) dairy products (milk, yogurt, and cheese), (4) flesh foods (meat, fish, poultry, and liver/organ meat), (5) eggs, (6) vitamin A-rich fruits and vegetables, and (7) other fruits, and (8) vegetables (RDHS, 2021). Out of the 385 respondents, approximately 34.8% were classified as having poor complementary feeding practices. Figure 3 provides further information.

Figure 2: Prevalence of poor complementary feeding practice among caregivers of babies in Nyamasheke district, Rwanda

3.3 Factors associated with poor complementary feeding

To determine factors associated with participants' characteristics, they were cross-tabulated with complementary feeding practice (CFP) status. Various characteristics were found to be statistically associated with Complementary Feeding Practices. To understand geographical differences, we classified sectors according to their location relative to the city, Lake Kivu, and Nyungwe forest. Nyabitekeri, Macuba, and Shangi sectors were grouped as sectors

around Lake Kivu; Karambi, Rangiro, Bushekeri, and Cyato were considered as those near Nyungwe forest, while Kagano and Kanjongo were considered as those located in the urban or city area of Nyamasheke district.

The analysis output shows that a significantly higher proportion of respondents with poor CFP were aged 19-29 years (42.7%), residing in sectors around Nyungwe forest (as compared to those around Lake Kivu or in urban areas), engaged in cultivation as an occupation (48.0%), had babies aged 6 to 12 months (50.7%), reported a declining amount of earned money per month, self-reported not being visited by a lay community health worker (59.1%), reported hardship in accessing health services (60.7%), and did not attend ANC as recommended (72.2%), among others. Furthermore, the prevalence of poor complementary feeding practices was higher among babies reported to have any disease at the time of data collection (65.2%), such as having low blood hemoglobin (54.5%).

Participants' knowledge of some characteristics of complementary feeding practices was also found to be associated with poor Complementary Feeding Practices. A larger proportion of participants with poor CFP were not aware of when CF starts (50.0%), perceived that milk CF involves milk (50.0%), and fruits at least more than 3 times (35.6%), among others (Table 2).

Table 2: Factors associated with poor complementary feeding among caregivers of babies in Nyamasheke district, Rwanda

All participant characteristics statistically associated with poor complementary feeding practices (CFP) in Table 2 were included in a logistic regression model to determine the strength and direction of the association. Out of these characteristics, only five were found to uniquely predict poor complementary feeding among the study participants. These predictors included the location of residence sector, the amount of money earned by the participant per month, the age of the baby in months, regular attendance of postnatal care (PNC), and whether the child had any diseases.

Participants residing in sectors near Nyungwe forest were 3.5 times more likely to exhibit poor CFP compared to those living around Lake Kivu (AOR: 3.5, 95% CI 1.5-7.7). Additionally, participants earning between 10,000 to 29,999 Rwandan francs per month were 2.3 times more likely to have poor complementary feeding practices than those earning 30,000 francs and above (AOR: 2.3, 95% CI: 1.1-4.9). Moreover, respondents with babies aged between 6 to 12 months were 2.0 times more likely to have poor complementary feeding practices compared to those with babies aged between 19-24 months (AOR: 2.0, 95% CI: 1.1-3.9). Participants who reported irregular attendance of postnatal care (PNC) were seven times more likely to exhibit poor CFP than those who attended PNC regularly.

Furthermore, babies with diseases were three times more likely to be raised by a caregiver with poor complementary feeding practices compared to those without any diseases (AOR: 3.0, 95% CI: 1.4-6.2). Table 3 provides additional information on the logistic regression analysis model of factors associated with poor complementary feeding among caregivers of children aged 6 to 24 months in Nyamasheke district, Western Rwanda. Table 3

Table 3:  Logistic regression on factors associated with poor complementary feeding practices