Factors associated with complementary feeding practices among children aged 6–23 mo in Malawi: an analysis of the Demographic and Health Survey 2015–2016

Abstract

Background

Optimal child complementary feeding practices are crucial for nutritional status, growth, development and health, and ultimately affect child survival. This is the first population-based study in Malawi that aimed to examine factors associated with complementary feeding practices among children aged 6–23 mo.

Methods

Utilizing data from the 2015–16 Malawi Demographic and Health Survey (MDHS), 4732 children aged 6–23 mo and their mothers were analysed. The MDHS produced a nationally representative sample using a multistage cluster sampling design that included sampling weights. The impact of child, maternal, household, community and health service utilization factors on complementary feeding practices was examined using the generalized estimating equation logistic regression.

Results

After controlling for a wide range of covariates, children from mothers with secondary or post-secondary education and from mothers working in agriculture and living in the central region were significantly more likely to have timely introduction to solid, semi-solid or soft food. Surprisingly, being >1 y of age was associated with reduced odds of achieving minimum meal frequency. In addition, children >1 y of age from mothers older than 24 y and from mothers with primary, secondary and post-secondary education were significantly more likely to achieve minimum dietary diversity. Children from rich households were more likely to achieve both minimum dietary diversity and minimum acceptable diet. Finally, exposure to mass media was significantly associated with increased odds of achieving minimum meal frequency, minimum dietary diversity and minimum acceptable diet.

Conclusions

Public health strategies aimed at reducing childhood undernutrition should focus on children from poor households whose mothers have no formal education and are unemployed. In addition, exposure to mass media had a positive impact on the three complementary feeding indicators. Therefore behaviour change communication messages through mass media aimed at promoting child nutrition are necessary to achieve optimal child complementary feeding practices.

Complementary feeding indicators, Infant nutrition, Malawi, Malnutrition, Risk factors

Introduction

Optimal child complementary feeding practices are crucial for nutritional status, growth, development and health, and ultimately affect child survival.^1,2 Complementary feeding is defined as the process of initiating other foods and liquids along with breast milk when breast milk alone is no longer adequate to meet the nutritional requirements of infants.³ Thus the World Health Organization (WHO) established guidelines with respect to infant and young child feeding (IYCF) practices for children 6–23 mo of age. These guidelines consider the introduction of solid, semi-solid or soft food, minimum dietary diversity, minimum meal frequency, minimum acceptable diet and consumption of iron-rich or iron-fortified foods as some of the eight core indicators for assessing IYCF practices.⁴ The transition period from exclusive breastfeeding to 2 y is a critical window for optimal growth and development of the child. Therefore appropriate, safe, adequately nutritious and frequent feeding is essential during this period.^5,6 It is well documented that children who are poorly fed during periods of growth spurts are at an increased risk of being malnourished.⁷ Undernutrition is a major contributor of global morbidity and mortality in children younger than 5 y and accounts for about 45% of all childhood mortalities.⁸ In addition, undernutrition has been associated with compromised child’s immunity and brain development issues that lead to delayed motor and cognitive development.⁷

Globally, only a few children are introduced to solid, semi-solid or soft food and receive nutritionally adequate, acceptable and diversified foods in a timely manner.⁹ In many countries, less than one-quarter of infants 6–23 mo of age are reported to meet the criteria for dietary diversity and feeding frequency appropriate for their age.⁹ Prior research has demonstrated that children who are introduced to solid, semi-solid or soft food before 6 mo of age as well as those who do not receive sufficient dietary diversity, meal frequency or adequate diet after 6 mo become stunted, develop micronutrient deficiencies and have common childhood illnesses.⁷ Therefore complementary feeding practices ought to be diverse, of adequate energy density and introduced at the right time in order to meet the growth needs of infants and young children. Previous studies have reported that factors such as the age of the child,⁵ employment status of the parents,⁵ mother’s marital status,¹⁰ mother’s educational attainment,^5,9 household monthly income/wealth,^6,9,11 mother’s knowledge/education,⁹ number of children who are 5 y of age or younger in the household,⁶ exposure to media,^6,11 geographic region and maternal health-seeking behaviour¹¹ have an influence on child complementary feeding practices.

In Malawi, malnutrition rates among infants and young children have consistently remained high over the past two decades. In 1992, the Malawi Demographic and Health Survey (MDHS) reported that 48.7, 27.2 and 5.4% of children younger than 5 y of age were stunted, underweight and wasted, respectively.¹² Similar trends of high levels of childhood malnutrition were reported between 2000 and 2014.^13,14 As of 2015–2016, the prevalence of undernutrition remained unacceptably high (37.1, 11.7 and 2.7% of children were stunted, underweight and wasted, respectively),¹⁵ indicating that undernutrition is a chronic public health problem in Malawi that needs urgent attention. The United Nations Children’s Fund states that the immediate cause of undernutrition is the result of low or a lack of dietary intake and recurrent infections.¹⁶

Despite the fact that the Malawi government has implemented a number of interventions since the early 2000s to guide and improve the levels of child complementary feeding practices,¹⁷ a large proportion of children younger than 2 y still lack nutritionally adequate and diversified foods.¹⁵ The 2016 MDHS revealed that 25% of breastfed children had an adequately diverse diet and had been given foods from the appropriate number of food groups, while 29% had been fed the minimum number of times appropriate for their age. In addition, only 8% of children 6–23 mo of age met the minimum standards for all three child complementary feeding practices.¹⁵ These percentages are low and there is a need to scale up optimal child complementary feeding practices in Malawi.

Although previous studies have assessed child complementary feeding practices in Malawi,^18,19 these studies did not use nationally representative samples. Furthermore, there are limited studies that have assessed the newly introduced WHO complementary feeding indicators.²⁰ Thus, using a nationally representative sample, we aimed to examine the factors associated with the introduction of solid, semi-solid or soft food, minimum dietary diversity, minimum meal frequency and minimum acceptable diet among children 6–23 mo of age in Malawi. The findings of the study will provide evidence for programme managers and policymakers to design and implement effective complementary feeding interventions. Specifically, the findings will help to address the levels of inadequate meal frequency, inadequate dietary practices and poor dietary diversity practices, and ultimately reduce the high prevalence of childhood malnutrition in Malawi.

Materials and methods

Study area

The study covered all three administrative regions of Malawi (northern, central and southern regions). Malawi is a sub-Saharan African country located south of the equator. It is bordered to the north and northeast by the United Republic of Tanzania; to the east, south and southwest by the People’s Republic of Mozambique; and to the west and northwest by the Republic of Zambia.²¹ The economy of Malawi is based primarily on agriculture, which accounts for about 30% of the gross domestic product.²²

Study data, design and sampling

This was a cross-sectional study that utilized 2015–2016 MDHS data. The Malawi Population and Housing Census, conducted in 2008, was used as the sampling frame for the MDHS. Comprehensive methods for the survey have been published elsewhere.¹⁵ Briefly, the 2015–2016 MDHS employed a two-stage probability sampling in the three regions of Malawi and produced a nationally representative sample. In the first stage, 850 standard enumeration areas (SEAs), comprising 173 SEAs in urban areas and 677 in rural areas, were selected with probability proportional to the SEA size and with independent selection in each sampling stratum. In the second stage, 30 households per urban cluster and 33 per rural cluster were selected with an equal probability systematic selection method.

Data collection and study sample

Data were collected from women aged 15–49 y with children younger than 5 y of age prior to the survey using face-to-face interviews. In the interviewed households, 24 562 of the 25 146 eligible women were interviewed, representing a 98% response rate.¹⁵ Information on sociodemographics and IYCF practices was collected via verbal reports. Data on IYCF practices were collected using a 24-h recall method, that is, mothers were asked to recall all foods and liquids consumed by their children in the day or night before the interview.²³ We restricted our analysis to the children currently living with the respondent and 6–23 mo of age.⁴ Figure 1 shows sample inclusion criteria.

Figure 1.

Flow chart showing inclusion of the study participants.

Measures

Dependent variables

Four outcome variables were considered in this study: introduction of solid, semi‐solid or soft foods (yes/no); minimum meal frequency (yes/no); minimum dietary diversity (yes/no); and minimum acceptable diet (yes/no). Introduction of solid, semi‐solid or soft foods was defined as the proportion of infants 6–8 mo of age who received solid, semi‐solid or soft foods in the previous day or night.⁴ Minimum meal frequency was defined as the proportion of breastfed and non-breastfed children 6–23 mo of age who received solid, semi-solid or soft foods (but also including milk for non-breastfed children). The minimum frequency was defined as twice for breastfed infants 6–8 mo of age, three times for breastfed children 9–23 mo of age and four times for non-breastfed children 6–23 mo of age.⁴ Minimum dietary diversity was defined as the proportion of children 6–23 mo of age who received foods from four or more food groups in the 24-h recall period. The food groups in the MDHS were regrouped to be consistent with the seven WHO-recommended food groups used in defining the minimum dietary diversity indicator for children as follows: grains, roots and tubers; legumes and nuts; dairy products; flesh foods (meats/fish/poultry); eggs; vitamin A-rich fruits and vegetables; and other fruits and vegetables.⁴ Minimum dietary diversity was calculated by summing up the reported number of food groups consumed in the 24-h recall period. Minimum acceptable diet was defined as the proportion of children 6–23 mo of age who received a minimum acceptable diet (apart from breast milk). This composite indicator was calculated from the following two fractions: breastfed children 6–23 mo of age who had at least the minimum dietary diversity and the minimum meal frequency during the previous day divided by breastfed children 6–23 mo of age; and non-breastfed children 6–23 mo of age who received at least two milk feedings and had at least the minimum dietary diversity not including milk feeds and the minimum meal frequency during the previous day divided by non-breastfed children 6–23 mo of age.⁴

Independent variables

A number of covariates were considered in this current study, including child, maternal, household, community and health service utilization factors. Child characteristics included the sex of the child (male or female), the age of the child (6–11, 12–17 or 18–23 mo) and perceived birth size (very small/smaller than average, average or very large/larger than average). Women’s characteristics were maternal age (15–24, 25–34 or 35–49 y), maternal education levels (no formal education, primary education or secondary and post-secondary education), marital status (unmarried vs married), parity (1, 2, 3 or 4) and maternal occupation (no employment, agriculture or non-agriculture); exposure to mass media (measured by access to newspapers, radio or television; respondents who had access to any of the three were considered to have media exposure [yes/no]); religion (Roman Catholic, Protestant, Muslim or other); and ethnicity (Chewa, Tumbuka, Lomwe, Yao, Ngoni or other). Household characteristics included household wealth (poorest, poorer, middle, richer or richest). The household wealth index is a composite measure of a household’s cumulative living standard and was calculated using easy-to-collect data on a household’s ownership of selected assets, such as televisions and bicycles. Household asset scores were generated through a principal component analysis. The resulting asset scores were standardized and categorized into quintiles.²⁴ Other household factors were the household size (<5 vs ≥5), the number of children younger than 5 y in the household (≤1, 2 or ≥3) and sex of the head of the household (male vs female). Community-level characteristics included the place of residence (urban vs rural) and geographic region (northern, central or southern). Health service utilization factors included vitamin A supplementation in the last 6 mo (no vs yes), place of delivery (non-health facility vs health facility), the number of antenatal care visits during pregnancy (inadequate vs adequate) according to WHO recommendations²¹ and baby postnatal check within 2 mo (no vs yes).

Statistical analysis

All analyses were performed using Stata version 15 (StataCorp, College Station, TX, USA). The ‘svy’ (survey) command was applied to take into account the complex survey design of the MDHS (i.e., to adjust for clustering and sample weights). The frequencies of study characteristics were reported as percentages with their 95% CIs. Using Pearson’s χ², the bivariate analysis was performed to test the differences in distribution between groups (introduction of solid, semi‐solid or soft foods [yes/no], minimum meal frequency [yes/no], minimum dietary diversity [yes/no], minimum acceptable diet [yes/no]). All variables that showed significance at p≤0.25 in the bivariable analyses were fitted into the final models of the regression analyses. The multivariable analyses were conducted using a series of logistic regression models fitting the generalized estimating equation (GEE) for estimating the effects of predictors on the outcomes after controlling for confounding factors. Due to the hierarchical structure of the dataset, the GEE was used to adjust for possible correlated individual responses.²⁵ In the GEE models, sampling weights were applied. The results of the multivariate analyses were presented as the adjusted odds ratio (aOR) with the corresponding 95% CI. Variables with p-values <0.05 were considered statistically significant.

Ethics statement

The survey protocol was reviewed and approved by the Malawi Health Sciences Research Committee and the Institutional Review Board of ICF Macro. However, the present study obtained approval from ICF Macro to analyse the data; therefore, no further ethical consideration was required. The data are publicly available at http://dhsprogram.com/data/available-datasets.cfm and can be obtained for free upon request.

Results

Sample characteristics

A total of 4732 children 6–23 mo of age were analysed in the current study. Figure 1 displays sample inclusion and exclusion criteria. A majority of children (85%) had been timely introduced to solid, semi-solid and soft foods. Furthermore, 29.8% of children had achieved minimum meal frequency, 24.9% had achieved minimum dietary diversity and 8.3% had achieved a minimum acceptable diet. Table 1 reveals the sociodemographic characteristics of the study participants.

Table 1.

Sociodemographic characteristics of the study participants, MDHS 2015–2016 (n=4732)

Characteristics	Numbera	Percentageb	95% CI
Child characteristics
Sex of the child
Male	2400	50.72	49.06 to 52.37
Female	2332	49.28	47.63 to 50.94
Age of the child (mo)
6–11	1648	34.82	33.13 to 36.55
12–17	1601	33.83	32.19 to 35.51
18–23	1483	31.35	29.76 to 32.97
Perceived birth size
Very small	715	11.12	13.69 to 16.66
Average	2395	50.61	48.66 to 52.56
Very large	1632	34.27	32.46 to 36.13
Maternal characteristics
Maternal age (y)
15–24	2079	43.94	42.07 to 45.81
25–34	1900	40.15	38.37 to 41.97
35–49	753	15.91	14.64 to 17.27
Maternal education level
No formal education	554	11.71	10.36 to 13.21
Primary	3180	67.20	65.17 to 69.18
Secondary and above	998	21.09	19.08 to 23.24
Marital status
Unmarried	785	16.60	15.25 to 18.04
Married	3947	84.40	81.96 to 84.75
Parity
1	1264	26.71	25.09 to 28.39
2	856	20.22	18.92 to 21.57
3	766	16.19	15.04 to 17.41
≥4	1746	36.89	34.99 to 38.82
Maternal occupation
No employment	1438	30.40	28.14 to 32.75
Agricultural	2111	44.61	42.06 to 47.20
Non-agricultural	1183	24.99	22.86 to 27.25
Media exposure
No	3151	66.59	64.64 to 68.49
Yes	1581	33.41	31.51 to 35.35
Religion
Catholic	850	17.98	16.21 to 19.89
Protestant	981	20.73	18.84 to 22.76
Muslim or other	2901	61.29	58.75 to 63.78
Ethnicity
Chewa	1662	35.13	31.92 to 38.48
Tumbuka	404	8.55	7.06 to 10.31
Lomwe	803	16.97	15.00 to 19.14
Yao	768	16.23	13.64 to 19.21
Ngoni	825	11.11	9.31 to 13.19
Household characteristics
Household wealth
Poorest	1208	25.53	23.68 to 27.48
Poorer	1090	23.02	21.46 to 24.67
Middle	921	19.46	18.02 to 20.99
Richer	785	16.59	15.21 to 18.07
Richest	728	15.39	13.29 to 17.75
Household size
<5	1824	38.54	36.65 to 40.46
≥5	2908	61.46	59.54 to 63.35
Number of children younger than 5 y
≤1	2229	47.10	45.16 to 49.04
2	2129	44.98	43.34 to 46.75
≥3	374	7.92	6.82 to 9.18
Sex of the head of the household
Male	3564	75.31	73.61 to 76.94
Female	1168	24.69	23.06 to 26.39
Community characteristics
Place of residence
Urban	639	13.50	11.04 to 16.42
Rural	4093	86.50	83.58 to 88.96
Geographic region
Northern	555	11.74	10.10 to 13.60
Central	2044	43.18	40.22 to 46.20
Southern	2133	45.08	42.23 to 47.96
Health service utilization
Vitamin A in the last 6 mo
No	1633	34.52	32.67 to 36.41
Yes	3099	65.48	63.59 to 67.33
Place of delivery
Non-health facility	325	6.88	5.85 to 8.06
Health facility	4407	93.12	91.94 to 94.41
Number of antenatal visits
Inadequate	2389	50.49	48.46 to 52.51
Adequate	2343	49.51	47.48 to 51.54
Postnatal (within 2 mo)
No	2631	55.60	53.07 to 58.10
Yes	2101	44.40	41.90 to 46.93
Outcomes
Introduction of solid, semi-solid or soft foods (n=774)
No	118	15.19	12.22 to 18.70
Yes	656	84.81	81.30 to 87.77
Minimum meal frequency
No	3322	70.20	68.40 to 71.93
Yes	1410	29.80	28.06 to 31.59
Minimum dietary diversity
No	3553	75.08	73.03 to 77.02
Yes	1179	24.92	22.98 to 26.97
Minimum acceptable diet
No	4337	91.64	90.54 to 92.63
Yes	395	8.36	7.37 to 9.46

^aWeighted frequency.

^bWeighted percentage.

Table 2 shows the percentage of food intake among children 6–23 mo of age according to food groups consumed in the previous 24 h. About 70% of children were fed grains, roots and tubers, while 26% of children were fed legumes and nuts. Few children (8%) were fed dairy products, while 33% of children consumed flesh foods. Furthermore, nearly 12% of children consumed eggs, 48% consumed vitamin A-rich fruits and vegetables and about 68% consumed other fruits and vegetables. The distribution of participants’ characteristics according to the four outcomes is shown in Table 3.

Table 2.

Percentage of food intake among children 6–23 mo of age according to food groups, MDHS 2015–2016 (n=4732)

Characteristics	Numbera	Percentageb	95% CI
Grains, roots and tubers
Not received	1455	30.75	28.91 to 32.66
Received	3277	69.25	67.34 to 71.09
Legumes and nuts
Not received	3497	73.90	71.79 to 75.89
Received	1235	26.10	24.10 to 28.21
Dairy products
Not received	4332	91.54	90.25 to 92.67
Received	400	8.46	7.33 to 9.75
Fresh
Not received	3254	68.76	66.92 to 70.55
Received	1478	31.24	29.45 to 33.08
Eggs
Not received	4168	88.07	86.75 to 89.27
Received	564	11.93	10.73 to 13.24
Vitamin A-rich fruits and vegetables
Not received	2481	52.43	49.98 to 54.87
Received	2251	47.57	45.13 to 50.02
Other fruits and vegetables
Not received	1522	32.15	30.26 to 34.10
Received	3211	67.85	65.89 to 69.74

^aWeighted frequency.

^bWeighted percentage.

Table 3.

Distribution of study participants’ characteristics according to childhood complementary feeding practices, MDHS 2015–2016

Characteristic	Introduction to solid, semi-solid or soft foods			Minimum meal frequency			Minimum dietary diversity			Minimum acceptable diet
	No, n (%)	Yes, n (%)	χ2 p-valuea	No, n (%)	Yes, n (%)	χ2 p-value	No, n (%)	Yes, n (%)	χ2 p-value	No, n (%)	Yes, n (%)	χ2 p-value
Sex of the child			0.660			0.920			0.243			0.116
Male	58 (49.4)	342 (52.1)		1687 (50.7)	713 (50.6)		1822 (51.3)	578 (49.0)		2219 (51.2)	181 (45.8)
Female	59 (50.6)	315 (47.9)		1635 (49.2)	697 (49.4)		1731 (48.7)	602 (51.0)		2118 (48.8)	214 (54.2)
Age of the child (mo)						<0.001			<0.0001			0.219
6–11	—b			1043 (31.4)	604 (42.9)		1368 (38.5)	280 (23.7)		1521 (35.1)	127 (32.2)
12–17				1153 (34.7)	448 (31.8)		1164 (32.8)	437 (37.0)		1476 (34.0)	125 (31.6)
18–23				1126 (33.9)	358 (25.4)		1021 (28.7)	463 (39.3)		1340 (30.9)	143 (36.3)
Perceived birth size			0.479			0.333			0.116			0.0.21
Very small	13 (11.7)	103 (15.6)		518 (15.6)	198 (14.0)		559 (15.7)	157 (13.3)		675 (15.6)	41 (10.3)
Average	68 (57.6)	335 (51.1)		1657 (49.9)	738 (52.4)		1809 (50.9)	586 (49.7)		2170 (50.0)	225 (56.6)
Very large	36 (30.7)	219 (33.3)		1147 (34.5)	474 (33.6)		1185 (33.4)	437 (37.0)		1492 (34.4)	130 (32.9)
Maternal age (y)			0.897			0.063			0.022			0.346
15–24	58 (49.8)	330 (47.0)		1417 (42.6)	663 (47.0)		1590 (44.7)	490 (41.5)		1904 (43.9)	175 (44.2)
25–34	44 (37.1)	251 (38.5)		1357 (40.9)	543 (38.5)		1376 (38.7)	525 (44.5)		1731 (39.9)	169 (42.8)
35–49	15 (13.1)	76 (14.5)		548 (16.5)	204 (14.5)		587 (16.6)	165 (14.0)		701 (16.2)	52 (13.0)
Maternal education level			0.069			0.049			<0.0001			<0.0001
No formal education	21 (18.3)	71 (11.0)		400 (12.0)	154 (10.9)		485 (13.6)	70 (59.1)		530 (12.2)	24 (6.1)
Primary	76 (64.4)	416 (63.3)		2263 (68.1)	918 (65.1)		2438 (68.6)	742 (62.9)		2937 (67.7)	244 (61.6)
Secondary and above	20 (17.3)	170 (25.7)		659 (19.9)	338 (24.0)		630 (17.7)	368 (31.2)		870 (20.1)	128 (32.4)
Maternal occupation			<0.0001			0.606			<0.0001			0.098
No employment	61 (52.0)	201 (30.6)		1009 (30.4)	430 (30.5)		1134 (31.9)	305 (25.8)		1317 (30.4)	122 (30.8)
Agriculture	32 (27.1)	293 (44.7)		1467 (44.2)	644 (44.7)		1666 (46.9)	446 (37.8)		1957 (45.1)	155 (39.1)
Non-agriculture	24 (20.9)	163 (24.7)		846 (25.5)	336 (28.3)		753 (21.2)	429 (36.4)		1063 (24.5)	207 (30.1)
Religion			0.125			0.612			<0.0001			0.048
Catholic	22 (19.2)	145 (22.0)		588 (17.7)	263 (18.7)		622 (17.5)	229 (19.4)		768 (17.7)	83 (21.0)
Protestant	14 (12.0)	131 (20.0)		680 (20.5)	301 (44.7)		679 (19.1)	302 (25.6)		883 (20.3)	99 (24.9)
Muslim or other	81 (68.8)	381 (58.0)		2054 (61.8)	846 (28.3)		2252 (63.4)	649 (55.0)		2687 (62.0)	2901 (54.1)
Exposure to mass media			0.231			<0.001			<0.0001			<0.0001
No	84 (71.3)	423 (64.4)		2288 (68.9)	863 (61.2)		2518 (70.9)	633 (53.7)		2977 (68.6)	175 (44.2)
Yes	33 (28.7)	234 (35.6)		1034 (31.1)	547 (38.8)		1035 (29.1)	547 (46.3)		1360 (31.4)	220 (55.8)
Marital status			0.260			0.266			0.211			0.045
Unmarried	27 (23.1)	116 (17.7)		567 (17.1)	219 (15.5)		607 (17.1)	178 (15.1)		735 (16.9)	51 (12.8)
Married	90 (76.9)	541 (82.3)		2755 (82.9)	1191 (84.5)		2946 (82.9)	1002 (84.9)		3602 (83.1)	345 (87.2)
Ethnicity			0.241			0.768			0.002			0.427
Chewa	38 (32.3)	242 (36.9)		1188 (35.8)	474 (33.6)		1230 (34.6)	433 (36.7)		1534 (35.4)	128 (32.4)
Tumbuka	10 (9.0)	47 (7.1)		281 (8.4)	124 (8.8)		309 (8.7)	96 (8.1)		367 (8.5)	38 (9.4)
Lomwe	17 (14.7)	131 (19.9)		564 (17.0)	239 (16.9)		608 (17.1)	195 (16.6)		738 (17.0)	66 (16.6)
Yao	30 (25.3)	97 (14.8)		530 (11.2)	238 (50.4)		632 (17.8)	136 (11.5)		715 (16.5)	53 (11.5)
Ngoni	7 (6.2)	63 (9.6)		375 (11.3)	151 (10.7)		372 (10.5)	154 (13.0)		474 (10.9)	51 (13.4)
Other	15 (15.2)	77 (11.7)		384 (11.6)	184 (13.0)		402 (11.3)	166 (14.1)		509 (11.8)	59 (14.9)
Parity			0.569			0.213			<0.0001	1534 (35.4)	128 (32.4)	0.014
1	26 (22.5)	193 (29.4)		858 (25.8)	406 (28.8)		901 (25.4)	364 (30.8)		1136 (26.2)	128 (32.3)
2	30 (25.8)	163 (24.9)		666 (20.1)	291 (20.6)		702 (19.8)	255 (21.6)		863 (19.9)	94 (23.8)
3	19 (15.8)	83 (12.6)		538 (16.2)	228 (16.2)		561 (15.8)	205 (17.4)		710 (16.4)	56 (14.2)
≥4	42 (35.7)	218 (33.1)		1260 (37.9)	485 (34.4)		1389 (39.1)	356 (30.2)		1628 (37.5)	117 (29.7)
Household wealth			0.764			0.057			<0.0001			<0.0001
Poorest	32 (27.2)	159 (24.3)		898 (27.0)	310 (22.0)		1021 (28.7)	187 (15.9)		1161 (26.8)	48 (25.5)
Poorer	29 (24.6)	166 (25.2)		760 (22.9)	329 (23.3)		861 (24.2)	229 (19.4)		1005 (23.2)	84 (23.0)
Middle	16 (13.9)	113 (17.3)		630 (19.0)	291 (20.6)		707 (19.9)	214 (18.2)		850 (19.6)	71 (19.5)
Richer	23 (20.1)	100 (15.3)		548 (16.5)	238 (50.2)		568 (16.0)	218 (18.4)		706 (16.3)	80 (21.1)
Richest	17 (14.2)	119 (17.9)		486 (14.6)	242 (51.3)		366 (11.2)	332 (28.2)		616 (14.2)	112 (24.4)
Household size			0.733			0.501			0.029			0.188
<5	45 (38.5)	265 (40.4)		1267 (38.2)	557 (39.5)		1328 (37.4)	496 (42.1)		1656 (31.2)	168 (42.4)
≥5	72 (61.5)	392 (59.6)		2055 (61.9)	853 (60.5)		2225 (62.6)	683 (57.9)		2681 (61.8)	228 (57.6)
Sex of the head of the household			0.168			0.128			0.001			0.037
Male	79 (67.3)	492 (74.8)		2477 (74.7)	1087 (77.1)		2630 (70.0)	934 (79.2)		3245 (74.8)	319 (80.7)
Female	38 (32.7)	165 (25.3)		845 (25.4)	323 (22.9)		922 (30.0)	246 (20.8)		1092 (25.2)	76 (19.3)
Number of children younger than 5 y in household			0.301			0.773			0.009			0.175
<1	48 (41.6)	256 (39.0)		1554 (46.8)	675 (47.8)		1613 (45.4)	616 (47.8)		2021 (46.6)	208 (52.5)
2	52 (51.7)	338 (51.4)		1509 (45.4)	620 (44.0)		1654 (46.6)	475 (40.3)		1974 (45.5)	155 (39.2)
≥3	17 (14.1)	63 (9.6)		259 (7.8)	115 (8.2)		286 (8.0)	89 (7.9)		342 (7.9)	33 (8.3)
Vitamin A supplementation in the last 6 mo			0.838			0.423			0.001			0.571
No	54 (46.7)	317 (48.1)		1131 (34.0)	503 (35.7)		1283 (36.1)	351 (29.7)		1503 (34.7)	130 (33.0)
Yes	63 (53.3)	340 (51.9)		2191 (66.0)	907 (64.4)		2270 (63.9)	829 (70.1)		2834 (65.3)	265 (67.0)
Place of delivery			0.879			0.837			0.043			0.152
Non-health facility	6 (5.5)	39 (5.9)		226 (6.8)	99 (7.0)		265 (7.5)	61 (5.2)		306 (7.1)	19 (4.8)
Health facility	111 (94.6)	618 (94.1)		3096 (93.2)	1311 (93.0)		3288 (92.6)	1119 (94.9)		4030 (92.9)	377 (95.2)
Number of antenatal visits			0.851			0.488			0.103			0.604
Inadequate	61 (52.1)	334 (50.9)		1691 (50.9)	698 (49.5)		1824 (51.4)	565 (47.9)		2196 (50.6)	193 (48.9)
Adequate	56 (47.9)	323 (49.1)		1631 (49.1)	712 (50.5)		1729 (48.6)	615 (52.1)		2141 (49.4)	202 (51.1)
Postnatal check within 2 mo			0.147			0.476			<0.0001			0.019
No	75 (64.4)	363 (55.2)		1862 (56.0)	769 (54.6)		2054 (56.0)	577 (54.6)		2440 (56.3)	190 (48.2)
Yes	42 (35.6)	294 (44.8)		1460 (44.0)	641 (45.5)		1499 (44.0)	603 (45.5)		1896 (43.7)	204 (51.8)
Place of residence			0.032			0.474			<0.0001			<0.0001
Urban	9 (7.4)	110 (16.7)		437 (13.1)	202 (14.4)		365 (10.3)	274 (23.2)		551 (12.7)	88 (22.2)
Rural	108 (92.6)	547 (83.3)		2885 (86.9)	1208 (85.6)		3188 (89.7)	906 (76.8)		3785 (87.3)	309 (77.8)
Geographic region			0.007			0.290			<0.0001			0.460
Northern	20 (17.6)	68 (10.3)		390 (11.7)	165 (11.7)		425 (12.0)	130 (11.0)		504 (11.6)	51 (13.0)
Central	32 (27.3)	302 (45.9)		1464 (44.1)	580 (41.1)		1452 (40.9)	592 (50.2)		1864 (43.0)	179 (45.4)
Southern	65 (55.1)	287 (43.8)		1468 (44.2)	665 (47.2)		1676 (47.1)	458 (38.8)		1969 (45.4)	165 (41.6)

All n (%) were weighted.

^aχ² P, Pearson’s χ² p-value.

^bSample for introduction of solid, semi-solid or soft foods (6–8 mo).

Factors associated with complementary feeding practices among children 6–59 mo of age

Introduction of solid, semi-solid or soft food

Table 4 displays the results of multivariable logistic regression. Children whose mothers had secondary and post-secondary education had increased odds of timely introduction to solid food, semi-solid food or soft food (aOR 2.46 [95% CI 1.12 to 5.39]) compared with those whose mothers had no formal education. Further, children whose mothers had agricultural employment (aOR 2.65 [95% CI 1.59 to 4.43]) and resided in the central region (aOR 2.84 [95% CI 1.17 to 6.90]) had increased odds of timely introduction to solid food, semi-solid food or soft food compared with children whose mothers were not employed and resided in the northern region.

Table 4.

Factors associated with childhood complementary feeding practices, MDHS 2015–2016

	Introduction to solid, semi-solid or soft foods		Minimum meal frequency		Minimum dietary diversity		Minimum acceptable diet
	aOR (95% CI)	p-value	aOR (95% CI)	p-value	aOR (95% CI)	p-value	aOR (95% CI)	p-value
Sex of the child
Male	—		—		1.00		1.00
Female	—		—		1.01 (0.85 to 1.19)	0.937	1.19 (0.91 to 1.56)	0.200
Age of the child (mo)
6–11	—		1.00		1.00		1.00
12–17	—		0.67 (0.55 to 0.80)	<0.001	1.90 (1.54 to 2.40)	<0.001	0.99 (0.71 to 1.40)	0.995
18–23	—		0.56 (0.46 to 0.68)	<0.001	2.30 (1.82 to 2.91)	<0.001	1.36 (0.99 to 1.87)	0.061
Perceived birth size
Very small	—		—		1.00		1.00
Average	—		—		1.10 (0.83 to 1.44)	0.508	1.64 (1.11 to 2.42)	0.013
Very large	—		—		1.20 (0.90 to 1.60)	0.213	1.33 (0.90 to 2.00)	0.155
Maternal age (y)
15–24	—		1.00		1.00
25–34	—		0.81 (0.64 to 1.02)	0.079	1.53 (1.17 to 1.20)	0.002	—
35–49	—		0.81 (0.58 to 1.13)	0.224	1.56 (1.08 to 2.25)	0.017	—
Maternal education level
No formal education	1.00		1.00		1.00		1.00
Primary	1.62 (0.87 to 3.00)	0.125	0.95 (0.74 to 1.23)	0.719	1.65 (1.20 to 2.27)	0.002	1.30 (0.77 to 2.19)	0.318
Secondary and above	2.46 (1.12 to 5.39)	0.025	1.14 (0.84 to 1.56)	0.404	1.67 (1.14 to 2.45)	0.008	1.40 (0.76 to 2.57)	0.283
Maternal occupation
No employment	1.00		—		1.00		1.00
Agriculture	2.65 (1.59 to 4.43)	<0.001	—		1.16 (0.93 to 1.45)	0.187	1.04 (0.75 to 1.44)	0.821
Non-agriculture	1.45 (0.86 to 2.46)	0.165	—		1.78 (1.39 to 2.26)	<0.001	1.08 (0.76 to 1.54)	0.651
Religion
Catholic		1.00		—		1.00		1.00
Protestant	1.38 (0.68 to 2.78)	0.372	—		1.12 (0.85 to 1.47)	0.437	0.92 (0.63 to 1.34)	0.652
Muslim or other	0.99 (0.58 to 1.68)	0.972	—		0.99 (0.79 to 1.24)	0.915	0.82 (0.59 to 1.14)	0.239
Exposure to mass media
No	1.00		1.00		1.00		1.00
Yes	1.28 (0.79 to 2.05)	0.317	1.31 (1.11 to 1.56)	0.002	1.37 (1.14 to 1.65)	0.001	2.08 (1.61 to 2.69)	<0.001
Marital status
Unmarried	—		—		1.00		1.00
Married	—		—		0.92 (0.69 to 1.21)	0.529	1.19 (0.81 to 1.75)	0.375
Ethnicity
Chewa	1.00				1.00
Tumbuka	1.53 (0.55 to 4.28)	0.415	—		0.91 (0.60 to 1.37)	0.642	—
Lomwe	1.49 (0.68 to 3.25)	0.321	—		1.10 (0.77 to 1.56)	0.607	—
Yao	1.13 (0.54 to 2.37)	0.754	—		0.78 (0.55 to 1.12)	0.175	—
Ngoni	1.18 (0.51 to 2.72)	0.693	—		1.10 (0.81–1.49)	0.538	—
Other	1.66 (0.75 to 3.70)	0.211	—		1.78 (0.84 to 1.64)	0.336	—
Parity		NS
1	—		1.00		1.00		1.00
2	—		0.99 (0.76 to 1.26)	0.934	0.80 (0.61 to 1.05)	0.105	1.01 (0.67 to 1.52)	0.954
3	—		1.10 (0.82 to 1.47)	0.520	0.67 (0.47 to 0.94)	0.021	0.74 (0.48 to 1.12)	0.158
≥4	—		1.07 (0.77 to 1.47)	0.691	0.52 (0.36 to 0.75)	<0.001	0.76 (0.50 to 1.14)	0.179
Household wealth
Poorest	—		1.00		1.00		1.00
Poorer	—		1.13 (0.89 to 1.43)	0.311	1.28 (0.97 to 1.69)	0.081	1.65 (1.05 to 2.58)	0.030
Middle	—		1.25 (0.97 to 1.61)	0.080	1.36 (1.03 to 1.80)	0.031	1.59 (1.05 to 2.40)	0.029
Richer	—		1.06 (0.82 to 1.38)	0.647	1.68 (1.23 to 2.29)	0.001	1.89 (1.20 to 2.97)	0.006
Richest	—		1.15 (0.85 to 1.56)	0.357	2.64 (1.81 to 3.87)	<0.001	2.27 (1.37 to 3.78)	0.002
Household size
<5	—		—		1.00		1.00
≥5	—		—		0.86 (0.68 to 1.10)	0.224	0.97 (0.69 to 1.38)	0.884
Sex of head of household
Male	1.00		1.00		1.00		1.00
Female	0.80 (0.50 to 1.28)	0.356	0.91 (0.76 to 1.09)	0.288	0.82 (0.66 to 1.02)	0.079	0.93 (0.65 to 1.34)	0.652
Number of children younger than 5 y in the household
<1	—		—		1.00		1.00
2	—		—		1.02 (0.83 to 1.26)	0.846	0.95 (0.69 to 1.31)	0.750
≥3	—		—		1.29 (0.91 to 1.84)	0.157	1.20 (0.71 to 2.03)	0.494
Vitamin A supplementation in the last 6 mo
No	—		—		1.00		—
Yes	—		—		1.23 (1.03 to 1.48)	0.024	—
Place of delivery
Non-health facility	—		—		1.00		1.00
Health facility	—		—		1.16 (0.80 to 1.66)	0.434	1.04 (0.61 to 1.78)	0.887
Number of antenatal visits
Inadequate	—		—		1.00		—
Adequate	—		—		1.01 (0.85 to 1.19)	0.944	—
Postnatal check within 2 mo
No	1.00		—		1.00		1.00
Yes	1.07 (0.68 –1.68)	0.777	—		1.24 (1.01 to 1.51)	0.040	1.21 (0.93 to 1.59)	0.154
Place of residence
Urban	1.00		—		1.00		1.00
Rural	0.77 (0.40 to 1.45)	0.414	—		0.94 (0.68 to 1.31)	0.722	1.01 (0.65 to 1.55)	0.979
Geographic region
Northern	1.00		—		1.00		—
Central	2.84 (1.17 to 6.90)	0.021	—		1.59 (1.09 to 2.32)	0.017	—
Southern	1.70 (0.79 to 3.65)	0.176	—		1.12 (0.78 to 1.60)	0.531	—

The Bold texts indicate a statistically significant association at a p-value less than 0.05.

Minimum meal frequency

Children aged 12–17 and 18–23 mo were 33% (aOR 0.67 [95% CI 0.55 to 0.80]) and 44% (aOR 0.56 [95% 0.46 to 0.68]) less likely, respectively, to achieve minimum meal frequency compared with children 6–11 mo of age. However, higher odds of achieving minimum meal frequency were observed in children whose mothers had exposure to mass media (aOR 1.31 [95% CI 1.11 to 1.56]) compared with those who were not exposed to mass media.

Minimum dietary diversity

Children aged 12–17 mo and 18–23 mo were 1.90 (95% CI 1.54 to 2.40) and 2.30 (95% CI 1.82 to 2.91) times, respectively, more likely to receive minimum dietary diversity compared with children 6–11 mo of age. Similarly, higher odds of receiving adequate dietary diversity were observed among children whose mothers had primary education (aOR 1.65 [95% CI 1.20 to 2.27]), secondary and post-secondary education (aOR 1.67 [95% CI 1.14 to 2.45]) and were in non-agricultural employment (aOR 1.78 [95% CI 1.39 to 2.26]) compared with children whose mothers had no formal education and were unemployed.

Furthermore, higher odds of receiving minimum dietary diversity were observed among children whose mothers were 25–34 y of age (aOR 1.53 [95% CI 1.17 to 1.20]) and 35–49 y of age (aOR 1.56 [95% CI 1.08 to 2.25]) compared with children whose mothers were younger than 25 y old. Compared with children whose mothers were not exposed to mass media, children whose mothers were exposed to mass media were more likely to achieve minimum dietary diversity (aOR 1.37 [95% CI 1.14 to 1.65]). Children who resided in the richest households (aOR 1.37 [95% CI 1.14 to 1.65]) received vitamin A supplements in the last 6 mo (aOR 1.23 [95% CI 1.03 to 1.48]), had a postnatal check within 2 mo (aOR 1.24 [95% CI 1.01 to 1.51]) and resided in the central region (aOR 1.59 [95% CI 1.09 to 2.32]) were more likely to receive the recommended minimum dietary diversity compared with their respective counterparts. Reduced odds of achieving minimum dietary diversity were observed in children whose mothers’ parity was 3 (aOR 0.67 [95% CI 0.47 to 0.94]) and ≥4 (aOR 0.52 [95% CI 0.36 to 0.75]) compared with children whose mother had only one child.

Minimum acceptable diet

Lastly, children who had average birth sizes (aOR 1.64 [95% CI 1.11 to 2.42]), who resided in the richest households (aOR 2.27 [95% CI 1.37 to 3.78]) and whose mothers had satisfactory media exposure (aOR 2.08 [95% CI 1.61 to 2.69]) had higher odds of achieving a minimum acceptable diet as compared with their respective counterparts.

Discussion

To our knowledge, this study is one of the first population-based studies to investigate the factors associated with complementary feeding practice (introduction of solid, semi-solid and soft foods; minimum dietary diversity; minimum meal frequency; minimum acceptable diet) in Malawi. The results demonstrate that an array of sociodemographic and health factors are associated with complementary feeding in Malawi. Specifically, children from women 25–49 y of age who had primary and post-primary education had increased odds of achieving minimum dietary diversity. Similarly, children whose mothers were exposed to mass media, resided in the richest households and used postnatal care services had increased odds of meeting minimum dietary diversity. Surprisingly, children older than 12 mo were less likely to meet the minimum meal frequency.

The proportion of children 6–23 mo of age who received the recommended minimum meal frequency, minimum dietary diversity, minimum acceptable diet and introduction of solid, semi-solid or soft food was 29.8, 24.9, 8.4 and 84.8%, respectively. Low prevalence was observed for minimum meal frequency, minimum dietary diversity and minimum acceptable dietary. These results suggest that complementary feeding practices remain a challenge in Malawi. Hence studies to understand factors affecting complementary feeding practices are warranted and this article contributes to that knowledge.

The present study found that children born to older women with secondary and post-secondary education, women working in the agriculture sector and women from the central region had higher odds of timely introduction to solid, semi-solid or soft foods. The findings are consistent with previous studies.²⁶ The possible reason might be that education is directly linked to women’s autonomy, changes in traditional beliefs and women’s control over household resources.²⁷ Hence educated mothers are more likely to understand information provided through health and nutritional programmes regarding complementary feeding practices.²⁸ Children whose mothers were in agriculture had higher odds of timely introduction to solid, semi-solid or soft food, probably due to the availability and use of adequate food the women harvested from their farms. Malawi’s central region has the highest percentage of employed women, who are largely working in agriculture.¹⁵ Consequently, this may explain why children from the central region were more likely to have a timely introduction of solid, semi-solid or soft foods. Women who work in agriculture might have easy access to different groups of foods and possibly this may have an influence on complementary feeding practices.

Consistent with studies conducted in Ethiopia,⁶ India,²⁹ Ghana,³⁰ Tanzania³¹ and Sri Lanka,³² the current study also found that children 12–23 mo of age had higher odds of meeting minimum dietary diversity, implying that the practice of minimum dietary diversity increases as the child’s age increases. Further, consistent with a study conducted in northwest Ethiopia,⁷ children whose mothers were older than 24 y were more likely to achieve minimum dietary diversity. Older mothers are more likely to be experienced and have knowledge on how to raise their children as compared with younger women. As expected, children whose mothers had primary education or secondary and post-secondary education were more likely to achieve minimum dietary diversity, owing to the fact that educated women are more likely to have access to quality health services and messages, which they may more easily comprehend and apply.³³ Children whose mothers were working in the non-agricultural sector were more likely to achieve minimum dietary diversity as compared with children born to non-working women. The possible explanation is that women with formal employment may be in a better socio-economic position to achieve food security, which in turn may influence their complementary feeding practices.²⁶ Therefore it is not surprising that even children from the middle, richer and richest households were more likely to achieve minimum dietary diversity. Children from the richest households might be fed with recommended diversified foods, as their families are more likely to achieve food security as compared with children from poor households.⁹ In addition, higher parity was associated with reduced odds of achieving minimum dietary diversity. The underlying mechanism could be that women with fewer children may commit themselves to taking better care of their children and feeding them more appropriately with respect to complementary feeding practice as compared with those with more children.³⁴ In line with prior studies,³⁵ we also found that children who had received vitamin A supplements in the last 6 mo and whose mothers had postnatal care within 2 mo had increased odds of achieving minimum dietary diversity. These factors might underlie women’s health-seeking behaviour and the value they place on their health as well as that of their children.³⁶ In addition, postnatal care has been proven to be a good platform to improve mothers’ knowledge and change unfavourable attitude towards the implementation of appropriate feeding practices.³⁷ Prior studies have reported that these positive effects mainly operate through counselling and behavioural changes and communication interventions provided by health professionals.³⁸

Surprisingly, children ≥12 mo of age were less likely to achieve minimum meal frequency. This is inconsistent with studies conducted in Ethiopia,^34,39 Ghana³⁰ and other countries in Asia.^32,40 The reasons why older children are less like to meet the minimum meal frequency are poorly understood. However, the possible explanation might be that when children grow, they are more likely to be given less attention compared with when they were younger. This may eventually result in failure to achieve minimum meal frequency as the children grow. However, there is a need for further research to understand this finding.

In terms of minimum acceptable diet, children with a perceived average birth size were more likely to achieve a minimum acceptable diet as compared with those perceived to have a very small birth size. Previous studies have indicated that birth sizes may reflect overall maternal health status as well as socio-economic status.⁴¹ Moreover, children from poorer, middle, richer and richest households were more likely to achieve minimum acceptable diet as compared with children from the poorest households. This finding emphasizes the role of socio-economic status on feeding practices. Households with high socio-economic status are more likely to be food secure, thus they can afford to provide the minimum acceptable diet to their children.^42,43

The finding of media exposure having an association with minimum meal frequency, minimum dietary diversity and minimum acceptable diet practices is consistent with previous studies.^6,39 The media is usually considered a credible source of health and nutrition information, hence its messages are more likely to be adopted.⁴⁴ Thus mothers who are exposed to mass media may acquire important information regarding complementary feeding practices.

Policy implication

The current study highlights the important factors influencing child complementary feeding practices in Malawi. Policymakers in nutritional projects should ensure that an integrated approach with other sectors, such as education, is adopted, as this study has revealed the significance of maternal education in child complementary feeding practices. Further, our results revealed that the use of healthcare services such as postnatal care influences complementary feeding practices. To effectively improve child complementary feeding practices, child nutrition programme planners should focus on younger women with low uptake of such healthcare services. Women from poorest households should be empowered economically and be given an education that would essentially help them to achieve food security. Lastly, mass media is essential in delivering important messages, thereby imparting necessary knowledge to mothers. Therefore programmes aimed at improving feeding practices should consider utilizing mass media. However, simultaneous efforts should be made to make mass media available to underprivileged households in order to increase exposure.

Strengths and limitations of the study

The present study was able to identify a wide range of determinants on the four components of complementary feeding practices in Malawi and provides important insights by which the most appropriate interventions can be designed. The results can be generalized in the Malawian context due to the use of a nationally representative sample. However, the results should be considered in light of several limitations. First, the cross-sectional nature of the study design limited our ability to draw causal inferences between the covariates and the outcome variables. Second, the use of secondary data limited us from including other variables that could explain our outcome variables. Third, as the study considered only a 24-h recall method, it might not accurately reflect participants' past feeding habits. Lastly, there might be a recall bias, and, being a self-reported study, it might not provide exact figures for minimum dietary diversity practices.

Conclusions

In the context of Malawi, our findings clearly demonstrate that a large proportion of children younger than 2 y of age are missing out on the three indicators of complementary feeding practices: minimum dietary diversity, minimum meal frequency and minimum acceptable diet. However, these findings show that most of the factors included in this study provided a better prediction of minimum dietary diversity than the other three outcomes. Hence future studies should explore more of the other three indicators. Overall, the age of the child, mother’s age, mother’s education, mother’s occupation, media exposure, parity, household wealth, vitamin A supplementation in the last 6 mo, postnatal care within 2 mo and region were associated with the components of complementary feeding practices. Identifying further determinants will help policymakers focus on improved and effective behaviour changes and communication, as well as improved child complementary feeding practices through micronutrient supplementation, food security, healthcare utilization and improved socio-economic status.

Authors’ contributions: PAMN, TGM and ON contributed to the conception and design of the study. PAMN acquired data. PAMN and ON conducted analysis and interpreted results. PAMN drafted the manuscript. ON and TGM critically revised the manuscript for important intellectual content. PAMN, ON and TGM reviewed and approved the final version of the manuscript.

Acknowledgements: Sincere thanks go to the National Statistical Office (NSO) and the Community Health Sciences Unit (CHSU) of Malawi for data collection. We also thank the Monitoring and Evaluation to Assess and Use Results Demographic and Health Surveys for providing us with the population-based dataset through their archives, which can be downloaded from http://dhsprogram.com/data/available-datasets.cfm.

Funding: This research received no funding from any agency in the public, commercial or not-for-profit sectors. The funding for the 2015–2016 MDHS was provided by the government of Malawi, the United States Agency for International Development, the United Nations Children’s Fund, the Malawi National AIDS Commission, the United Nations Population Fund, UN WOMEN, Irish Aid and the World Bank.

Competing interests: None declared.

Ethical approval: Not required.