1. Introduction
Undernutrition is one of the major reasons for child deaths worldwide. According to the results of the Global Burden of Disease study 2013, about 200 deaths per 100,000 in children under five years of age are attributable to childhood undernutrition, which is 21% of total deaths in this age group [
1]. It continues to remain one of the alarming public health concerns in South Asian countries, especially Pakistan, Bangladesh and India [
2]. In Pakistan, 428 deaths per 100,000 children aged below five years are attributable to childhood undernutrition [
1], and 96% of the children under the age of two years do not receive adequate diet recommended by the World Health Organization (WHO) [
3]. Adequate diet for children under the age of two years includes breastfeeding, intake of solid, semisolid and soft foods, dietary diversity, such as vegetables, fruit, dairy products, protein-rich foods and grains, and consumption of three to four meals a day [
4], The lack of appropriate diet affects child health and results in increasing child mortality [
5]. Poor child nutrition is strongly associated with high functional impairment, slow intellectual development and low work capability [
6,
7].
Child health is considered an important indicator of socioeconomic development [
8]. This is the reason that Millennium Development Goal 4 aimed to reduce under-five mortality by two-thirds by 2015. Nevertheless, no significant improvement has been witnessed in Pakistan [
9]. According to the Child Nutrition Report, the prevalence of stunting, wasting and underweight in children under the age of five years in Pakistan was 42%, 14% and 31%, respectively, in 2013 [
10]. Studies reveal that stunting in Pakistan has increased from 31% in 2001 to 43.7% in 2011 [
11]. Similar trends were observed for wasting, which increased in Pakistan from 11.6% in 2001 to 15.1% in 2011 [
11]. The WHO (2012–2013) estimates that the incidence of stunting, wasting and underweight in Pakistan for children under the age of two years was 37.6%, 16.6% and 32.4%, respectively [
12].
Consideration of the nutritional status is an important factor to improve child health [
13]. There are many factors which can influence the nutritional status of children. Studies reveal that these factors range from maternal literacy, household income and utilization of health care services [
14,
15,
16,
17,
18,
19,
20] to economic equality, effective decision-making and migration [
14,
21,
22]. Among other factors, micronutrient consumption has been reported to be one of the major factors to improve child health [
23]. Iron deficiency anemia and vitamin A deficiency are key nutritional problems in developing countries like Pakistan, resulting in stunting, wasting and underweight [
24].
There is a need for carrying out studies that could highlight factors affecting child health so that effective policy could be devised to help Pakistan achieve its Sustainable Development Goal 3. The prevalent literature on child undernutrition in Pakistan is based on either specific health facilities, small sample size targeting specific regions and few health factors [
25,
26,
27,
28,
29]. These studies considered limited sociodemographic determinants of child undernutrition, such as gender, maternal education, father’s income, family size and access to information. The findings of these studies suggested that these characteristics were associated with child undernutrition, however, the only anthropometric indices used to measure undernutrition in most of these studies was stunting (wasting and underweight children were not included). Furthermore, the nutritional and health-related factors were not included in these studies as determinants of child undernutrition. Keeping in mind the limitations of these studies, the purpose of this study was to identify the sociodemographic, nutritional and health-related factors that are associated with undernutrition in children under the age of two years in Pakistan using a countrywide representative sample.
2. Materials and Methods
We conducted a secondary data analysis based on the Pakistan Demographic and Health Survey (PDHS) 2012–2013, a publicly accessible dataset generated by ICF International, Fairfax, VA, USA. The PDHS 2012–2013 was a third national survey as part of the Demographic and Health Survey initiative funded by US Agency for International Development [
30,
31]. The PDHS is one of the very few sources that deal with sociodemographic and health data as a nationally representative sample. The 2012–2013 survey provided detailed information on issues related to maternal and child health.
The PDHS entails countrywide representative data collected through stratified cluster sampling in two stages on 14,000 households in 498 areas, of which 248 were urban areas involving 6944 households and 250 were rural areas totaling 7056 households. Of the 14,000 households sampled, 14,569 women of the reproductive ages (15–49) were selected for interviews and 13,558 women were interviewed. The remaining 1011 respondents could not be interviewed due to their non-availability despite various visits from the PDHS staff. The data were limited to four major provinces (Punjab, Sindh, Khyber Pakhtunkhawa and Baluchistan), Gilgit Baltistan and Islamabad. However, the exacerbated security concerns excluded Azad Jammu and Kashmir, Federally Administered Tribal Areas, few areas in Baluchistan and other areas controlled by the army from the sample. The complete details of the PDHS including sampling, design, data collection, management and administration are described elsewhere [
30].
The PDHS comprised data on numerous components, such as domestic violence, fertility, reproductive health and the empowerment of women. Our study focused on components dealing with sociodemographic characteristics, child health and the nutritional status of children. A total of 5300 children under the age of five years were initially selected to study the factors that affect their health. The reason for this selection was a detailed literature review that had studied children in the same age group [
14,
32,
33,
34]. The data was later limited to children under the age of two years, as the complete sociodemographic information, data on nutritional intake and anthropometric indicators for measuring child health were only available for this age group. Furthermore, a review of literature suggested that this age group is more vulnerable, as most of the irremediable harm due to undernutrition occurs in the first two years of life [
35]. All the children in this age group were selected, resulting in a sample size of 984 children.
The independent variables were selected by conducting a literature review [
5,
15,
19,
20,
21,
36,
37,
38,
39,
40,
41,
42,
43,
44,
45,
46] and looking for the appropriate variables in the PDHS 2012–2013. The variables selected included sociodemographic characteristics such as parental education level, wealth quintile, region and type of residence (rural vs. urban). The characteristics of mothers included their current age, age at the time of first childbirth, body mass index (BMI), consanguineous marriage (marriages between cousins), mother’s access to information, intake of iron supplements during pregnancy, source of drinking water and type of toilet facility. The characteristics of children included their sex, age, birth order, size at birth, prelacteal feeding, immunization and status of being breastfed (it was categorized as “yes” if the child was being breastfed). Additionally, consumption of micronutrients by children was also included as two independent variables (iron and vitamin A). Existing categories in the PDHS for all independent variables were used as such, except for a few variables, such as the mother’s age at the time of first birth, her access to information, access to water and sanitation facilities, and BMI, in addition to the consumption of micronutrients by children. The BMI of the mothers was calculated using their weight and height. The variable was recoded in four categories (underweight, normal, overweight, and obese) based on the WHO standard. The variable “source of drinking water” had more than fifteen categories and it was recoded in three categories (poor, medium and good). The variable “iron consumption by children” was computed using five food consumption variables in the PDHS. The responses were in two categories “no” and “yes”, which were computed to constitute iron consumption. The responses were recoded as “no consumption of iron” if the child had no intake of any of these diets and “consumption of iron” if one or more of the diets stated above were given. The variable “consumption of vitamin A” was computed and recoded using four variables (four types of vitamin A-rich foods/supplements) in the PDHS. The responses to these questions as “no” and “yes” were computed and recoded in three categories: “no consumption” if the child had no intake of vitamin A, “low consumption” if up to two foods/supplements were being consumed and “high consumption” if three vitamin A-rich foods/supplements were being eaten.
Nutritional status was measured using anthropometric indices, such as stunting (if children’s height is less, relative to their age), wasting (if children weigh less, relative to their height) and underweight (if they weigh less, relative to their age), which were identified by a review of literature [
5,
33,
34,
38,
40]. Stunting (height/age), wasting (weight/height) and underweight (weight/age) were defined considering the standard of WHO. As per WHO child growth standards, a child was considered stunted if their height for age Z-score was below minus two standard deviations (−2SD) from the median of the WHO reference population [
28]. Wasting was categorized if a child’s weight for height Z-score was below −2SD from the median reference population, and a child was considered “underweight” if their weight for age was below −2SD from the reference population median. The variables of stunting, wasting and underweight were recoded as “stunted” and “not stunted”, “wasted” and “not wasted”, and “underweight” and “not underweight”, respectively.
The data were administered and analyzed using SPSS version 22 (IBM, Armonk, NY, USA). Descriptive statistics of independent and dependent variables were presented as frequency distributions and percentages. Simple binary logistic regression analysis was performed to show the association between sociodemographic and other independent factors that affect stunting, wasting and underweight in children. Additionally, multivariable logistic regression analyses were performed using variables that were significant at the 0.05 level in binary analysis. The variables regarding mother’s age, wealth quintile and education were kept constant and adjusted, as previous studies in Pakistan have considered these factors as major determinants of child health [
25,
26,
27,
28,
29,
30]. The purpose for holding these variables constant was to see the effect of other important factors on undernutrition after controlling for these variables. Multicollinearity was tested between highly correlated variables before entering them in the multivariate analysis. Multicollinearity was seen between the mother’s current age and mother’s age at the time of first birth, father’s education and mother’s education, and mother’s education and her access to information by using variance inflation factors. All these variables had a variance inflation factor of 1 and, hence, the mother’s age at the time of first birth, father’s education and mother’s access to information were considered in the multivariate analysis. All the variables that were significant at 0.05 levels in the bivariate analysis were then entered separately against these adjusted variables to determine their effect on dependent variables.
The findings of this study should be considered keeping in mind the following limitations. The analysis was not based on case control methods, therefore, drawing causal relationships between the variables is not possible. The data was limited to children under two years of age, giving a sample of 1360, which was further reduced to 984, due to no availability of responses in 376 cases. The data pertinent to consumption of micronutrients, size at birth and breastfeeding was based on self-reports of mothers and not confirmed by clinical tests/reports. Therefore, there is a possibility that certain biases, such as giving socially desirable responses in addition to inability to recall properly, could have affected the data. Additionally, it was difficult to calculate the consumption of micronutrients due to limited data.
4. Discussion
Child undernutrition is a major public health issues in Pakistan leading to a high infant mortality rate [
1]. Child undernutrition has become an important concern in South Asian and African countries despite many steps taken to reduce it [
47]. Studies point out that the level of stunting in the Indian states of Bihar, Uttar Pradesh, and Jharkhand is 49.4%, 50.6%, and 47.3%, respectively, whereas the incidence of stunting in African countries such as Mozambique, Liberia, and Sierra Leone is 43%, 42%, and 45%, respectively [
47]. The factors responsible for undernutrition in these countries, for example, poverty, dietary intake, low BMI, exclusive breastfeeding or no breastfeeding at all, teenage marriages, and education, are similar to those which have been found in this study [
47,
48]. The present study has highlighted many factors in multivariable analysis, such as size at birth, maternal BMI, consanguinity, mother’s intake of iron supplements during pregnancy, and the birth order of the child, that contribute significantly to stunting and underweight in children. However, the only significant factors found for wasting were the age of the child and current status of breastfeeding in children.
One of the findings of the study was that the likelihood of stunting, wasting and underweight increased with age. The reasons pointed out by various studies are numerous, ranging from improper diet administration and lack of supplementary feeding to exposure to hazardous environmental conditions [
2,
5]. Most children were currently breastfeeding which, alone, cannot meet their nutritional needs, particularly after six months of age [
49]. The lack of supplementary feeding can help explain why child undernutrition increased with age. Most children in various rural areas of Pakistan, due to poverty and no access to a nutrition rich diet, are given wheat bread which, alone, cannot satisfy their nutritional needs [
50].
One of the major though not significant findings in the multivariable analysis was that micronutrient deficiency in children, that is, those with a low consumption of iron and vitamin A, are at a higher risk of stunting and being underweight compared to those who have a diet rich in micronutrients [
2,
4,
31,
32,
34]. Literature on micronutrients highlight that iron supplementation in the antenatal stage can significantly reduce the likelihood of low birth weight [
51]. Furthermore, iron deficiency during pregnancy can enhance the likelihood of maternal and child mortality [
52]. Micronutrient deficiency in children can instigate a vicious circle of issues by weakening their immunity, which results in increased vulnerability to diseases, which are responsible for a further depletion of nutrients. Many studies highlight that vitamin A and iron deficiency are one of the major causes of avertable blindness, infections, anemia, low immunity and child mortality [
53,
54,
55,
56]. Micronutrient deficiency in children and mothers affects almost half of the global population, impeding the Millennium Development Goal of better health [
57]. Despite various studies pointing out the reversible and irreversible health damages that micronutrient deficiency can cause in children and mothers, no policy-making has yet been witnessed to target the potential sufferers. Pakistan’s Maternal and Child Health Policy (2005–2015) lacks the interventions required to target this potential source of child undernutrition.
A number of studies provide evidence that a lower level of health literacy and autonomy result in low nutritional intake [
58,
59,
60,
61,
62,
63], which may contribute significantly to stunting, wasting, and underweight in children. Consistent with these findings, our results show that the chances of stunting and underweight were greater in children if their mothers did not take iron supplements during pregnancy. Several studies have developed a causal link between maternal iron deficiency anemia, underweight children, and early childhood mortality [
64,
65,
66]. Evidence from studies suggests that iron intake by anemic women during pregnancy improved their hemoglobin levels [
67]. One of the reasons behind the low health awareness in pregnant women can be early marriages, as our study found that the mother’s age at the time of first birth contributes significantly to child health. The risk of stunting, wasting and underweight is greater in children who have teenage mothers (13–17 years) compared to those who have older mothers. Similarly, teenage mothers are more likely to be undernourished during pregnancy due to the depletion of nutrients and, subsequently, due to breastfeeding [
61,
62,
63,
68,
69,
70]. This argument gains strength from another finding of this study, which presents that the low BMI of mothers was a significant factor in the stunting and underweight of their children. Several studies indicate that mothers who are underweight are more likely to produce preterm births, which can result in stunting, incomplete growth and development, or death [
71,
72]. One of the major reasons behind mothers being underweight can be the low economic status [
62,
63,
68,
69,
70,
71,
72], as a finding of this study was that children belonging to poor families are more likely to be stunted, wasted, and underweight. Low economic status can contribute significantly to the poor nutritional status of mothers by restricting their access to nutrition-rich foods and that, in turn, will affect child health [
69,
70,
71,
72,
73]. Education and access to information can play a vital role in improving child health-related issues in Pakistan. Consistent with the evidence from literature [
37,
39,
40], the study found that the mother’s education and access to information is an important determinant in child health, as the chances of stunting, wasting, and underweight were greater in those children whose mothers had no access to information and no formal schooling. Lack of knowledge related to child and maternal health can result in overlooking the symptoms related to child health [
68]. The study identified another factor affecting child health which was related to the father’s education. Males are the major wage earners in patriarchal societies such as Pakistan. Therefore, having low education would mean jobs paying less which can contribute to household food shortages. Another significant finding of the study was the relationship between consanguinity, stunting, and underweight. The incidence of stunting and underweight was greater in children who were born of consanguineous marriages. Studies point out that such children are at a higher risk of postnatal mortality, exposure to multiple diseases and autosomal recessive disorders [
74].
The findings show that the incidence of stunting was higher in a few provinces compared to others. The Baluchistan and Sindh region had an alarming rate of stunting compared to other provinces. The reasons are multiple, ranging from the worsening law and order situation, natural disasters, poverty, insufficient health services, food shortages, and fewer awareness campaigns in rural areas to the lack of governmental will. This situation can also be attributed to the decentralization of the health care system in Pakistan. Decentralizing the responsibility of health was a good move, as various studies point out the perceived benefits of such a system. Nevertheless, it was poorly executed in Pakistan [
75,
76,
77]. The devolution has actually overstrained the provinces that were not structurally and strategically ready for such a change due to previous health care inequalities. Furthermore, even after delegation of power to provinces, the districts are not yet self-sufficient to actually implement the system due to issues such as patronage, corruption, deputing incompetent staff, rifts between federal and provincial bureaucracies, and political intervention. Despite all these constraints, an effort has been made by the provincial governments of Baluchistan and Sindh by enacting the Protection and Promotion of Breastfeeding and Child Nutrition Act. The study found that children who are not breastfeeding are more likely to suffer from wasting compared to those who are breastfeeding. In light of this finding, it can be argued that ensuring breastfeeding through legislative reforms was a good move, but lack of health interventions targeted at mothers can even increase wasting, as mothers who are involved in breastfeeding require nutritional support for themselves [
61,
69,
70]. International humanitarian organizations, such as the United Nations Children’s Fund and their sister organizations in Pakistan, are continuously mobilizing and providing resources to these undernourished provinces. However, their efforts are limited, as health issues are not prioritized by the government [
78].
This study has highlighted many important sociodemographic and nutritional factors that affect child undernutrition, especially in developing countries, such as Pakistan. Contrary to the previous studies [
25,
26,
27,
28,
29,
30], which either studied few anthropometric indices, had small sample sizes, or relatively fewer predictors of undernutrition, the present study has outlined many factors using a nationally representative sample that, if targeted properly, can reduce the odds of stunting, wasting, and underweight in children.