Abstract
Malnutrition is a primary cause of child morbidity and mortality. The effects of maternal depressive symptoms on children’s health, especially their nutritional status, have received less attention in developing countries but needs to be evaluated to understand the public health implications of maternal depression. The aim of the current study was to investigate the association between maternal depressive symptoms and children’s nutritional status using data from low socioeconomic community in Pakistan. Maternal depressive symptoms defined as Aga Khan University Anxiety and Depression Scale score of 20 or greater was assessed for mothers with children under two years of age. Logistic regression models estimated the association between maternal depressive symptoms and stunting and underweight. Of 325 mothers, 40% scored positive on the depressive scale. The prevalence of stunting and underweight in children under two years was 36.6% and 35.4%, respectively. Maternal depressive symptoms were significantly associated with children being stunted and underweight. Mothers with depressive symptoms were more than three times likely to have stunted (odds ratio (OR) 3.15, 95% confidence interval (CI) 1.91–5.18, p value < .001) and underweight (OR 3.26, 95% CI 1.99–5.34, p value < .001) children compared to non-depressed mothers. Maternal-related factors such as poor education, unemployment, and low household income were found to be significantly associated with higher odds of children’s short stature and underweight.
Malnutrition is a primary cause of child morbidity and mortality, accounting for almost 35–45% annual deaths worldwide (Bhutta and Black, 2013; Liu et al., 2012). Child stunting, the chronic slowing of physical and cognitive growth, prevents over 150 million children in developing countries from attaining their developmental potential (De Onis et al., 2012). Children in South Asia are markedly at a greater risk than those in other Asian countries with 30% of children being underweight and 37% of them stunted (Onis et al., 2013). Although genetics play a role, many socio-demographic variables and less favorable environments in terms of economic resources, food security, and access to hygiene and health-care facilities are identified as risk factors of poor child growth and development. Furthermore, insufficiencies in care during early childhood are most likely to manifest in child’s physical and social well-being later in life. Stunted childhood growth can result in poor schooling, low educational performance, followed by reduced work capacity and wasted adult physique leading to low economic productivity and greater risk of disease in adulthood. The consequences of impairment in child development are likely to be intergenerational, indicating that the impact of childhood stunting is likely to be seen in the next generation of children (Victora et al., 2008; Walker et al., 2015). In developing countries, children are exposed to multiple risk factors including poverty, malnutrition, poor health, and unstimulating home environments. Among these risk factors, suboptimal maternal care potentially carries larger risk to the physical and mental growth of a child (Rahman et al., 2002, 2007; Surkan et al., 2011). Recently, the role of parental practices and behaviors in the etiology of child growth has been examined and particular attention has been directed to the role of maternal depression in child growth.
Maternal depressive symptoms defined as major or minor depressive episodes that occur during pregnancy or postpartum (Gelaye et al., 2016) are believed to impede child developmental potential during initial growth years as well as long-term disorder of emotional and cognitive development of the infants (Raposa et al., 2014; Surkan et al., 2011). Maternal depressive symptoms include fatigue, social withdrawal, sadness, changes in sleeping and eating patterns, anxiety and guilt of inability to care for the infant, and crying. Clinical symptoms include depressed mood, fear, anxiety, despondency, and may include thoughts of harming self or the infant (De Castro et al., 2015). It is reported in various studies that children of depressed mothers are at a higher risk of poor development as compared to non-depressed mothers (Surkan et al., 2014). Untreated maternal depression in mothers has negative and long-lasting health consequences for the woman and the infant. Evidence suggests that psychological problems among mothers may be associated with less sensitive parenting practices and nonresponsive feeding styles (Mallan et al., 2015). In a more direct pathway, depressive symptoms in mothers increase children’s susceptibility to the physical growth, poor general health, and behavioral and emotional problems across the developmental period. A wide range of psychological problems can make mothers physically and emotionally less capable of providing their children’s basic need, thus affecting child’s health in an indirect way (Avan et al., 2010).
Maternal depressive symptoms are becoming increasingly prevalent in both developed and developing countries, with rates ranging from 0.5% to 60% (Halbreich and Karkun 2006; Valentine et al., 2011). Marital and socioeconomic status, unplanned pregnancy, low self-esteem, limited social support, and history of depression are some of the well-known risk factors of maternal depression. In low- and middle-income countries, being a victim of gender-based violence, being socially disadvantaged, adverse life events, disappointment with the sex of the baby, poor relationship with a mother-in-law or partner, and experiencing an unplanned pregnancy are common risk factors related to depression in women (De Castro et al., 2015). A recent report from low- and middle-income countries suggested antepartum depression in one of four women and postpartum depression in one of five women (Gelaye et al., 2016).In Pakistan, a limited number of maternal depression prevalence studies have been conducted. Likewise, there are few studies conducted in Pakistan that examine association of maternal depression with child’s malnutrition status. A cross-sectional study reported a prevalence of psychological distress in 36% Pakistani women (Husain et al., 2006). While, Rahman et al.’s community cohort study reported that infants with depressed mothers had twice the risk of diarrheal illnesses and four times higher risk of undernutrition in their first year of life (Rahman and Creed, 2007). Given the impact of maternal depression on child outcomes, it is plausible that maternal depression may also lead to poor childhood growth parameters, especially in low socioeconomic settings where the environment is frequently more hostile. Recognizing the relative scarcity of evidence in local context, this study aimed to assess the prevalence of maternal depressive symptoms in a low-socioeconomic status community using Aga Khan University Anxiety and Depression Scale (AKUADS) and to examine whether maternal depressive symptoms are associated with children’s growth, particularly stunting and underweight, in the first two years.
Methods
Study setting
Karachi is the largest economic hub of Pakistan with a total population of over 21.2 million (http://www.kmc.gos.pk/). The Gadap town is located in the district Malir of Karachi which covers over 1200 km2 and includes a population of 500,000 with majority belonging to low socioeconomic status. It is one of the most backward towns of city with diverse background lacking resources and basic health facilities.
Sampling frame and study participants
This was a cross-sectional study undertaken in a community with eight union councils (UC). Each UC on average contains 50 villages. Multistage random sampling technique was used. Initially, two UCs were selected based on simple random technique by lottery method and at the second stage of sampling, 10 villages were selected at random and finally the houses were chosen by systematic random sampling technique. The initial house was picked arbitrarily, and then every Kth house was selected. Selection of the house was based on simple arithmetic method used for systematic random sampling.
The sample consisted of all mothers aged 18–44 with at least one child less than two years of age. All identified mothers suffering from any serious physical disorder and who were unable to answer the interview questions were excluded. Children with any congenital disease interfering with their growth were also excluded from the study. All subjects were provided with detailed information about the study, and both written and verbal consent were obtained. Trained staff read out the study information to subjects who could not read and took consent on their behalf. In order to maintain the confidentiality and anonymity of the respondent, a unique Id was allocated to each individual in the data file and data analyst, and other relevant investigators had no access to the confidential information. For ethical considerations, the institutional ethical review committee (IRB) of the Dow University of Health Sciences, Karachi, Pakistan reviewed and approved the proposal. The study was carried out from August 2014 to September 2014.
Material and procedure
Sample size estimation
Sample size was calculated by the World Health Organization (WHO) sample size calculator using a proportion of 30% malnourished children, with 5% margin of error and 95% confidence interval (CI) (Patel et al., 2003). Thus, the sample size estimated for our study was 322.
Assessment of maternal depressive symptoms
Maternal depressive symptom assessed in 12 months postpartum was the main predictor variable. All mothers completed the AKUADS, a validated 25-item questionnaire with 12 somatic and 13 psychological items. This scale is designed for the assessment of psychiatric morbidity in the general population. Each item in the scale is based on four options: never, sometimes, mostly, and always; scored from 0 to 3, whereas the total score is obtained by adding the scores for each of the responses. With a cut point of ≥20 used in this study, a sensitivity of 64% and a specificity of 77% have been reported for postnatal depression using the tool (Kasi et al., 2012). None of the mothers in our study were clinically diagnosed for maternal depression.
Assessment of maternal and socio-demographic factors
Data were obtained on the following important confounding variables which were ‘recoded’ in the questionnaire on the basis of (1) mothers’ age (18–31 years as younger maternal age and 32–44 years as older maternal age); (2) ethnicity (to describe their ethnicity, subjects were categorized as Sindhi, Urdu, Punjabi, Balochi, Pashto, and others); (3) education level (subjects highest level of education was noted; response options included no education, grade 1 to 5 as primary, grade 5 to 10 as secondary, and grade 11 onward as higher); (4) occupational status (working or nonworking); (5) feeding practices (whether the child was breastfed vs. bottle-feeding); (6) family structure (families were categorized as nuclear vs. joint); (7) income (for poverty assessment, family monthly income obtained was noted based on the household’s ability to buy food and basic household needs): subjects with household income <12,000 Pakistani Rupees (PKRs) were considered relatively poorer as compared to those whose household income was ≥12,000 PKRs; and (8) number of children under seven years of age (two or more children under seven years vs. less than two children under seven years of age).
Assessments of anthropometric characteristics of child
Child characteristics included child gender and age at interview. Weight and height were measured by trained study staff using standard ‘infant weighing scale’ and ‘infantometer’. Weight and height were taken to the nearest 0.1 kg and 0.1 cm, respectively. Child weight and height data obtained were extrapolated onto WHO child growth standard scale. Children were classified as underweight or stunted if their weight-for-age or length-for-age was less than third percentile respectively on the basis of WHO multicenter growth reference study (Bhan and Norum, 2004).
Statistical analysis
We used Statistical Package for Social Sciences version 21.0 to analyze the data. Pearson χ 2 and binary logistic regression with p values of .05 were used to investigate the association of maternal depressive symptoms with underweight and stunting among their children. Independent variables with a p value ≤ .25 were included in the multivariate analysis. Odds ratio (OR) with 95% CI were reported.
Firstly, prevalence of the depressive symptoms among respondents based on the AKUADS scores was reported. Subsequently, differences in prevalence of child growth patterns in terms of stunting and low weight were estimated. Thirdly, multivariate analysis was conducted taking physical growth status as an outcome and maternal depressive symptoms as independent variable after adjusting for other predictor variables, including mothers age, ethnicity, education, occupation, household income (PKRs), family status, number of children under seven years of age, feeding practices, and child’s gender and age at interview.
Results
Descriptive characteristics of the women and their children under two years included in the study are presented in Table 1. Of 325 mothers, 71% (n = 231) were aged <30 years and almost 17% had no formal education. More than 93% (n = 304) subjects were unemployed and 78% (n = 253) belonged to low socioeconomic status. Among the subjects, 9% (n = 30) had two or more children less than seven years of age, while 53.8% (n = 175) followed bottle-feeding practices.
Maternal and child factors in relation to the weight and height of children.
More than half (n = 193) of the children included in the study were less than 12 months of age. A high level of malnutrition was observed among children under two years: 36.6% were stunted and 35.4% were underweight for their age (Table 1).
The overall prevalence of depressive symptoms among mothers was 40%; 50% of these depressed mothers either had underweight or stunted child (Table 1).
Stunting in children was significantly associated with maternal depressive symptoms (p value < .001), low education of mothers (p value .012), low household income (p value .034), and female child (p value .010). Status of being underweight in children was also significantly associated with maternal depressive symptoms (p value < .001), low education of mothers (p value .048), and female child (p value .018; (Table 1).
Association between maternal depressive symptoms and children’s stunting
The association of maternal depressive symptoms with child’s nutritional outcome is presented in Tables 2 and 3. Univariate analysis indicated that symptomatic mothers were more likely to have stunted children (OR 2.70, 95% CI 1.69–4.30, p value < .001). The association remained significant (OR 3.15, 95% CI 1.91–5.18, p value < .001) after adjusting the mother’s age, education, and occupation, household income, and child’s gender. Furthermore, univariate analysis indicated that uneducated mothers (OR 2.21, 95% CI 1.21–4.04, p value < .010) and mothers with household income ≤12,000 PKRs (OR 1.76, 95% CI 1.04–3.00, p value .035) were more likely to have stunted children. Similarly, female children (OR 1.81, 95% CI 1.14–2.87, p value .011) were more likely to be stunted as compared to male children. Multivariate logistic regression analysis showed slight difference in OR for all the factors that were significant in the univariate analysis. In addition, nonworking mothers (OR 3.37, 95% CI 1.01–11.20, p value .047) as compared to working mothers were likely to have children with short height (Table 3).
Association between maternal depressive symptoms and children’s underweight
Mothers with depressive symptoms were three times more likely to have underweight children as compared to non-symptomatic mothers (OR 3.00, 95% CI 1.874–4.81, p value < .001; Table 2). The difference between the groups remained statistically significant in multivariate model (OR 3.26, 95% CI 1.99–5.34, p value < .001), after adjusting for mother’s age and education, household income, and child’s gender (Table 3). Univariate analysis (Table 2) also indicated that mothers with no education (OR 2.07, 95% CI 1.13–3.78, p value < .018) were more likely to have underweight children. Furthermore, female children (OR 1.74, 95% CI 1.09–2.76, p value < .018) were more likely to be underweight. In addition, in multivariate analysis, a girl child was found to be at a higher risk of being underweight (OR 1.90, 95% CI 1.16–3.10, p value .011) as compared to their counterpart.
Univariate analysis of stunted and underweight children with child and maternal factors.
Note: OR: odds ratio; CI: confidence interval; cut off .25.
Multivariate analysis of stunted and underweight children with child and maternal factors.
Note: OR: odds ratio; CI: confidence interval; PKR: Pakistani Rupees.
*Adjusted for mother’s age, Education, income (PKR), occupation, and child’s gender.
*Adjusted for mother’s age, Education, income (PKR) and child’s gender.
Discussion
Our study results indicate that in Pakistan 4 out of 10 mothers belonging to low socioeconomic status suffer from depressive symptoms during the postpartum period. Maternal depressive symptoms remain an independent predictor of child stunting and underweight. Interestingly, inclusion of potential confounders in the multivariate analysis did not change the effect of maternal depressive symptoms on stunting and underweight status. The independent association between maternal depressive symptoms and child nutritional status suggests that there might be some important underlying mechanisms for this association. In a typical male-oriented society, women remain more vulnerable to low self-esteem, anxiety, and depression (Kathree et al., 2014). They perceive themselves as insecure, less competent, and incapable without men. Further, they are also dependent on men for social support and economic resources (Rabbani et al., 2008). In such situations, most mothers are not vocal about their feelings and consider their depressed mood as a personal weakness rather than mental illness, and thus depression becomes difficult to detect. Such social-contextual factors are well-established risks for maternal depression (Rahman and Creed, 2007). Thereby, it is more important to explore and gain understanding of the effects of traditional societal norms on stress and psychiatric problems in postnatal women in relation to child growth.
Regarding the prevalence of the maternal depressive symptoms among Pakistani women, the results are comparable with findings of previous studies in the country and the South Asian region (Rahman et al., 2002; Nasreen et al., 2015; Patel et al., 2003). Furthermore, it has been noted that depressive symptoms were more common in young mothers. One possible explanation may be the changing caregiving demand that exacerbate parenting difficulties. The new role and routine with their baby may put additional stress on mothers who are experiencing depressive symptoms (Goyal et al., 2010). This suggests that young mothers may begin motherhood with poor interpersonal skills which may increase vulnerability to postnatal depression. Thus, it calls for deeper understanding of the interpersonal skills of mothers and other stressors in relation to the ability to provide good quality parenting (Ashaba et al., 2015).
Additionally, malnutrition status, both stunting and underweight among children of mothers with depressive symptoms, differs markedly from those of children of no symptomized mothers. This finding is also consistent in low-income and middle-income contexts (Nasreen et al., 2013, 2015; Patel et al., 2003; Rahman and Creed 2007; Rahman et al., 2002; Surkan et al., 2011). A notable gender difference emerged in this study: compared with boys, girls indicated higher odds of being stunted and underweight. Although some studies have reported that the male child among low- and middle-income countries, were more likely to be stunted during postpartum; the evidence to date is mixed (Medhin et al., 2010; Nasreen et al., 2015). However, our study found girls to be more vulnerable to malnutrition than boys among subjects belonging to low socioeconomic status. The observed sex-specific malnutrition effect indicates that the preference toward a male child enhances the risk of child stunting or underweight in female child in a cultural setting with deep-rooted gender disparities (Abuya et al., 2012).
We also found that maternal-related factors such as poor education, unemployment, and low household income (used as a proxy for economic deprivation) put children at increased risk of being stunted. It is noted that an overall lower socioeconomic profile may promote a sense of hopelessness among mothers and may adversely influence their performance in the parenting role (Surkan et al., 2011). Therefore, we can suggest that socioeconomic factors may have a greater role in the etiology of undernutrition among children.
Our results reinforce the significance of preventing maternal depression to help place children on optimal growth trajectories. None of the mothers in our study were clinically diagnosed for maternal depression. The impact of clinically diagnosed maternal depression on child growth has been frequently examined and found to be extensively significant as well as in clinical settings (Nasreen et al., 2015). Thus, it is possible that even undiagnosed or untreated maternal depression can lead to child’s hampered development, possibly mediated by alterations of mother warmth and affection.
Study limitations
Our study has some limitations. Within the realm of the study, participants are representative of more socially disadvantaged than the general population. Thus, the results may need to be interpreted with caution. Secondly, the precise estimate of the prevalence of maternal depression in Pakistan is difficult to obtain because of limited access to health-care facilities and cultural and social taboos may result in women underreporting depressive disorders. More extensive research to explore this issue is further needed. Thirdly, this study is basically correlational, and therefore causal associations cannot be assumed. Moreover, because of the study design depressive symptoms in mothers were not confirmed through clinical psychology criteria, severity of the depressive symptoms cannot be standardized in relation to diagnostic criteria against diagnostic scales either. Nevertheless, standardized and validated AKUDAS screening scale was used in this study and was found valid and reliable in general population settings and perhaps was the best available tool locally. Notwithstanding these limitations, this study is an addition to the existing literature that confirms the relationship of maternal depressive symptoms in the postnatal period with child’s physical health. Moreover, we recommend integration and collaboration of women and child health issues into prevailing health systems to improve not only mental health programs but also infant health and development outcomes.
Conclusion
Our study suggests that the postnatal depressive symptoms among mothers have detrimental effects on child physical growth and development. The screening and earlier detection and treatment of depressive symptoms in mothers may alter the risk of poor physical growth parameters among children.
Footnotes
Acknowledgements
The authors would like to thank all involved in conducting the survey and the study participants. We also like to acknowledge the DUHS for all the logistic support provided.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.