Maternal depression and its association with responsive feeding and nutritional status of infants: A cross-sectional study from a rural medical college in central India.

Context: Globally, around half of all under-5 deaths are attributable to undernutrition. The magnitude of child undernutrition in India is one of the highest in the world. Responsive feeding that has the potential to optimize nutrition and development depends on a "healthy mother".
Objectives: The objective of this study was to find out the magnitude and determinants of maternal depression among mothers of infants who attended immunization clinic in a rural medical college in central India and observe its association with infant-feeding practices and nutritional status of infants. Study Design: This was an analytical cross-sectional study. Materials and
Methods: Study was conducted among consecutively chosen 300 mothers and their infants who visited the immunization clinic of hospital during May-September, 2016. Data were collected individually by interviewing each mother, and anthropometry of her infant was done. The mothers with depression were diagnosed according to the 10-item Edinburgh Postnatal Depression Scale. Magnitude of depression has been reported using frequency and percentage, whereas association has been studied using prevalence odds ratio with 95% confidence interval (CI) and logistic regression at 0.05 significance level.
Results: The magnitude of maternal depression was 19% (95% CI 14.4-23.5). Mothers with depression had significantly higher odds for not indulging in responsive feeding. Maternal education had significant association with maternal depression after adjusting for other variables. Conclusions: Burden of maternal depression is high among mothers of infants who attended immunization clinic in a rural medical college in central India and is significantly associated with nonresponsive feeding practices.

Introduction

Inadequate nutrition in the first 1000 days of a child's life can lead to undernutrition, impaired growth, and development. Almost 50% of all deaths in children under 5 years are attributable to undernutrition.[1] The magnitude of child undernutrition in India is one of the highest in the world. Around 39% of under-5 children in India are stunted, 15% are wasted, and 29% are underweight. In India, along with the burden of macronutrient deficiencies, 72% of infants suffer from anemia.[2]

World Health Organization recommends that feeding should be both responsive and sensitive. Responsive feeding that has the potential to optimize nutrition and development is an important component of infant feeding.[34] One of the key determinants for responsive feeding is a “healthy mother.” A depressed mother is not able to identify signs of hunger in her child and respond appropriately. Available evidence suggests significant association between maternal depression and impaired child growth in developing countries.[5]

Globally, approximately 10–20% of women experience depression either during pregnancy or in the first 12-month postpartum, whereas in India, maternal depression has been reported to range from around 11% to 23%.[67]

The present study was conducted as magnitude of maternal depression is not exactly known among mothers in our area. The primary objective of the study was to find out the magnitude and determinants of maternal depression among mothers of infants who attended immunization clinic in a rural medical college in central India. The secondary objectives included to study the association of maternal depression with responsive feeding and nutritional status of infants.

Study design

This was an analytical cross-sectional study.

Materials and Methods

This hospital-based study was conducted between May and September, 2016, with mothers of infants who visited the immunization clinic of a single hospital. The sample did not include antenatal mothers or mothers of under-5 children beyond infancy.

The study was initiated after prior approval from the Institutional Ethics Committee for Human subjects. Only mothers who gave written informed consent for them and their infants' participation were included in the study.

Maternal depression was defined as the spectrum of depressive conditions that can affect mothers up to 12 months postpartum and mothers to be.[6] Estimated sample size using OpenEpi software was 300 assuming a 95% confidence level, prevalence of maternal depression (p) as 23%,[7] an absolute precision of 5%, and a nonresponse rate of 10%. The sample size of 300 was sufficient to capture a 50% prevalence of responsive feeding and undernutrition (underweight, wasting, and stunting) with an absolute precision of 6% (post-hoc assessment). Infants with any diagnosed congenital problems, and who came for immunization with other relatives such as father and grandparents, were excluded. Mothers who were on treatment for any concurrent chronic illness (e.g., hypertension, epilepsy, diabetes, etc.) were also excluded from the study. We used consecutive sampling and offered participation to each successive mother–infant pair that visited the immunization clinic during the study period until the desired sample size of 300 was recruited.[89] Data were collected individually by interviewing each participant maintaining confidentiality and privacy.

A pretested semistructured interview schedule was used to collect information pertaining to sociodemographic profile. It included details on age, education, working status, economic status, husband's occupation, age, and gender of the child and family type. Alcohol by husband, being blamed for girl child, and domestic violence were either considered as “yes” or “no” as reported by the mother herself based on her experience during the preceding week. A mother was classified as literate (if she could read and write with understanding in any language) or illiterate (can neither read nor write or can read but cannot write in any language) and education level was the highest level of education completed. Occupation of the mother was categorized as employed if she was engaged in any kind of economically productive work or a homemaker if she was unemployed. In India, public distribution system issues ration card to families by assessing their economic status. In this study, the type of ration card possessed by her family was taken as proxy indicator of her economic status, that is, yellow and orange/white-colored cards for below poverty line (BPL) and above poverty line families, respectively. Those who did not have ration card were considered as non-BPL.

The mothers were screened using the 10-item Edinburgh Postnatal Depression Scale (EPDS) that has been used and validated in India.[10] It is a 10-item self-reported scale based on 1-week recall, designed to screen postnatal depression (PND), yielding a total score of 0–30. Responses are scored 0, 1, 2, and 3 based on the seriousness of the symptom. Items 3 and 5–10 are reverse scored (i.e. 3, 2, 1, and 0). The total score is found by adding together the scores for each of the 10 items. For this study, a cutoff score of ≥10 was used to diagnose maternal depression.[11] Diagnosis was confirmed by careful clinical assessment by a trained physician, and the mothers detected with maternal depression were referred to psychiatry for counseling and/or treatment. Standard definitions of underweight, stunting, and wasting were considered as defined in the WHO Child Growth Standards.[12] For responsive feeding assessment, a tool used in Bangladesh by Aboud et al. was adapted and used.[13] The responsive feeding tool included information on who fed the child, child eats meal alone or together with someone, talking to the child while feeding, refusal to eat, forced feeding, self-feeding, and adult feeding. Responses were marked 1 or 0 depending on the response; a mother scoring 3 or more was considered to be indulging in responsive feeding. Data were entered and analyzed using Epi-info software. Magnitude is reported using frequency and percentage, whereas association has been studied using prevalence odds ratio (POR) with 95% confidence interval (CI) and the prevalence ratio (PR). We used logistic regression for estimating the predictors for maternal depression and get the adjusted odds ratio (AOR). Mothers with undernourished children were also counseled regarding proper infant and young child feeding practices and responsive feeding.

Results

Table 1 gives the background characteristics of study participants. The mean age of mother's was around 25 years with around 3/4th staying in nuclear families. Very few of the parents (<2%) were illiterate. Majority (74%) of the mothers were homemakers and did not indulge in any paid work, while all the father's indulged in one or other type of paid work. Around 22% of the mothers belonged to BPL families. Around 25% of the newborns were of low birth weight. The magnitude of underweight, stunting, and wasting was around 33%, 45%, and 22%, respectively.

Table 1

Background characteristics of study participants

Variable (n=300)	Category	Frequency (%)
Age of mother	25.17±3.39*
	19-24 years	149 (49.7)
	25-35 years	151 (50.3)
Type of family	Nuclear	219 (73)
	Joint	81 (27)
Education of mother	Graduate	37 (12.3)
	Higher secondary	83 (27.7)
	High school completed	103 (34.3)
	Primary	75 (25)
	Illiterate	2 (0.7)
Education of father	Graduate	47 (15.7)
	Higher secondary	75 (25)
	High school completed	95 (31.7)
	Primary	77 (25.7)
	Illiterate	6 (2)
Occupation of mother	Homemaker	222 (74)
	Daily wage laborer	35 (11.7)
	Farming	34 (11.3)
	Service	9 (3)
Occupation of father	Farming	110 (36.7)
	Daily wage laborer	144 (48)
	Service	44 (14.7)
	Self-employed	2 (0.7)
Color of ration card	White	89 (29.7)
	Orange	130 (43.3)
	Yellow	68 (22.7)
	Do not have	13 (4.3)
Alcoholic husband	Yes	20 (6.66%)
	No	280 (92.3%)
Domestic violence	Yes	15 (5%)
	No	285 (95%)
Blamed for girl child	Yes	35 (11.6%)
	No	265 (88.3%)
Age of child	0-5 months	221 (73.7)
	6-11 months	79 (26.3)
	(Age range 3-344 days)
Birth weight of child	<2500 g	77 (25.7)
	≥2500 g	223 (74.3)
Sex of child	Male	161 (53.7)
	Female	139 (46.3)
Underweight	Yes	101 (33.6%)
	No	199 (66.4%)
Stunting	Yes	134 (44.6%)
	No	166 (55.3%)
Wasting	Yes	68 (22.6%)
	No	232 (77.3%)
RF	Yes (RF score=3-5)	187 (62.3)
	No (RF score=0-2)	113 (37.7)

RF:Responsive feeding, *Mean±SD

The proportion of maternal depression among mothers of infants who attended immunization clinic was 19% (57 mothers) as depicted in Figure 1. The 95% CI for proportion of maternal depression among mothers of infants was 14.4–23.5. The mean age of mothers with depression was 25.39 ± 3.57 years and without depression was 25.12 ± 3.36 years. Similarly, the mean birth weight of infants of depressed mothers was 2729 ± 539 g, whereas that of those who were not depressed was 2659 ± 482 g.

Figure 1

Magnitude of maternal depression

Table 2 shows the association of maternal depression with the probable determinants. Only the association between education of mother and father with maternal depression was found to be statistically significant. The proportion of mothers with depression was 2.3-fold (CI 1.2–4.3) greater if she had received primary or lesser level of education. Similarly, proportion of mothers with depression was twofold (CI 1.1–3.7) greater if the father of the child had received primary or lesser level of education. Other variables like birth weight, type of family, sex of the child, alcoholism of husband, domestic violence, poverty, being blamed for girl child, age of mother, age of child, and occupation of mother were not found to be statistically significant as depicted by CIs that contain the null value.

Table 2

Determinants of maternal depression

		EPDS score		Total	POR	95% CI
		Depression	No depression			Lower	Upper
Sex of child	Male	32 (19.9%)	129 (80.1%)	161 (53.6)	1.131	0.633	2.022
	Female	25 (18.0%)	114 (82%)	139 (46.3%)
Alcoholic husband	Yes	5 (25%)	15 (75%)	20 (6.66%)	1.462	0.508	4.202
	No	52 (18.6%)	228 (81.4%)	280 (92.3%)
Domestic violence	Yes	5 (33.3%)	10 (66.7%)	15 (5%)	2.240	0.735	6.831
	No	52 (18.2%)	233 (81.8%)	285 (95%)
Socioeconomic status	BPL	17 (25%)	51 (75%)	68 (22.6%)	1.600	0.839	3.053
	Non-BPL	40 (17.2%)	192 (82.8%)	232 (77.3%)
Blamed for girl child	Yes	8 (22.9%)	27 (71.1%)	35 (11.6%)	1.306	0.560	3.049
	No	49 (18.5%)	216 (81.5%)	265 (88.3%)
Mother’s occupation	Homemaker	41 (18.5%)	181 (81.5%)	222 (74%)	0.878	0.460	1.674
	Any paid work	16 (20.5%)	62 (79.5%)	78 (26%)
Mother’s education	Primary or less	23 (29.9%)	54 (70.1%)	77 (25.6%)	2.368*	1.287	4.355
	Secondary or more	34 (15.2%)	189 (84.8%)	223 (74.33%)
Father’s education	Primary or less	23 (27.7%)	60 (72.3%)	83 (27.6%)	2.063*	1.128	3.775
	Secondary or more	34 (15.7%)	183 (84.3%)	217 (72.3%)
Birth weight	LBW	18 (23.4%)	59 (76.6%)	77 (25.7%)	1.439	0.766	2.705
	Normal	39 (17.5%)	184 (82.5%)	223 (74.3%)
Age of mother	19-24	30 (20.1%)	119 (79.9%)	149 (49.7%)	1.158	0.650	2.063
	>25	27 (17.9%)	124 (82.1%)	151 (50.3)
Age of child	0-5 months	41 (18.6%)	180 (81.4%)	221 (73.7%)	0.897	0.471	1.710
	6-11 months	16 (20.3%)	63 (79.7%)	79 (26.3%)
Type of family	Nuclear	40 (17.5%)	189 (82.5%)	229 (76.3%)	0.672	0.353	1.279
	Joint	17 (23.9%)	54 (76.1%)	71 (23.7%)

Bold values mean significant p value or 95% Confidence interval, EPDS:Edinburgh Postnatal Depression Scale, POR:Prevalence odds ratio, BP L:Below poverty line, *Significant, LBW:Low birth weight

We used logistic regression analysis to determine predictors of maternal depression as shown in Table 3. P value of 0.05 was considered as significant. The predictor for maternal depression in study participants was education of mother at P < 0.01. Mothers who had received primary or less education were associated with higher maternal depression after adjusting for other factors. Mothers who had received primary or less education were having 1.7 times more likelihood of having maternal depression than who had received secondary or higher education. (AOR 1.717, 95% CI 1.174–2.490).

Table 3

Predictors of maternal depression

Variables	Odds ratio	95% CI		Coefficient	SE	Z-statistic	P
		Lower	Upper
Birth weight	1.384	0.504	3.802	0.325	0.516	0.397	0.529
Age of child	1.104	0.504	2.421	0.099	0.401	0.061	0.805
Age of mother	0.641	0.272	1.508	−0.445	0.437	1.038	0.308
Blame for girl child	0.720	0.211	2.462	−0.329	0.627	0.274	0.600
Socioeconomic status	1.361	0.693	2.670	0.308	0.344	0.801	0.371
Type of family	1.286	0.786	2.102	0.251	0.251	0.286	0.317
Education of father	1.369	0.695	2.011	0.314	0.346	0.826	0.363
Education of mother	1.717	1.174	2.490	0.536	0.192	0.907	0.005
Occupation of mother	1.156	0.584	2.287	0.145	0.348	0.174	0.677
Domestic violence	2.454	0.745	8.077	0.898	0.608	2.180	0.140
Alcoholic husband	0.805	0.239	2.706	−0.217	0.619	0.123	0.726
Sex of child	1.143	0.625	2.089	0.134	0.308	0.189	0.664
Constant				−1.549	0.366	17.921	0.837

Bold value means significant p value

Table 4 shows the association between maternal depression and responsive feeding. Mothers who were depressed had significantly lower odds (POR = 0.297) of practicing responsive feeding. Only 22 (38.6%) mothers who were depressed were indulging in responsive feeding as compared to 165 (67.9%) who were not depressed. The association between responsive feeding and maternal depression was found to be statistically significant (95% CI 0.163–0.540). The PR was 0.568, meaning that the proportion of mothers indulging in responsive feeding was around 44% less if a mother was having maternal depression.

Table 4

Association between responsive feeding and maternal depression

		Responsive feeding		Total	POR	95% CI
		Yes	No			Lower	Upper
EPDS score	Depression	22 (38.6)	35 (61.4)	57 (19)	0.297* (PR=0.568)	0.163	0.540
	No depression	165 (67.9)	78 (32.1)	243 (81)
Total		187 (62.3)	113 (37.7)	300

*Significant, Bold values mean significant p value or 95% Confidence interval

Association of maternal depression was studied with all the three indicators for nutritional assessment in children, that is, weight for age (underweight), length for age (stunting), and weight for length (wasting) as shown in Table 5. POR and PR were higher for stunting among infants of depressed mothers; however, this was not statistically significant as depicted by CIs that contain the null value.

Table 5

Association between nutritional status and maternal depression

		Weight for age		POR	95% CI
		Underweight	Normal		Lower	Upper
Depression	Yes	20 (19.8%)	37 (18.6%)	1.081 (PR=1.052)	0.590	1.981
	No	81 (80.2%)	162 (81.4%)
		Height for age
		Stunting	Normal
Depression	Yes	30 (22.4%)	27 (16.3%)	1.485 (PR=1.229)	0.833	2.649
	No	104 (77.6%)	139 (83.7%)
		Weight for height
		Wasting	Normal
Depression	Yes	13 (19.1%)	44 (19.0%)	1.010 (PR=1.007)	0.508	2.009
	No	55 (80.9%)	188 (81.0%)

Discussion

This hospital-based analytical cross-sectional study was conducted with mothers of infants who visited the immunization clinic of rural medical college. We preferred to use analytical cross-sectional design as it is useful for examining the association between exposure and disease onset for chronic diseases where researchers specifically lack information on time of onset. The association has been studied using POR as for chronic disease studies or studies of long-lasting risk factors, POR is the preferred measure of association in cross-sectional studies. Of the 300 mothers, those scoring 10 or more on EPDS were diagnosed to have depression.

We found that 57 (19%, 95% CI 14.4–23.5) participants had maternal depression. Our results are in accordance with the findings of other studies and reflect that the magnitude of maternal depression in the study area is similar to other regions of India and Southeast Asia. A similar survey by the World Health Organization showed the prevalence of maternal depression to be 10–15%.[14] The study by Surkan et al. showed 15–57% prevalence of depression in mothers,[5] whereas study by Shidhaye found that in India, 10–20% mothers face postpartum depression (PPD).[6]

Shivalli et al. in their study in a rural hospital from Karnataka found poverty, birth of female baby, and pregnancy complications or known medical illness as independent predictors of PND.[15] Similarly, studies from other countries on PPD or PND found place of residence, form of marriage, mother's job, house income level, no family support, mothers' education, lack of social support, and history of violence as predictors for maternal depression.[1617181920] A study from Sudan found that older age of mothers, exclusive breast feeding, and regular prenatal vitamins during pregnancy result in decreasing odds of PND. In the present study, only mothers' and fathers' education was found to be statistically significant on bivariate analysis, while on logistic regression, only maternal education was found to be a significant predictor for maternal depression. Therefore, along with the identification and treatment of maternal depression, it is imperative to act on distal determinant like maternal education if this important public health problem is to be tackled. Other variables like birth weight, type of family, sex of the child, alcoholism of husband, domestic violence, poverty, being blamed for girl child, age of mother, age of child, and occupation of mother were not found to be statistically significant. The difference in findings could be attributed to the different sociocultural context in which these studies were conducted. Other factors found to be significant in these studies were preexisting depression or psychiatric conditions, chronic health problems, primiparity, vaginal delivery, women experiencing complications at time of delivery, unintended pregnancy, and obesity.[1617181920] In the present study, antenatal and intranatal factors were not considered as we presumed that they would be subjected to recall bias. Moreover, presence of known chronic conditions in mother was an exclusion criteria.

In Indian culture, preference to male child is deeply entrenched. Birth of a girl child acts as a family and social stressor and hence was hypothesized to be a determinant for maternal depression. However, the present study did not find a significant association with either the gender of the baby or being blamed for girl child. The not-significant results for being blamed for girl child or some of the other variables like domestic violence may be due to the smaller sample size in these subgroups in our study. Nonetheless, the higher odds are indicative of plausible association with negative mileu in which a mother may have to live and should be evaluated further with studies involving adequate power for respective subgroup analysis.

In our study, a significant statistical association was found between responsive feeding and maternal depression. Similar results were found in another study by Hurley et al. which stated that depressed mothers had a tendency to view their infant's behavior negatively, which were a reflection of them having nonresponsive feeding styles.[21] A study done by Wachs et al. found that negativity and lesser responsiveness by the caregiver contribute to very high rates of insecure attachment as seen in infants of mothers undergoing depression.[22] Similarly depressed mothers in our study showed nonresponsiveness for their infants and lack of proper feeding practices. Post-hoc power analysis showed that study was adequately powered to find association with responsive feeding (power: 97.9%).

The systematic review, conducted by Hurley et al., found the association between nonresponsive feeding and high birth weight of infants.[23] Similarly, the study conducted by Worobey et al. found that infants whose mothers did not identify their children's satiety cues gained more weight significantly from 6 to 12 months than infants of sensitive mother.[24] In our study, we found that the mean birth weight of infants of depressed mothers (2729.6 g) was higher than the standard normal weight of 2500 g.

Maternal depression was shown to have higher odds for having undernutrition among infants, especially stunting (POR = 1.5) which was not statistically significant. Post-hoc power analysis showed that study was not adequately powered to find association with nutritional status (power: <70%). The long-term effects of maternal depression on nutritional status of children should be evaluated further with studies involving adequate power.

In a study conducted by Bentley et al., it was proved that infants who are fed with responsiveness show significantly less proportion of stunting, wasting, and underweight compared to those who were not fed actively.[25] In our study, we found significant association of maternal depression with nonresponsive feeding. The long-term consequences of nonresponsive feeding warrant further evaluation through valid prospective observational studies to confirm these findings in the Indian context.

The present study has several limitations. The study was conducted with mothers of infants (age range 3–344 days) who visited the immunization clinic of a single hospital. As a hospital-based sample, it is likely to be a biased one, especially prone to selection bias with limited external generalizability. Moreover, the magnitude of maternal depression determined may be an underestimate as the study was conducted among mothers of infants (0–11 months) and did not include assessment during antenatal period. As the questionnaire included some sensitive questions like blame for girl child, alcoholism of husband, domestic violence, and suicidal tendencies, a possibility of underreporting may not be denied despite our best efforts. In general, analytical cross-sectional studies are useful for establishing only a preliminary evidence for a causal relationship as these studies are affected by the antecedent–consequent bias. Therefore, although a significant association was found with maternal education and responsive feeding interpretation of results requires caution, the findings should be confirmed with prospective study designs of adequate power that will not be affected by antecedent–consequent bias or survival bias.

Conclusion

The present study demonstrates that burden of maternal depression is high among mothers of infants who attended immunization clinic in a rural medical college in central India and is significantly associated with nonresponsive feeding practices.

Financial support and sponsorship

This study was funded by the Indian Council of Medical Research (ICMR) under its Short-term Studentship program (ICMR-STS). The STS Reference ID was 2016-01258.

Conflicts of interest

There are no conflicts of interest.