Prevalence of postpartum depression & anxiety among women in rural India: Risk factors & psychosocial correlates

Postpartum is a period of remarkable transition in the life of women with multiple challenges associated with assuming the role of a mother, thus making her vulnerable to psychological distress. Postpartum depression (PPD) is generally described as depressive symptoms experienced within four weeks after childbirth, commonly manifesting as crying spells, insomnia, depressed mood, fatigue and poor concentration. The mothers can experience these symptoms any time after delivery to a year post-delivery1. Up to 20 per cent of women worldwide reportedly experience PPD2. Higher rates of PPD are most commonly associated with high postpartum anxiety (PPA) as well3, which may present alone or as comorbidity with PPD. While the overall prevalence of PPA at 1-24 wk post birth is reportedly 15 per cent, the situation is particularly alarming in low- and middle-income countries (LMICs), especially India4,5. However, scant research has been generated from the LMICs to establish its prevalence. Existing studies vary widely depending on methodology, variance in the data collection timeline, different screening tools, cut-off scores and the nature of the population5-7. Moreover, most PPD/A prevalence studies are hospital based and primarily include urban populations8. Thus, there is a wide gap in attaining a clear picture of PPD/A in the community setting.

The postpartum period is associated with multiple challenges related to breastfeeding, baby care, self-care, changing roles to bear various responsibilities and adjustments in couple and family relationships. A systematic review of women with PPD reported profound consequences on social and marital relationships, further impacting infant care and mother-infant bonding9. However, corroboratory evidence from the community setting on the same is presently lacking from LMICs. Some of the existing data show maternal depression to be linked to underweight and under nutrition during the first year of life and higher rates of diarrhoeal diseases, incomplete immunization and poor cognitive development in young children4. However, detailed investigation and corroboratory evidence are required to explore the impact of PPD/A on mothers’ mental health, family relationships, marital relationships, functionality and attachment to the infant.

Thus, the present study aimed to study the prevalence and psychosocial correlates of PPD/A in women from the rural community of northern India. Furthermore, the risk factors of PPD/A and their impact on infant bonding, the functional status of mothers and parenting stress were also assessed.

Material & Methods

This cross-sectional study was undertaken over a period of June 2014 to June 2016, after receiving approval from the Institutional Ethics Committee. It was conducted in 28 villages in the Ballabhgarh district of northern India and designated under the Comprehensive Rural Health Services Project (CRHSP), All India Institute of Medical Sciences, New Delhi, India. CRHSP caters to 28 villages by networking 12 sub-centres and two Primary Health Centres. A unique identification number was provided to all the residents, and prospective demographic and health records were maintained in Health and Management Information System (HMIS)10. With due permission from concerned authorities, a comprehensive list of all potential subjects from HMIS were obtained.

Study participants: Women aged 18-40 yr, between 4 wk to 6 months postpartum, having alive and healthy baby, available and willing to participate were included in the study. Those women with a stillborn or congenitally malformed baby, having a concurrent diagnosis of any other major psychiatric disorder (other than depression and anxiety), substance dependence and receiving pharmacotherapy for the same were excluded from the study. Further, women were also excluded if they had a history of significant medical illness, intellectual disability, cognitive decline or sensory impairment that could interfere with the assessment process.

Sample size: An earlier community-based study from India found a prevalence of 11 per cent for PPD, whereas, a 23 per cent prevalence was reported in another hospital-based study11,12. For the present study, presuming a prevalence of 15 per cent and allowing 20 per cent error (80% precision) the required sample with a 95 per cent confidence level was 567. Expecting a 20 per cent attrition rate, the total sample was calculated to be 680.

Study tools: A sociodemographic and clinical datasheet was constructed as part of the study to collect information on sociodemographic information and the obstetric history of the women. In addition, the Udai Pareek Scale13 was used separately to assess socioeconomic status (SES).

Edinburgh Postnatal Depression Scale (EPDS)14, a self-report measure to assess depressive symptoms across 10 items. Scores above 10 were considered as screened positive.

State-Trait Anxiety Inventory-Form Y (STAI-Y-1)15, a self-report measure to assess anxiety across 40 items was used in this study. The current study used STAI form Y-1 to assess state anxiety only. Scores above 40 were considered to be indicating significant anxiety.

Mini International Neuropsychiatric Interview (MINI)-6^th Edition16 was used as a diagnostic tool to assess if women screened positive for EPDS. Furthermore, STAI was administered for the diagnosis of the same. In addition, MINI was also used to determine if women had any comorbid disorders.

Postpartum Social Support Questionnaire (PSSQ)17, a 32 item scale for assessing the present and expected social support to women from their partner, parents-in-law, extended family and friends was used. A higher score on the test items indicates higher social support.

Postpartum Bonding Questionnaire (PBQ)18 was used to assess four forms of difficulties in bonding, i.e. impaired bonding (IB), rejection and pathological anger (RP), infant-focussed anxiety (IFA) and incipient Abuse (IA). A score above 11 indicated IB, above 16 indicated RP, equal or higher than 9 indicated IFA and above 2 indicated IA.

Inventory of Functional Status After Childbirth19 assessed the following domains of functionality after childbirth, namely, the ability to provide infant care, resume household responsibilities, social and community activities and lifestyle in the last two weeks. In addition, scoring of overall functional status, as well as sub-domain mean scores were calculated.

Parental Stress Index-Short Form (PSI-SF)20 assesses parental distress across 36 items. The present study evaluated the domain of Parental Distress (PD), which deals with newborns. Scoring has been done without reference cut-off scores based on the sub-domains mean and standard deviation (SD) values.

HITS (Hurt, Insult, Threaten, Scream) Scale21 measure of Intimate Partner Violence (IPV) is a short scale that measures IPV on four items. A score value above 10 is considered an indication towards domestic abuse.

Kansas Marital Satisfaction Scale (KMSS)22 is a short scale of three items that assesses women’s marital satisfaction with marital relationships. A score greater than 17 is indicative of poor marital satisfaction.

All the scales were translated into Hindi using the World Health Organization (WHO) guidelines for scale translation and adaptation using WHODAS 2.0 translation package (https://terrance.who.int/mediacentre/data/WHODAS/Guidelines/WHODAS% 202.0%20Translation%20guidelines.pdf)

Procedure for assessment: A list of all 876 postpartum women was obtained from HMIS. Medical social workers coordinated with Anganwadi workers to locate these 876 women, explain the study and obtain informed consent. Women were then screened for inclusion in the study by two trained clinical psychologists. Of the 876 women identified, 680 were recruited (162 were not available at time of home visit, 6 had lost their baby and 28 refused to participate). So, 680 women were administered EPDS and STAI, followed by assessing psychosocial factors using PSSQ, IFSCB (Inventory of Functional Status after Childbirth), PSI-SF, PBQ, KMSS and HITS. It took approximately 40 min to complete the entire evaluation for each woman. Fifty nine women obtained scores above the cut-off points on EPDS and, STAI and MINI was applied to confirm the diagnosis of major depressive episode (MDE) and general anxiety disorder (GAD). The diagnosis was confirmed in 30/59 women using MINI. The 29 screen-positive women, who tested negative on MINI, were screened again after one month for the persistence of symptoms using EDPS and STAI. Women who persisted in screening positive were administered MINI.

A schema of the flow of participants in the study is provided in the Figure.

Close

Figure

Procedural steps followed during the study. CRHP; comprehensive rural health services project; EPDS- Edinburgh postnatal depression Scale; STAI- state-trait anxiety inventory; PSSQ- postpartum social support questionnaire; PBQ- postpartum bonding questionnaire; KMSS- Kansas marital satisfaction Scale; HITS- hurt, insult, threaten and scream; PSI- parental stress index; IFSCB- inventory of functional status after childbirth

Export to PPT

Statistical analysis: The data were analysed using the SPSS version 22.0 (IBM Corp., Armonk, NY, USA). The prevalence was calculated based on the criteria that women should be positive on screening scales (EPDS and STAI) and MINI assessment. Further, the analysis has been conducted across two groups: (i) Women with PPD/A/both categorized as case group, and (ii) another group consisting of women without a diagnosis of PPD/A/both. Continuous variables were tested for normality assumption using the Kolmogorov–Smirnov test. For normally distributed variables, descriptive measures such as mean, SD and range values were calculated and mean values were tested using Student’s t test. The categorical variables were expressed as frequency and per cent values. Frequency data were compared using Chi-square/Fisher’s exact test as appropriate. To assess the relationship between the study variables and background characteristics, univariate and multivariable logistic regression analysis was performed. Unadjusted and adjusted odds ratios with 95 per cent confidence limits were calculated. A two-sided probability of P<0.05 was considered significant.

Results

The overall prevalence of depression and anxiety disorder in the study population was 38/680 [5.59%; 95% confidence interval (CI): 4-7.6%] with the specific prevalence of 2.2 per cent of PPD, 0.74 per cent of PPA and 2.79 per cent of both disorders. Two women had suicidal tendencies and were referred to a psychiatrist for treatment.

The mean age of women was 24 yr. Most women were multiparous (70%), lived in a joint family (83%) and belonged mostly to the upper middle class (49%). Comparative analysis of sociodemographic factors (Table I) between women with and without PPD/A/both indicated that illiteracy in women (OR: 3.2; CI: 0.9-9.2) and spouse (OR: 9.2; CI: 1.9-37.9) were both independent risk factors for PPD/PPA. Women from lower SES had a higher risk of PPD/A/both, with odds of 2.7 (CI: 1.2-5.9). Other significant risk factors were a history of substance abuse in the family (OR: 3.2; CI: 1.6-6.4) and primiparity (OR: 2.6; CI: 1.2-7.8). Most women had an uneventful pregnancy followed by vaginal delivery in both groups. More women in the PPD/A/both groups had an unplanned pregnancy and did not desire the present pregnancy, but the numbers were not significant.

Comparative analysis (Table II) indicated that women with PPD/PPA/both experienced significantly higher levels of parenting stress (P<0.001), poor lifestyle in the last two weeks (P<0.001), significantly less present (P<0.001) and expected (P=0.03) support from their partner and significantly less present support from parents-in-law (P=0.02) and expected support from parents (P=0.02). On analysis of categorical variables, it was found that women with PPD/A/both reported less marital satisfaction (P<0.001), significantly high IPV (P<0.001), poor bonding with their infants (P<0.001) and had more IFA (P=0.01).

Multivariate regression analysis (Table III) revealed that the significant factors associated with reducing the risk of PPD/A/both were found to be higher education (OR: 0.08, 95% CI: 0.02-0.35), high marital satisfaction (OR: 0.1, 95% CI: 0.05-0.19), more support from partner (OR: 0.4, 95% CI: 0.17-0.97) and more support from parents-in-law (OR: 0.4, 95% CI: 0.18-0.87). In addition, the univariate analysis found that the risk for development of PPD/A/both increased with higher intimate-partner violence (OR: 5.78, 95% CI: 2.92-11.4) and less support from friends (OR: 0.51, 95% CI: 0.26-1). However, these values were not found to be significant.

Discussion

The overall prevalence of depressive and anxiety disorder in the postpartum period in our population was 5.6 per cent, which is lower than in previous studies. A meta-analysis2 of 291 studies across 56 countries reported the global pooled prevalence of PPD as 17.7 per cent (3-38%), with considerable heterogeneity2. Hospital- and community-based studies in India have reported a prevalence ranging from 6 to 23 per cent23,24. The present study attempted to include the population with clinically significant depression and selectively exclude the population with depression-like symptoms. Most studies have diagnosed PPD/A through symptom-based measures that usually produce higher prevalence estimates. This is partly due to the inability to differentiate between clinically significant symptoms of depression with commonly reported difficulties. For this reason, self-reported questionnaires tend to estimate a higher prevalence rate than diagnostic tools7.

The present study excluded women with a previous history of psychiatric illness to ensure that symptoms were due to PPD rather than other repetitive depressive or bipolar affective disorder. Silverman et al7 found a relative risk of about 21 for PPD in women with known depressive disorders. In the present study, women were screened from four weeks to six months postpartum. This was done particularly to changes in maternal mood at different points after delivery, and marked variability in prevalence estimation was recorded. Studies have shown that the rate of PPD was higher immediately after birth and gradually decreased over time6. Moreover, since most women had uneventful pregnancies followed by vaginal delivery of healthy babies, a lower prevalence was reported in our population.

The present study shows that maternal anxiety and depression can co-exist. Although extensive research has been done on PPD, recent interest has shifted to PPA. It may present alone or as comorbidity with PPD. The prevalence of anxiety disorder is reported as 15-20 per cent25. It has been estimated that two out of three women suffering from depression in the postpartum period have comorbid anxiety disorder26. Thus, our results are aligned with these previous research findings.

Further, in developing countries, various other factors could contribute to PPD. For example, a comprehensive systematic review of studies published on PPD from India reported that risk factors associated with PPD were financial difficulties, domestic violence, history of psychiatric illness, marital conflict, lack of support from husband, female baby, lack of support from family, family history of psychiatric illness, sick baby or death of baby, substance abuse by husband, low birth weight or preterm baby, low maternal education, current medical illness, a complication in the current pregnancy, unwanted or unplanned pregnancy5. In addition, some factors, such as IPV and decreased support from parents-in-laws, were also found to be significant risk factors in the present study.

In general, 47 to 57 per cent of postpartum women achieve pre-pregnancy functional levels by six weeks postpartum and 76 per cent by 12 week17. In the present study, although the personal lifestyle of the woman was affected, women with PPD/PPA carried out their household responsibilities and cared for their babies.

Social support gives woman emotional and social resources to cope with stress and adjust to her new role as a mother. In the present study, women with PPD/A perceived getting less social support from partners and parents-in-law but expected more support from partners and parents alone reflecting on the cultural factors. It has been shown previously that perceived support from a partner lowers the odds of depression (OR: 0.23)27. Support from both partners and others are important. Women who receive no social support from either partners/others, have social support from a partner only and have social support from others only were 7.22, 2.34 and 3.13, respectively times more likely to have PPD, reported in a study from Japan28.

Further, studies have shown poor marital relationships as a consistent psychosocial predictor of PPD29. Major psychosocial stresses occurred in postpartum marital adjustment when partners were unsupportive and not involved in child-rearing29. Supporting this evidence, the present study also showed that marital satisfaction reduces the chance of PPD/A, while lack of support from the partner or IPV increases the risk.

Parenting stress is perceived discomfort or distress resulting from demands of childcare and interaction. A consistent association between PPD and parenting stress has been identified in various studies30, and the present study reflects on similar lines.

Poor bonding is characterized by a lack of maternal feeling, feelings of irritability, hostility, aggressive impulses or rejection towards the infant. Women with PPD report difficulty forming an affectionate relationship with the child and sometimes express a lack of love, rejection, neglect and impulse to harm the child31. These are known to have long-term implications on the cognitive development of the child. The present study and another study32 from Bangladesh are few to report on infant bonding in mothers with PPD/PPA.

The study presents one of the few pieces of research available on PPD/A conducted in a community setting which provides a comprehensive picture of not only the prevalence of PPD/A but also the various psychosocial factors as well. One of the main strengths of this study is the attempt to accurately study the PPD/A prevalence, which relied not only on the screening questionnaires but also the diagnostic tool. One of the critical limitations of the present study, however, was the inability to follow up with the women longitudinally to observe the course of the disorder and the long-term impact of PPD/A on the mental health of women and the care provided to the infant. Furthermore, despite sufficient sample size, since the obtained positive cases were limited a separate analyses to understand the specific psychosocial factors associated with each diagnostic group could not be carried out. Since, various new scales were employed for which neither translation nor adaptation was available, efforts towards scale translation could only be made using the WHO process of translation and adaptation of instruments. Finally, a comparative study with hospital-based prevalence could have added a more interpretational frame to our findings.

Overall, this study elaborates on the effect of mental health on a mother’s functional status, relationships and infant bonding in the Indian context. Therefore, it is essential to detect and treat depression and anxiety postpartum to avoid harmful effects on the mother and infant.