Postpartum depression in India: a systematic review and meta-analysis

1Postpartum psychiatric disorders can be divided into three categories: postpartum blues; postpartum psychosis and postpartum depression., Postpartum blues, with an incidence of 300‒750 per 1000 mothers globally, may resolve in a few days to a week, has few negative sequelae and usually requires only reassurance. Postpartum psychosis, which has a global prevalence ranging from 0.89 to 2.6 per 1000 births, is a severe disorder that begins within four weeks postpartum and requires hospitalization. Postpartum depression can start soon after childbirth or as a continuation of antenatal depression and needs to be treated. The global prevalence of postpartum depression has been estimated as 100‒150 per 1000 births.

2Postpartum depression can predispose to chronic or recurrent depression, which may affect the mother‒infant relationship and child growth and development.,– Children of mothers with postpartum depression have greater cognitive, behavioural and interpersonal problems compared with the children of non-depressed mothers., A meta-analysis in developing countries showed that the children of mothers with postpartum depression are at greater risk of being underweight and stunted. Moreover, mothers who are depressed are more likely not to breastfeed their babies and not seek health care appropriately. A longitudinal study in a low- and middle-income country documented that maternal postpartum depression is associated with adverse psychological outcomes in children up to 10 years later. While postpartum depression is a considerable health issue for many women, the disorder often remains undiagnosed and hence untreated.,

3The current literature suggests that the burden of perinatal mental health disorders, including postpartum depression, is high in low- and lower-middle-income countries. A systematic review of 47 studies in 18 countries reported a prevalence of 18.6% (95% confidence interval, CI: 18.0‒19.2). Scarcity of available mental health resources, inequities in their distribution and inefficiencies in their utilization are key obstacles to optimal mental health, especially in lower resource countries. Addressing these issues is therefore a priority for national governments and their international partners. The impetus for this will come from reliable scientific evidence of the burden of mental health problems and their adverse consequences.

4Despite the launch of India’s national mental health programme in 1982, maternal mental health is still not a prominent component of the programme. Dedicated maternal mental health services are largely deficient in health-care facilities, and health workers lack mental health training. The availability of mental health specialists is limited or non-existent in peripheral health-care facilities. Furthermore, there is currently no screening tool designated for use in clinical practice and no data are routinely collected on the proportion of perinatal women with postpartum depression.

5India is experiencing a steady decline in maternal mortality, which means that the focus of care in the future will shift towards reducing maternal morbidity, including mental health disorders. Despite the growing number of empirical studies on postpartum depression in India, there is a lack of robust systematic evidence that looks not only at the overall burden of postpartum depression, but also its associated risk factors. Our current understanding of the epidemiology of postpartum depression is largely dependent on a few regional studies, with very few nationwide data. The current review was done to fill this gap, by providing an updated estimate of the burden of postpartum depression in India, to synthesize the important risk factors and to provide evidence-based data for prioritization of maternal mental health care.

6Two authors (RPU and AP) independently searched PubMed®, Google Scholar and Embase® databases for articles on the prevalence of postpartum depression in India, published until 31 March 2016. The search strategy (Box 1) used subject headings and keywords with no language restrictions. Any discrepancy in the search results was planned to be discussed with a third author (AKR). We also searched the bibliographies of included articles and government reports on government websites to identify relevant primary literature to be included in the final analysis. For studies with missing data or requiring clarification, we contacted the principal investigators.

7(“depression” OR “depressive disorder” OR “blues” OR “distress” OR “bipolar” OR “bi-polar” OR “mood disorder” OR “anxiety disorder”)

8(“postpartum” OR “postnatal” OR “perinatal” OR “post birth” OR “after delivery” OR “after birth” OR “puerperium” OR “puerperal”)

9(“prevalence” OR “incidence” OR “burden” OR “estimate” OR “epidemiology”)

10(“India” OR “South East Asia”)

11(#1 AND #2 AND #3 AND #4)

12(Addresses[ptyp] OR Autobiography[ptyp] OR Bibliography[ptyp] OR Biography[ptyp] OR pubmed books[filter] OR Case Reports[ptyp] OR Congresses[ptyp] OR Consensus Development Conference[ptyp] OR Directory[ptyp] OR Duplicate Publication[ptyp] OR Editorial[ptyp] OR Systematic reviews OR Meta analysis OR Festschrift[ptyp] OR Guideline[ptyp] OR In Vitro[ptyp] OR Interview[ptyp] OR Lectures [ptyp] OR Legal Cases[ptyp] OR News[ptyp] OR Newspaper Article[ptyp] OR Personal Narratives [ptyp] OR Portraits[ptyp] OR Retracted Publication[ ptyp] OR Twin Study[ptyp] OR Video-Audio Media[ptyp])

13(#5 NOT #6) Filters: Original research; published in the past 15 years; humans

14For a study to be included in the systematic review, it had to be original research done in India, within a cross-sectional framework of a few weeks to 1 year post-birth. We excluded research done in a specific population, such as mothers living with human immunodeficiency virus; research including mothers with any current chronic disease. To have a fairly recent estimate of the burden of postpartum depression, we considered only studies published from the year 2000 and later. After initial screening of titles and abstracts, we reviewed the full text of eligible publications. Decisions about inclusion of studies and interpretation of data were resolved by discussion among the reviewers. Data from all studies meeting the inclusion criteria were extracted and tabulated.

15We used the Newcastle‒Ottawa quality assessment scale adapted for cross-sectional studies., The scale is used to score the articles under three categories: (i) selection (score 0‒5); (ii) comparability (score 0‒2 ); and (iii) outcome (score 0‒3); total score range 0‒10. The selection category consists of parameters, such as representativeness of the sample, adequacy of the sample size, non-response rate and use of a validated measurement tool to gather data on exposure. The comparability category examines whether subjects in different outcome groups are comparable based on the study design and analysis and whether confounding factors were controlled for or not. The outcome category includes whether data on outcome(s) were collected by independent blind assessment, through records or by self-reporting. The outcome category also includes whether the statistical tests used to analyse data were clearly described and whether these tests were appropriate or not. Two authors (RPU and KS) made separate quality assessments of the included studies. In case of any discrepancy, a third author (AP) was consulted. We grouped the studies into those with quality scores ≤ 5 and > 5.

16We did a meta-analysis of the reported prevalence of postpartum depression in the included studies. Heterogeneity between studies was quantified by the I2 statistic. We considered I2 values > 50% to represent substantial heterogeneity. The degree of heterogeneity among the studies was high (> 95%), and thus we used a random effects model to derive the pooled estimate for postpartum depression in mothers. The final estimates of prevalence were reported as percentages with 95% CI.

17We did a subgroup analysis by excluding articles in which depression was assessed within 2 weeks postpartum,,, since some researchers argue that it is difficult to differentiate postpartum depression from postpartum blues within 2 weeks of birth. In addition, the Edinburgh postnatal depression scale, which was used in the majority of studies we identified, can give false-positive results in the early postpartum period.

18We also did separate subgroup analyses on each of the following factors: place of study (geographical location; rural or urban; hospital or community); study instrument used; quality score of the articles; time of publication; and age of mothers. Not all the studies provided data on the mean age of the study participants that was required for subgroup analysis; however, the proportion of mothers in specific age ranges were available. Using this information, we estimated the mean age of the study participants. For studies that reported the prevalence of postpartum depression in mothers at different time points, we used the prevalence reported in the earliest time point to reduce the effect of lost to follow-up. We used meta-regression analysis to identify factors contributing to the heterogeneity in effect size, i.e. the pooled proportion of mothers with postpartum depression.

19We assessed publication bias with the Egger test and used a funnel plot to graphically represent the bias. Finally, we listed the risk factors for postpartum depression. We used Stata software, version 14 (StataCorp. LLC, College Station, United States of America) for all analyses.

20Of the 1285 articles we identified in our search, we screened 1248 titles of unique articles. Out of these, we reviewed 211 relevant abstracts, assessed 62 full-text articles for eligibility and included 38 articles in our final analysis.– (Fig 1). These 38 studies included data from 20 043 mothers in total. More of the articles (26 studies) were published in the most recent five-year period 2011‒2015 than in the earlier periods 2000–2005 (6) and 2006–2010 (6). The majority of studies were from south India (16 studies), followed the western (9) and northern regions (7) of the country; 24 studies were done in an urban setting and 29 in hospitals (Table 1; available at: http://www.who.int/bulletin/volumes/94/10/17-192237). In 19 studies, the mean age of the study mothers was ≤ 25 years. The Edinburgh postnatal depression scale was the most commonly used study instrument (29 studies). The median quality score for the studies was 5 (21 articles had a score of ≤ 5 and 17 had a score > 5).

21Based on the random effects model, the overall pooled estimate of the prevalence of postpartum depression in Indian mothers was 22% (95% CI: 19–25; Fig. 2). Eight studies included women reporting depression within 2 weeks of delivery. After excluding these, the pooled prevalence for the remaining 30 studies (11 257 women) was 19% (95% CI: 17–22; Fig. 3).

22The estimated overall pooled prevalence was highest in the southern region of the country (26%; 95% CI: 19–32), followed by eastern (23%; 95% CI: 12-35), south-western (23%; 95% CI: 19–27) and western regions (21%; 95% CI: 15–28; Table 2). The northern region of India had the lowest prevalence (15%; 95% CI: 10–21). The pooled prevalence was higher, but not significantly so, for studies conducted in hospital settings (23%; 95% CI: 19–28) than in community settings (17%; 95% CI: 13–22); Fig. 4; Table 2) and in urban versus rural areas (24%; 95% CI: 19–29 versus 17%; 95% CI: 14–21). Prevalence was 20% (95% CI: 16–24) and 21% (95% CI: 16–26) when studies with mean maternal age of ≤ 25 years and > 25 years were pooled respectively.

23Pooling of studies that used the Edinburgh postnatal depression scale as the study instrument produced a prevalence of 24% (95% CI: 20–28) compared with 17% (95% CI: 13–22) in those that used other study instruments (Table 2).

24Studies with a quality score ≤ 5 had a pooled prevalence of 22% (95% CI: 18–27) and those with a score > 5 had a prevalence of 21% (95% CI: 18–25).

25The studies had a high degree of heterogeneity (I2 = 96.8%). Both the Egger plot (Egger bias = 2.58; 95% CI: 0.83–4.33; Fig. 5) and the funnel plot (Fig. 6) showed evidence of publication bias.

26A total of 32 studies reported risk factors for postpartum depression. The risk factors most commonly reported were financial difficulties (in 19 out of 21 studies that included this variable), domestic violence (6/8 studies), past history of psychiatric illness in the mother (8/11 studies), marital conflict (10/14 studies), lack of support from the husband (7/11 studies) and birth of a female baby (16/25 studies). Other commonly reported risk factors were lack of support from the family network (8/14 studies), recent stressful life event (6/11 studies), family history of psychiatric illness (7/13 studies), sick baby or death of the baby (6/13 studies) and substance abuse by the husband (4/9 studies). Preterm or low birth-weight baby, high parity, low maternal education, current medical illness, complication in current pregnancy and unwanted or unplanned pregnancy and previous female child, were some of the other reported risk factors (Table 3).

27The pooled prevalence of postpartum depression in India in our meta-analysis was 22% (95% CI: 19–25). A systematic review of studies in 11 high-income countries showed that, based on point prevalence estimates, around 12.9% (95% CI: 10.6–15.8) of mothers were depressed at three months postpartum. Data from 23 studies conducted in low- and middle-income countries, which included 38 142 women, was 19.2% (95% CI: 15.5–23.0). Another systematic review from 34 studies found that the prevalence of common mental disorders in the postpartum period in low- and lower-middle income countries was 19.8% (95% CI: 19.2–20.6). These estimates in low- and middle-income countries are similar to ours and, taken together, they support an argument for placing greater importance on maternal mental health as part of overall efforts to improve maternal and child health.

28Although facility-based deliveries are increasing in many low- and middle-income countries, a high proportion of pregnant mothers still deliver at home. Beyond the lack of awareness of postpartum depression by health professionals, there are issues that may be barriers to prompt recognition and management of the illness.– In India, women who deliver at a health facility often stay for less than 48 hours after delivery. This leaves little opportunity for health personnel to counsel the mother and family members on the signs and symptoms of postpartum depression and when to seek care. In low- and middle-income countries, the proportion of women who visit the health facility for postpartum visits is generally low and consequently mental disorders often remain undetected and unmanaged, especially for those delivering at home. Analysis of demographic and health survey data from 75 countdown countries showed that postnatal care visits for mothers have low coverage among interventions on the continuum of maternal and child care Postnatal traditions, such as the period of seclusion at home observed in many cultures, can negatively affect care-seeking behaviour in the postpartum period. Furthermore, mothers may be reluctant to admit their suffering either because of social taboos associated with depression or concerns about being labelled as a mother who failed to deliver the responsibilities of child care. In the current public health system in most low- and middle-income countries, including India, primary-care workers are supposed to be in regular contact with recently delivered mothers. However, at postnatal visits community health workers tend to focus on promoting essential infant care practices, with lower priority given to the mother’s health., These factors might explain, to some extent, the lack of availability of reliable, routine data on the burden of postpartum depression in low- and middle-income countries.

29A strength of our study is the large sample of recently delivered mothers included in the review. This is probably the first review that documents the overall estimated prevalence of postpartum depression in India. The study has its limitations as well. Most of the studies included in the review did not provide effect sizes against the risk factors for postpartum depression and this precluded pooling of risk factors to provide an estimate. Most of the studies included in the review used the Edinburgh postnatal depression scale and the cut-offs used to label postpartum depression varied among studies. This could limit the internal validity of our findings. We observed significant heterogeneity in the results and performed subgroup analysis and meta-regression. The meta-regression analysis was able to explain < 10% of the heterogeneity and suggests that unidentified factors were causing such heterogeneity.

30Among the studies included in our review, risk factors for postpartum depression included financial difficulties, birth of a female child, marital conflict, lack of support from the family, past history of psychiatric illness, high parity, complications during pregnancy and low maternal education. Previous studies from low- and middle-income countries report similar risk factors.,

31We found relatively higher pooled proportion of postpartum depression in mothers residing in urban than in rural areas. This may be due to factors such as overcrowding, inadequate housing, breakdown of traditional family structures leading to fragmented social support systems, increased work pressure, high cost of living and increased out-of-pocket expenditure on health care. Pooling of hospital-based studies found comparatively higher estimates of postpartum depression than studies in community settings. It is likely that mothers suffering from any illness during the postnatal period, including postnatal depression, will seek care at a health facility, compared to physically healthy mothers and babies who may not visit a facility at all. Moreover, being in a hospital environment provides an opportunity for the mother to express her concerns and problems to the health personnel, but when interviewed at her home she may not admit to having depressive symptoms, owing to the presence of other family members or neighbours and the social stigma attached to mental health conditions.

32On subgroup analysis, we found a slightly higher proportion of postpartum depression in mothers who were aged > 25 years compared with those aged ≤ 25 years. Moreover, high maternal age emerged as a risk factor for depression in 4/28 studies which included this variable compared with 3/28 studies reporting low maternal age as a risk. Older mothers may suffer more from depression because they lack peer support or because they have more obstetric complications and multiple births or greater use of assisted reproductive technologies.– On the other hand, it is possible that depression among older mothers is simply a biological phenomenon.

33In our meta-analysis, geographical variation in the prevalence of postpartum depression was observed, with the highest prevalence in the southern regions. The observed differences in prevalence were not statistically significant on meta-regression and therefore more data are needed to document any significant geographical variations. The southern parts of the country have high literacy rates, which could lead to increased awareness about this health issue and therefore increased care-seeking. Moreover, the health system in southern India is more organized and there is comparatively better primary health-care provision than in other parts of the country and this could be a factor in greater care-seeking. South India also has a higher proportion of people living in urban slums compared with the northern parts of the country and greater rates of intimate partner violence.,

34We found that the number of studies on postpartum depression has seen an upward trend in the last five years. There were 26 published studies between 2011‒2016, compared with six each in the periods 2000‒2005 and 2006‒2010. This reflects a recent interest of the medical research community towards this important issue.

35There are a lack of data on perinatal mental health problems from low- and middle-income countries and this gap in the evidence hinders the process of establishing interventions to promote maternal psychosocial health. Gathering data on perinatal mental health issues will be essential in these countries, not only to gauge the magnitude of the problem, but also to inform policy-makers. Such evidence can stimulate governments to allocate resources for capacity-building in maternal mental health care, such as developing and implementing guidelines and protocols for screening and treatment, and setting targets for reducing the burden of postpartum depression.