In the most recent analyses of the Global Burden of Disease by the World Health Organization, mental and behavioural disorders are the largest contributors to the years lived with disability (YLD), accounting for 22.7 % of all YLD [1]. Depression is the most common of the mental health challenges [2] with the highest YLD (3.7 %), and is about twice as common among women compared to men [3]. Yet, to date, mental illness has not achieved commensurate visibility or funding compared to other diseases, particularly in low and middle income countries (LMIC) [4]. Despite the high prevalence of mental disorders and the known association with poverty [5], data on the consequences of depression, particularly maternal depression, is limited for LMIC [6].

Maternal depression has repeatedly been demonstrated to negatively impact both mothers and their children [7, 8]. Globally, children of depressed mothers show deficits in cognitive, physical, and social development, with the most significant deficits demonstrated among children in LMIC [9]. Children of depressed mothers consistently demonstrate mental health problems [10], yet the infant developmental outcome most affected by maternal depression remains unclear [11, 12].

This study focuses on infants of antenatally depressed mothers in a community sample of pregnant women in South Africa over the first six months of life. Though less well documented than postpartum depression, depression in pregnancy is significantly associated with adverse child outcomes [13–17], including poor child growth [18, 19]. However, depression also co-varies with a range of related stressors. Perinatal depressed mood is also associated with poverty and life stress [20] which are in themselves strongly related to poor child growth. In South Africa, up to 47 % of women have been diagnosed with major depression across the perinatal period [21–23]. Key aspects of infant development (e.g., growth) are compromised in the context of parental poverty and mental health problems [9, 24]. Concurrently, depressed mothers in LMIC are likely to experience overcrowding, food insecurity and poor sanitation [25]. Infant growth is more complex than being based solely on food security and caloric intake. Factors such as the health environment, availability of health care, and feeding and care practices of the mother influence infants’ growth.

In the context of antenatal depression and poverty, retarded infant growth has consistently been linked to maternal depression in Asia, with multiple studies in Pakistan [26] and Vietnam [27]. In Asia, the studies have been conducted with both clinical and community samples, and cross-sectional and longitudinal data. However, there have been inconsistent results regarding the impact of maternal depression on growth in South Africa [28], Ethiopia [27] and Malawi [29].

Identifying sustainable and effective strategies to improve maternal and child outcomes in LMIC is a high priority [30]. Intervention trials with postnatally depressed mothers have hypothesized very different mechanisms by which children are influenced by maternal depression [31]. Depressed mothers may provide sub-optimal care which has detrimental effects on the health of her child [6]. Depression may compromise care-giving behaviours (e.g., ensuring adequate hygiene, optimal nutrition through breastfeeding, immunization, recognizing illness and seeking care), as well as responsive parenting that is needed for physical and mental development of a child. In Pakistan the children of antenatally depressed mothers showed growth retardation in the first year of life [25]. When the mothers received home visits by CHW, maternal depression was reduced, but infants did not improve weight- or length-for-age over the first year of life [32].

Maternal and child health intervention strategies have tended to address only one health risk at a time. As a result many studies provide little insight on the confluence of challenges faced by mothers in LMIC. Our project evaluated a community-based, home visiting program utilising CHW who were trained to address multiple health challenges facing pregnant South African mothers and their newborn infants. The CHW were trained to conduct home visits that addressed nutrition, alcohol, and HIV among pregnant township women, but not maternal depression. The training approach aimed to teach principles of behaviour change [33], based on cognitive-behavioral principles (CBT). CHW supported the mother to problem solve her own situation, rather than to provide solutions for her daily stressors and health challenges. While the CHW did not focus specifically on depressed mood, many of the factors related to positive child outcomes were addressed by the CHW. We report the child growth outcomes over the first six months of life, based on intervention status and maternal depression.

The Institutional Review Boards of University of California Los Angeles (UCLA), Stellenbosch University, and Emory University approved the study, whose methods have previously been published [34]. Three independent teams conducted the assessment (Stellenbosch), intervention (Philani Maternal, Child Health, and Nutrition Project, hereafter referred to as Philani), and data analyses (UCLA).

Participants

In 2009, 26 township neighborhoods were matched on size (450–600 households), density, public utilities (water, electricity, toilets), distance of primary health care, and the number of alcohol bars. UCLA randomized matched pairs of neighborhoods to either the PIP or SC condition. The minimum number of pregnant women needed per neighbourhood to achieve 80 % power to detect a standardized effect size of 0.40 set the sample size; the original size was 1238. Participant flow through each phase of the study can be seen in Fig. 1.

The children of antenatally depressed mothers in the intervention condition were more often in the normative range for length for their age over time compared to children of depressed mothers in the control condition. Children of antenatally depressed children in the control condition tended to be significantly shorter at birth. Because only 16.5 % were significantly depressed (EPDS < 18), the power to detect significant differences in growth is substantially reduced. The rate of stunting (being two SD lower in height) was more than double at two weeks post birth (OR = 2.36) and similar 6 months later (OR = 1.8) compared to children of depressed mothers in the intervention condition. However, with clustering of neighbourhoods and the reduced sample size, the confidence intervals are too large to have precision in evaluating if this is a significant difference. These rates suggest that CHW may be able to reduce stunting of depressed mothers by focusing on encouraging depressed mothers to maintain caretaking. Our findings contribute to a growing body of literature examining the impact of community based interventions on child health in the context of perinatal depression [32, 43].

The intervention’s negative effect on WFHZ score is a reflection of the positive intervention effect on height (HAZ). Intervention children are taller and significantly less likely to be stunted, their weight for height is at the desirable mean score of 0 (normative weight) at 6 months post birth. Children of antenatally depressed mothers in the control condition were shorter, but of similar weight compared to the children of antenatally depressed mothers in the intervention condition and children of non-depressed mothers in both conditions. Therefore, when calculating WFHZ, the children are relatively heavier at their height, above the desirable mean of 0 and similar to the children of non-depressed mothers in both conditions.

The PIP intervention’s benefit in height was achieved in the context of a study where CHW were not specifically trained to address depression (through screening or treatment) and in which the prevalence of depressive symptoms were highly similar in the intervention and control conditions over time. A study by Rahman and colleagues [32] examined whether or not delivering a cognitive behavioural therapeutic intervention to depressed mothers using village CHW women would improve infant growth [32], and found that directly addressing depression had a large impact on reducing maternal depression, but had no impact on improving infant growth. The work of Tripathy and colleagues in Nepal, on the other hand [43] found significant reductions in neonatal mortality amongst moderately depressed women in a group- based intervention programme aimed at improving social support and problem solving [43]. Similar to findings of our South African study, this result was achieved without specifically targeting perinatal depression. It is likely that the psycho-educational activities of the PIP intervention providing skills to address alcohol, nutrition, and HIV enabled women to be better caretakers of their infants, in spite of being depressed.

Increasingly, antenatal depression is being seen as an important public health concern. The link between postnatal depression and compromised child development is well established [20, 44], but this is less the case with antenatal depression although this relationships is increasingly a target of investigation. Our data suggest that interventions during the antenatal period are crucial. Our findings provide support for training CHW as generalists, in order to address the highest priority health issues in a local community. Training of CHW to address caretaking, problem solving and support, appears to benefit child growth [41]. Treating depression does not necessarily remove the conditions that frequently cause or exacerbate depression – such as poverty and interpersonal violence. On the other hand, providing mothers the tools to help their infants thrive (something most mothers are highly motivated about) allows them to act toward a powerfully motivated goal. Indeed, an inability to help their child thrive may be a real driver of maternal depression. While our findings also provide evidence that there are benefits for child development in the context of perinatal depression, the primary preventive nature of community -based interventions, such as PIP, may not ameliorate moderate to severe depression.

Intervening in multiple health risks concurrently, even by paraprofessional CHW, appears to benefit child growth, even when mothers are depressed during pregnancy and remain depressed. The mechanism by which the caretaking influences height and stunting, but not weight gain is a mystery. It is surprising that the children of PIP mothers who were antenatally depressed had children who were less stunted but did not weigh as much as the children in the SC condition who had depressed mothers. In other analyses we have conducted [40], PIP mothers were significantly more likely to breastfeed their infants and to breastfeed longer. The children of antenatally depressed PIP mothers also breastfed their children longer. Breastfeeding often results in lower weights, which may explain that the intervention children were of normal weight per height.

More research is needed in order to establish the optimal mix of interventions that are sufficiently targeted (screening for perinatal depression and providing referral or treatment), while also being sufficiently generalised and horizontally integrated given the resources constraints of LMIC. Given the explanatory models of mental illness in many countries [11, 45] which link disturbance to social adversity and social problems rather than individualistic explanations, interventions targeted at these social determinants of health may be more acceptable to local communities. Given what is known about the causes of depression, it also may be more appropriate. As a clinical condition, depression is often highly stigmatized as an individual deficiency. In some contexts, interventions may be more efficacious when framed as addressing community challenges, rather than as a treatment program for individuals [11, 45]. The PIP model did not target individual depression, but aimed to improve parenting and child health. Every household was visited in the neighbourhood, regardless of such risks as HIV and alcohol use. The strategic approach adopted in the PIP program is similar to the “Thinking Healthy” programme of Rahman and colleagues [32] and uses cognitive behavioural change principles as a generic method of intervention for a broad range of social problems, rather than depression alone.

The relationship between maternal mental health and child growth is complex [12]. It is possible that interventions focussed on maternal depression may not fully address this complexity [46], while generalist models of implementation are not sufficiently targeted to make a significant impact on moderate to severe perinatal maternal depression. Evidence is needed on what the optimal implementation models are. The improvements we have seen in this study warrant continued tracking to examine sustainability and durability of an intervention’s impact when delivered by CHW.

Finally, Philani is a NGO which has survived 30 years. Over that time, it has developed core relationships with staff in the Department of Health, community leaders, church officials, clinics and a broad range of key community roles. Unlike many other NGOs, it has well-developed accountability and training systems for both CHW and supervisors. Diffusion of this model to other NGO without these strengths is likely to face many obstacles. The strengths of this study are that it is a cluster randomized controlled trial of mothers monitored during pregnancy and the similarity of the mothers and infants and the high retention in the intervention and the control condition are additional evidence of the strengths of this design. We did not conduct full diagnostic interviews for depression, which is a limitation of our current design.

We showed benefits to infant growth in the context of antenatal depression when community health workers intervened to address concurrent community health risks, but not specifically trained to address depression through screening or treatment. This adds to a growing body of evidence identifying sustainable and effective strategies to improve maternal and child outcomes in LMIC.