In a large study involving >40,000 pregnant women, Milgrom et al. (2008) reported factors found to have predictive value for postnatal depression. Six variables were found: antenatal depression, antenatal anxiety, major life events, practical/emotional support, partner support and previous history of depression.

Whist these factors emerge consistently in studies of psychosocial risk factors in depressed mothers, other important risk factors that will impact on a woman’s parenting capacity include a mother having had a poor relationship with their own mother, any history of or current domestic or family violence (Howard et al. 2013a), childhood abuse (Banyard et al. 2003; Pears and Capaldi 2001) or drug and/or alcohol abuse (Ross and Dennis 2009). These experiences, especially those in early life, may contribute to a reduced competence in parental care and interactive abilities (Raphael-Leff 1991). In addition both depression and parenting need to be understood in the context of a woman’s cultural and religious beliefs, major traumas (including migration due to war) as well as the experience of growing up with family members struggling with their own mental health problems (Stein et al. 2014).

Howard et al. (2014) in reviewing the evidence conclude that a history of any psychopathology and psychosocial adversities are key predictors of mental health disorders in the perinatal period, but few systematic reviews exist, and there is a need for more studies using standardised diagnostic measures, longitudinal approaches and comparison groups. The impact of psychosocial risk factors on the infant warrants further investigation.

A wide range of stressors including a difficult or premature birth, illness or an infant with a difficult temperament (e.g. infants who are very reactive) may reduce a mother’s capacity to respond to her infant as well as her feelings of competence (Teti et al. 1990). Feelings and symptoms of depression may be precipitated in vulnerable mothers.

The extent of the impact of depression will be moderated by a number of factors that can both accentuate difficulties or be protective.

Maternal cognitive style may moderate the effects of depression on parenting (Whitton et al. 1996; Cutrona and Troutman 1986). For example, maternal self-efficacy in particular may influence how a mother interacts with her infant and behaves in ways that lead to low rates of reinforcement from their infants.

In addition, a mother with a poor sense of competence may think her baby would be better off without her and feel a sense of failure as a parent, further contributing to her depression (Cox 1988). Cognitions regarding close relationships have been found to raise the risk for depression (Murray 2014; Gjerde 1995; Gore et al. 1993). On the other hand, high maternal self-efficacy could serve as a protective factor for coping with a demanding infant making it less likely she will experience depressed mood.

Other mediating variables may include parenting skills and resources. Stein et al. (2014) recently suggested that ‘maternal programming’ may differentiate some mothers who have a predisposition to positive parenting. Maternal reflective functioning (RF) capacity or ‘mentalisation’ has also been proposed as a mediating factor. Fonagy and colleagues (Fonagy et al. 2002) define the construct as referring to the mother’s capacity to understand her baby’s behaviour whilst considering the baby’s underlying mental states and intentions. There is a growing literature suggesting that maternal reflective capacities make it possible to accurately recognise and sensitively respond to the baby’s internal states of arousal and are the key to sensitive care giving and ultimately secure attachment. Further, a parent high in RF can see their child as different from themself in its needs, feelings, desires and intentions. Maternal representations of the baby and maternal caregiving behaviour are thought to be the manifestations of reflective functioning (Slade et al. 2005; van Ijzendoorn 1995). Grienenberger et al. (2005) have proposed that reflective functioning is a mechanism in the transgenerational transmission of attachment.

Similarly, social support including parenting alliance may have a protective function partly by enhancing maternal self-efficacy beliefs (Abidin 1995). Satisfaction with partners has been found to be correlated with positive mother-infant interactions (Klaus et al. 1982).

Impaired mother-infant relationship may be a prime mechanism whereby postnatal depression and anxiety exert an impact on child development.

As described earlier, there is evidence that poor sensitivity to infant distress is a predictor of a child’s early social, emotional, cognitive and behavioural self-regulation (Parfitt et al. 2014; Leerkes 2011; Clark et al. 2008; Cirulli et al. 2003; Stams et al. 2002). In a study designed to test the mediating role of early interactions in a sample of women with PND, poor child cognitive and behavioural outcomes at 4 years of age were found to be mediated by maternal responsiveness at 3 months postpartum (Milgrom et al. 2004, 2006).

Recently, Murray et al. (2015) have speculated that whilst maternal responsiveness (or contingency) appears to be important for child cognitive development, other interactional behaviours may also influence later cognitive performance, such as the mother’s ability to support the infant’s attention and engagement with the environment (Eshel et al. 2006; Slater 1995; Fagen and Ohr 1985). On the other hand, parental sensitivity to the infant’s emotions may assist in developing emotional regulation, and being affectively attuned to the child’s behaviour may provide ‘emotional scaffolding’ where an infant’s difficult emotions such as anxiety are contained.

A direct effect on brain development is also implicated. There is a growing awareness of the importance of early experiences in shaping infant brain development, based on animal studies as well as emerging evidence from human infants. During the first 3 years of life, brain development is at its fastest and the brain is at its most malleable showing evidence of plasticity and susceptibility to stress. Stressful experiences may influence brain organisation, and disrupted mother-infant interaction may be an important source of stress (Mustard 2008; Cirulli et al. 2003; Perry et al. 1995). Brain development in turn underpins cognitive, behavioural and social development. Recently, we were able to demonstrate that sensitivity training for parents in the Neonatal Intensive Care Nursery, involving assisting parents to recognise subtle premature infant cues, resulted in improved brain connectivity on MRI at 40 weeks gestation (Milgrom et al. 2010).

Dysfunctional mother-infant interactions may be resistant to change due to ongoing negative vicious cycles of repeated interactive failures (Meager and Milgrom 1996). Interactional difficulties appear to persist even when depression improves.

‘A mother preoccupied with her sadness finds it difficult to engage with her infant who in turn may learn to gaze avert and may be less responsive. These infant behaviours then lead to feelings of rejection in the mother and further withdrawal on both parts’ (Milgrom et al. 1999). As early as 3 months of age, infants of depressed mothers appear to generalise their depressed style of interaction to non-depressed adults (Field et al. 1985). Mothers may depend on the responses they get from their babies for a sense of their own competence or effectiveness as a parent (Beebe 2010).

Importantly, chronicity of depression appears a key factor. Postnatal depression has shown consistent association with a range of poorer cognitive outcomes in children, with the persistence of the depression of particular importance. Other factors that may influence the relationship between depression and child development include socio-economic status (SES), gender, antenatal complications, illness, prematurity and genetic factors (Stein et al. 2014).

Currently, treatment of postnatal depression generally targets maternal mood without a focus on the infant. Given the accumulating evidence that maternal depression impacts on child cognitive, behavioural and socio-emotional development and the likelihood that sensitivity in the mother-infant interaction is important, a number of studies have explored whether addressing the mother’s depression alone improves outcomes for the infant. In a randomised trial of specialised treatments for PND (Milgrom et al. 2005), 162 mothers with PND were randomly assigned to a number of treatments including CBT (n = 46) and routine care (n = 33). Despite the success of these interventions in treating the depressive episode, the mother-infant relationship was not substantially changed (Milgrom et al. 2006). Whilst 73 % of depressed women had a dysfunctional relationship with their baby before treatment, following treatment 56 % continued to have dysfunctional relationships (compared to 3 % of non-depressed women; Milgrom et al. 2006). This is despite a significant reduction in depression below clinical thresholds. It appears that once relationship difficulties are triggered by PND, they may persist, despite improvement in maternal depression. Other studies report similar findings, using a range of depression treatments, and all show that a substantial proportion of disturbed mother-infant interactions persist without direct intervention (Forman et al. 2007; Cooper and Murray 1997).

As a result, there is growing interest in the evidence for the effectiveness of mother-infant interventions. A number of reviews and meta-analyses of mother-infant treatments have been conducted (Barlow et al. 2015; Doughty 2007; Bakermans-Kranenburg et al. 2003) although many of these interventions have not been developed or evaluated in the context of PND.

Existing programmes include interactional coaching (Field 1997); ‘Wait, Watch and Wonder’ (Muir 1992); ‘Brazelton Neonatal Assessment Protocol’ (Brazelton et al. 1974); parent-infant psychotherapy (McDonough 1993; Fraiberg 1980); and interventions targeting maternal reflective functioning such as ‘Minding the Baby’ (Slade et al. 2005; Marvin et al. 2002) based on theories and developmental research describing the elements of a ‘good-enough’ parent interaction (Fonagy et al. 1995; Stern 1985; Winnicott 1965).

In a major systematic review of interventions that included mother-infant outcomes in the context of PND, only eight trials met inclusion criteria, but all interventions showed some improvements in mother-infant relationships (Poobalan et al. 2007). A number of studies have also shown beneficial effects in child cognitive, emotional and social development (e.g. Field et al. 2000; Cohen et al. 1999). However, only two studies have evaluated brief interventions in RCTs and a recent Cochrane Review describes these latest developments in parent-infant psychotherapy for improving parent and infant well-being (Barlow et al. 2015; Murray et al. 2003; Horowitz et al. 2001).

Surprisingly, many evaluations of mother-infant treatments have not specifically combined treatment of maternal mental health with mother-infant relationship difficulties. This is despite the finding that chronicity of depression is a key factor in later child outcomes.

The HUGS programme (Milgrom et al. 2006) is a brief 4-session intervention addressing mother-infant difficulties. This approach is in line with the conclusions by Bakermans-Kranenburg et al. (2003) that short-term mother-infant interventions (i.e. less than five sessions) appear to be as effective as 5–16 session interventions and more effective than long-term (16+ sessions) interventions.

The HUGS (Happiness, Understanding, Giving and Sharing) programme is novel in that mother-infant work is combined with treatment of maternal depression and anxiety postnatally. Although full remission of depression is not necessary to benefit from mother-infant intervention, some reduction in depressive symptoms is desirable as emotional unavailability is likely to seriously interfere with successful interactions. The 4-session HUGS programme is added to a well-evaluated 9-week cognitive-behavioural therapy intervention for PND (Milgrom et al. 2005) and builds on the therapeutic ‘momentum’ and skills learnt in the 9-week programme. Two essential elements central to the mother-infant relationship are targeted: (i) parent skills in communication, observation and responsiveness and (ii) distorted maternal internal representations. Exercises build on the pioneering work of Field (1997), who showed behavioural-level skills, can profoundly change interactions. Even small changes (around 10 %) in maternal responsiveness directly mediate substantial differences in outcome at 4 years (e.g. IQ) for children of depressed mothers (Milgrom et al. 2004). Thus, key behaviours are targeted for change in a brief intervention in order to begin a trajectory of positive interactions and break the cycle of negativity.

Initial findings suggest HUGS is a promising intervention for improving mother-infant relationship difficulties associated with depression (Milgrom et al. 2006). The weekly rate of improvement during HUGS (4.9 points per week) on the Parenting Stress Index was more than threefold higher than during the 9-week PND programme (1.6 points per week) and fell below threshold for dysfunctional interactions. In a playgroup adaptation of HUGS (Community HUGS), elements of the maternal depression treatment, play and movement were combined with HUGS to form a 10-week playgroup for women with adjustment difficulties. Post-treatment scores on the PSI short form dropped below the threshold (<90) for clinically significant levels of parenting stress (pretreatment: M = 91.56, SD = 17.67; post-treatment: M = 81.00, SD = 16.72) (Ericksen et al. in preparation).

In summary, although encouraging, existing mother-infant intervention studies have been limited methodologically, most notably non-randomised designs and small sample sizes. Further research with larger sample sizes and infant follow-up is needed, and a current study is underway (Milgrom and Holt 2014).

Research into the role of fathers’ mental health in outcomes for infants and the impact on their relationship with their partners is a growing field (Ramchandani et al. 2011, 2013) but outside the scope of this chapter. A brief description of pertinent issues is provided below.

In the perinatal period, around 5 % of new and expectant fathers will experience depression, anxiety and other forms of emotional distress (Condon 2004; for a more detailed review of the incidence of paternal perinatal depression, see Fletcher et al. 2015). Paternal mental health disorders (depression, anxiety and other forms of distress) are also associated with increased risk of emotional and behavioural disorders in offspring (Gutierrez-Galve et al. 2015; Velders et al. 2011) including language development (Paulson et al. 2009) and depression at 18 years (Pearson et al. 2013).

The review by Stein and colleagues (Stein et al. 2014) concludes that maternal and paternal postnatal depression have similar effects on the child’s behavioural outcomes but that maternal depression has a greater effect on the child’s emotional development. The younger the child in the study, the greater the effect of maternal depression, with older children more affected by paternal depression (Connell and Goodmans 2002).

Depression in fathers may also contribute to maternal distress, just as maternal depression can affect their partner’s mental health. Living with a depressed partner is itself a risk factor for depression (Burke 2003). Thus, given the reciprocal relationship between maternal and paternal depression/distress, it is important to also assess whether the partner’s mental health is an issue that needs to be addressed, either in the context of the mother’s presentation or independently. In addition the relationship between the couple may be at risk (Milgrom and McCloud 1996).

Psychotic illnesses, mainly schizophrenia and bipolar mood disorder, but also those related to substance misuse, clearly make a major impact on women who are unfortunate enough to suffer from these conditions which as a group are traditionally called the severe mental illnesses (SMI). For many decades it has been observed that deleterious impacts are highly likely to extend to their infants and families (Sved Williams 2004). Until relatively late in the twentieth century, genes and environment were dichotomised in debates about causation of problem outcomes in offspring. As science has expanded on many fronts including genetic, molecular, neuroscientific, radiological and epidemiological, more sophisticated models focus on the interaction between heredity and environment and how one influences the other in patterns which are likely to begin antenatally and then flow through into early postnatal life (Saffery and Novakovic 2014). In this section some of these more general issues will be summarised before providing more specific information relating to each of the SMIs mentioned already and also to borderline personality disorder.

Winnicott’s goal of good-enough parenting is generally used to consider parents tuning in sensitively to their infants enough of the time to meet their needs (Winnicott 1965), as described earlier. Sadly, for parents with SMI, Maslow’s hierarchy of needs (Maslow 1943) may provide a better framework to consider whether outcomes for infants of parents with mental illness (IOPMI) are good enough, as it is clear that SMI can interfere with provision of adequate parenting at all levels including at quite fundamental levels of need, such as the need for safety, adequate food and sleep and appropriate accommodation.

In broad general terms, relevant individual factors determining outcome for infants include genetic inheritance of both parents, risk and protective factors for physical and emotional health during pregnancy and birth and early life experiences related to attuned parenting. Beyond this, epidemiological studies clearly show that socio-economic status, parental education and single-parent status have a cumulative effect on child health (Bauman et al. 2006), and these risk factors are frequently present in marginalised women with SMI. Cutler et al. describe the crucial role that early life plays in the co-evolution of socio-economic status and adult health (Cutler et al. 2008). People with SMI are more likely to be unemployed, marginalised by society and living in poverty, and so these general factors are likely to impact substantially. The complex patterns of factors involved at individual and societal levels are extremely well summarised diagrammatically by Leight et al. (2010) and are consistent with Maslow’s (1943) hierarchy of needs beginning with basic physiological needs (food, water, etc.), to safety (resources, health, property, etc.), love (friendship, family, etc.), esteem (self-esteem, respect by others, etc.) and self-actualisation (creativity, acceptance, etc.). Leight et al. (2010) draw on the original concepts from Misra et al. (2003) and present an integrated perinatal health and mental health framework beginning with distal factors (genetic, physical and societal) to proximal factors (biomedical and behavioural/psychological) and processes. Perinatal events and outcomes then follow (such as perinatal mental health issues), as also described by Milgrom et al. in Fig. 4.1.

Studies of pregnant women with SMI often focus on medical outcomes of pregnancy, in particular low birth weight, prematurity and neonatal deaths. There is no doubt that prematurity adds to the potential for negative outcomes on a large range of fronts including perinatal death, neurodevelopmental difficulties, breastfeeding problems, cerebral palsy, asthma in childhood, poor school performance, schizophrenia and young adult diabetes (Machado et al. 2014; Allen et al. 2011).

Neurodevelopment is increasingly better understood through a range of complex investigations and studies across several disciplines. Marques et al. (2013) review maternal nutrition perinatally and its interrelationship with immune system development and neurodevelopmental disorders. Sherman et al. (2015) focus on gut microbiota specifically and the interplay with neurodevelopment. The focus for Swain et al. (2014) is the oxytocin system postnatally in both mothers and fathers. Functional magnetic resonance imaging (fMRI) provides insights into the brain circuits in fronto-limbic systems involved in parental responses to infants and the interrelationship with oxytocin in parental responsiveness to infant cues. They propose that these factors may be modifiable, for instance, with the administration of oxytocin.

Thus, there is increasing knowledge at micro- and macrolevels of some of the potential mechanisms by which healthy pregnancy and early life and its antithesis prematurity, low birth weight and poor maternal care may influence life-long developmental patterns, as first suggested by Barker (1990). For women with SMI, risk factors of prematurity and low birth weight are further complicated by higher rates of smoking (Judd et al. 2014; Nguyen et al. 2013), substance use and psychotropic medication as well as other lifestyle factors of intimate partner violence (Frayne et al. 2014), presumably with higher levels of risk. Much is still unknown; for instance, a recent review (McColl et al. 2013) found insufficient research on the nutritional state of pregnant women with SMI to form any conclusions regarding status. Research to provide more factual information and further elucidate the interplay between these many individual and general factors will provide more answers in the next decade about pathways to better outcomes.

For many perinatal women, the effects of anxiety and depression make for lives with great unhappiness and infants whose developmental pathways may be compromised. For women with ongoing chronic depression and anxiety, the effects are greater. Nevertheless, researchers are continually progressing effective treatments for these conditions in combination with parent-infant interventions with promising results. The role of fathers in both the genesis of problems and protection from problem outcomes is being elucidated. At the societal level, beneficial effects are being seen and approaches to stigma have been partially successful in some countries (Reavley and Jorm 2012), and this in turn aids campaigns focused on early interventions generally in families with young children (1001 Critical Days on http://​www.​1001criticaldays​.​co.​uk). In many families targeted by this campaign, high stress, anxiety and depression are likely to be present. Finally, emphasis must be given to the usually positive outcomes for infants, both with and without intervention, in families with risk factors but also strengths, reflective capacity and resilience. Risk factors are not synonymous with poor outcomes, and research will continue to define both the minimisation of risk and the maximisation of resilience to change the balance of outcomes more.

For women with severe maternal mental illnesses (SMIs), there are clearly increased and varied effects on infants. The most severe include the effects of compromised pregnancy, in particular LBW and prematurity, to heightened chances of early death as adults from a large range of mental and physical health problems. In the first postnatal year, many authors have found deficits in parenting, including direct harm to infants, problems in attunement with compromised attachment relationships and life patterns of developmental dysregulation.

What hope is there for families with a member with SMI? Older notions of stigma abound. Historically, a mad woman with a child has not been treated kindly. Much of the literature on women with SMI is deficit based and problem focused. There are however more positive ways to view women with SMI and their offspring, beginning with the rapid inroads which science is making to elucidate the mechanisms by which problems occur. Some are beginning to propound changes during pregnancy, e.g. to diet (Prescott 2015) and other health-related behaviours (e.g. Davis et al. 2014). Others are moving to ensure best care of women with SMI perhaps beginning preconceptually (Temel et al. 2014) and then in pregnancy ensuring coordinated care (Nguyen et al. 2013) as well as psychoeducational interventions directed towards ceasing or reducing toxins such as cigarettes, alcohol and illicit substances. Postnatally treatment will vary depending, for instance, on the availability of mother-baby units or other intensive postnatal evidence-based treatments both for the illness itself and for psychosocial support. Negative attitudes towards mental illness generally must be tackled perhaps particularly in low-income countries where outcomes are more compromised as well as ensuring appropriate treatment for sufferers (Stein et al. 2014). A focus from forward thinking world leaders such as Seeman and Howard et al. on kindly and humane treatment of mothers with SMI provides inspiration towards improving health and social circumstances worldwide for this compromised group.