Mental health spillovers and the Millennium Development Goals: The case of perinatal depression in Khayelitsha, South Africa

Mental illness currently ranks among the top ten causes of burden of disease in low-income countries [1]. In the African region specifically, neuropsychiatric disorders account for approximately 5% of disability-adjusted life years lost, with nearly one-quarter of this burden attributable to unipolar depressive disorders [1]. Furthermore, this burden is projected to increase by 2030 [2]. There is accumulating evidence on the potential public health impact of scalable mental health treatments involving non-psychiatrists [3-5], with more studies under way [6-8], but overall the prevention and treatment of mental disorders have been relatively neglected in the global agenda [9,10].

A substantive portion of the burden of mental disorders in low-income countries is thought to be attributable to many of the failures of human development as targeted through the Millennium Development Goals (MDGs), including poverty, HIV, and gender inequality. The evidence on depressive disorders and depressed mood is most well developed in this respect (see Figure 1). Depression has been associated with economic deprivation, especially in low-income countries and with regards to specific indicators of deprivation such as food insecurity [12,13]. Depression is also a known consequent of poor physical health [14]. And finally, gender inequality [15], often manifested starkly as violence against women in low-income countries [16], is commonly conceptualized as a risk factor for poor mental health among women [17].

Figure 1.

Conceptual framework of multilevel influences on depression and corresponding types of interventions. Adapted from McKinlay & Marceau [11].

If these relationships were causal and unidirectional, then interventions targeting MDG indicators related to poverty, HIV, and gender inequality would be expected to reduce the burden of disease from mental disorders. However, some of these relationships are bidirectional, suggesting that scaling up interventions to improve mental health may support efforts to achieve the MDGs. Emphasizing these spillover effects on other health outcomes of greater political interest may be one effective strategy to build support for mental health programming [18]. For example, depressive disorders and depressed mood are associated with significant psychosocial disability resulting in reduced economic productivity [19]. Depressed mood among women in the postnatal period has been associated with elevated risks for diarrhea and poorer growth among their newborn infants [20-23]. And, among persons living with HIV/AIDS, psychological stress and poor mental health have been associated with reduced adherence to HIV antiretroviral therapy [24] and worsened HIV-related outcomes [25].

Photo: Courtesy of Dr Mark Tomlinson, personal collection

ADDRESSING PERINATAL DEPRESSION TO IMPROVE CHILD HEALTH

In order to concretely illustrate the potential contribution of mental health programming to achieving MDG targets, we sought to estimate the total burden of poor child health attributable to perinatal depression. To do this, we drew on our own experience conducting research on perinatal depression in Khayelitsha, a high-risk, peri-urban setting near Cape Town, South Africa (Table 1). In several studies we have conducted in this community, the prevalence of women meeting screening criteria for clinically significant depressive symptoms has ranged from 32–47% in the antenatal period [7,26-28] and 16–35% in the postnatal period [29-32]. Other researchers have employed similar methodologies and have obtained similar prevalence estimates [33,34]. The relevance of maternal mental health for child health has been demonstrated in a series of longitudinal studies showing that probable depression among mothers is associated with an approximately 2-fold increased risk of underweight status among their children [20-23].

Table 1.

Prevalence of perinatal depression in a peri-urban settlement near Cape Town

Source	Sample and timing	Findings
Antenatal assessment
Honikman et al., 2012 [26]	5402 women assessed during antenatal care	32% were referred to a counselor on the basis of EPDS screening and a risk factor assessment tool
Tsai et al., 2012 (personal communication)	461 women assessed during antenatal care	43% screened positive for significant depressive symptoms (EPDS≥13)
Rotheram-Borus, et al. 2011 [7,27]	1239 women assessed during second or third trimester antenatal care	42% screened positive for significant depressive symptoms (EPDS≥13)
Rochat et al., 2011 [28]	109 women assessed during antenatal care (third trimester)	47% met DSM-IV criteria for major depressive disorder

Postnatal assessment
Tomlinson et al., 2004 [29,30]	147 women assessed at two months postnatally	35% met DSM-IV criteria for major depressive disorder (18% with onset subsequent to delivery)
Cooper et al., 2002 [31]	32 women assessed at six months postnatally	28% met DSM-IV criteria for major depressive disorder
Cooper et al., 2009 [32]	184 women assessed at six months postnatally	16% met DSM-IV criteria for major depressive disorder

Given the high prevalence of perinatal depression and the strong association between perinatal depression and child underweight, it is clear that perinatal depression constitutes a substantial contributor to the burden of child underweight in peri-urban Cape Town. If, borrowing from the previously cited studies, we assume that perinatal depression and child underweight are associated with a relative risk of 2 and that the prevalence of perinatal depression ranges from 16–47% (Table 1), then we can apply standard formulas to obtain a population attributable risk (PAR) estimate ranging from 14–32%. If perinatal depression is causally related to child underweight, these estimates suggest that it is responsible for up to 14–32% of cases of child underweight in this community.

Further extrapolation to estimate the child mortality burden in South Africa that could be eliminated through successful scale-up of prevention or treatment of perinatal depression would require additional assumptions about the relationships between underweight and mortality, as well as about intervention efficacy in this context. However, given that approximately one-half of deaths of children under the age of five can be attributed to underweight [35-37] and that less than one-third of persons in South Africa with a severe mental disorder are estimated to be receiving needed care [38], we anticipate that scale-up efforts could potentially result in large gains relative to the status quo. The pace of progress toward MDG 4 has stalled in South Africa [39], further underscoring the potential for perinatal depression interventions to contribute toward achieving MDG 4 goals.

STRENGTHENING THE EVIDENCE BASE

While suggestive, these estimates are not conclusive, and more work needs to be done to confirm that these potential benefits could be realized in real-world settings. As shown in Figure 1, both the causes of depressed mood and the potential targets for mental health interventions can be conceptualized at several different levels [11]. Structural, psychological, and biological factors have all been shown to exert varying influences on mood [40]. Structural interventions aim to alter social structures or local contextual influences [41] that in some cases may be directly related to the MDGs. Individually targeted interventions aim to alleviate suffering that is rooted in psychological or somatic influences at the individual level, such as dysfunctional schemas or interpersonal difficulties. Mental health, in turn, influences access to and use of these bio-psycho-social resources [42], consistent with the spillover effects described in this essay.

In general, few mental health intervention studies have emphasized both mental health and non-mental health outcomes. Even fewer have assessed the extent to which improvements in non-mental health outcomes might be mediated by improvements in mental health [43]. For individual-level interventions, the results of randomized or econometric studies have been somewhat equivocal with regards to the spillover effects of depression treatment on MDG-related outcomes such as income generation and poverty reduction (MDG 1) [44], child health (MDG 4) [45,46], and ART adherence [47,48] and HIV acquisition risk [49] (MDG 6). Few systems-level interventions have been tested, but one recently published study showed that an innovative method of organizing the delivery of care by specialist and non-specialist health care workers can have beneficial impacts on both depression and economic productivity [50].

Even were the evidence base on mental health spillovers to be strengthened overnight, additional questions would need to be answered in order to determine how best to deliver these interventions in different contexts. Given the present lack of adequate mental health care systems financing and lack of adequate human resources for mental health in low-income countries, a scaled-up response will likely involve integration of treatment for mental disorders into primary health care settings [51]. Screening for mental disorders will need to be implemented at some level (eg, in the community, among primary health care attendees, etc.), but little evidence exists to inform programming in this area. In high-income countries, screening and case-finding interventions implemented in isolation (ie, without additional organizational enhancements) have not resulted in improved diagnostic or management outcomes [52]. Screening may potentially have benefits if integrated into wider enhanced-care programs [53,54], but few studies in low-income countries have incorporated these strategies into their design [4]. Screening instruments developed using study participants living in high-income countries will need to be adapted and validated in low-income countries [55], and separate evaluations of their test properties will be needed in order to ensure that screening yields a locally appropriate referral volume. Simply adding to the responsibilities of medical officers working within already overburdened primary health care systems is a non-starter. In order to address some of these needs, we are currently engaged in research on the use of lay health workers in community-based, perinatal care interventions [6-8,56].

CONCLUSIONS

Significant strides have been made in ensuring a greater prominence for mental health on the global agenda, reflected in the Lancet’s Global Mental Health series in 2007 [57] and 2011 [58], the PLoS Medicine Packages of Care series in 2009 [59], and the Grand Challenges in Global Mental Health initiative [60]. As of yet, however, significant commitments from global funding agencies such as the Bill and Melinda Gates Foundation have not been forthcoming. Clear priorities for mental health research in low-income countries have been identified [61]. In low-income countries, however, there are many barriers to the conduct and dissemination of mental health research [62], and there is a critical need to build organizational structures for research governance [63]. A comprehensive approach to the prevention and treatment of mental disorders would include interventions aimed at the multilevel influences on mental health and will require collaborative, interdisciplinary efforts involving both mental health and public health professionals.

In the years leading up to 2015, we hope that mental health advocacy will be intensified to ensure that programming and funding for prevention and treatment of mental disorders are not sidelined in future initiatives as they have been to date with regards to the MDGs [64] and non-communicable diseases [65]. Estimating the extent to which prevention and treatment of mental disorders potentially increase the probability of achieving indicators of political importance can capitalize on greater support for these other health goals [9,18,64]. Doing so, however, has the unattractive potential for instrumentalizing the alleviation of mental suffering and undermining concern for mental suffering for its own sake. We must not lose sight of our human development and public health priorities while also appreciating the human rights implications of taking action to mitigate one of the most common and disabling sources of human suffering worldwide.