global mental health

ETIOLOGY
ORIGINAL RESEARCH PAPER
Household water insecurity, missed schooling, and the mediating role of caregiver depression in rural Uganda
C. E. Cooper-Vince1,2*, B. Kakuhikire3, D. Vorechovska1, A. Q. McDonough4, J. Perkins1, A. S. Venkataramani1,2, R. C. Mushavi2, C. Baguma3, S. Ashaba3, D. R. Bangsberg3,5 and A. C. Tsai1,2,3
1 Massachusetts General Hospital, Boston, MA, USA 2 Harvard Medical School, Boston, MA, USA 3 Mbarara University Science and Technology, Mbarara, Uganda 4 Bloomberg School of Public Health, Johns Hopkins University, Baltimore, Maryland, USA 5 School of Public Health, Oregon Health Sciences University-Portland State University, Portland, Oregon, USA
Global Mental Health (2017), 4, e15, page 1 of 9. doi:10.1017/gmh.2017.14
Background School attendance rates in sub-Saharan Africa are among the lowest worldwide, placing children at heightened risk for poor educational and economic outcomes. One understudied risk factor for missed schooling is household water insecurity, which is linked to depression among women and may increase children’s water-fetching burden at the expense of educational activities, particularly among children of depressed caregivers. In this study conducted in rural Uganda, we assessed the association between household water insecurity and child school participation and the mediating pathways behind these associations.
Method We conducted a population-based, cross-sectional study of female household heads (N = 257) and their children ages 5–17 (N = 551) in the rural regions surrounding the town of Mbarara, in southwestern Uganda. We used multivariable linear regressions to estimate the association between water insecurity and missed schooling. We then assessed the extent to which the association was mediated by caregiver depression.
Results Among children, water insecurity had a statistically signiﬁcant association with the number of missed school days (a standard deviation increase in water insecurity resulted in 0.30 more missed school days in the last week). The estimated association was partially mediated by caregiver depression. When stratiﬁed by sex, this mediating pathway remained signiﬁcant for boys, but not among girls.
Conclusions Water insecurity is a risk factor for missed schooling among children in rural Uganda. Caregiver depression partially mediated this relationship. Also addressing caregiver mental health in water insecure families may more fully address the needs of sub-Saharan African families and promote educational participation among youth.
Received 6 March 2017; Revised 3 June 2017; Accepted 20 June 2017
Key words: Child, depression, schooling, water insecurity.

* Address for correspondence: C. E. Cooper-Vince, Ph.D., Chester M. Pierce MD Division of Global Psychiatry, Massachusetts General Hospital, 5 Longfellow Place, Suite 215, Boston, MA 02114, USA.
(Email: ccooper-vince@mgh.harvard.edu)

Fifty-nine million primary school-aged children worldwide are out of school (UNESCO, 2015), with the lowest rates of attendance in sub-Saharan Africa (UNICEF, 2015a, b). School attendance is important for healthy

© The Author(s) 2017. This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution, and reproduction in any medium, provided the original work is properly cited.

global mental health

child development and is linked to increased earnings and health of oﬀspring later in life (Psacharopoulos & Alam, 1991; Hargreaves et al. 2008; Henley et al. 2010; Wood et al. 2012; Cortina et al. 2013; Abdullah et al. 2015; Devkota & Panda, 2016; Kanté et al. 2016; Mehta & Preston, 2016; Richter et al. 2017). In sub-Saharan Africa, missed schooling is associated with child engagement in water fetching (Burke & Beegle, 2004; Hemson, 2007; Dreibelbis et al. 2013). This task that is often completed by children, particularly girls, in support of women (Sorenson et al. 2011; Graham et al. 2016). In one study conducted in rural Ethiopia, nearly 20% of households reported that girls missed school to help with water fetching, while only 5% of households reported that boys missed school for this reason (Stevenson et al. 2012). These ﬁndings suggest that girls may be at elevated risk for missed schooling in East Africa. As one-third of families in sub-Saharan Africa dedicate over an hour a day to water fetching (Sorenson et al. 2011; Graham et al. 2016), insecure water access poses a signiﬁcant risk to child educational outcomes.
Little is known about how household water access drives problems with child schooling attendance. In sub-Saharan Africa, nearly 350 million people lack reliable access to safe drinking water (WHO & UNICEF, 2014). The limited or uncertain availability of safe water acquired via culturally acceptable means, termed water insecurity, is linked to depression and anxiety among women in low- and middle-income countries (LMICs), as they bear the primary water fetching burden globally and in sub-Saharan Africa (Wutich & Ragsdale, 2008; Hadley & Wutich, 2009; Sorenson et al. 2011; Stevenson et al. 2012; Graham et al. 2016; Tsai et al. 2016). This distress is in part associated with women’s failure to fulﬁll domestic roles (e.g. washing, cleaning, and cooking), acts of hospitality, and hygiene due to limited water (Stevenson et al. 2012), which mirror and may further exacerbate any observed domestic impairments among depressed women (Bolton et al. 2002).
In high-income countries, maternal depression is linked to poor child school attendance and academic performance (Guevara et al. 2013; Claessens et al. 2015). In both high- and low-income countries, poor maternal mental health is a well-established risk factor for adverse child developmental outcomes, which are central to school-readiness and academic performance (Petterson & Albers, 2001; Parke et al. 2004; Tsai & Tomlinson, 2012; Choe et al. 2014; Verkuijl et al. 2014; Ng et al. 2015; Tomlinson & Morgan, 2015; Bennett et al. 2016; Familiar et al. 2016; Huang et al. 2017). Further, the cognitive and psychomotor symptoms of depression (i.e. fatigue, psychomotor slowing, impaired concentration/decision-making) in combination with the domestic and childrearing impairments

(e.g. cooking meals, bathing children, and laundering clothes) found among depressed mothers in sub-Saharan Africa are likely to undermine their children’s ability to arrive at school on time appropriately fed, bathed, and dressed (Cooper et al. 1999; Bolton et al. 2002) and inhibit parental involvement which is known to support child academic engagement and success (Wang & Sheikh-Khalil, 2014).
No studies have attempted to integrate these disparate strands of literature and assess the extent to which caregiver depression mediates the association between household water insecurity and adverse child schooling outcomes. To address this gap in the literature, we analyzed data from a cross-sectional, populationbased sample of female household heads and schoolage children in a rural community in southwestern Uganda. We hypothesized that household water insecurity would be associated with child school attendance and this relationship would be mediated by caregiver depression. We also hypothesized that the mediating eﬀect would be stronger for girls.
Method
Participants
The present study was conducted in Nyakabare Parish, Mbarara District, Uganda, which is located approximately 260 km southwest of the capital city of Kampala. The Parish is located approximately 20 km outside of Mbarara Town, the district’s primary commercial hub. The economy is based primarily on subsistence agriculture, and both food and water insecurity are common (Tsai et al. 2011, 2016).
Among the 758 households in the Parish, we identiﬁed all 358 households in which there was a child younger than 5 years of age and a woman of reproductive age (18–49 years, or emancipated minors aged 16–18 years) who considered Nyakabare her primary place of residence, and who was available and capable of providing consent for interviewing. In households with multiple eligible women, we interviewed the oldest woman in that age range. Therefore, all women included in this study were female caregivers of the children assessed, though not all were the biological mother of the child. Potential participants meeting these eligibility criteria were approached by a research assistant ﬂuent in the local language (Runyankole) to request their participation in the study. Written informed consent for study participation was obtained from those expressing interest. Those who were not able to provide written consent due to literacy concerns were permitted to indicate consent with a thumbprint. Once enrolled, each study participant completed an individual

global mental health

interview in a private location beyond the earshot of others. Interviews were completed with 257 women who provided data on 551 school-age children (5–17 years) living in their households.
Ethical approval for all study procedures was obtained from the Partners Human Research Committee, Massachusetts General Hospital; and the Institutional Review Committee, Mbarara University of Science and Technology. Consistent with national guidelines, we received clearance for the study from the Uganda National Council for Science and Technology and from the Research Secretariat in the Oﬃce of the President.
Measures
Water Insecurity (WI) was measured with the Household Water Insecurity Access Scale (HWIAS; Tsai et al. 2016). The HWIAS is an eight-item self-report measure of household water insecurity, with possible total scores ranging from 0 to 24. The items of the HWIAS were developed based on the items of the Household Food Insecurity Access Scale (Swindale & Bilinsky, 2006), and assess perceptions of insuﬃcient quantity or quality of water, feelings of uncertainty or anxiety over water access, and strategies for coping with insuﬃcient water for completing water-based tasks. The HWIAS has been validated for use among Runyankore-speaking populations in Uganda (Tsai et al. 2016).
Caregiver Depression was assessed with the Hopkins Symptom Checklist-Depression Subscale (HSCL-D). The HSCL-D is a 15-item self-report of depressive symptoms taken from the short form of the HSCL-25, which assesses both anxiety and depression (Derogatis et al. 1974). HSCL-D assess the frequency of symptoms of depression in the past 7 days, such as ‘felt weak and low in energy, been blaming yourself when things go wrong.’ The version of the HSCL-D used in our study was modiﬁed for the local context; one item from the original scale (‘feeling trapped’) was dropped because it did not perform well in this context, and one item was added (‘don’t care what happens to your health’; Bolton & Ndogoni, 2001). This version of the HSCL-D has shown strong evidence of reliability and validity in the local context (Tsai et al. 2012; Psaros et al. 2015).
Missed schooling was assessed for children aged 5–17 years via caregiver report. For children enrolled in school, caregivers were asked ‘In the past 7 days, how many school days did this child miss for any reason?’ and ‘In just the past 7 days, how many days was this child late to school because of fetching water?’
In addition to these measures, the survey asked female household heads about basic demographic

characteristics including their age, marital status, level of education, number of children under 18 living in the home, child age and sex, and family assets in the form of consumable goods (e.g. owns a working radio, has a house with walls made of cement). The asset data were used to calculate an index of household asset wealth (Filmer & Pritchett, 1999); absolute values of the asset index cannot be interpreted directly but can be used to rank households by overall wealth in relative terms, with higher values indicating higher levels of asset wealth.
Statistical analysis
All statistical analyses were conducted using the Stata/ SE software package (version 14.0; StataCorp, 2015). We followed the procedures described by Imai et al. (2010a, b) to estimate the association between household water insecurity and missed schooling, and the extent to which caregiver depression mediated this association. The parametric algorithm described in Imai et al. (2010a) computes two quantities of interest: the average causal mediation eﬀect, which is the average change in missed schooling corresponding to a change in caregiver depression under less v. more exposure to household water insecurity; and the average direct eﬀect, which is the average of all other causal mechanisms linking household water insecurity to missed schooling. As the exposure, outcome, and mediator were all continuous variables, we used standard linear regression. Separate models were used to evaluate the outcomes of number of school days missed (see Fig. 1) and number of days late to school (see Fig. 2) in all children, and then in girls and boys separately. In all analyses, standard errors were corrected for clustering at the household level; and we adjusted for caregiver age, education, marital status, household asset wealth, and number of children in the household.
Results
Sample characteristics
Participating women were on average 33.5 years old (S. D. = 7.9). Each woman, on average, lived in a household with three children under the age of 18 (range, 1–8). The mean WI score was 9.9 (S.D. = 6.9). The mean depression score among female household heads was 1.76 (S.D. = 0.49), with 44% scoring above the 1.75 point cut-oﬀ used to denote probable depression (Winokur et al. 1984). Most women were married or cohabiting with a partner (82%) and had not completed primary education (52%). Among the children described by the caregivers, 49% were girls, and the mean age across all children was 9.2 years (S.D. 3.4).

global mental health

Fig. 1. Unstandardized models of caregiver depression as a mediator between water insecurity and missed schooling. Note: *p < 0.05, **p < 0.01

Overall, 42% of children missed one or more school days in the last week and 20% were late to school one or more days in the last week. There were no statistically signiﬁcant diﬀerences in number of school days missed (t = 0.54, p = 0.59) or number of school days tardy (t = −0.83, p = 0.40) between boys and girls.
Number of missed school days
Among all children, there was a statistically signiﬁcant association between household water insecurity and the number of missed school days (b = 0.04; 95% CI 0.02–0.06). Evaluated at the mean values of the covariates, a one standard deviation (S.D.) increase in

household water insecurity was associated with a 0.30 increase in the number of missed school days in the last week (0.30/1.31 = 0.23 S.D. increase). Household water insecurity also had a statistically signiﬁcant association with depression symptom severity among female household heads, which partially mediated the association between water insecurity and missed school days: 16% of that association was mediated by caregiver depression (see Table 1 and Fig. 1).
When analyses were stratiﬁed by child sex, among boys, there was a statistically signiﬁcant association between household water insecurity and the number of missed school days (b = 0.05; 95% CI 0.01–0.08).

global mental health

Fig. 2. Unstandardized models of caregiver depression as a mediator between water insecurity and school tardiness. Note: *p < 0.05, **p < 0.01, ***p < 0.001.

Evaluated at the mean values of the covariates, a 1 S.D. increase in household water insecurity was associated with a 0.31-day increase in the number of missed school days in the last week (0.23 S.D. units when compared with the sample mean). Household water insecurity also had a statistically signiﬁcant association with depression symptom severity among female household heads, which partially mediated the association between water insecurity and missed school days (see Fig. 1); 17% of that association was mediated by caregiver depression (see Table 1).
Among girls, there was also a statistically signiﬁcant association between household water insecurity and number of missed school days (b = 0.04; 95% CI 0.01–0.06).

Evaluated at the mean of the covariates, a 1 S.D. increase in household water insecurity translated into a slightly smaller eﬀect size of 0.26/1.26 = 0.21 S.D. units. However, the average causal mediated eﬀect was not statistically signiﬁcant (see Table 1 and Fig. 1).
School tardiness
Household water insecurity had a statistically signiﬁcant association with days of school late in the entire sample (b = 0.06; 95% CI 0.04–0.09) and translated into slightly larger eﬀect size of 0.44/1.42 = 0.31 S.D. units. When stratiﬁed by sex, this ﬁnding was signiﬁcant for both boys (b = 0.06; 95% CI 0.04–0.09; 0.32 S.D. units)

global mental health

Table 1. Mean effects for caregiver depression mediating the relationship between water insecurity and missed schooling

Missed days of school All children Boys only Girls only
Days late to school All children Boys only Girls only

Average causal mediated effect (95% CI)
0.006 (0.001–0.013) 0.007 (0.001–0.016) 0.007 (−0.0001 to 0.016)
0.004 (−0.002 to 0.013) 0.004 (−0.006 to 0.015) 0.004 (−0.003 to 0.015)

Direct effect (95% CI)
0.036 (0.014–0.057) 0.038 (0.007–0.067) 0.031 (0.004–0.056)
0.059 (0.033–0.083) 0.059 (0.030–0.086) 0.062 (0.026–0.097)

Total effect (95% CI)
0.042 (0.020–0.065) 0.045 (0.014–0.077) 0.037 (0.010–0.064)
0.063 (0.039–0.088) 0.062 (0.037–0.090) 0.067 (0.032–0.103)

% of total effect mediated (95% CI)
0.155 (0.099–0.318) 0.168 (0.095–0.522) 0.182 (0.102–0.608)
0.070 (0.050–0.112) 0.059 (0.041–0.100) 0.068 (0.043–0.137)

and girls (b = 0.07; 95% CI 0.03–0.10, 0.31 S.D. units). However, because there was no statistically signiﬁcant association between number of days late and depression among the female household heads, no statistically signiﬁcant average causal mediation eﬀects were observed (see Table 1 and Fig. 2).
Discussion
In this cross-sectional, population-based study of female household heads in rural Uganda and their children, we found that household water insecurity had a statistically signiﬁcant association with missed schooling for children. The estimated association was large and statistically signiﬁcant, and the eﬀect was partially mediated by caregiver depression. These ﬁndings are consistent with the literature linking water insecurity to increased depression among women in LMICs (Wutich & Ragsdale, 2008; Stevenson et al. 2012; Tsai et al. 2016). Building on a previous work from high-income countries, our study establishes an intergenerational spillover eﬀect leading to poor school attendance in a rural Ugandan community (Guevara et al. 2013; Claessens et al. 2015). Our analyses adjusted for household asset wealth and other socioeconomic characteristics, so the eﬀects of water insecurity and depression on missed schooling are unlikely to be better explained by lack of ﬁnancial means for paying school fees. The mediating role of depression among female household heads may potentially be explained by the domestic role impairments associated with water insecurity and depression in eastern Africa (e. g. preparing family meals, washing clothing, and bathing children; Bolton et al. 2002; Stevenson et al. 2012). Such impairments may increase the chore burden, including water fetching, placed on children, and time spent completing these tasks is likely to delay departure for school and require costly tradeoﬀs with school attendance. The present study does not allow for examining diﬀerences between biological and

non-biological caregivers. Findings from qualitative work underway at this site suggest did not identify differences in water procurement demands for biologically and non-biologically related caregivers, though themes suggest that non-biologically related children may disproportionally bear the burden of chores including water fetching.
Notably, when analyses were stratiﬁed by sex, we only found evidence of mediation for boys but not girls. This discrepancy could potentially be explained by the increased levels of instrumental support in water fetching that girls tend to provide their mothers, relative to boys (Sorenson et al. 2011; Graham et al. 2016). This distribution of labor between boys and girls was described by some mothers interviewed in our qualitative investigation at this site, which is consistent with local cultural norms of girls being assigned household chores. However, other mothers described a more equal division of water fetching among young boys and girls. Such support may buﬀer women against the emotional stress of experiencing water insecurity but at the cost of lowering school attendance for girls. It is possible that the gender speciﬁc ﬁnding is driven by statistical power. Depression among the female household heads was signiﬁcantly associated with water insecurity in the mediation models evaluating the outcome of school tardiness, consistent with the previous work (Burke & Beegle, 2004; Hemson, 2007; Dreibelbis et al. 2013), but the magnitude and statistical signiﬁcance were smaller in the model for girls. However, there were an approximately equivalent number of girls and boys, so a diﬀerence in statistical power is unlikely to be a suﬃcient explanation for the divergent ﬁndings.
Though we hypothesized that caregiver depression would delay departure for school among their children, we did not ﬁnd evidence of this phenomenon. It is possible that caregiver depressive role impairments cause more signiﬁcant disruption to the family routine, resulting in outcome diﬀerences along the

global mental health

extensive margin (i.e. children missing an entire day of school) rather than along the intensive margin (i.e. children not missing school but arriving late). Further, unpublished data from our ongoing qualitative study of water insecurity in this setting suggest that Ugandan children who are tardy for school often face harsh punishment and shaming at school, and therefore may prefer to miss an entire day rather than arrive late and face these undesirable consequences. Such treatment at school is consistent with social norms supporting the use of corporal punishment and low mental health literacy found among educators in other parts of sub-Saharan Africa (Lansford et al. 2010; Kutcher et al. 2015).
The ﬁndings of this investigation must be considered in light of several limitations. First, the cross-sectional study design does not allow for the evaluation of temporal precedence in the mediational models. Therefore, these ﬁndings should be regarded as preliminary support for the mediating role of caregiver depression, and further evaluation of these models using a longitudinal study design is warranted. Second, the data were all based on self-report and therefore carry an inherent reporting bias. However, it should be noted that in the local setting, because child school attendance is a source of pride for families, at times families may actually under report their child’s missed schooling; thus, we expect that our ﬁndings potentially underestimate the magnitudes of the eﬀects observed. Third, our measure of depression was based on a symptom checklist, and these are known to overestimate the prevalence of major depressive disorder due to false positives (Kagee et al. 2013; Tsai, 2014). Given the study design, we were not able to determine which women were experiencing a major depressive episode. Fourth, the study design did not allow for evaluating diﬀerences between biological and non-biological caregivers in the relationships examined. Lastly, we were not able to account for the potential mediating eﬀects of intergenerational transmission of depression from caregiver to child (Bowlby, 1988; Manczak et al. 2017; Weissman et al. 2016) or of physical illness associated with water insecurity (Fewtrell et al. 2005; Hunter et al. 2009), which may have also inﬂuenced child school attendance.
Taken together, these ﬁndings advance our understanding of the processes through which water insecurity undercuts children’s school participation in LMICs, with attendant implications for interventions to enhance children’s developmental outcomes. Foremost, these ﬁndings suggest that structural interventions alone are unlikely to fully alleviate the deleterious eﬀects of water insecurity on child schooling and should consider also seeking to address the mental health needs of caregivers in addition to improving

reliable access to clean water. For example, supported parenting and counseling interventions (e.g. Interpersonal Therapy and Cognitive Behavioral Therapy; Bolton et al. 2003; Nakimuli-Mpungu et al. 2014) delivered alongside structural water interventions could more fully address depressive symptoms and functional impairments associated with water insecurity, and help women engage in problem solving and more adaptive parenting behaviors (Stein et al. 2014). In turn, they may be better able to manage domestic role demands, including their family’s water collection. Such integrated interventions have the promise to support children’s school attendance and achievement to reduce the intergenerational transmission of adverse water insecurity-related outcomes.
Acknowledgements
This study was funded by NIH K23 MH096620, the Paiko Foundation, and was also supported by T32MH093310.
Declaration of Interest
None.
Ethical Standards
The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008.
References
Abdullah A, Doucouliagos H, Manning E (2015). Does education reduce income inequality? A meta-regression analysis. Journal of Economic Surveys 29, 301–316.
Bennett IM, Schott W, Krutikova S, Behrman JR (2016). Maternal mental health, and child growth and development, in four low-income and middle-income countries. Journal of Epidemiology and Community Health 70, 168–173.
Bolton P, Bass J, Neugebauer R, Verdeli H, Clougherty KF, Wickramaratne P, Speelman L, Ndogoni L, Weissman M (2003). Group interpersonal psychotherapy for depression in rural Uganda. JAMA 289, 3117.
Bolton P, Ndogoni L (2001). Cross-Cultural Assessment of Trauma-Related Mental Illness (Phase II). Baltimore, MD. (http://pdf.usaid.gov/pdf_docs/Pnacm033.pdf).
Bolton P, Neugebauer R, Ndogoni L (2002). Prevalence of depression in rural Rwanda based on symptom and functional criteria. The Journal of Nervous and Mental Disease 190, 631–637.
Bowlby J (1988). Developmental psychiatry comes of age. American Journal of Psychiatry 145, 1–10.

global mental health

Burke K, Beegle K (2004). Why children aren’t attending school: the case of Northwestern Tanzania. Journal of African Economies 13, 333–355.
Choe DE, Olson SL, Sameroﬀ AJ (2014). Eﬀortful control moderates bidirectional eﬀects between children’s externalizing behavior and their mothers’ depressive symptoms. Child Development 85, 643–658.
Claessens A, Engel M, Curran FC (2015). The eﬀects of maternal depression on child outcomes during the ﬁrst years of formal schooling. Early Childhood Research Quarterly 32, 80–93.
Cooper PJ, Tomlinson M, Swartz L, Woolgar M, Murray L, Molteno C (1999). Post-partum depression and the mother-infant relationship in a South African peri-urban settlement. The British Journal of Psychiatry: The Journal of Mental Science 175, 554–558.
Cortina MA, Fazel M, Hlungwani TM, Kahn K, Tollman S, Cortina-Borja M, Stein A (2013). Childhood psychological problems in school settings in rural Southern Africa. PLoS ONE 8, e65041.
Derogatis LR, Lipman RS, Rickels K, Uhlenhuth EH, Covi L (1974). The Hopkins Symptom Checklist (HSCL): a self-report symptom inventory. Behavioral Science 19, 1–15.
Devkota S, Panda B (2016). Socioeconomic gradients in early childhood health: evidence from Bangladesh and Nepal. International Journal for Equity in Health 15, 78.
Dreibelbis R, Greene LE, Freeman MC, Saboori S, Chase RP, Rheingans R (2013). Water, sanitation, and primary school attendance: a multi-level assessment of determinants of household-reported absence in Kenya. International Journal of Educational Development 33, 457–465.
Familiar I, Nakasujja N, Bass J, Sikorskii A, Murray S, Ruisenor-Escudero H, Bangirana P, Opoka R, Boivin MJ (2016). Caregivers’ depressive symptoms and parent-report of child executive function among young children in Uganda. Learning and Individual Diﬀerences 46, 17–24.
Fewtrell L, Kaufmann RB, Kay D, Enanoria W, Haller L, Colford JM (2005). Water, sanitation, and hygiene interventions to reduce diarrhoea in less developed countries: a systematic review and meta-analysis. The Lancet Infectious Diseases 5, 42–52.
Filmer D, Pritchett L (1999). The eﬀect of household wealth on educational attainment: evidence from 35 countries. Population and Development Review 25, 85–120.
Graham JP, Hirai M, Kim SS (2016). An analysis of water collection labor among women and children in 24 sub-Saharan African countries. PLoS ONE 11, e0155981.
Guevara JP, Mandell D, Danagoulian S, Reyner J, Pati S (2013). Parental depressive symptoms and children’s school attendance and emergency department use: a nationally representative study. Maternal and Child Health Journal 17, 1130–1137.
Hadley C, Wutich A (2009). Experience-based measures of food and water security: biocultural approaches to grounded measures of insecurity. Human Organization 68, 451–460.
Hargreaves JR, Morison LA, Kim JC, Bonell CP, Porter JDH, Watts C, Busza J, Phetla G, Pronyk PM (2008). The association between school attendance, HIV infection and sexual behaviour among young people in rural South

Africa. Journal of Epidemiology & Community Health 62, 113– 119. Hemson D (2007). ‘The Toughest of Chores’: policy and practice in children collecting water in South Africa. Policy Futures in Education 5, 315. Henley R, Mcalpine K, Mueller M, Vetter S (2010). Does school attendance reduce the risk of youth homelessness in Tanzania? International Journal of Mental Health Systems 4, 1–8. Huang KY, Abura G, Theise R, Nakigudde J (2017). Parental depression and associations with parenting and children’s physical and mental health in a sub-Saharan African setting. Child Psychiatry & Human Development 48, 517–527. Hunter PR, Zmirou-Navier D, Hartemann P (2009). Estimating the impact on health of poor reliability of drinking water interventions in developing countries. Science of The Total Environment 407, 2621–2624. Imai K, Keele L, Tingley D (2010a). A general approach to causal mediation analysis. Psychological Methods 15, 309– 334. Imai K, Keele L, Yamamoto T (2010b). Identiﬁcation, inference and sensitivity analysis for causal mediation eﬀects. Statistical Science 25, 51–71. Kagee A, Tsai AC, Lund C, Tomlinson M (2013). Screening for common mental disorders in low resource settings: reasons for caution and a way forward. International Health 5, 11–14. Kanté AM, Nathan R, Jackson EF, Levira F, Helleringer S, Masanja H, Phillips JF (2016). Trends in socioeconomic disparities in a rapid under-ﬁve mortality transition: a longitudinal study in the United Republic of Tanzania. Bulletin of the World Health Organization 94, 258–266. Kutcher S, Gilberds H, Morgan C, Greene R, Hamwaka K, Perkins K (2015). Improving Malawian teachers’ mental health knowledge and attitudes: an integrated school mental health literacy approach. Global Mental Health 2, 1–1. Lansford JE, Pe L, Al-hassan S, Bacchini D, Bombi AS, Bornstein MH, Chang L, Deater-deckard K, Giunta L, Dodge KA, Oburu P, Pastorelli C, Runyan DK, Skinner AT, Sorbring E, Tapanya S, Maria L, Tirado U, Zelli A (2010). Corporal punishment of children in nine countries as a function of child gender and parent gender. International Journal of Pediatrics 2010, 1–12. Manczak EM, Williams D, Chen E (2017). The role of family routines in the intergenerational transmission of depressive symptoms between parents and their adolescent children. Journal of Abnormal Child Psychology 45, 643–656. Mehta N, Preston S (2016). Are major behavioral and sociodemographic risk factors for mortality additive or multiplicative in their eﬀects? Social Science & Medicine 154, 93–99. Nakimuli-Mpungu E, Wamala K, Okello J, Alderman S, Odokonyero R, Musisi S, Mojtabai R (2014). Developing a culturally sensitive group support intervention for depression among HIV infected and non-infected Ugandan adults: a qualitative study. Journal of Aﬀective Disorders 163, 10–17. Ng LC, Kirk CM, Kanyanganzi F, Fawzi MCS, Sezibera V, Shema E, Bizimana JI, Cyamatare FR, Betancourt TS (2015). Risk and protective factors for suicidal ideation and

global mental health

behaviour in Rwandan children. The British Journal of Psychiatry: The Journal of Mental Science 207, 262–268. Parke RD, Coltrane S, Duﬀy S, Buriel R, Dennis J, Powers J, French S, Widaman KF (2004). Economic stress, parenting, and child adjustment in Mexican American and European American families. Child Development 75, 1632–1656. Petterson SM, Albers AB (2001). Eﬀects of poverty and maternal depression on early child development. Child Development 72, 1794–1813. Psacharopoulos G, Alam A (1991). Earnings and education in Venezuela: an update from the 1987 household survey. Economics of Education Review 10, 29–36. Psaros C, Haberer JE, Boum Y II, Tsai AC, Martin JN, Hunt PW, Bangsberg DR, Safren SA (2015). The factor structure and presentation of depression among HIV-positive adults in Uganda. AIDS and Behavior 19, 27–33. Richter L, Daelmans B, Lombardi J, Heymann J, Boo FL, Behrman JR, Lu C, Lucas JE, Perez-Escamilla R, Dua T, Bhutta ZA, Stenberg K, Gertler P, Darmstadt GL (2017). Investing in the foundation of sustainable development: pathways to scale up for early childhood development. The Lancet 398, 103–118. Sorenson SB, Morssink C, Campos PA (2011). Safe access to safe water in low income countries: water fetching in current times. Social Science and Medicine 72, 1522–1526. StataCorp LP (2015). Stata/SE 14.0. StataCorp LP: College Station, TX. Stein A, Pearson RM, Goodman SH, Rapa E, Rahman A, McCallum M, Howard LM, Pariante CM (2014). Eﬀects of perinatal mental disorders on the fetus and child. The Lancet 384, 1800–1819. Stevenson EGJ, Greene LE, Maes KC, Ambelu A, Tesfaye YA, Rheingans R (2012). Water insecurity in 3 dimensions: an anthropological perspective on water and women’s psychosocial distress in Ethiopia. Social Science & Medicine 75, 392–400. Swindale A, Bilinsky P (2006). Development of a universally applicable household food insecurity measurement tool: process, current status, and outstanding issues. The Journal of Nutrition 136, 1449S–1452S. Tomlinson M, Morgan B (2015). Infant Mental Health Research in Africa: a call for action for research in the next 10 years. Global Mental Health 2, e7. Tsai AC (2014). Reliability and validity of depression assessment among persons with HIV in sub-Saharan Africa: systematic review and meta-analysis. JAIDS Journal of Acquired Immune Deﬁciency Syndromes 66(5 PG-503–511), 503–511. Tsai AC, Bangsberg DR, Emenyonu N, Senkungu JK, Martin JN, Weiser SD (2011). The social context of food insecurity among persons living with HIV/AIDS in rural Uganda. Social Science & Medicine 73, 1717–1724.

Tsai AC, Bangsberg DR, Frongillo EA, Hunt PW, Muzoora C, Martin JN, Weiser SD (2012). Food insecurity, depression and the modifying role of social support among people living with HIV/AIDS in rural Uganda. Social Science & Medicine 74, 2012–2019.
Tsai AC, Kakuhikire B, Mushavi R, Vořechovská D, Perkins JM, McDonough AQ, Bangsberg DR (2016). Population-based study of intra-household gender diﬀerences in water insecurity: reliability and validity of a survey instrument for use in rural Uganda. Journal of Water and Health 14, 280–292.
Tsai AC, Tomlinson M (2012). Mental health spillovers and the Millennium Development Goals: the case of perinatal depression in Khayelitsha, South Africa. Journal of Global Health 2, 10302.
UNESCO (2015). Rapid acceleration of progress is needed to achieve universal primary education (http://data.unicef.org/ education/primary.html).
UNICEF (2015a). Education: Primary net attendance ratio (http:// data.unicef.org/education/primary.html).
UNICEF (2015b). Education: Survival rate to last primary grade (http://data.unicef.org/education/primary.html).
Verkuijl NE, Richter L, Norris SA, Stein A, Avan B, Ramchandani PG (2014). Postnatal depressive symptoms and child psychological development at 10 years: a prospective study of longitudinal data from the South African Birth to Twenty cohort. The Lancet Psychiatry 1, 454–460.
Wang MT, Sheikh-Khalil S (2014). Does parental involvement matter for student achievement and mental health in high school? Child Development 85, 610–625.
Weissman MM, Berry OO, Warner V, Gameroﬀ MJ, Skipper J, Talati A, Pilowsky DJ, Wickramaratne P (2016). A 30-year study of 3 generations at high risk and low risk for depression. JAMA Psychiatry 73, 970.
WHO & UNICEF (2014). Progress on sanitation and drinkingwater – 2014 update (http://apps.who.int/iris/bitstream/ 10665/112727/1/9789241507240_eng.pdf?ua=1).
Winokur A, Winokur DF, Rickels K, Cox DS (1984). Symptoms of emotional distress in a family planning service: stability over a four-week period. The British Journal of Psychiatry 144, 395–399.
Wood JJ, Lynne-Landsman SD, Langer DA, Wood PA, Clark SL, Mark Eddy J, Ialongo N (2012). School attendance problems and youth psychopathology: structural cross-lagged regression models in three longitudinal data sets. Child Development 83, 351–366.
Wutich A, Ragsdale K (2008). Water insecurity and emotional distress: coping with supply, access, and seasonal variability of water in a Bolivian squatter settlement. Social Science and Medicine 67, 2116–2125.