Policy Research Working Paper 10975 Maternal Mental Health and Its Influence on Children’s Early Development Evidence from Khyber Pakhtunkhwa, Pakistan Mahreen Tahir-Chowdhry Elizabeth Hentschel Heather Tomlinson Amna Ansari Amer Hasan Aisha Yousafzai Naveed Hussain Education Global Practice A verified reproducibility package for this paper is November 2024 available at http://reproducibility.worldbank.org, click here for direct access. Policy Research Working Paper 10975 Abstract This paper reports on the prevalence of three facets of mental how exposure to stressors such as food insecurity, financial health—depression, anxiety, and parenting stress—among insecurity, being impacted by flooding, community crime, mothers of children ages 0–6 years in Khyber Pakhtunkhwa, discrimination, and domestic violence exacerbate both Pakistan. Data from mother-child dyads were analyzed to maternal mental health and child outcomes. The regression examine differences in maternal mental health and early analyses indicate a significant and negative compounding childhood development outcomes by maternal educational interaction of maternal depression, anxiety, and parenting attainment, urban versus rural setting, and refugee versus stress on early childhood development for younger (0–3 non-refugee status. The analysis finds a higher prevalence years) and older (3–6 years) children, even after controlling of self-reported mental health concerns among refugee, for stressors and other covariates. Policy improvements are less-educated, and rural mothers relative to non-refugee, needed that focus on at-risk communities, providing mental more-educated, and urban mothers. Maternal mental health services and reducing exposure to stressors within health concerns are significantly associated with lower levels communities and households. of early childhood development. This paper also analyzes This paper is a product of the Education Global Practice. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://www.worldbank.org/prwp. The authors may be contacted at ahasan1@worldbank.org. A verified reproducibility package for this paper is available at http://reproducibility.worldbank.org, click here for direct access. RESEA CY LI R CH PO TRANSPARENT ANALYSIS S W R R E O KI P NG PA The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Maternal Mental Health and Its Influence on Children’s Early Development: Evidence from Khyber Pakhtunkhwa, Pakistan Mahreen Tahir-Chowdhry * Elizabeth Hentschel*- Heather Tomlinson* - Amna Ansari* Amer Hasan* Aisha Yousafzai • Naveed Hussain* Keywords: Early childhood development (ECD); Depression; Anxiety; Parenting stress; Fragile, Conflict and Violence-affected (FCV); refugees JEL Classification: I14, I15, I24, I25 * Education Global Practice, World Bank - Yale Child Study Center • Harvard T.H. Chan School of Public Health Introduction Fragile, conflict and violence-affected (FCV) settings present complex challenges for societies, families and children by contributing to greater mental health burdens. FCV areas are characterized by low levels of per capita income, high levels of food insecurity, vulnerability to climate hazards, and increasing intersections with internally displaced and refugee populations (The World Bank, 2024). This can create a vicious cycle wherein people who are least equipped with the resources to manage stressful environments will be most likely to live in areas defined by reduced access to services and greater risks. Young children often bear the heaviest health burdens in the wake of fragility and conflict (Murthi & van der Heijden, 2024), and the compounding effects of various facets of maternal mental health concerns found at high rates in FCV contexts may be especially detrimental to early childhood development (ECD). In areas of conflict and forced displacement, there is evidence of high rates of depression and stress-related disorders (80% and 88%, respectively) (Bogic et al., 2015; Morina et al., 2018), and multi-generational impacts of conflict on child development outcomes and beyond (Chatterjee et al., 2023). Parents in Khyber Pakhtunkhwa, Pakistan, confront a multitude of risks to child development that have worsened in the last decade. Longstanding challenges include high rates of poverty, low access to high-quality education (including preprimary education), and incomplete and inadequate coverage of social assistance cash transfers (Alam et al., 2022; Cheema et al., 2020; Shah et al., 2022). Added to these risks are maternal depression, anxiety and parenting stress, which have been found to be consistently negatively impacted by fragility, conflict and violence (Al-Abri et al., 2023; Chen et al., 2017). Yet little is known about the prevalence of maternal mental health concerns in Khyber Pakhtunkhwa and how recent events may have exacerbated them. Various studies confirm that higher levels of parental depression, anxiety, and parenting stress respectively predict higher levels of children’s internalizing behaviors (e.g., withdrawal, anxiety), higher levels of externalizing behaviors (e.g., anger, aggression), and reduced social-emotional competence (e.g., ability to self-regulate, follow instructions, get along with peers) (De Oliveira et al., 2019; Fang et al., 2022; Mäntymaa et al., 2012). When multiple parental mental health problems co-occur, early childhood functioning is at increased risk (Watkeys et al., 2022). In Pakistan, rates of comorbidity of anxiety and depression have been found to range between 25% to 34% and higher rates are seen in women than men (Farooq et al., 2019; Mirza & Jenkins, 2004). While these risks occur across the socioeconomic spectrum, women in low-income households experience disproportionate levels of mental health concerns, such as depression, and display fewer characteristics that foster favorable ECD, such as maternal supportiveness and cognitive stimulation (Reeves & Krause, 2019). For instance, exposure to community-level stressors such as flooding or other natural disasters, and community-level crime have been shown to increase stress and anxiety (Kerker et al., 2023; Maselko et al., 2018; Sitwat et al., 2015; Yousuf et al., 2023). Risks and stressors at the household and individual levels, including poverty, food insecurity, discrimination, exposure to intimate partner and other domestic violence, and fear, related to safety in the home or community, are also correlated with depression and anxiety (Adamu & Adinew, 2018; Iqbal & Ali, 2021; Wei et al., 2024). In addition, research on chronic depression shows correlations with impaired social-emotional functioning among young children in Pakistan (De Oliveira et al., 2019). Nationally representative studies in Pakistan also show a decline in the likelihood of developing on track as exposure to risk factors increases in number (Hentschel et al., 2024; Tomlinson et al., 2023). Further, evidence suggests that living with parents 2 who experience poor mental health has multi-generational impacts, putting young children at risk of poor cognitive, emotional and behavioral development and affecting outcomes not only in the early years but also as they get older (Jagtap et la., 2023; Merced-Nieves et al. 2021). Vulnerable groups typically experience markedly higher rates of mental health concerns. Depression and anxiety during pregnancy have been shown to have a higher incidence among women from rural areas compared to those from urban areas in Pakistan (Waqas et al., 2015). A cross-sectional study in Khyber Pakhtunkhwa found an association between women with low educational attainment and higher levels of maternal depression and anxiety (Wazir et al., 2023). Other predictive factors included poverty and illiteracy (Shams et al., 2021). In addition, among refugee populations, a systematic review looking at 40 studies across low- and middle-income countries found that 32% of refugees met diagnostic criteria for a major depressive episode and 31% for post-traumatic stress disorder (Patanè et al., 2022). Refugee families experience worry and stress, compounded by multiple stressors that may include lack of documentation, loss of income, discrimination, reduced access to education, and histories of trauma, violence, and displacement (Hazer & Gredebäck, 2023; Leyendecker et al., 2018). Consequently, refugee parents experience high levels of depression and stress disorders (Blackmore et al., 2020; Chen et al., 2017). Pakistan is one of the largest refugee-hosting countries in the world, according to the United Nations High Commissioner for Refugees (UNHCR, 2024c). Pakistan hosts 3.20 million Afghans, of whom 1.52 million are registered as either refugees or asylum seekers (UNHCR, 2024a; UNCHR, 2024b). 1 Of these, 72% are women and children and 12% are children under age 5 (UNICEF, 2024). Khyber Pakhtunkhwa hosts the majority (716,099 or 53%) of Afghan nationals in the country (UNHCR, 2024c) and nationally, most refugee households are situated in urban or semi-urban areas (69%), while 31% are in refugee villages (UNHCR, 2024a). Whereas data from Pakistan shows correlations with maternal depression in the antenatal and postpartum periods, there is less research showing relationships between mothers’ mental health status and ECD through the preschool-age years. Pakistan has one of the highest estimates of maternal postpartum depression in Asian countries, with rates ranging from 28% to 63%—as well as evidence of paternal post-partum depression (Khabir et al., 2022)—and underreporting is likely, due to cultural norms (Gulamani et al., 2013). In Khyber Pakhtunkhwa specifically, 28% of women and 58% of pregnant women seen at health facilities have been found depressed (Irum et al., 2022; Wazir et al., 2023). Evidence from Pakistan indicates that infants exposed to maternal depression during pregnancy had higher levels of stress-related disease at one-year of life (Hagaman et al., 2020). In terms of maternal anxiety, prenatal exposure has been linked to less positive child emotional development and has been shown to impact long-term developmental trajectories (Huizink et al., 2014; Rees et al., 2019). An estimated 21% of women in a study in Khyber Pakhtunkhwa were found to have generalized anxiety disorder (Wazir et al., 2023). In terms of parenting stress, globally, about 5% of parents report parenting stress or burnout (Roskam et al., 2021), although data specific to Pakistan are not available. In a study from Canada, children aged 3 years whose parents reported high parenting stress during the child’s infancy were more likely than peers to show mental health problems (Hattangadi et al., 2020). Despite links between facets of maternal mental health and ECD, there is little research on these relationships in Pakistan 1 For the purposes of this paper, the term refugee includes registered and unregistered Afghan nationals. 3 beyond the postpartum phase, generally considered to be the period up to eight weeks after childbirth. One exception is the Hentschel et al. (2024) study of parenting during the COVID-19 pandemic period showing that 57% of parents of children 0-3 years reported feeling distressed and distress was one of the strongest predictors of reduced ECD among those children. This study aims to address the need for a more nuanced understanding of the contributing factors that influence caregiving in FCV settings, as called for in previous research (e.g., Burgund Isakov et al., 2022). More specifically, this study examines whether maternal depression, anxiety or parenting stress independently or jointly predict ECD, and additionally considers whether potential relationships remain even after accounting for stressors that could be expected to influence outcomes. The study also evaluates whether young children in certain vulnerable groups—namely, those living in rural areas, those with caregivers with less than a primary school education, or refugees, respectively—have mothers with more severe mental health conditions and are more at- risk developmentally. These topics have not yet been well-explored in FCV settings in general, nor in Pakistan in particular. Therefore, this study aims to answer the following questions: (i) How do levels of maternal mental health—specifically depression, anxiety and parenting stress—differ by education level, refugee status, and living in an urban vs. rural area? (ii) What is the compounding influence of multiple maternal mental health conditions, namely depression, anxiety and parenting stress, on child development for children 0-6 years old? (iii) How, if at all, is this association altered by controlling for stressors and other confounders? This study builds on a nascent literature in the following ways: (i) it examines multiple maternal mental health outcomes for caregivers at the same time; (ii) it differentiates between maternal mental health outcomes for hosts and refugees; (iii) it assesses the severity of various maternal mental health conditions, not just the onset of the condition itself; (iv) it documents differences in child outcomes – both from caregiver reports and from a direct assessment of children; (v) it examines outcomes separately for children between 0-35 months of age and those aged 36-72 months. To the best of our knowledge this is the first study to assess the compounding association of multiple maternal mental health conditions on child development for children 0-72 months in Khyber Pakhtunkhwa, Pakistan. Methods Research Design and Data Collection Strategy This household survey was commissioned by The World Bank and carried out by the Center for Evaluation and Development between December 2023 and February 2024 in Khyber Pakhtunkhwa, Pakistan. Households were selected based on being located in the catchment area of one of 200 public schools across Khyber Pakhtunkhwa (excluding newly merged districts). 2 In most cases, the boundaries of a catchment area represented a maximum of a 30-minute walk from the proximate school. Households living in a given catchment area who had at least one child under the age of 72 months (for either of the 0-35-month or 36-72-month age groups) available to be 2 The sample of 200 public schools was drawn for a different World Bank survey in 2022. These 200 schools are a representative sample of public schools across all districts of Khyber Pakhtunkhwa (excluding newly merged districts). The households in question are thus representative of those in the catchment area of a representative sample of schools. 4 interviewed (i.e., on the premises when the enumerators began their interviews) were included in this survey. Enumerators selected households within the catchment area randomly. Among households with an age-eligible child present (i.e., 0-35 months or 36-72 months), enumerators explained the purpose of the study and invited the target child’s primary caregiver to be interviewed. Respondents were supported to withdraw their consent anytime during or after the interview and such cases were dropped from analyses. For children 0-35 months of age, assuming a minimum power level of .8 and an alpha of .05, the study design predetermined that the minimum desired detectable difference in average ECD Z- score measured by the Caregiver Reported Early Development Instrument (CREDI) was a 0.2 standard deviation (SD) difference, which has been shown to be the average difference between urban and rural Pakistani children assessed using the CREDI (Hentschel et al., 2024). For children between 36 and 72 months of age, assuming a minimum power level of .8 and an alpha of .05, the study design predetermined that the minimum detectable difference in average ECD score (measured by the AIM-ECD score) is a 0.23-point difference (the difference between males and females in Hentschel et al., 2024). As such, we sampled a total of 1,397 children aged 0-35 months and 705 children aged 36-72 months. Survey weights were created at the household, caregiver, and child levels. The survey employs a stratified, two stage clustered design. There are five strata 3 and units are clustered at the level of the school catchment area. Area of residence (rural and urban), roof material, cooking fuel, tenancy, and language were used for the model-based adjustments. 4 The design weights for households were calculated as follows: 1 ℎ ℎ = ∗ Where represents the empirical probability of selecting cluster j and ℎ , represents the number of households from cluster j in the sample and the estimated total number of eligible households in cluster j, respectively. is estimated as: = ∗ � ℎℎ � is an Where represents the number of students enrolled in the selected school and ℎℎ estimate for the household size where k indexes whether the school is rural or urban. M is defined as: 3 The five strata include: area of residence, roof material, cooking fuel, tenancy, and language. 4 Area of residence, roof material, cooking fuel and tenancy were used to rake on the household margins and language was used to rake on the child margins. “Raking” indicates an iterative process that uses the sample design weight as the starting weight and stops once the model converges. 5 1 ℎ = � 2 ℎ 3 ℎ Household weights are subsequently calibrated using census data from the 2017 census. Let ℎ () represent these model-adjusted weights for households. Design-based weights for children are calculated as 1 1 = ℎ ∗ () min� , 3� Where is the number of eligible children in the household. Weights were adjusted via iterative proportional fitting to reflect population-level margins for Khyber Pakhtunkhwa on the categorical variables taken from the 2017 census described above (e.g., area of residence, roof material, etc.). Survey weights were applied to all analyses, with household weights for the descriptives and child weights for the regressions. Key Variables and Tools Basic demographic data were collected from each respondent for child gender, age, maternal educational attainment, birth registration (respondents confirming the child was registered were asked to provide proof of the document), and area of residence (i.e., urban, rural, semi-urban, inner-city). All instruments were administered in Urdu. In circumstances where children were unable to speak Urdu, enumerators effectively engaged children using various methods, such as teaching them Qaida in Madrasa. Outcome Variables Early Childhood Development Measures Children 0-35 months. The Caregiver-Reported Early Development Instrument (CREDI) long- form was used for children under 36 months of age. The CREDI is a globally-validated developmental assessment tool (Waldman et al., 2021), and the short-form version of the CREDI has previously been validated and used in Pakistan (Hentschel et al., 2024; McCoy et al., 2018). The tool consists of up to 100 caregiver-reported items developed according to typical abilities by age within the birth to age 3 range in 6-month increments (e.g., for children 0-6 months, 6-12 months, and so forth) and administration ends based on 5 consecutive incorrect answers so as not to distress the child. Items measured children’s developmental status across four primary domains: motor, language, cognition and social-emotional development. An age-standardized Z-score, created by uploading deidentified data to the CREDI application, allows comparison of a given sample to a global, advantaged, sample based on children from 15 low- and middle-income countries. 5 A child scoring more than 2 standard deviations below the mean is typically considered to be developmentally “off-track." 5 The reference sample contains 19,165 children who all have a mother who completed secondary school or higher education and live in a household where least one adult had engaged in 4 or more of the 6 “Play activities” from the Family Care Indicators with the child. 6 Children 36-72 months The AIM-ECD Direct Assessment (DA) was used for children 36-72 months (Pushparatnam et al., 2021). The AIM-ECD measures ECD in four domains: early literacy, early numeracy, executive functioning, and social emotional development. The AIM-ECD DA is structured into 14 subtasks, with a total of 77 questions/tasks, administered in a series of activities directly with the child. The AIM-ECD has previously been used in Khyber Pakhtunkhwa as well as other parts of Pakistan (Hentschel et al., 2024; Seiden et al., 2024). Items were designed to be used in various cultural or linguistic contexts with minimal adaptation (i.e., inclusion of literacy in both English and Urdu). Items are scored based on percentage correct with enumerators giving credit if items were responded to correctly in Pashto. The AIM-ECD Caregiver Report (CR), made up of 29 core items across early literacy, early numeracy, executive functioning, and social emotional development, was also utilized. Given the high correlation between the AIM-ECD DA and the AIM-ECD CR (corr=.71, p<.05), the AIM-ECD DA results were used as the primary outcome and the AIM-ECD CR results are available upon request. Maternal Mental Health Measures Depression. Maternal depression was measured by the Self-Reporting Questionnaire (SRQ-20), which has previously been used in Pakistan and has been found to be valid and reliable for screening for depressive disorders in rural Pakistan (Husain et al., 2006). The SRQ-20 contains 20 yes/no questions, with higher scores indicating higher levels of depression, and was found to have a high internal consistency reliability, with a Cronbach’s alpha of .89 in this sample. Anxiety. Maternal anxiety was assessed using the General Anxiety Disorder 7-item questionnaire (GAD-7). This 7-item questionnaire has been found to be valid in the Gilgit province of Pakistan (Ahmad et al., 2017) and has subsequently been used in other assessments of anxiety in Pakistan (Yasmin et al., 2021). The GAD-7 was found to have a high internal consistency reliability, with a Cronbach’s alpha of .80 in this sample. Parenting Stress. Parenting stress was measured using an adapted version of the Parental Stress Scale (PSS), which has previously been validated in Urdu in Punjab (Bilal et al., 2021). The adapted version used in this survey was designed by Ugarte et al., 2024 for Rohingya refugees in Bangladesh. The 10-item parenting stress scale used in this sample was found to have a high internal-consistency reliability, with a Cronbach’s alpha of .85 in this sample. Independent (Stressor and Other Covariate) Variables A variety of known stressors and correlates of maternal mental health and child development were also assessed. Stressors, including food insecurity, poverty, flooding impact, experiences of domestic violence, felling unsafe in one’s community, discrimination, living in a rural area, lack of enrollment in ECE, and low maternal educational attainment have consistently been associated with lower levels of maternal mental health conditions and child development across contexts (e.g., Hentschel et al., 2023; Iqbal & Ali, 2021; Reed et al., 2012: Sitwat et al., 2015; Yousuf et al., 2023) Other covariates, such as refugee status and child age, were also assessed. Refugee Status. At the time of the survey, sensitivities around “refugee status” were particularly heightened. As such enumerators were asked to keep notes on interviews to help identify whether 7 a respondent was likely a refugee – rather than to ask explicitly about citizenship status. Because many Afghan nationals living in Pakistan lack documentation and are not registered with UNCHR as either asylum seekers or confirmed refugees (UNCHR, 2024b), the study did not differentiate between verified refugees, asylum seekers and unregistered Afghan nationals, the group of which are collectively referred to as “refugees” within this paper. Enumerators determined refugee status of the household by confirmation of at least one of the following factors: the primary caregiver or her parents were born in Afghanistan; the use of Afghani Pashto, Persian and/or Dari as the primary language in the home; respondent declined to share identification, or the target child’s date of birth card or other documentation was valid for Afghanistan but not Pakistan; community leaders had identified the household as Afghani; or enumerator-reported knowledge of the household’s refugee/migrant status based on previous survey engagement. A “yes” response to any one of these indicators placed the respondent and household in the refugee category for analytical purposes. Food Insecurity. Food insecurity was assessed by the Food Insecurity Experience Scale (FIES), an 8-item questionnaire comprised of yes/no questions relating to experiences of food insecurity. It has been previously used in Pakistan and found to be reliable (Afridi et al., 2021). The Cronbach’s alpha of the scale was .91, indicating high internal consistency reliability. Asset Index. An asset index was calculated by a principal components analysis of a series of questions related to household materials, access to water, and home ownership, among other resource-related factors. For each household, based on respondent report, an asset quintile was created to categorize by quintile from lowest asset index values (quintile 1) to highest asset index values (quintile 5). Flooding Impact. Respondents were asked a series of questions about the impact of flooding on their day-to-day life in both 2022 and 2023. Households were categorized as being “impacted by flooding” if they answered “yes” to at least one question about whether flooding impacted their day-to-day life in 2022 or 2023. Domestic Violence. Domestic violence experiences and attitudes were measured based on self- report using three questions: (a) In your opinion, is a husband justified in hitting or beating his wife? (b) Did your husband ever hit or beat you? And (c) Are (were) you afraid of your husband: most of the time, sometimes, or never? These questions have been used in the Government of Pakistan Demographic and Health Survey 2017-2018. For the regression analyses, fear of husband was controlled for as it was the only variable that was correlated with all mental health outcomes and child development. Fear of husband was categorized from 1 to 3, with 1 indicating never afraid, 2 indicating sometimes afraid, and 3 indicating afraid most of the time. Community Safety. Perceptions of and experiences of safety in the community (or lack thereof) were determined by exposure to crime (i.e., if anyone has taken or tried taking something from you by using force or threatening to use force in the last 12 months) and perceptions of safety in the community—specifically, feeling safe while home alone at night or while walking through the neighborhood alone after dark. Discrimination. Respondents were asked if, in the past 12 months, they felt discriminated against or harassed on the basis of ethnic origin, immigration or refugee status, sex, sexual orientation, 8 age, religion or belief, disability, or some other reason. Individuals were classified as experiencing discrimination in the past 12 months if they reported experiencing discrimination for any of the mentioned reasons. School/ECE Enrollment. ECE enrollment was determined using an open-ended item administered only if the selected child was between 3 and 6 years old. Response options included nursery, prep, ECE, Katchi, kindergarten or grade 1. Area of residence. Enumerators categorized households as rural or urban based on recorded household location. Maternal Educational Attainment. Mothers indicated their highest level of educational attainment. Birth Registration. Mothers were asked if their child has a birth certificate. If they said yes, they were asked to provide proof. As such, there were three potential options: not registered (answered no to if their child has a birth certificate), birth certificate (includes those with and without proof of registration), and proof of registration (mothers that answered yes and provided proof). Child Age. Age of the child was recorded by either checking the child's birth registration or vaccination card, or where neither a birth certificate nor a vaccination card was available the child's age was recorded based on the mother's recall. Statistical Analyses Examining Maternal Mental Health by Refugee Status, Maternal Education Level, and Area of Residence First, descriptive characteristics of the sample were presented for child age group (i.e., 0-35 months, 36-72 months), child gender, area of residence (e.g., urban, rural), enrollment in school/ECE, school/ECE class (e.g., playgroup, katchi), school type (e.g., public, private), frequency of ECE attendance, birth registration, refugee status (“yes” inclusive of documented and undocumented individuals), maternal education level, household size and head of household age. Next, descriptive characteristics of the sample in terms of the prevalence of stressors (food insecurity, flood exposure, domestic violence, crime, discrimination, and asset index) were provided. Third, bivariate associations between levels of ECD, parenting stress, anxiety, and depression by refugee status, education level (less than a primary education v. primary education or more), and area of residence (urban v. rural) were calculated and Chi squared p-values were computed to assess differences. Fourth, pairwise Pearson correlations between stressors and maternal parenting stress, anxiety, and depression were computed. 9 Examining the Triple Interaction of Maternal Depression, Anxiety, and Parenting Stress on ECD First, the average ECD level as measured by the CREDI Z-score for children 0-35 months, and AIM-ECD for children 36-72 months (percent correct) was calculated for varying levels of maternal depression, anxiety and parenting stress respectively. Six unadjusted OLS linear regression analyses were run for each mental health variable (i.e., parenting stress, depression, and anxiety) on ECD for children aged 0-35 months and 36-72 months. Six adjusted OLS linear regression analyses were then run for each mental health variable (i.e., parenting stress, depression, and anxiety) on ECD for children 0-35 months and 36-72 months), controlling for stressors and other key confounders (i.e., refugee status, food insecurity, flood impact in 2022, fear of husband, asset index, less than primary maternal education, feeling unsafe while at home alone at night, child gender, and child age) that were correlated with both maternal mental health and ECD. In the 36-72 months regression analysis, enrollment in school/ECE was also controlled for. Next, an unadjusted OLS linear regression analysis was run on the triple interaction of maternal depression, anxiety, and parenting stress on ECD for 0-35-months. The same regression was run on ECD for 36-72 months, controlling for child age. Next an adjusted OLS linear regression analysis was run on the triple interaction of mental health conditions on ECD controlling for stressors (and other key confounders (refugee status, food insecurity, flood impact in 2022, fear of husband, asset index, maternal education, feeling safe while at home, child gender, and child age) that were correlated with both maternal mental health and ECD. The same OLS adjusted linear regression was run on ECD in the 36-72 month age group, where enrollment in ECE was also controlled for. Ethical Approval The study was approved by the Institutional Review Board of Research and Development Solutions (RADS) based in Islamabad, Pakistan on December 5th, 2023. RADS is registered with the Office for Human Research Protection at the U.S. Department of Health and Human Services (IORG0009092). Results Descriptive Overall, 2,102 mother-child dyads living in Khyber Pakhtunkhwa, Pakistan were sampled (Table 1). In only four households, men self-identified as the primary caregiver for the child; these dyads were dropped from the sample. Grandmothers and aunts comprised less than 1% of the sample. The majority of the sample lived in a rural area (77%) and most mothers (58%) had no education. Among heads of household, the mean age was 47 years. The average age of caregivers in the sample was 31.5 years, with a range of 15-72. Enumerators identified 6% of households as refugee 10 households. 6 The average household housed 8.2 people, with a range of 3 to 59 people within the house. The sample included twice as many children under age 3 (n=1,387 or 66.7%) as children aged 36-72 months (n=705 or 33.3%), weighting results accordingly. Table 1 Sample Descriptive Characteristics Percentage All children (n=2102) (weighted) n (raw) Child Age 0-35 months 66.67 1397 36-72 months 33.33 705 Child Gender Male 49.99 1076 Female 50.01 1026 Location Urban* 23.28 749 Rural 76.71 1353 Birth registration Birth certificate 69.72 1436 Proof of registration 21.65 455 Not registered 16.01 324 Refugee Status Yes 5.62 135 No 94.38 1967 Maternal Education No education** 58.14 1135 Completed pre-primary 4.60 106 Completed primary 12.32 258 Completed middle 14.87 347 Completed Secondary 4.63 120 Completed Bachelor’s Degree 2.45 65 Completed Graduate Degree 2.72 66 Religious Education .27 5 All children (n=2102) Mean (weighted) Mean (range) raw Child Age (months) 30.90 31.00 (0.00, 71.88) Household Size 8.25 8.13 (3, 59) Caregiver Age (years) 31.58 31.22 (15, 72) Household Head Age (years)**** 46.41 46.41 (18, 94) 6 Official estimates on registered and unregistered refugees in Khyber Pakhtunkhwa suggest this might be an overestimate and, as such, caution in interpreting these findings is warranted. 11 Percentage Children aged 36+ months ((n=705) (weighted) n (raw) ECE enrollment Yes 62.28 435 No 37.72 270 ECE class*** Playgroup 32.48 156 Nursery 24.28 103 Katchi 31.69 125 Other ECE/Missing 11.55 51 School Type Government 72.11 288 Private 25.37 136 Refugee School .79 2 Other [deeni madaris, non-formal basic education, distance education/privately] 1.74 9 Frequency of ECE attendance (days/week)***** 1-3 days 10.57 38 4-5 days 24.20 116 6 days 65.23 276 Frequency of ECE attendance (hours/day) 1-3 hours 12.00 43 4-5 hours 52.29 232 6 hours+ 35.71 155 *Urban includes urban, semi-urban, and inner city **No education= never attended school, some primary= completed grades 1-4, completed primary= completed grades 5-7, includes those who completed some middle school, completed middle= completed grades 8-10, includes those who completed some secondary school, completed secondary= completed F.A/ F. Sc/ I. Com/ ICS/A-Level or Polytechnic Diploma, completed Bachelor’s degree= completed B.A/B.Sc./ B. Com/ B. Ed/ Degree in Engineering/Accounting/Law/Agriculture, completed graduate degree= completed M.A/M. Sc or M.Phil or PhD, religious education=completed religious education or classified as “Islamic scholar” ***Other ECE/Missing includes Class 1+, Prep, Madrasa, and Kindergarten; ****Data was missing for 3 household heads indicating that these household heads did not live in the household *****Data was missing for 5 children on days per week and hours per day attendance in ECE CREDI Z-Scores are norm-referenced and standardized against a global population 7 of children from low-and middle-income countries who are from an “advantageous environment.” An advantageous environment is defined as children with maternal educational attainment of secondary or higher and adequate levels of early stimulation (indicated as participated in 4 or more 7 A total of 4,652 children from Bangladesh, Brazil, Cambodia, Chile, Colombia, Ghana, Guatemala, India, Jordan, the Lao People’s Democratic Republic, Lebanon, Nepal and Pakistan were included in the CREDI reference population (24% of the full cross-country sample). 12 stimulating activities with any caregiver in the previous three days). A Z-score of 0 would indicate that a child in this sample has a similar developmental status to that of a child in the CREDI reference sample who is the same age. As demonstrated by Figure 1, the average child in this sample has a developmental status almost 2 standard deviations (1.72) below the mean, indicating much lower levels of child development in this sample compared to the global population. Distributions varied by various characteristics of interest, such as maternal education, living in a rural area and refugee status (Figures 2-4). Figure 1 Sample Distribution Compared to Global Distribution (n=1397) Figure 2 Sample Distribution for Children with Mothers with Less than a Primary Education Compared to Global Distribution (n=824) 13 Figure 3 Sample Distribution for Refugee Children Compared to Global Distribution (n=92) Figure 4 Sample Distribution for Children Living in Rural Areas Compared to Global Distribution (n=902) AIM-ECD direct assessment scores were created by calculating the percent correct out of the 77 direct assessment tasks and questions asked. There were four early literacy tasks (listening comprehension, letter identification, initial sound discrimination and name writing), six numeracy 14 tasks (receptive spatial vocabulary, producing a set, simple addition, number identification, number comparison and shape identification), three executive functioning tasks (head, toes, knees, shoulders task, forward digit span, and backward digit span) and one social emotional learning tasks (perspective taking/empathy). The average child in KP scored a 33% on the AIM-ECD (95% CI: 40.9%, 34.9%). The distribution of AIM-ECD direct assessment scores is reported in Figure 5. Figure 5 Sample Distribution of AIM-ECD Direct Assessment Scores (n=635) 8 The prevalence of stressors is presented in Table 2. For example, 26% of the sample reported having to skip a meal, with 14% of the sample reporting they were hungry but did not eat. Approximately 8% of the sample was impacted by flooding in 2022, 6% of the sample experienced flooding impacts in 2023, and 8% of the sample were impacted by flooding in either 2022 or 2023. Of those impacted by flooding, the most common disruptions were to household members schooling, essential travel, and household members well-being. The majority of the sample reported being afraid of their husband sometimes or most of the time, and 22% of the sample reported ever being beaten by their husband. Over half of caregivers reported feeling safe or very safe at home after dark (63%), and 53% reported feeling safe or very safe walking in their neighborhood after dark. Approximately 5% of caregivers reported experiencing some form of discrimination, with ethnic origin and sex as the most common forms of discrimination. 8 Only 635 individuals were sampled for the AIM-ECD initial sound discrimination sub-domain due to a coding error. As a result, for the overall AIM-ECD score and AIM-ECD literacy sub-domain we only have data from 635 children instead of 705. 15 Table 2 Prevalence of Stressors (n=2,102) Variable Percentage (weighted) n (raw) Food Insecurity Worried you would not have enough food to eat 36.90 782 Unable to eat nutritious foods 51.36 1050 Only ate a few kinds of foods 49.54 1027 Had to skip a meal 26.20 543 Ate less than you thought you should 30.93 661 Ran out of food 28.51 608 Were hungry but did not eat 14.49 333 Whole day without eating 7.45 182 Flood Exposure Impacted by Flooding in 2022 or 2023 8.17 173 Impacted by 2023 flooding 6.22 132 Impacted by 2022 flooding 7.73 165 Type of 2023 flooding impact Eviction from dwelling 7.15 6 Destruction of house 14.39 16 Damage to house 11.28 18 Disruption of work 13.72 21 Lost substantial income 14.69 23 Disruption of essential travel 18.18 24 Impact on household members well-being 21.15 35 Disruption of household members schooling 20.20 39 Do not know 28.31 37 Type of 2022 flooding impact* Eviction from dwelling 4.68 5 Destruction of house 15.38 26 Damage to house 12.50 25 Disruption of work 18.15 33 Lost substantial income 12.08 24 Disruption of essential travel 23.40 42 Impact on household members well-being 19.82 42 Disruption of household members schooling 21.79 51 Do not know 29.76 46 Afraid of Husband Never 20.81 458 Sometimes 53.99 1078 Most of the time 22.65 520 16 Variable Percentage (weighted) n (raw) Prefer not to answer/not maternal respondent 2.55 46 Believe that husband is justified in beating wife No 94.16 1989 Yes 2.49 53 Prefer not to answer/not maternal respondent 3.36 60 Husband has ever beat you No 74.24 1576 Yes 21.71 453 Prefer not to answer/not maternal respondent 4.04 73 Seek help No 92.87 1943 Yes 5.24 119 Prefer not to answer/not maternal respondent 1.89 40 How safe do you feel when you are at home alone after dark? Very Safe 16.17 422 Safe 47.11 932 Unsafe 10.67 241 Very Unsafe 3.62 69 Never Alone After Dark 22.43 438 How safe do you feel walking alone in your neighborhood after dark Very Safe 8.75 200 Safe 43.89 877 Unsafe 11.56 265 Very Unsafe 4.09 82 Never Walk Alone After Dark 31.71 678 Experience of Discrimination Yes 4.66 109 No 95.34 1993 Discrimination Reason** Ethnic Origin 20.27 26 Immigration/refugee status 10.91 17 Sex 21.96 26 Sexual Orientation 3.96 4 Age 19.32 17 Religion 3.86 6 Disability 2.67 2 Other 38.91 38 Asset Index 17 Variable Percentage (weighted) n (raw) Poorest 27.24 421 Second 24.28 420 Third 19.51 421 Fourth 16.07 420 Richest 12.90 420 Note. *Denominator is out of those who experienced some flooding impact in either 2023 or 2022 **Denominator is out of those who reported experiencing some level of discrimination Refugee households, households with a mother with less than a primary education, and households in rural areas on average had children with significantly lower levels of child development for children 0-35 months, measured by CREDI Z-scores, compared to a global population (Table 3). As such, the average non-refugee child in this population would be considered to be developmentally on track (CREDI Z-score= -1.69 SDs) while the average refugee child in this population would not be considered to be developmentally on track (CREDI Z-score= -2.35). In terms of child development for children 36-72 months, measured by percentage correct on the AIM-ECD, the same differences were seen but only some relationships were statistically significant. Specifically, there were no significant differences by refugee-status, but mothers with less than a primary education had children with significantly lower AIM-ECD scores on all domains except for executive functioning, where the same trend was seen but it was not a statistically significant difference. Mothers living in rural areas had children with significantly lower AIM-ECD scores on all domains of development. Across both age groups, mothers in refugee households experienced significantly higher levels of depression, on average, and no significant differences were seen for depression levels by maternal education status (Table 4a, Table 4b). Mothers in urban areas with a selected child 0-3-years-old experienced significantly higher levels of maternal depression. Mothers in refugee households experienced significantly higher levels of anxiety, on average, and no significant differences were seen for anxiety by area of residence. Mothers with less than a primary education experienced significantly higher levels of maternal anxiety in households with a selected child 3-6-years-old. Mothers in refugee households experienced significantly higher levels of parenting stress, on average, and no significant differences were seen for parenting stress by maternal education or area of residence. 18 Table 3 Child Development by Refugee Status, Education, and Area of Residence; Weighted Percentages Non- Less than Primary or Variable Refugee Refugee p-value primary more p-value Rural Urban p-value Average CREDI Score (Z- Score) (n=1397) n=92 n=1305 n=824 n=573 n=902 n=495 Overall -2.35 -1.69 <.001 -1.81 -1.59 .001 -1.83 -1.42 <.001 Language -1.84 -1.29 <.001 -1.36 -1.26 .074 -1.40 -1.09 <.001 Motor -1.96 -1.64 .010 -1.75 -1.50 <.001 -1.73 -1.44 .010 Social Emotional -2.20 -1.63 <.001 -1.75 -1.53 .002 -1.78 -1.32 <.001 Cognitive -2.43 -1.81 <.001 -1.92 -1.72 .005 -1.96 -1.51 <.001 Average AIM-ECD Score DA (% Correct) (n=705)* n=43 n=662 n=422 n=283 n=451 n=254 Overall 31.52 33.01 .666 30.59 36.76 .003 30.99 38.48 <.001 Numeracy 51.10 49.08 .652 46.08 54.35 .001 47.63 53.76 .011 Literacy 33.65 36.40 .537 32.74 42.00 .001 34.31 41.76 .003 Executive Functioning 27.82 27.82 .999 25.77 31.20 .010 26.71 31.02 .041 Social Emotional 12.83 17.17 .311 16.48 17.62 .673 14.66 23.47 .003 Note. *3 individuals dropped in overall score and literacy score with missing data for all literacy questions; remaining missing data assumed to be incorrect *68 individuals dropped in overall score and early literacy score with missing data for all of initial sound identification questions, 4 individuals dropped in overall score and early numeracy with missing data for all addition and/or number comparison questions, 6 individuals dropped in overall score and executive functioning score with missing data for all executive functioning questions Table 4a Maternal Mental Health by Refugee Status, Education, and Area of Residence for Households with Child 0-3-years-old; Weighted Percentages Non- Less than Variable Refugee Refugee p-value primary Primary or more p-value Rural Urban p-value n=92 n=1305 n=824 n=573 n=902 n=495 Experience of Depression (SRQ-20>=9) 56.58 25.21 <.001 27.04 26.63 .881 24.72 33.27 .002 Experience of Anxiety (GAD-7) <.001 .357 .377 Minimal Anxiety (0-4) 19.67 57.99 54.26 58.64 54.82 59.25 Mild Anxiety (5-9) 38.32 33.87 35.73 31.52 35.33 30.52 Moderate Anxiety (10-14) 37.11 7.01 8.37 9.01 8.43 9.16 Severe Anxiety (15-20) 4.91 1.13 1.64 0.84 1.42 1.06 Parenting Stress (PSS)*; mean (se) 7.03 (.34) 6.05 (.16) .008 6.24 (.21) 5.87 (.23) .231 5.99 (.19) 6.44 (.25) .151 Note. The maximum parenting stress score was 33, with 0 indicating no parenting stress and 33 indicating maximum parenting stress. In this sample, scores ranged from 0-25. Table 4b Maternal Mental Health by Refugee Status, Education, and Area of Residence for Households with Selected Child 3-6- years-old; Weighted Percentages Non- Less than Variable Refugee Refugee p-value primary Primary or more p-value Rural Urban p-value n=43 n=662 n=422 n=283 n=451 n=254 Experience of Depression (SRQ-20>=9) 52.34 32.56 .026 35.61 30.61 .248 33.08 35.66 .545 Experience of Anxiety (GAD-7) <.001 .028 .068 Minimal Anxiety (0-4) 25.25 54.22 48.37 59.36 53.52 49.49 Mild Anxiety (5-9) 48.07 37.11 41.70 31.21 36.23 42.25 Moderate Anxiety (10-14) 17.93 7.12 7.26 8.60 8.77 4.84 Severe Anxiety (15-20) 8.75 1.55 2.66 0.84 1.48 3.42 Parenting Stress (PSS)*; mean (se) 7.73 (.60) 6.37 (.24) .031 6.53 (.27) 6.31 (.40) .658 6.45 (.27) 6.45 (.35) .992 Note. The maximum parenting stress score was 33, with 0 indicating no parenting stress and 33 indicating maximum parenting stress. In this sample, scores ranged from 0-25 Several stressors showed significant relationships with maternal mental health concerns. Higher levels of food insecurity, being impacted by flooding in 2022 or 2023, being afraid of one’s husband, seeking help in the context of domestic violence, and experiencing discrimination were significantly associated with higher levels of maternal depression, parenting stress and anxiety (Table 5). Believing that one’s husband is justified in beating his wife and reporting ever being beaten by one’s husband was correlated with higher levels of maternal depression and anxiety, but not parenting stress. Feeling unsafe when home alone after dark was correlated with higher levels of maternal anxiety and parenting stress, but not maternal depression, while feeling unsafe when walking alone in the neighborhood after dark was correlated with higher levels of maternal depression but not maternal anxiety or parenting stress. Neither being robbed in the previous 3 years nor asset index was significantly correlated with maternal parenting stress, depression, or anxiety. Table 5 Pairwise correlations between stressors and maternal parenting stress, depression, and anxiety (n=2,102) Variable Depression (0-20) Anxiety (0-20) Parenting Stress (0-33) Food Insecurity .26* .16* .09* Worried you would not have enough food to eat .25* .15* .06* Unable to eat nutritious foods .08* -.01 -.01 Only ate a few kinds of foods .16* .08* .09* Had to skip a meal .18* .09* .07* Ate less than you thought you should .25* .17* .13* Ran out of food .23* .15* .06* Were hungry but did not eat .31* .24* .10* Whole day without eating .26* .22* .11* Impacted by Flood Impacted in 2022 .16* .17* .07* Impacted in 2023 .12* .14* .08* Domestic Violence Afraid of Husband .23* .15* .10* Believe that husband is justified in beating wife .11* .08* .05* Husband has ever beat you .19* .12* -.03 Seek help .20* .19* .10* Crime Robbed in last 3 years .08* .05* .06* Feel unsafe when home alone after dark (1= very safe, 4= very unsafe) .03 .16* .26* Feel unsafe when walking alone after dark (1= very safe, 4= very unsafe) -.05* -.04 .03 Discrimination (yes=1, no=0) .09* .06* .15* Asset Quintile (5= richest, 1= poorest) .02 -.03 .02 Note. *Indicates significance at alpha=.05 level 21 Examining the Triple Interaction of Maternal Depression, Anxiety, and Parenting Stress on ECD First, depression, anxiety and parenting stress were dichotomized. Individuals who scored a 9 or above on the SRQ-20 were considered to be depressed based on a cut-off previously established on a validation study of the SRQ-20 in Pakistan (Husain et al., 2006). Individuals who scored a 10 or above on the GAD-7, indicating moderate to severe anxiety, were considered to be anxious. An adapted version of the PSS-20 was used in our sample, to create a cut-off for parenting stress we considered anyone scoring a 10 or above on the PSS-11, which was above the 75th percentile, to be stressed. Based on these three dichotomous variables, 40% of the sample who had a child between 0-35 months of age experienced either depression, anxiety, or parenting stress or a combination of the three (Figure 6, panel a). Most individuals experienced depression only (13%) or parenting stress only (11%). The most common co-occurrence was depression and parenting stress (6%). 45% of the sample who had a child between 36-72 months of age experienced either depression, anxiety, or parenting stress or a combination of the three (Figure 6, panel b). Most individuals experienced depression only (17%) or parenting stress only (11%). The most common co-occurrence was depression and parenting stress (9%). Figure 6 Prevalence and co-occurrence of maternal mental health conditions by age group (0-35 months and 36-72 months, respectively) Panel a: Mothers of 0-35 month olds Panel b: Mothers of 36-72 month olds The average child development CREDI Z-Score (for children 0-35 months) and percent correct AIM-ECD score (for children 36-72 months) was calculated for varying levels of maternal mental health (Table 6). For children 0-35 months, those that lived in households with mothers experiencing both anxiety and parenting stress had the lowest average CREDI Z-Score (2.02 standard deviations below the mean). For children 36-72 months, those that lived in households with mothers experiencing anxiety and parenting stress only had the lowest average AIM-ECD Score (27.4% correct) followed by those that experienced depression, anxiety or parenting stress (27.7% correct). For both the 0-35 month age group and the 36-72 month age group, those in 22 households that experienced only parenting stress had on average the highest child development scores (1.51 standard deviations below the mean and 47.5% correct, respectively). Table 6 Average child development by maternal mental health Average ECD for 0-3-year- Average ECD for 3-6-year- olds (CREDI Z-Score) olds (% correct AIM-ECD n (95% CI) (n=1397) n DA) (n=635) No depression, anxiety or parenting stress 777 -1.76 331 30.05% (-1.84, -1.67) (27.39, 32.71) Depression only 210 -1.71 100 30.78% (-1.91, -1.51) (26.41, 35.14) Anxiety only 16 -1.83 2 N/A* (-2.42, -1.24) Parenting stress only 165 -1.51 78 47.52% (-1.68, -1.33) (41.82, 53.22) Depression & anxiety 51 -1.82 25 38.88% (-2.19, -1.45) (25.97, 51.79) Depression & parenting stress 102 -1.68 62 38.38% (-1.92, -1.42) (31.60, 45.16) Anxiety & parenting stress 14 -2.02 9 27.40% (-2.46, -1.57) (8.76, 46.04) Depression, anxiety & parenting stress 62 -1.84 28 27.72% (-2.14, -1.55) (19.91, 35.53) Note. *Only 2 individuals in the 36-72 month age group lived in a household that experienced only anxiety For children under 36 months of age, individual linear regression analyses were run with each maternal mental health condition (depression, anxiety, parenting stress) on child development as measured by the CREDI Z-score (Table 7). The unadjusted model regressed the exposure variable on ECD. The following model was fit: � = 0 + 1 ( ℎ ) + The adjusted regression analysis regressed the exposure variable on ECD, controlling for food insecurity, flooding impact, fear of husband, asset index, maternal education, feeling safe when home after dark, child gender, refugee status, and child age. Individually, there was no significant association of parenting stress or depression on ECD. A significant negative association of anxiety on ECD was found in the unadjusted model, but once relevant community level stressors and confounders were controlled for, the association, while still negative, was no longer significant. 23 Table 7 Individual Linear Regression Results of Parenting Stress, Depression, and Anxiety on ECD (0-35 months) (n=1397) Linearized Beta Standard R- Outcome Predictor Coefficient Error t-value p-value Squared ECD Unadjusted Parenting Stress .005 .007 .800 .427 .001 Depression .004 .007 .630 .527 .000 Anxiety -.016 .009 -1.910 .057 .004 ECD Adjusted Parenting Stress .008 .007 1.210 .225 .071 Depression .010 .007 1.380 .168 .072 Anxiety -.006 .009 -.670 .505 .070 Note. Anxiety is a continuous score, 0= least, 21= most anxious Parenting Stress is a continuous score, 0=least, 25= most stressed Depression is a continuous score, 0=least, 20= most depressed (+) controlled for food insecurity (continuous 0-8; higher indicates more insecurity, flood impact (impacted in 2022 or 2023), fear of husband (never afraid=0, sometimes afraid=1, always afraid=2), asset index (1= poorest, 5= wealthiest), maternal education (1=less than primary up to 8= graduate degree), feeling safe while home (1= very safe up to 4= very unsafe), child gender (0=male, 1=female), refugee status and child age (continuous in months) Individual linear regression analyses were run with each exposure variable (depression, anxiety, parenting stress) on the AIM-ECD sum score, measured as percentage correct (Table 8). The unadjusted model regressed the exposure on ECD for 3- to 6-year-olds controlling only for child age, while the adjusted regression analysis regressed the exposure on ECD for 3- to 6-year-olds controlling for food insecurity, flooding impact, fear of husband, asset index, maternal education, feeling safe when home after dark, child gender, refugee status, child age, and enrollment in early childhood education. A significant positive association (p<.05) of parenting stress on ECD for children aged 36-72 months was found in the unadjusted and adjusted models. A significant negative association was seen between maternal anxiety and ECD for 3- to 6-year-olds in the adjusted model, but not the unadjusted model. No association was seen between maternal depression and ECD in either the unadjusted or adjusted model. However, the coefficient in both instances is negatively signed. Table 8 Individual Linear Regression Results of Parenting Stress, Depression, and Anxiety on ECD (36- 72 months) Direct Assessment (n=635) Beta Linearized R- Outcome Predictor Coefficient Standard Error t-value p-value Squared ECD Unadjusted Parenting Stress 0.007 0.002 3.560 <.001 .092 Depression -0.001 0.002 -0.480 .629 .063 Anxiety -0.004 0.003 -1.510 .133 .068 ECD Adjusted Parenting Stress 0.004 0.002 1.980 .048 .278 Depression -0.003 0.002 -1.46 .144 .273 Anxiety -0.006 0.003 -2.250 .025 .278 24 Next, two linear regression analyses were run to understand the association with the triple interaction of maternal depression, anxiety, and parenting stress on child development for children 0-35 months. An unadjusted linear regression analysis was run of the triple interaction of maternal depression, anxiety and parenting stress on child development, and an adjusted regression analysis was run controlling for food insecurity, flooding impact, fear of husband, asset index, maternal education, feeling safe when home after dark, child gender, refugee status, and child age. The following model was fit: � = 0 + 1 ( ) + 2 ( ) + 3 ( ) + 4 ( ∗ ) + 5 ( ∗ ) + 6 ( ∗ ) + 7 ( ∗ ∗ ) + A significant negative interaction was found between depression, anxiety, and parenting stress on child development for 0-35 months in both the unadjusted and adjusted models at the p<.1 level (Table 9). That is, as levels of depression, anxiety, and parenting stress increased, there was a significant negative association with levels of child development. Table 9 Linear Regression Results of Depression, Anxiety, and Parenting Stress Interaction on ECD (0- 35 months) (n=1,397) Beta Linearized Outcome Predictor Coefficient Standard Error t-value p-value ECD Unadjusted (R Squared= .0207) Intercept -1.7375 .094 -18.550 <.001 Anxiety -0.0717 .030 -2.410 .016 Depression 0.0445 .019 2.360 .019 Parenting Stress -0.0070 .015 -0.460 .643 Anxiety*Depression -0.0001 .002 -0.070 .942 Parenting Stress*Depression 0.0005 .002 0.250 .802 Anxiety*Parenting Stress 0.0075 .003 2.180 .029 Anxiety*Parenting Stress*Depression -0.0005 .000 -1.870 .061 ECD Adjusted (+) (R Squared= .0849) Intercept -2.2235 .160 -13.930 <.001 Anxiety -0.0505 .030 -1.710 .088 Depression 0.0407 .018 2.300 .022 Parenting Stress -0.0000 .015 -0.00 1.000 Anxiety*Depression 0.0001 .002 0.070 .946 Parenting Stress*Depression 0.0004 .002 0.220 .827 Anxiety*Parenting Stress .0060 .003 1.760 .079 25 Anxiety*Parenting Stress*Depression -.0005 .000 -1.860 .062 Note. (+) controlled for food insecurity (continuous 0-8; higher indicates more insecurity, flood impact (impacted in 2022 or 2023), fear of husband (never afraid=0, sometimes afraid=1, always afraid=2), asset index (1= poorest, 5= wealthiest), maternal education (1=less than primary up to 8= graduate degree), feeling safe while home (1= very safe up to 4= very unsafe), child gender (0=male, 1=female), refugee status, and child age (continuous in months) The same two linear regression models were run on ECD for 3- to 6-year-olds. In the unadjusted model only child age was controlled for, and in the adjusted model, the stressors and confounders were controlled for, with the addition of enrollment in early childhood education. A significant negative interaction was found between depression, anxiety, and parenting stress on child development for 36-72 months in both the unadjusted model at the p<.05 level and in the adjusted model at the p<.1 level (Table 10). That is, as levels of depression, anxiety, and parenting stress increase there is a significant negative association with levels of child development for 36-72 months. Table 10 Linear Regression Results of Depression, Anxiety, and Parenting Stress Interaction on ECD Direct Assessment (36-72 months) (n=635) Linearized Beta Standard Outcome Predictor Coefficient Error t-value p-value ECD Unadjusted (R Squared= .0795) Intercept 0.3299 .036 9.08 <.001 Anxiety -0.0291 .009 -3.24 .001 Depression -0.0135 .006 -2.15 .032 Stress 0.0095 .006 1.72 .086 Anxiety*Depression 0.0036 .001 3.86 <.001 Stress*Depression 0.0010 .001 1.43 .152 Anxiety*Stress 0.0012 .001 1.09 .274 Anxiety*Stress*Depression -0.0003 .000 -2.54 .011 ECD Adjusted (+) (R Squared= .2492) Intercept -0.1893 .0724 -2.62 .009 Anxiety -0.0178 .0084 -2.11 .035 Depression -0.0143 .0059 -2.41 .016 Stress 0.0114 .0047 2.41 .016 Anxiety*Depression 0.0030 .0009 3.34 .001 Stress*Depression 0.0009 .0007 1.26 .209 Anxiety*Stress 0.0001 .0011 0.06 .950 Anxiety*Stress*Depression -0.0002 .0001 -1.85 .065 Note. (+) controlled for food insecurity (continuous 0-8; higher indicates more insecurity, flood impact (impacted in 2022 or 2023), fear of husband (never afraid=0, sometimes afraid=1, always afraid=2), asset index (1= poorest, 5= wealthiest), maternal education (1=less than primary up to 8= graduate degree), feeling safe while home (1= very safe up to 4= very unsafe), child gender (0=male, 1=female), refugee status, ECE attendance (0= not currently attending, 1= currently attending) and child age (continuous in months) 26 Next, linear predictions of child development were calculated at varying levels of depression, anxiety, and parental stress (Table 11). For example, the average predicted level of child development for a child under 3 years old in a household with no maternal anxiety, depression, or parenting stress is 1.74 standard deviations below the mean, compared to the average predicted level of child development for a child in a household with the sample’s median level of maternal anxiety (equivalent to a score of 4), no depression, and no parenting stress of 2.02 standard deviations below the mean. Similarly, for children 3-6-years-old, the average predicted child development in a household with no maternal anxiety, depression, or parenting stress is 33%, compared to the average predicted level of child development for a child in a household with the sample’s median level of maternal anxiety and depression, and no parenting stress of 22%, over 10 percentage points lower. This can also be visualized by the Margins plot (Figure 7). For children under 3-years-old, as demonstrated by the yellow line, at median levels of depression (depression=6) and parenting stress (parenting stress=6), but minimum levels of maternal anxiety, a child’s predicted child development Z-score is 1.49 SDs below the mean. However, as levels of anxiety increase to median levels of anxiety, the predicted child development Z-score moves to 1.68 SDs below the mean. Figure 8 describes the same phenomenon for children 3-6-years-old. For children 3-6-years- old, as demonstrated by the yellow line, at median levels of depression and parenting stress, but minimum levels of maternal anxiety, a child’s predicted child development score is 34%. However, as levels of anxiety increase to median levels of anxiety, the predicted child development Z-score moves to 30%. Table 11 Linear predictions of child development by maternal mental health Average ECD for 0-3-year- Average ECD for 3-6-year-olds (% olds (CREDI Z-Score) correct AIM-ECD) (95% CI) Direct (95% CI) (n=1397) p-value Assessment (n=635) p-value No depression, anxiety or parenting stress -1.74 (-1.92, -1.55) <.001 32.99% (25.85, 40.13) <.001 Median depression only -1.47 (-1.68, -1.26) <.001 24.91% (19.35, 30.47) <.001 Median anxiety only -2.02 (-2.24, -1.81) <.001 21.35% (15.12, 27.59) <.001 Median parenting stress only -1.78 (-1.92, -1.63) <.001 38.70% (33.90, 43.50) <.001 Median depression & anxiety -1.76 (-1.89, -1.64) <.001 21.80% (18.43, 25.18) <.001 Median depression & parenting stress -1.49 (-1.63, -1.36) <.001 34.32% (30.67, 37.96) <.001 Median anxiety & parenting stress -1.89 (-2.02, -1.75) <.001 30.00% (26.22, 33.77) <.001 Median depression, anxiety & parenting stress -1.68 (-1.75, -1.60) <.001 30.42% (28.37, 32.48) <.001 27 Figure 7 Adjusted Prediction of Child Development (0-3-years-old) at Varying Levels of Maternal Mental Health Figure 8 Adjusted Prediction of Child Development (3-6-years-old) at Varying Levels of Maternal Mental Health 28 Discussion This study finds that refugee households, households with a mother with a less than a primary school education, and those in rural areas had children with significantly lower levels of child development, on average. Mothers in refugee households and those with less than a primary school education experienced significantly higher levels of anxiety on average than non-refugee mothers, and those in refugee and urban households experienced significantly higher levels of depression on average than non-refugee and rural mothers. Looking at the compounding association of multiple maternal mental health conditions on ECD, a significant negative triple interaction between anxiety, parenting stress, and depression on ECD was seen for both 0-3-year-olds (0-35 months) and 3-6-year-olds (36-72 months), which remained significant even after controlling for stressors and other confounders. The study’s findings shed light on the importance of reducing mental health concerns and exposure to stressors that burden mothers caring for young children and thus impact ECD. Firstly, the study’s findings point to an urgent need to consider how best to support vulnerable mothers in the face of few supports and resources. Access to health care, especially mental health care, is low for many women in Pakistan, particularly those living in rural or remote areas, in part because the ratio of providers to patients is very low for psychologists, psychiatrists, and other mental health providers (Sikander, 2020). In addition, the quality of services is rarely optimal. Lady Health Workers who provide home visits provide perinatal care and family planning services for many women but are often not equipped in terms of knowledge, time or support to provide preventative mental health services. In terms of treatment for women’s mental health concerns, these are also quite limited in Pakistan, especially regarding quality of programming, but extant evaluations point to successful treatment protocols. For instance, in a systematic review of relevant publications examining the effectiveness of healthcare interventions on women’s health outcomes in Pakistan, just 18 were considered by the authors to be relevant to mental health (i.e., counseling, awareness interventions, and social and psychological interventions); of those, 14 of them showed significant and positive results, and the remaining four showed partially favorable results (Rizvi Jafree & Barlow, 2022). These findings underscore the limited scale and efficacy of the few existing interventions when they are evaluated. The finding that mothers in refugee households experienced significantly higher levels of depression, anxiety, and parenting stress reinforce previous studies which identify refugees as a particularly vulnerable group when it comes to mental health and well-being, as they often experience cumulative stressors over the course of displacement experiences, including premigration, flight, exile, and resettlement periods (Porter & Haslam, 2005); and refugee or refugee-like contexts can extend not only over years but across generations, compounding chronic concerns. Previous studies have established how interventions rarely provide adequate support for refugee mothers’ specific mental health needs and challenges (Abi Zeid Daou, 2022). Similarly, existing evidence indicates high maternal parenting stress among under-educated groups, as well as associations between displaced status and higher parenting stress (Parkes et al, 2015). While both host-country populations and refugees are at risk for unfavorable outcomes in FCV settings, where resources tend to be scant and needs tend to be high, refugees who lack documentation face uncertainty around eligibility for services, high levels of anxiety about deportation, reduced access 29 to health care, employment and education, and high risks for discrimination and exploitation (UNHCR 2024b). Even with the coverage provided by asylum seeker and refugee status, emotional, physical and financial reserves tend to be low relative to the general population, with spillover effects between parents and young children being a concern, as indicated in current analyses. Secondly, the study demonstrated exposure to stressors increases the likelihood of experiencing high levels of depression, anxiety, and parenting stress. Community-level stressors such as low access to basic resources like affordable food, disruptions from flooding or other natural disasters, and neighborhood crime increase maternal stress and anxiety, according to global literature and studies within Pakistan (Iqbal & Ali, 2021; Reed et al., 2012: Sitwat et al., 2015; Yousuf et al., 2023). Moreover, risks and stressors occurring at the household and individual levels, including poverty, discrimination, exposure to intimate partner and other domestic violence, and fear related to safety in the home or community further correlate with depression, anxiety and parenting stress (Adamu & Adinew, 2018; Irum et al., 2022; Maselko et al., 2022; Wei et al., 2024). With reduced access to public services, social capital, health care, and familiar routines – all of which are protective factors – in the face of expanded stressors, mental health concerns are more likely to manifest (El Gemayel & Rigon, 2020). The additive effects of cumulative and prolonged risk exposure and low access to beneficial supports and resources can jeopardize health and well-being for vulnerable mothers such as refugees or under-educated women (Bogic et al., 2015; Chen et al., 2024; Fazel et al., 2005). Households with young children in Pakistan experience multiple risks that put stress on caregivers and children in the home, and the more stresses a household experiences, the greater the decline in outcomes for young children across physical, cognitive and social-emotional domains (Hentschel et al., 2023; Tomlinson et al., 2023). This study adds to the knowledge base by showing that maternal mental health concerns are relatively prevalent and should be considered in future studies considering stressors in family systems in FCV and under-resourced contexts. While individual associations between various stressors on women’s well-being have been previously documented, this study additionally suggests the need to consider families’ contexts comprehensively—mothers with young children face multiple stressors in many cases, compounding risks to depression and anxiety and possibly attenuating the ability to provide nurturing care. While mental health is a concern for all caregivers, the high and intergenerational cost for caregivers and children in at-risk households is salient, especially in post-conflict settings (East et al., 2018). Thus, based on the strength of current analyses and previous research (De Oliveira et al., 2019; Maselko et al., 2018) vulnerable mothers would benefit from both reduced exposure to risks and stressors and tailored mental health interventions (Parkes et al., 2015; Rahman et al., 2019), as would their young children. Across both age groups high levels of maternal anxiety had the strongest negative association with ECD. The results illustrate the crucial role of maternal anxiety, which remains relatively underexamined in Pakistan, although the World Health Organization (2023) has noted extraordinarily high prevalence rates of anxiety disorders worldwide. The effects of anxiety on child outcomes vis-à-vis maternal symptomatology are apparent in multiple domains of ECD (e.g., socio emotional, language, cognitive, gross motor and fine motor development) (Ali et al., 2013). 30 The current results provide confirmation that more work is needed to develop and implement effective interventions to reduce maternal anxiety, with potential to improve ECD. Finally, the study offers unique insights into the prevalence of various and potentially co-occurring mental health concerns among mothers and their concomitant association with children’s development. A significant negative interaction effect between continuous measures of maternal depression, anxiety and parenting stress was observed on ECD across the birth to age 6 span in this study, an effect that remained significant even after controlling for stressors and confounders. While previous studies have explored comorbidity and some have established clear links between parental mental health and children’s internalizing and externalizing behaviors (Kingston et al., 2018), few or no studies have previously explored these in relation to mothers and young children’s development in Pakistan. Despite the presence of enabling policy and programming characteristics, the overall climate for maternal and child well-being for households with children is lacking, based on the poor to dire average outcomes seen across ECD levels, and maternal depression and anxiety rates. Policy improvements are needed that focus on at-risk communities and programs that afford young families with better opportunities to access and utilize services to directly benefit women and children. Cross-sectoral programming designed to meet families’ holistic needs would also increase the odds of maternal and child well-being, including increasing provision of cash transfers through BISP, expanding access to ECE programs, ensuring food availability and affordability for mothers, and implementing emergency-response programs for continuity of basic services in the context of floods and other disruptive disasters to protect vulnerable mothers and young children. Both direct interventions that provide mental health services and indirect interventions that reduce exposure to stressors within communities and households are approaches to alleviate mental health concerns among mothers of young children in the FCV setting of the current study. In addition to individual-level impacts, societal and economic impacts result from reduced maternal depression, anxiety and stress. For example, costing estimates from Pakistan suggest that there is a $16.6 billion cost per cohort of unaddressed maternal mental health concerns (Bauer et al., 2024). Limitations The data collection coincided with a politically sensitive time in Khyber Pakhtunkhwa, inclusive of security concerns for enumerators, ongoing election campaigns, and occasional disruptions to data collection, which could have impacted the accuracy and consistency of data collected. Further, repatriation efforts being implemented during the data collection period led researchers to avoid self-report determination of refugee (documented and undocumented) status and rely on enumerator report using observable indicators as outlined above. This may have therefore impacted the accuracy of the data on identifying refugee households. There are also several measurement constraints to be considered. The tools selected to measure maternal mental health may not cover all relevant dimensions of symptoms and experiences specific to the region of Pakistan studied, even though tools were specifically selected based on previous adaptation, validation and use in the province. 31 It is important to note that the cross-sectional approach to the survey prohibits any implications of causality between variables. Additionally, there is a wide range of stressors likely to associate with maternal mental health and ECD status, and the current study examined a sampling of stressors. Conclusion Few studies in Pakistan have used direct observation measures to gauge children’s ECD levels, making this study an important contribution to the body of knowledge on contributing factors for child development in Pakistan. The results overall fortify previous research showing the powerful connection between maternal mental health and ECD and highlight how exposure to stressors can jeopardize maternal and child outcomes. While these findings provide insight into maternal mental health in Khyber Pakhtunkhwa, research on maternal mental health concerns in Pakistan is somewhat limited, especially on the quality of available programming. Data on both the quality and accessibility of services for mothers and children that go beyond physical health and growth metrics to incorporate mental health and human development needs will fill gaps in the evidence base and inform public policy and programming. Research is needed on how to design sustainable, community-based service delivery to redress some of the unique challenges faced by both parents and young children in FCV contexts, as well as enhanced programming that is socially acceptable, physically accessible, and financially affordable can be designed and implemented at scale. Longitudinally, these approaches have the potential to improve human capital gains as well as reduce the toll on maternal well-being and child development. 32 References Abi Zeid Daou, K. R. (2022). Refugee Mothers Mental Health and Social Support Needs: A Systematic Review of Interventions for Refugee Mothers. Europe’s Journal of Psychology, 18(3), 337-349. Adamu, A. F. & Adinew, Y. M. (2018). Domestic violence as a risk factor for postpartum depression among Ethiopian women: Facility based study. Clinical Practice and Epidemiology in Mental Health, 14, 109-119. Afridi, G.S., Jabbar, A., Khan, S., & Akmal, N. (2021). An Analysis of Food Insecurity in Pakistan: Prevalence of Undernourishment (PoU) and Food Insecurity Experience Scale (FIES). Journal of Applied Economics and Business Studies. Ahmad, S., Hussain, S., Shah, F. S., & Akhtar, F. (2017). Urdu translation and validation of GAD- 7: A screening and rating tool for anxiety symptoms in primary health care. The Journal of the Pakistan Medical Association, 67(10), 1536–1540. Alam, J. & Ashraf, M. A., Tsegay, S. & Shabnam, N. (2022). Early childhood between a rock and a hard place: Early childhood education and students’ disruption in Khyber Pakhtunkhwa province, Pakistan. International Journal of Environmental Research and Public Health, 19, 4486. Al-Abri, K., Edge, D., & Armitage, C. J. (2023). Prevalence and correlates of perinatal depression. Social psychiatry and psychiatric epidemiology, 58(11), 1581–1590. Ali, N. S., Mahmud, S., Khan, A., & Ali, B. S. (2013). Impact of postpartum anxiety and depression on child's mental development from two peri-urban communities of Karachi, Pakistan: a quasi-experimental study. BMC psychiatry, 13, 274. Bauer, A., Knapp, M., Alvi, M., Chaudhry, N., Gregoire, A., Malik, A., Sikander, S., Tayyaba, K., Waqas, A., & Husain, N. (2024). Economic costs of perinatal depression and anxiety in a lower middle income country: Pakistan. Journal of Affective Disorders,15, 357:60-67. Bilal, A., Khalid, Z., & Fatima, S. (2021). Socio demographic factors of parental stress during COVID-19 pandemic related lockdown in Pakistan. Pakistan Journal of Physiology, 17(3), 58–61. Blackmore, R., Boyle, J. A., Fazel, M., Ranasinha, S., Gray, K. M., Fitzgerald, G., Misso, M., & Gibson-Helm, M. (2020). The prevalence of mental illness in refugees and asylum seekers: A systematic review and meta-analysis. PLoS Med, 17(9). e1003337. Bogic, M., Njoku, A. & Priebe, S. (2015). Long-term mental health of war-refugees: a systematic literature review. BMC International Health and Human Rights, 15, 29. Burgund Isakov, A., Krasić, B., Marković, V., Milic, N., Ristić, T., & Žegarac, N. (2022). Wherever We Go, Someone Does Us Harm: Violence against Refugee and Migrant Children Arriving in Europe through the Balkans. Save the Children and Center for Interdisciplinary Studies of the University of Sarajevo. Cheema, I., Farhat, M., Binci, M., Asmat, R., Javeed, S. & O’Leary, S. (2020). Benazir Income Support Programme Evaluation Report. Oxford Policy Management Report. Chen, W., Hall, B. J., Ling, L., & Renzaho, A. M. (2017). Pre-migration and post-migration factors associated with mental health in humanitarian migrants in Australia and the moderation effect of post-migration stressors: findings from the first wave data of the BNLA cohort study. The Lancet Psychiatry, 4(3), 218-229. 33 Chatterjee, P., Chen, J., Yousafzai, A., Kawachi, I., & Subramanian, S. V. (2023). Area level indirect exposure to extended conflicts and early childhood anthropometric outcomes in India: a repeat cross-sectional analysis. Conflict and health, 17(1), 23. De Oliveira, C. V. R., Rasheed, M., & Yousafzai, A. K. (2019). Chronic maternal depressive symptoms are associated with reduced socio-emotional development in children at 2 years of age: Analysis of data from an intervention cohort in rural Pakistan. Frontiers of Psychiatry, 10, 859. East, P. L., Gahagan, S. & Al-Delaimy, W. K. (2018). The impact of refugee mothers’ trauma, posttraumatic stress, and depression on their children’s adjustment. Journal of Immigrant Minority Health, 20, 271–282. El Gemayel, S. & Rigon, A. (2020). Working with Children Affected by Displacement. DeCID Thematic Brief #1, London: University College London. Fang, Y., Luo, J., Boele, M., Windhorst, D., van Grieken, A., & Raat, H. (2022). Parent, child, and situational factors associated with parenting stress: A systematic review. European Child & Adolescent Psychiatry, 25, 1–9. Farooq, S., Khan, T., Zaheer, S., et al., (2019). Prevalence of anxiety and depressive symptoms and their association with multimorbidity and demographic factors: a community-based, cross-sectional survey in Karachi, Pakistan. BMJ Open, 9, e029315. Fazel, M., Wheeler, J., & Danesh, J. (2005). Prevalence of serious mental disorder in 7000 refugees resettled in western countries: a systematic review. The Lancet, 365(9467), 1309-14. Gulamani, S. S., Shaikh, K., & Chagani, J. (2013). Postpartum depression in Pakistan. Nursing for Women's Health, 17(2), 47-152. Hagaman, A. K., Baranov, V., Chung, E., LeMasters, K., Andrabi, N., Bates, L. M., Rahman, A., Sikander, S., Turner, E., & Maselko, J. (2020). Association of maternal depression and home adversities with infant hypothalamic-pituitary-adrenal (HPA) axis biomarkers in rural Pakistan. Journal of affective disorders, 276, 592–599. Hattangadi, N., Cost, K.T., Birken, C.S. et al. (2020). Parenting stress during infancy is a risk factor for mental health problems in 3-year-old children. BMC Public Health 20, 1726. Hazer, L. & Gredebäck, G. (2023). The effects of war, displacement, and trauma on child development. Humanities and Social Sciences Communications, 10, 909 (2023). Hentschel, E. L., Tomlinson, H. B., Hasan, A., Yousafzai, A., Ansari, A., Tahir-Chowdhry, M., & Zamand, M. (2024). Risks to child development and school readiness among children under six in Pakistan: Findings from a nationally representative phone survey. International Journal of Early Childhood, 56,129-167. Huizink, A. C., Menting, B., Oosterman, M., Verhage, M. L., Kunseler, F. C., & Schuengel, C. (2014). The interrelationship between pregnancy-specific anxiety and general anxiety across pregnancy: a longitudinal study. Journal of Psychosomatic Obstetrics & Gynecology, 35, 92–100. Iqbal, S. & Ali, I. (2021). Maternal food insecurity in low-income countries: Revisiting its causes and consequences for maternal and neonatal health. Journal of Agriculture and Food Research, 3. 34 Irum, S., Khan, A. A., Rabbani, U., Lodhi, F. S., & Elsous, A. (2022). Frequency and risk factors of anxiety and depression among pregnant women in Abbottabad, Pakistan: A facility- based cross-sectional study. Pakistan Journal of Medical & Health Sciences, 16(7). Jagtap, A., Jagtap, B., Jagtap, R., Lamture, Y., & Gomase, K. (2023). Effects of prenatal stress on behavior, cognition, and psychopathology: A comprehensive review. Cureus, 15(10). Johnson, M. S., Skjerdingstad, N., Ebrahimi, O. V., Hoffart, A. & Johnson, S. U. (2022). Parenting in a pandemic: Parental stress, anxiety and depression among parents during the government-initiated physical distancing measures following the first wave of COVID-19. Stress and Health, 38(4), 637–652. Kerker, B. D., Willheim, E., & Weis J.R. (2023). The COVID-19 pandemic: Implications for maternal mental health and early childhood development. American Journal of Health Promotion 37(2), 265-269. Khabir, Y., Amir, M. A., & Syed, B. U. (2022). The neglect of male post-partum depression in Pakistan. The Lancet Psychiatry, 19(12): e52. Kingston, D., Kehler, H., Austin, M. P., Mughal, M. K., Wajid, A., Vermeyden, L., Benzies, K., Brown, S., Stuart, S., & Giallo, R. (2018). Trajectories of maternal depressive symptoms during pregnancy and the first 12 months postpartum and child externalizing and internalizing behavior at three years. PLoS One, 13(4). Leyendecker, B., Cabrera, N., Lembcke, H., Willard, J., Kohl, K., & Spiegler, O. (2018). Parenting in a new land: Immigrant parents and the positive development of their children and youth. European Psychologist, 23(1), 57–71. Mäntymaa, M., Puura, K., Luoma, I., Latva, R., Salmelin, R. K., & Tamminen, T. (2012). Predicting internalizing and externalizing problems at five years by child and parental factors in infancy and toddlerhood. Child Psychiatry & Human Development, 43, 153–170. Maselko, J., Bates, L., Bhalotra, S., Gallis, J. A., O'Donnell, K., Sikander, S., & Turner, E. L. (2018). Socioeconomic status indicators and common mental disorders: evidence from a study of prenatal depression in Pakistan. Social Science & Medicine (SSM) - Population Health, 4, 1-9. McCoy, D.C., Waldman, M., Team, C.F., Fink, G. (2018). Measuring early childhood development at a global scale: Evidence from the Caregiver-Reported Early Development Instruments. Early childhood research quarterly, 45, pp.58-68. Merced-Nieves, F. M., Dzwilewski K. L., Aguiar, A., Lin, J., Schantz, S. L. (2021). Associations of prenatal maternal stress with measures of cognition in 7.5-month-old infants. Developmental Psychobiology, 63, 960–972. Mirza, I. & Jenkins, R. (2004). Risk factors, prevalence, and treatment of anxiety and depressive disorders in Pakistan: systematic review. BMJ, 328(7443):794. Morina, N., Akhtar, A., Barth, J. & Schnyder, U. (2018). Psychiatric disorders in refugees and internally displaced persons after forced displacement: A systematic review. Frontiers in Psychiatry, 9. Murthi, M., & van der Heijden. (2024). Investing in the future: Empowering children and youth in fragile contexts. World Bank Blogs, Development for Peace: Solutions to Tackle Fragility, Conflict, Violence. 35 Parkes, A., Sweeting, H., & Wight, D. (2015). Parenting stress and parent support among mothers with high and low education. Journal of Family Psychology, 29(6), 907-18. Patanè, M., Ghane, S., Karyotaki, E., Cuijpers, P., Schoonmade, L., Tarsitani, L.,& Sijbrandij, M. (2022). Prevalence of mental disorders in refugees and asylum seekers: a systematic review and meta-analysis. Global Mental Health (Cambridge), 9, 250-263. Porter, M. & Haslam, N. (2005). Predisplacement and postdisplacement factors associated with mental health of refugees and internally displaced persons: A meta-analysis. JAMA, 294(5), 602 – 612. Pushparatnam, A., Luna Bazaldua, D. A., Holla, A., Azevedo, J. P., Clarke, M., & Devercelli, A. (2021). Measuring Early Childhood Development Among 4-6 Year Olds: The identification of psychometrically robust items across diverse contexts. Frontiers in Public Health, 9, 569448. Rahman, A., Khan, M. N., Hamdani, S. U., Chiumento, A., Akhtar, P., & Nazir, H. (2019). Effectiveness of a brief group psychological intervention for women in a post-conflict setting in Pakistan: a single-blind, cluster, randomised controlled trial. The Lancet, 393(10182), 1733-1744. Reed, R. V., Fazel, M., Jones, L., Panter-Brick, C., & Stein, A. (2012). Mental health of displaced and refugee children resettled in low-income and middle-income countries: risk and protective factors. The Lancet, 379(9812), 250-265. Rees, S., Channon, S., & Waters, C. S. (2019). The impact of maternal prenatal and postnatal anxiety on children's emotional problems: a systematic review. European Child and Adolescent Psychiatry, 28, 257–280. Reeves, R. V. & Krause, E. (2019). The effects of maternal depression on early childhood development and implications for economic mobility. Washington, DC: The Brookings Institution. Rizvi Jafree, S. & Barlow, J. (2023). Systematic review and narrative synthesis of the key barriers and facilitators to the delivery and uptake of primary healthcare services to women in Pakistan. BMJ Open, 13. Roskam, I., Aguiar, J., Akgun, E. et al. (2021). Parental Burnout Around the Globe: a 42-Country Study. Affective Science, 2, 58–79 (2021). Shah, S. M., Sun, T., Xu, W., Jiang, W., & Yuan, Y. (2022). The mental health of China and Pakistan, mental health laws and COVID-19 mental health policies: a comparative review. General Psychiatry, 35(5):e100885. Shams, L., Kakakhel, M., Majeed, S., Khan, H., Khan, H., & Sajid, S. (2021). Depression in females: Comparison of patients visiting urban and rural teaching hospitals in Peshawar, Khyber Pakhtunkhwa. Journal of Rehman Medical Institute, 7(2). Sikander, S. (2020). A Letter from Pakistan. The Lancet Psychiatry, 7(10): 845. Sitwat, A., Asad, S., & Yousaf, A. (2015). Psychopathology, psychiatric symptoms and their demographic correlates in female adolescent flood victims. Journal of the College of Physicians and Surgeons, 25(12), 886-890. PMID: 26691364. Tomlinson, H. B., Hentschel, E., Chowdry, M. T., Ansari, A., Zamand, M., Yousafzai, A. K., Hasan, A. (2023). Fostering Early Childhood Development. In Pakistan Human Capital 36 Review: Building Capabilities throughout Life (Eds. Lire Ersado, Amer Hasan, Koen Martijn Geven, et al.). Washington, DC: World Bank. Ugarte, E., Michael, D., Hiott, C., Tofail, F., & Wuermli, A. (2024). Psychometric analysis of a revised version of the Parenting Stress Scale for the Rohingya context (Unpublished internal report). NYU Global Ties for Children. UNHCR. (2024a). Pakistan Overview of Refugee and Asylum-Seekers Population as of June 30, 2024. Operational Data Portal. UNHCR. (2024b). Regional Refugee Response Plan for Afghanistan Situation Summary 2024. Operational Data Portal. https://data.unhcr.org/en/documents/details/106210. UNCHR. (2024c). Registered Afghan Nationals in Pakistan. Operational Data Portal. UNICEF. (2024). Humanitarian Action for Children, Pakistan. Waldman, M., McCoy, D. C., Seiden, J., Cuartas, J., CREDI Field Team, & Fink, G. (2021). Validation of motor, cognitive, language, and socio-emotional subscales using the Caregiver Reported Early Development Instruments: An application of multidimensional item factor analysis. International Journal of Behavioral Development, 45(4), 368–377. Waqas, A., Raza, N., Lodhi, H.W., Muhammad, Z., Jamal, M., & Rehman, A. (2015). Psychosocial factors of antenatal anxiety and depression in Pakistan: is social support a mediator? PLoS One, 10(1), e0116510. Watkeys, O., O'Hare, K., Dean, K., Laurens, K., Harris, F., Carr, V., & Green, M. (2022). Early childhood developmental vulnerability associated with parental mental disorder comorbidity. The Australian and New Zealand Journal of Psychiatry, 57. Wazir, M. N. K, Kakakhel, S., Gul, A. N., Awan, Q., Khattak, A. F., Yousaf, N., & Wahid, F. (2023). Psychiatric illnesses, somatic complaints, and treatments in a tertiary care hospital in Khyber Pakhtunkhwa, Pakistan: A cross-sectional study. Cureus,15(8), e43151. Wei, X., Wang, W., Law, Y. W., & Zhang H. (2024). The impacts of intimate partner violence on postpartum depression: An updated meta-analysis. Trauma Violence Abuse 25(2):1531- 1550. World Bank. (2024). Overview. Fragility, Conflict & Violence. World Bank Group, Understanding Poverty Topics. World Health Organization. (2023). Anxiety Disorders. Yasmin, F., Dapke, K., Khalid, M. R., Naz, H., Naz, F., Admani, B., Asghar, M. S., Awan, S., Phadke, R., & Saleem, M. (2021). Generalized Anxiety Disorder and Depressive Symptoms among Pakistani Population during the Second Wave of the COVID-19 Pandemic: A Regression Analysis. The American journal of tropical medicine and hygiene, 105(4), 915–923. Yousuf, J., Mehmood, H., Aquil, S., Rija, A., Syyeda, Rahmat, Z., & Malikzai, A. (2023). Effects of floods on the mental health of Pakistanis: a commentary. Annals of Medicine & Surgery 85(5), 2253-2255. 37