Review of policies, data, and interventions to improve maternal nutrition in Afghanistan

Abstract Malnutrition contributes to direct and indirect causes of maternal mortality, which is particularly high in Afghanistan. Women's nutritional status before, during, and after pregnancy affects their own well‐being and mortality risk and their children's health outcomes. Though maternal nutrition interventions have documented positive impact on select child health outcomes, there are limited data regarding the effects of maternal nutrition interventions on maternal health outcomes globally. This scoping review maps policies, data, and interventions aiming to address poor maternal nutrition outcomes in Afghanistan. We used broad search categories and approaches including database and website searches, hand searches of reference lists from relevant articles, policy and programme document requests, and key informant interviews. Inclusion and exclusion criteria were developed by type of source document, such as studies with measures related to maternal nutrition, relevant policies and strategies, and programmatic research or evaluation by a third party with explicit interventions targeting maternal nutrition. We abstracted documents systematically, summarized content, and synthesized data. We included 20 policies and strategies, 29 data reports, and nine intervention evaluations. The availability of maternal nutrition intervention data and the inclusion of nutrition indicators, such as minimum dietary diversity, have increased substantially since 2013, yet few nutrition evaluations and population surveys include maternal outcomes as primary or even secondary outcomes. There is little evidence on the effectiveness of interventions that target maternal nutrition in Afghanistan. Policies and strategies more recently have shifted towards multisectoral efforts and specifically target nutrition needs of adolescent girls and women of reproductive age. This scoping review presents evidence from more than 10 years of efforts to improve the maternal nutrition status of Afghan women. We recommend a combination of investments in measuring maternal nutrition indicators and improving maternal nutrition knowledge and behaviours.


| INTRODUCTION
From 1990 to 2015, the global maternal mortality ratio (MMR) declined by 44%. However, many women continue to die from preventable and treatable complications, especially in low-and middleincome countries (LMIC), in which the MMR was 436 deaths versus 216 globally per 100,000 live births in 2015 (UNICEF, 2017). Malnutrition and micronutrient deficiencies contribute directly and indirectly to maternal morbidity and mortality. Evidence indicates that maternal malnutrition, measured through mid-upper arm circumference (MUAC), is strongly associated with the risk of maternal mortality (Christian et al., 2008;Sikder et al., 2014). Anaemia is a risk factor for maternal mortality and morbidity due to exacerbated effects of haemorrhage and impaired clotting capacity, and calcium deficiency increases the risk of hypertensive disorders of pregnancy, principally eclampsia Daru et al., 2018;Maternal and Child Nutrition Study Group, 2013;Rush, 2000). Christian et al., 2000 also found that night blindness, a symptom of vitamin A deficiency, was predictive of a 3.8 times greater risk of maternal death in Nepal. A recent nested study compared reported maternal complications among postpartum women in intervention and control areas for a cluster-randomized trial of intensified maternal nutrition counselling and micronutrient supplementation added to an existing home-based antenatal/postnatal care (ANC/PNC) programme in Bangladesh (Todd et al., 2019). Odds of reported retained placenta, postpartum bleeding, and postpartum infection or fever were significantly higher among women living in the control areas. Anaemia, micronutrient deficiencies, and attendant pregnancy complications also contribute to delayed maternal recovery from delivering a child, low birth weight, and neonatal morbidity, which in turn contribute to intergenerational malnutrition from mothers to their babies (Christian, Mullany, Hurley, Katz, & Black, 2015).

The 2013 Lancet Maternal and Child Nutrition Series
reemphasized global commitments to focus on the first 1,000 days of life for improving nutrition outcomes among children (Black et al., 2013). Maternal nutrition-defined as the nutrition needs of women during antenatal and postnatal periods and sometimes also to the period prior to conception, even as early as adolescence in contexts where early childbearing is normative (The Manoff Group, 2012)-affects the birth outcomes, growth, and development of their children as well as mothers' own well-being and mortality risk (Black et al., 2013;Goudet, Murira, Torlesse, Hatchard, & Busch-Hallen, 2018).
Despite associations between adequate maternal nutrition and positive maternal and child health outcomes, few studies have documented the impact of nutrition interventions focused on maternal preconception and pregnancy. Limited high-quality evidence exists for multisectoral nutrition interventions, particularly on the incorporation of nutrition-sensitive approaches such as water, sanitation, and hygiene (WASH); agriculture; and food security into existing nutritionspecific programming (Maternal and Child Nutrition Study Group, 2013). Additionally, most maternal nutrition interventions measure newborn and child health indicators as primary outcomes with relatively fewer assessing maternal health indicators (Christian et al., 2015). A systematic review of evidence-based nutrition programmes taken to scale found that the following components improve birth outcomes and overall household nutrition: micronutrient supplementation, food fortification and supplements, nutrition education and counselling, and conditional cash transfers (Victora et al., 2012). The extent to which these interventions reduce maternal complications is less rigorously documented. Goudet et al. (2018) identified nine studies in their systematic review measuring the effectiveness of the approaches to improve the coverage of maternal nutrition interventions in South Asia. They found that a range of individual and health service delivery barriers, such as education and access to services, affects coverage by nutrition interventions. The detected range of interventions (e.g., iron and folic acid [IFA] and calcium supplementation and maternal nutrition counselling) and evidence regarding their efficacy were both limited. Multisector nutrition studies were not included, and no studies from Afghanistan were deemed eligible for inclusion. Recent studies on WASH and nutrition have shown

Key messages
• Policies and strategies to improve nutrition more recently have not only shifted towards multisectoral efforts but also specifically targeted adolescent girls and women.
There is a dearth of evidence on the effectiveness of interventions targeting maternal nutrition in Afghanistan.
• Ensuring access to quality nutrition care for women and adolescent girls will have intergenerational health and nutrition gains. However, measures of nutrition outcomes and intervention/service coverage of pregnant women, WRA, and adolescent girls are rarely collected and reported.
• Maternal nutritional status is critical to the welfare of the population, and programmes need to invest in systematic implementation research and more robust outcome monitoring and evaluation to understand feasible delivery mechanisms and programme effectiveness. limited effects on reducing child malnutrition outcomes, indicating that WASH is necessary but not sufficient, because malnutrition has various underlying causes (Coffey & Spears, 2018;Humphrey et al., 2018). Thus, we believe a multisectoral approach is needed to address malnutrition.
The prevalence of malnutrition among the general population, especially women and children, is significantly higher among fragile and conflict-affected states, because factors such as forced migration, food insecurity, and restricted access to humanitarian aid can affect nutrition status and exacerbate poor health outcomes (Kinyoki et al., 2017;Seal, 2018;Taylor, Perez-Ferrer, Griffiths, & Brunner, 2014). Similarly, maternal morbidity and mortality are highest in conflict-affected countries (Urdal & Che, 2013). Afghanistan, in South Asia, is among the most fragile countries (Fund for Peace, 2018;World Bank, 2019) and has the highest MMR in South Asia (638 deaths/100,000 live births; Kassebaum et al., 2016;UNICEF, 2017). In Afghanistan, effectively addressing maternal malnutrition could result in health sector gains by reducing maternal morbidity and mortality and child stunting and by improving overall survival, growth, and development of children. Understanding the available maternal nutrition data in Afghanistan may improve knowledge sharing, evidence generation, policy development, and programme implementation in the country and other fragile contexts facing similar challenges.
This scoping review maps policies, data, and interventions aiming to address maternal anaemia and poor maternal nutrition outcomes in Afghanistan and to determine the extent of multisectoral involvement in these efforts. We build on a previous multisector nutrition review that was conducted in 2010 and systematically focused on maternal nutrition (Levitt, Kostermans, Laviolette, & Mbuya, 2010). Our objectives were to (1) identify relevant maternal nutrition policy and programmes and describe the degree to which multisectoral approaches were engaged; (2) describe main findings across policy, strategy, programming, and monitoring within available peer-reviewed evidence and grey literature on maternal nutrition in Afghanistan; and (3) critically appraise data summarizing maternal nutrition status and relevant programme coverage, knowledge, and gaps in Afghanistan.

| Study design
Overall, we followed the scoping review framework proposed by Arksey and O'Malley (Arksey & O'Malley, 2005) and enhanced by Levac, Colquhoun, and O'Brien (2010). A scoping exercise was deemed appropriate rather than a systematic review to map the range of published and grey literature, policy, and data sources on maternal nutrition in Afghanistan and to better define gaps within existing practice and research. We applied a two-step process for identifying and selecting sources and engaging stakeholders (Levac et al., 2010) and Web of Science. Example search terms included (anaemia OR "micronutrient deficiency" OR "iron deficiency" OR anaemia, iron-deficiency[Mesh] OR "dietary diversity" OR "MUAC" OR "middle upper arm circumference" OR "nutritional status" OR "pregnancy complications" OR "pelvic contracture" OR "knowledge" OR "behaviour change) AND (pregnancy OR pregnant OR pregnant women[Mesh] OR maternal OR "women of reproductive age" OR fertility) AND

| Source selection
Before selecting sources, we agreed on broad inclusion criteria applied at the first search stage to maximize the initial range of documents.
Documents included in the review had or addressed a maternal nutrition measure, policy, or strategy in Afghanistan. We then refined inclusion criteria by discussing source selection based on the first phase of abstract review, which delineated available and relevant maternal nutrition content and sources (Levac et al., 2010

| Citation management and data extraction
All citations were imported into Endnote X7 (Clarivate Analytics, Philadelphia). Duplicate citations were removed manually during screening. Data were extracted into a spreadsheet, where descriptive information was extracted based on specified categories.

| Analysis
We first collated and summarized the scope of available sources. We then described any evaluated nutrition programmes with at least one maternal nutrition component and appraised the extent to which multisectoral approaches were used. We synthesized information identifying changes to priority programme areas and outcomes over time in data sources and policy documents, respectively, with iterative identification of gaps.  in 1997, 2000, and 2003. A household food security study was also conducted in all regions except for eastern Afghanistan, and a focused study was conducted on vitamin A deficiency in Kabul city (Lautze, 2002;Mihora et al., 2004). Province (Hadi, Mujaddidi, Rahman, & Ahmed, 2007), and a secondary analysis of pilot national nutrition surveillance data (Johnecheck & Holland, 2005). We organized maternal nutrition-related measures into five categories: maternal nutrition status, micronutrients, food security, use of maternal health services, and WASH.   (Akseer et al., 2018). The prevalence of obesity among women overall and in an eastern city, Jalalabad, was 27.4% and 35.9%, respectively (Saeed, 2015). Among 55 peri-urban women, 20% were underweight (Oberlin et al., 2006). In a secondary analysis of 2013 NNS data, maternal BMI was associated with many factors, including household wealth status, education level, region, food insecurity, unimproved water and sanitation facilities, age, ethnolinguistic status, and parity (Akseer et al., 2018). Low maternal height was not reported by any of the data sources. This presents a knowledge gap considering the associations found between low birth weight and poor linear growth in children born to women with low maternal height (Addo et al., 2013;Britto et al., 2013). In Afghanistan, the NNS measured height as part of BMI calculations, but in its analysis and in secondary analyses of NNS data, there are no mentions of data presented on low (<145 cm) maternal height. Further, nutrition status of WRA based on MUAC scores was only done in small area SMART surveys (Habib, 2017).

| Micronutrient deficiencies
The 2013    although differences were found between communities in the summer, such as meat consumption no more than once a week in the rain-fed zone compared with twice a week in the irrigated zone, with the increased consumption tied to larger social occasions as a means of honouring guests. Households in the irrigated zone owned more food animals and livestock holdings than households in the rain-fed zone. In a shared plate culture, where men and women in a household typically eat together unless entertaining guests, it is unclear to what extent women consume meat even when it is available. Although women reported having three main meals, the default was bread and tea, especially for breakfast. In a qualitative study, Newbrander, Natiq, Shahim, Hamid, and Skena (2014) found that women reported postpartum diet restrictions due to "cold" (e.g., beef and watermelon) and

| Food security
"hot" (e.g.. beans) food characteristics, which were alleged to cause body pains or were perceived as difficult to digest, respectively, and thus were avoided in the postpartum period, limiting dietary diversity.
Although dietary diversity is a proxy indicator for food insecurity, this study found limited dietary diversity was based on social norms rather than actual inaccessibility of food, the latter of which was not directly explored (Hoddinott & Yohannes, 2002).

| Use of maternal health and nutrition services
In 2013

| Policy responses to maternal malnutrition
A timeline of policy documents developed since 2003 to improve nutrition outcomes in Afghanistan is presented in Figure 4 alongside  Table 2 summarizes relevant policies and strategies.

| Nutrition interventions through the health system
Two studies had health education interventions through communitybased health actors, such as community health workers (CHWs) and Health Centers, which are the facilities predominantly providing primary care in Afghanistan (Nasrat, 2014). Every province has one Provincial Nutrition Officer (PNO) responsible for coordinating and monitoring nutrition-related activities and training BPHS staff in nutrition services. However, the assessment found that due to limited resources, PNOs were unable to provide the required support to facilities, and many health facility staff were unaware of PNOs at the provincial level. Some providers reported integrating nutrition education into other services for mothers ad-hoc, but it was unclear whether message content and delivery among providers was consistent. This lack of consistency complicates the ability to assess effects. Some providers conducted education sessions in waiting rooms, some conducted cooking demonstrations with nutritious foods, whereas others promoted and distributed conventional information, education, and communications materials. IFA supplement stockouts were also reported by health facilities. Notably, the lack of clarity and standardization of IFA supplement use in the policy was also echoed by participants in this study.  (1) MoPH (3)MRRD (1) Government (2) MAIL (1)MoPH (4) ticipants misunderstood messages about handwashing meant to be done before a task (e.g., preparing food) compared with after specific tasks (e.g., using the toilet), but no quantified behaviour or knowledge change data were presented, and nutrition outcomes were not measured (The World Bank Group, 2014). More than two contacts were recommended to ensure that people retained the messages.

T A B L E 2 Summary of policies and strategies on maternal nutrition in Afghanistan
One published article by Alim and Hossain (2018)  ipe books for mothers, school and kitchen gardens, and food processing training. Interventions specifically targeting women were kitchen gardens, food processing, and food conservation with marketing training. Overall, they found challenges with the projects' ability to demonstrate change due to lack of consistent indicators across projects at baseline and endline and insufficient data collected. More data on intervention effects on the target groups and models used were needed. Programme implementation challenges included the lack of land ownership and access to resources for women to implement projects with livelihood components; thus, actual ability to achieve diverse food production, processing, and conservation was unclear.
Additionally, although nutrition education was included in the overall programme, it was not specifically integrated with women's food production components, resulting in a missed opportunity for combined messaging on nutrition, practical livelihoods, and diverse food con-   The 2013 Maternal and Child Nutrition Lancet Series identified key interventions to rapidly improve nutrition outcomes. Globally, the evidence base has expanded regarding effective nutrition interventions in the preconception period and during pregnancy to improve maternal and fetal outcomes (Vaivada, Gaffey, Das, & Bhutta, 2017).
Of the interventions included in our review, four focused on micronutrient supplementation, and the remaining five were related to education about healthy food consumption during pregnancy and hygiene or food-based programmes involving agriculture and livelihoods.
Although micronutrient deficiencies among Afghan women are widespread, nutrition education focuses predominantly on IFA supplementation, with less emphasis on other essential micronutrients for pregnant women and adolescent girls. Similar to other LMICs, in Afghanistan the adolescent period has historically been neglected, yet increased nutrients are essential for adequate growth during adolescence (Christian & Smith, 2018). An increased focus on adolescent girls' nutritional needs may ensure that effective interventions are delivered, such as IFA supplements and increased nutritious food intake through school gardens (Vaivada et al., 2017). Further, protein energy deficiency data are collected in household surveys, but we did not find interventions providing protein energy supplementation or promoting increased protein energy consumption during pregnancy in our search. Additionally, there was no messaging on the dangers of overweight or obesity. Obesity is a growing public health concern in South Asia due to dietary changes, particularly in urban areas, and a consideration of the multifaceted burdens of poor nutrition should be made for populations at risk for obesity. The multisector programmes confirm previous studies revealing the importance of using community-based channels for implementation, such as volunteer CHWs and Family Health Action Group members Goudet et al., 2018).
Evidence for effectiveness of large-scale multisector programmes was scarce because evaluations were designed to assess vertical interventions, not the programme overall. Also, the evaluations used methods that were feasible in the challenging research environment of Afghanistan, which frequently precluded randomized trials. As a result, the data from these studies were excluded from systematic reviews (Goudet et al., 2018 (Suresh et al., 2019).
A meta-analysis of maternal micronutrient supplement interventions, including studies from Bangladesh, India, Indonesia, Nepal, and Pakistan, found that multiple micronutrient supplements, particularly when started before 20 weeks gestation, were more effective in reducing incidence of preterm birth, low birth weight, and small for gestational age and also reduced maternal mortality among anaemic women (Smith et al., 2017).

CONFLICTS OF INTEREST
The authors declare that they have no conflicts of interest.

CONTRIBUTIONS
CK contributed to study conception and design, collected, extracted, and analysed the data, and drafted the manuscript. GFM contributed to study conception, data extraction and interpretation, and critical review. MOM contributed to data extraction and critical review. PMP, MJA, and MHL contributed to critical review. CST contributed to study conception and design, data extraction, and interpretation and critically revised the draft. All authors read and approved the final version submitted.