Maternity Care Services and Culture: A Systematic Global Mapping of Interventions

Background A vast body of global research shows that cultural factors affect the use of skilled maternity care services in diverse contexts. While interventions have sought to address this issue, the literature on these efforts has not been synthesised. This paper presents a systematic mapping of interventions that have been implemented to address cultural factors that affect women's use of skilled maternity care. It identifies and develops a map of the literature; describes the range of interventions, types of literature and study designs; and identifies knowledge gaps. Methods and Findings Searches conducted systematically in ten electronic databases and two websites for literature published between 01/01/1990 and 28/02/2013 were combined with expert-recommended references. Potentially eligible literature included journal articles and grey literature published in English, French or Spanish. Items were screened against inclusion and exclusion criteria, yielding 96 items in the final map. Data extracted from the full text documents are presented in tables and a narrative synthesis. The results show that a diverse range of interventions has been implemented in 35 countries to address cultural factors that affect the use of skilled maternity care. Items are classified as follows: (1) service delivery models; (2) service provider interventions; (3) health education interventions; (4) participatory approaches; and (5) mental health interventions. Conclusions The map provides a rich source of information on interventions attempted in diverse settings that might have relevance elsewhere. A range of literature was identified, from narrative descriptions of interventions to studies using randomised controlled trials to evaluate impact. Only 23 items describe studies that aim to measure intervention impact through the use of experimental or observational-analytic designs. Based on the findings, we identify avenues for further research in order to better document and measure the impact of interventions to address cultural factors that affect use of skilled maternity care.


Introduction
Global strategies to reduce maternal and newborn mortality and health have emphasised the need for scaling up the use of skilled maternity care [1]. Yet, experience has shown that provision of skilled care and availability of maternity care facilities does not necessarily lead to increased utilisation. A large body of literature describes how cultural factors affect women's use of services [2][3][4], including those resulting from differing 'cultures' of maternity care between service providers and populations served [5].

What is culture?
There is no one agreed definition of culture, but a focus on culture means emphases placed on aspects such as shared norms, beliefs and expectations, spoken language and behavioural customs [6]. In reality, it is difficult to separate out culture from social, economic and geographical context [7]. For example, members of a cultural group might not use a particular health service because they are too poor or because they know they will be discriminated against -highlighting the danger of conflating poverty with culture. Culture includes components that are both explicit and implicit. Hall [8] describes different levels of culture: a level that is explicit or manifest to outsiders (e.g., language, rituals, dress), a level of rules and norms that are known to group members but rarely shared with outsiders, and a level that is known and followed but not stated. Most societies have more than one culture within them. These cultural sub-divisions might take the form of social groups or strata (e.g., ethnic groups, religious groups, social classes, castes, ranks) marked by distinctive cultural attributes (e.g., beliefs, behaviour, perceptions, attitudes to illness and health, religion, language, manners, dress, housing, diet) alongside social and economic attributes (e.g., wealth, power, gender, education).

Culture and maternity care services
Childbirth, and the time around birth, is a social and cultural event that is often governed by norms. However, in most societies, the dominant culture, expressed through social institutions such as the health care system, regulates how health issues are both perceived and addressed. Differences between the cultures of health care services and service users have been recognised as a major issue in service delivery. Perceived or actual cultural insensitivity or incompetence of professionals can lead to perceptions of poor quality care by users or discrimination of certain users by providers, resulting in a lack of trust in services and service providers [9][10][11].
Many authors have recommended that cultural factors should be taken into account in the planning and delivery of services in order to effectively encourage service uptake as an important step in reducing maternal and newborn mortality [5,[12][13][14][15][16]. Intercultural approaches to the design and delivery of national policies are well-established in some countries, particularly in Latin America [17]. The need for 'culturally-appropriate' health facilities is core to the World Health Organization's (WHO) mandate on 'health for all' [18] and its strategy for improving maternal and newborn health [5]. However, research reveals the complexity of such endeavours. [19]. It is known that some interventions have been implemented in different world regions to address cultural factors that affect the use of maternity care services. However, the literature has not been synthesised.

Aims and objectives
This systematic mapping of the literature aims to understand the range of interventions that have been implemented to address cultural factors affecting women's use of skilled maternity care services. The study's objectives were to methodically identify and develop a map of the literature; categorise the range of interventions, the type of literature and the study designs included; and identify knowledge gaps.

Methods
The methodology for systematic mapping used in this study was developed from work at the Evidence for Policy and Practice Information and Co-ordinating Centre and is increasingly used in a range of social sciences [20][21][22][23][24]. The scope and types of literature included in a systematic mapping are normally broader than in a systematic review. The aim in this mapping is to describe as widely as possible all of the literature relating to the topic without limiting to studies that assess the strength or direction of the relationship, or even to empirical studies. We developed a protocol that was reviewed by an advisory group composed of content and method experts.

Inclusion/exclusion criteria
Potentially eligible studies included journal articles, and published and unpublished information from governments and other agencies, whether available in print or online, published in English, French and/or Spanish. Since the aim is to describe the nature and coverage of the literature, quality was not assessed and was not a criterion for inclusion. Multiple references based on the same sample were also not excluded (as would be the case in a systematic review in order to avoid bias). Table 1 provides the specification of the items to be mapped. Whilst intervention aims and outcomes relate only to women's use of services during pregnancy, childbirth and after birth, intervention recipients may include, for example, household or family members, community leaders or maternity care providers. Skilled care is defined in this mapping as those services provided by a skilled attendant: an ''accredited health professional -such as a midwife, doctor or nurse -who has been educated and trained to proficiency in the skills needed to manage normal (uncomplicated) pregnancies, childbirth and the immediate postnatal period, and in the identification, management and referral of complications in women and newborns. Traditional birth attendants, trained or not, are excluded'' [25]. Where the distinction between skilled and unskilled care is not clear, an inclusive approach is adopted. However, interventions that focused primarily on traditional birth attendants' (TBAs) roles in the direct provision of childbirth services were excluded, as the focus of the paper is on skilled maternity care services. Interventions concerned with improving satisfaction, but not also the use of services, are excluded.
Items must describe an implemented intervention in which a primary, focused aim or strategy is to address cultural factors as a vehicle to change use of maternity care services. This criterion excludes the following: N Interventions that exclusively address economic or geographical access barriers for a defined cultural group, although overlaps between cultural factors on the one hand and economic and geographical factors on the other are acknowledged [7]; N Generic quality improvement interventions that consider and/ or accommodate cultural factors explicitly or implicitly, but not as a primary, focused aim or strategy. Whilst it is acknowledged that an ''important question may be what combination of interventions and ways of incorporating culture into generic quality improvement are most likely to improve quality of care and outcomes'' [6], this question is beyond the scope of this mapping; and N Items in which 'cultural appropriateness' is evaluated, but not incorporated in intervention design or implementation.

Search strategy
Electronic databases, registers and websites were assessed for their availability, relevance and likely coverage of the eligible literature. Ten electronic databases and two targeted websites were included and searched for items published from 1 January 1990 to 28 February 2013. Combinations of relevant search terms were developed and tested in a sample of databases for sensitivity to a list of references that were known to the research team and judged to be potentially relevant. Table 2 presents the final combinations of search terms used. Electronic searches were adapted to each database using appropriate truncations/wildcards. Titles and abstracts were normally searched, and Medical Subject Headings (MeSH) terms were included where possible. A call for papers was sent by the WHO to various organisations and topic experts. The references received were combined with the electronic search results.

Screening process
All items identified through the search were screened initially on the basis of title and abstract. Where inclusion or exclusion could not be determined on the basis of title and abstract, the full text was screened. Figure 1 illustrates the screening process. EC, EJ, EH and SK contributed to screening the items individually. The following measures were taken for quality assurance: 1/all team members involved in screening independently screened the first 100 items. The whole team subsequently met to compare results, and discuss and resolve any differences in understanding of the inclusion/exclusion criteria. The criteria were further elaborated where necessary; 2/at the full text screening stage, any items that a team member considered borderline or problematic were noted. At the end of the process, EJ and EH independently screened all of the problematic items. Again, the whole team met to compare results, debate and resolve any differences. Decisions were made in favour of an inclusive approach where questions remained.

Analysis
Data were extracted for analysis from all items in the map, including background information; a description of the intervention; details of the type of literature (and study if relevant); and details of relevant outcomes measured. Based on the map, we inductively developed and defined intervention categories. It was unfeasible to create perfectly discrete categories since many interventions are complex. Where an intervention could have Table 1. Specification of systematic mapping.

Population
The outcomes apply to women during pregnancy, childbirth and after birth in any world region. The intervention population may be broader, including other community members or health providers.

Intervention
An implemented intervention that is primarily and explicitly designed to accommodate or address a cultural group's shared norms, values and/ or beliefs; behavioural customs; and/or spoken language/s. Cultural group is defined broadly to include any form of group or social stratum that is (considered to be) marked by its own distinctive cultural characteristics [7,128], such as ethnic groups, groups defined by religion, language groups, indigenous groups, tribal communities, caste-based groups/strata or geographically defined populations.
Aim/outcome Items must state an aim to increase, or measure as an outcome, women's use of skilled care during pregnancy, childbirth and after birth. fitted into more than one category, did not fit neatly into any category, or was not described in sufficient detail to understand its content, the item was placed in the category in which it was deemed to fit best. Data are presented in tables along with a narrative synthesis.

Results
After removing duplicates, the electronic database searches and website searches, combined with references suggested by experts, generated a total of 33,227 items for screening. The majority of items were not relevant. Following the screening, a total of 96 items were included in the map. The research team was unable to retrieve or screen a further 16 items.
Interventions addressing cultural factors affecting the use of skilled maternity care are not confined to a specific type of country. The map includes items from 35 countries across all world regions and the whole range of country income levels. Thirty-nine of the items were based in high-income countries, of which the majority were from the United States of America (USA), followed by Australia, Canada and the United Kingdom (UK). Items from the USA and UK predominantly included interventions targeting specific immigrant or ethnic minority groups, whilst items from Australia and Canada predominantly targeted indigenous communities. Only one item was based in continental Europe, although this may partly reflect the systematic mapping's restricted language coverage.
Of the remaining items, 25 were from low-income countries and 29 from middle-income countries, with three items located in multiple countries that cut across these categories. Sixteen items described interventions across nine countries in Eastern and Western Africa, and 14 items described interventions in Latin America. Twenty-seven items from Asia were dominated by literature from southern Asia.

Description of interventions
The five intervention categories we developed were: (1) service delivery models, (2) service provider interventions, (3) health education interventions, (4) participatory approaches, and (5) mental health interventions. We define the categories and describe the range of interventions within each below.
1. Service delivery models. This category includes 24 items describing models of service delivery specifically designed (or adapted from existing models of service delivery) to provide culturally-appropriate services for targeted groups (Table 3). These models are largely complex interventions taking a broad range of measures, often including elements of the 'service provider interventions' and 'participatory approaches' described below as well as adaptations to the service setting, practices, materials and/or language. The category is dominated by interventions from Australia and Latin America.
Several items from Australia describe comprehensive service delivery models for Aboriginal communities, implemented through a targeted health service [26][27][28][29][30]. For example, Nel and Pashen describe an indigenous medical centre managed by a community board and staffed by indigenous service providers [28]. Service users are seen in familiar surroundings, and Aboriginal health workers visit them to ensure attendance. Moreover,   pregnant women are allowed to bring children and family members, in recognition of the cultural importance of extended family links. Several items in this category, particularly from Latin America, describe the introduction of culturally-appropriate 'maternity waiting homes' or 'birthing houses' [31][32][33]. Moreno and Lopez [34] describe this model of service delivery as part of a broad strategy to adapt a national plan in Ecuador for indigenous communities. Different types of culturally-appropriate models of service delivery have been designed for indigenous communities [35,36], ethnic groups [37,38], refugees [39] and the general population [40]. For example, Gabrysch et al. [35] describe a delivery care model that was developed in Peru in cooperation with Quechua indigenous communities and health professionals, featuring a rope and bench for vertical childbirth, inclusion of family and TBAs in the service delivery process and during childbirth, and use of the Quechua language. One item in this category [41] is an outlier; in the item from Kazakhstan, the aim of a new culturally-appropriate service delivery approach is to introduce less medicalised care.
Some interventions refer to the provision of skilled home-based childbirth services explicitly in order to accommodate cultural norms [42,43]. Several further interventions in which a culturebased rationale was cited for training TBAs were excluded since the TBA was solely responsible for the direct provision of childbirth services [44][45][46][47][48][49]. However, these interventions also aimed to harness the TBAs' cultural role to facilitate linkages with the formal health system and improve referral for obstetric complications.
2. Service provider interventions. This category encompasses a range of interventions characterised by their focus on service providers. Analysis within this category revealed several sub-categories. One sub-category refers to three interventions in which the service provider is selected to match service users in terms of cultural characteristics [50][51][52] (Table 4). For example, Bilenko et al. [50] described an intervention in Israel in which a clinic for Bedouin families was staffed by an Arabic speaking Bedouin public health nurse. Few items describe this type of intervention as the sole focus. However, several items placed in other categories, particularly in the 'service delivery model' category, include this as one element of a broader intervention [26,27].
A large sub-category comprises 14 interventions in which people who share cultural characteristics with a target service user group are employed to bridge the cultural gap between this group and service providers (Table 4) [53][54][55][56][57][58][59][60][61][62][63][64][65][66]. They may fulfil various roles including encouraging and helping women to access care [57,59]; assisting women in communicating with healthcare providers [55,56]; and advocating on their behalf [62]. They are sometimes referred to as 'linkworkers,' 'peer health workers' or 'cultural brokers'. The category is dominated by items from the USA and the UK, where interventions have been implemented with specific immigrant or ethnic minority groups. This type of intervention has a long history, with many interventions implemented as early as the 1980s and 1990s. Again, several items placed in other categories, particularly in the 'service delivery model' category, include this position as one element of a broader intervention [26,32].
The final sub-category refers to eight interventions with existing staff to enhance their cultural awareness or sensitivity. These interventions have been implemented across diverse contexts and include various teaching and/or learning approaches [67][68][69][70][71][72][73][74] ( Table 4). For example, Smith and Davies [74] describe a knowledge exchange intervention in Canada that brings "communities together to enable them to understand each other's goals and cultures." 3. Health education interventions. Within this category, 22 items from diverse settings describe a wide range of strategies employed with the aim of designing culturally-appropriate health education activities (Table 5). Interventions may use one or more strategies. Several interventions use preliminary studies investigating cultural factors to inform health education activities, although they differ in the extent to which they describe how these cultural factors were addressed in the actual intervention [75][76][77][78][79][80]. This finding may reflect different levels of emphasis on these factors in intervention design, but may also simply be a reporting issue. For example, Opoku et al. [79] and Olaniran et al. [80] both refer to the same intervention in Nigeria. However, whilst Opoku et al. [79] make a brief mention of using research on cultural factors to inform design, Olaniran et al. [80] describe in more detail the measures taken to address cultural factors. Culturally-appropriate health education messages were developed, trainers were selected who were fluent in the local language, and campaigns were carried out in churches.
Related to a sub-category within 'service provider interventions,' one strategy within the category of health education interventions is to employ facilitators or trainers who share cultural characteristics with the relevant population [80][81][82][83][84]. For example, both Gennaro et al. [82] and Yeshi et al. [84] describe interventions employing village leaders as trainers.
Several items describe developing culturally-appropriate health education materials and practices, in terms of messages, language, modalities and/or images [80,[85][86][87][88][89][90]. For example, Omer et al. [89] describe using an education tool in Pakistan that reflects local materials and skills in embroidery, and DeStephano et al. [86] describe a culturally-tailored health education video series for Somali women in the USA. Finally, health education interventions have used participatory approaches [75,84,[91][92][93][94]. These are distinguished from the 'participatory approaches' category on the basis that their primary focus is the delivery of education interventions.
4. Participatory approaches. An inclusive approach was adopted for participatory interventions in this systematic mapping because a population's participation in intervention design and implementation may be considered inherently to address or accommodate cultural factors. Thus, all items in which the population's participation in intervention design and implementation was a primary and explicit strategy were included. Analysis of items in this category indicates different levels of explicit emphasis on 'cultural' factors, a distinction that is highlighted with sub-categories ( Table 6).
The first sub-category comprises four items that describe the participatory approach adopted as an explicit strategy to address or accommodate cultural factors affecting the use of skilled maternity care [95][96][97][98]. For example, Hounton et al. [96] describe their intervention in Burkina Faso as one that involved ''investment in communities through an understanding of their social structure and health seeking behaviours, through identification and partnership with credible community leaders, and through identification of culturally-sensitive and locally-acceptable approaches to address transport and referrals.'' Jewell and Russell [97] describe another approach implemented in the USA of forming a state-wide network of grassroots county minority health coalitions to develop projects to eliminate cultural barriers to prenatal care for minority women.
Manandhar et al. [110] and Osrin et al. [116] describe a participatory, community-based women's group intervention with a marginalised population with low access to services in Nepal, facilitated through a community action cycle. They do not frame the intervention as one designed to address cultural factors. Yet, Morrison et al. [98], referring to the same intervention in Nepal, was placed in the subcategory of 'participatory approaches that focus explicitly on cultural factors,' because the item describes formative research that was used to design the intervention to be acceptable and sensitive to the local culture. The implication is that the distinction between the three sub-categories may not always be a reflection of differences in intervention content; it may in some cases simply be a result of the emphasis in authors' reporting. This attests to the elusiveness of how culture is incorporated and addressed in many health interventions.
5. Mental health interventions. This small category includes three fairly recent interventions focusing on perinatal depression, all aiming to overcome treatment barriers for Latina women in the USA [117][118][119] (Table 7). All three items describe adaptations to existing interventions and models. They use strategies for addressing cultural factors that are similar to those used in other categories, such as service providers sharing cultural characteristics of the target group, a more appropriate service setting, and/or more appropriate materials or therapy in terms of language or content.

Description of types of literature and study designs
The map includes a wide range of literature, from narrative descriptions of interventions to studies using randomised controlled trials (RCTs) to evaluate impact. Eleven items provide only narrative descriptions of the intervention, including details of the design process, the content or its implementation. Some items Guatemala Three health education interventions, including a clinic-based intervention using images of indigenous women with staff using buttons to identify the language they speak.
Observational analytic study (repeat cross-sectional studies) Prenatal care utilisation

Ratnaike and
Chinner 1992 [83] Cambodia, Vietnam, Laos A community health education system that seeks to address the unique cultural beliefs concerning medicine and health in Indochina. Cultural sensitivity and community participation are sought through minimal involvement of non-national health professionals.

Narrative description of intervention
Sibley et al. 2001 [93] India and Ethiopia A community and competency-based program taking into consideration the social context of childbirth. A community self-assessment is conducted for community mobilisation and to provide information about local norms.  Nepal A participatory, community-based women's group intervention informed by formative research. Aiming for sensitivity to local needs and beliefs, it addresses negotiation within families, and tailors information to the needs of local groups and stakeholders such as mothers-in-law and traditional healers. It aims to be a forum in which to discuss cultural ideas.

Narrative description of intervention
Descriptive study (survey, health service audit + qualitative component)

Approaches that refer to cultural factors affecting use only
Orozco-Nuñ ez et al. [102] A key action was the coordination of a network to strengthen the links between communities, local authorities and health services.
Qualitative study Kaur 1994 [99] Malaysia The intervention is informed by an in-depth survey of the community's culture, beliefs and health status. Participatory planning meetings are held with relevant agencies. report on studies used to inform the design of an intervention now being implemented. All other items present some type of evaluation data, whether for monitoring, outcome evaluation or impact evaluation. The majority of items present various forms of quantitative or qualitative data used for monitoring the intervention and/or to evaluate outcomes. Only 23 items describe studies that aim to measure intervention impact through the use of experimental or observational-analytic designs. Of these, 14 use experimental designs including nine RCTs and five non-RCTs, and nine use observational-analytic designs including seven cohort studies.

Narrative description of intervention
Differences in study designs are evident across intervention categories. A large proportion of the studies with designs that aim to measure impact, particularly those with experimental designs, are in the 'participatory approaches' category. This is followed by a smaller proportion in the 'health education interventions' and 'service provider interventions' categories, and very few in the other categories combined.

Discussion
Global recognition of the need to address cultural factors affecting the use of skilled maternity care has led to a wide range of interventions being implemented across diverse settings in all world regions and across all country income levels. Overall, the map provides a rich source of information on the types of intervention options that have been attempted to address or accommodate cultural factors that affect the use of skilled maternity care. The articles included in the systematic mapping and others indicate a growing awareness of the need to incorporate culture into the design of appropriate care to improve maternal and newborn outcomes [120,121] and to deliver more responsive, effective maternity care services.
The mapping reveals examples of good practice and success stories. However, some of the literature identified, whilst stating a clear aim of addressing cultural factors, provides insufficient detail to understand exactly how they were addressed. It is clear that the complexities of formulating and implementing culturally-responsive programmes remain [16,[122][123][124][125][126]. In some of the literature found, culture is positioned as a barrier to maternal health service use, rather than an attribute of the population that services seek to serve. The type of cultural groups in a setting and the nature of cultural factors that affect use of skilled maternity care are context-specific, and give rise to a need for different kinds of intervention approaches that both anticipate and respect a community's culture, values and beliefs. This finding is reflected in the map by the clustering of items from specific contexts in specific categories of intervention. Cultural beliefs and behaviour are impossible to isolate from the social and economic context in which they occur [7]. Interventions research around culture and maternal health service use is heavily weighted in favour of evidence focusing on sub-populations in high-income countries. We know that 99% of maternal deaths occur in low income settings and that scaling up the use of skilled maternity care will reduce global maternal morbidity and mortality [18]. We know that focusing only on the supply side of maternity care does not necessarily lead to increased use, and that culture is often invoked as an explanation for this supply-demand gap. Our systematic mapping shows that there is a critical need for better documentation of interventions, with an emphasis on lowerincome countries, and better study methods to evaluate the ways in which cultural factors can be systematically mainstreamed into programmes to increase maternity care service use. Literature on interventions addressing cultural factors as evaluated models of practice remains limited. This finding may be to some extent related to the small-scale, context-specific nature of many interventions of this type.
We exclude several interventions with TBAs providing childbirth services because our focus is on use of skilled maternal health care (see category 1). However, within these items, there are several examples where the TBAs cultural role is also harnessed to facilitate linkages with formal health services. Although WHO guidance has moved away from promoting approaches that involve TBAs in the direct provision of care at birth, it also emphasises the importance of building links with TBAs and finding new roles so that this valuable resource can continue to support women during pregnancy, childbirth and after birth, and serve as an important link between services and communities [127]. Future interventions incorporating WHO guidance on the need for women to be attended by health care workers with the appropriate skills and training may consider employing traditional health workers in alternative roles. Going beyond public health programmes, there is also a need to incorporate cultural knowledge and responsiveness into health education, the development of health policies, and the delivery of culturally-competent health care.

Limitations
Limitations arising from the inevitable conceptual challenges of any attempt to map or categorise items focusing on 'culture' are acknowledged. Culture is a complex, elusive concept that is challenging to define, leading to diverse definitions and operationalisations in the literature. Usages of the term are not always helpful, and the concept is susceptible to assumptions and overgeneralisation. Making distinctions between interventions that address cultural factors and those that address other factors that affect use of skilled care was not a straightforward task. Distinguishing between interventions that address cultural factors as the primary and explicit aim on the one hand, and those in which cultural sensitivity is incorporated as part of generic quality improvement on the other, inevitably involves subjective judgement.
One particular challenge was the research question's focus not on what was actually 'done' in the intervention, but rather on what the intervention sought to address, which was reflected in the search terms. However, what the intervention seeks to address is not always carefully described by authors.
The above conceptual challenges all had implications for what was eventually identified and included. Identifying relevant items depended in part on the research team's judgement of where the line should be drawn around those to be included (outlined in Table 1). This effort was the first attempt of its kind to map such interventions; however, the range of interventions was unknown at the design stage, making it necessary for the research team to confer frequently and to continue drawing lines throughout the screening process. Moreover, at both the searching and screening stages, identifying relevant items depended in part on how authors of potentially relevant items framed their reporting, and also on the research team's ability to overcome challenges to identifying (from the text available) those that meet the criteria. The research team was aware of the challenges at the outset and continuously sought to address them from the design stage through to the final synthesis. Nevertheless, any shortcomings remain a limitation of this systematic mapping. Relevant items may have been missed through method limitations. More specifically, any relevant literature in languages other than English, French or Spanish would have been missed. Because searches were not designed to focus on participatory approaches, some literature on these interventions may have been missed. Also, since it would have been unfeasible to search all of the infinite potential sources of grey literature, the possibility that some items of this type were missed is high. The possibility that some in-service reports on small-scale interventions that were not intended or prepared for wider circulation were missed is particularly high. Finally, we were unable to retrieve 16 items for screening that may have been eligible for inclusion. These 16 items largely represented interventions implemented in low-and middle-income countries.

Conclusion
The map provides a rich source of information on interventions attempted in diverse settings that might have relevance elsewhere. However, many sources lack sufficient description or robust designs that allow us to draw firm conclusions. This may be to some extent related to the small-scale, context-specific nature of interventions of this type. Addressing the impacts of interventions to address cultural factors affecting the use of maternity care services is an issue of importance for researchers, programmers, and policy makers. It requires an inter-disciplinary approach and active dialogue with communities in order to understand their cultural systems, health beliefs, health practices and preferences. In order to better serve the varied needs of communities with culturally-diverse populations, the following recommendations are made for future research and reviews: N The interventions in this map are inherently context-specific.
Nevertheless, further intervention studies with harmonised outcomes, appropriate research methods and robust designs are warranted, which may provide valuable evidence on the impact (including benefits or potential harms) of a type of intervention model. N Where an intervention is designed explicitly to address cultural factors, sufficient detail should be provided in reporting for the audience to understand how they were addressed (i.e., by specifying the links between the cultural factors identified and the content of the intervention).
N A full systematic review may be warranted of the more cohesive set of interventions designed to provide culturallyappropriate skilled maternity care for defined ethno-linguistic or religious groups. This would allow both the quality and outcomes of intervention studies to be examined.