Maternal Social Support and Determinants of Health Facility Delivery in Rural Ethiopia


 Background: Maternal mortality continues to decrease in the world but remains the most important health problems in low-income countries. Although evidence indicates that social support is important factor influencing health facility delivery, it has not been extensively studied in Ethiopia. Therefore, this study aimed to assess social support in influencing health facility childbirth in Ethiopia. Methods: A cross-sectional survey data on 3304 women aged 15–49 years from an Innovating for Maternal and Child Health in Africa project in three districts of Ethiopia, were analyzed, including data on social support and important factors. Using multivariate logistic regression, we examined the association between health facility birth, social support, and included variables. Adjusted odds ratios with 95% confidence intervals were used to identify statistically significant associations at 5% alpha level.Result: Overall, 46.9% of women delivered at health facility in their last pregnancy. Average travel time from closest health facility (AOR: 1.51, 95% CI 1.21 to 2.90), mean perception score of health facility use (AOR: 1.83, 95% CI 1.44 to 2.33), involvement in final decision to identify their place of childbirth (AOR: 2.12, 95% CI 1.73 to 2.58) had significantly higher odds of health facility childbirth. From social support variables, women who perceived there were family members and husband to help them during childbirth (AOR: 3.62, 95% CI 2.74 to 4.79), women who received continuous support (AOR: 1.97, 95% CI 1.20 to 3.23), women with companions for facility visits (AOR: 1.63, 95% CI 1.34 to 2.00) and women who received support from friends (AOR: 1.62, 95% CI 1.16 to 3.23) had significantly higher odds of health facility childbirth.Conclusions: Social support was critical to enhance health facility delivery, especially if women’s close ties help facility delivery. An intervention to increase facility delivery uptake should target not only the women’s general social supports, but also continuous support during childbirth from close ties including family members and close friends as these are influential in place of childbirth and also actions that increase women's healthcare decision could be effective in improving health facility delivery.

the 2019 Ethiopian mini Demographic and Health Survey (EDHS), 72% of births among urban populations in Ethiopia occurred in health facilities, compared to only 40% of births among rural populations where over 80% of Ethiopians living, which highlights the need to improve use of facility delivery in underutilization (9).
Major factors that have been studied as contributing factors to the low utilization of existing maternal health services in SSA countries like Ethiopia are related to distance to a facility, means of transportation, maternal socio-demographics (e.g. age, parity, education, and wealth), health decision-making autonomy, and perceived needs/bene ts of facility birth and attitude (6,16). Most of these factors have been studied extensively in Ethiopia and elsewhere (10,15,16). Further, recent studies examined the in uence of social support during pregnancy and childbirth on birth weight, length of gestation including fewer pregnancy complications for women with high levels of stress, reduce fear/anxiety, improve outcomes, and stimulate healing but not for those with low levels of stress (17,18), however, these studies do not fully account for the interpersonal, family and community interactions thought to facilitate or constrain service use as well as decision making; a social context in which individual determinants are known to operate. For example, the importance of social networks in Ethiopia have long recognized and exhibit different kinds of social interaction and rely more on family and friend ties, and often consult family and friends in case of illness and service use (17)(18)(19)(20). However, maternal social support as direct determinants of health facility delivery including the different dimensions of maternal social support and its complex interactions have not been extensively studied in Ethiopia. Therefore, the aim of the present study was to assess the maternal social support and related factors on health facility childbirth, and investigating the effects of social support on health facility childbirth may provide insights into additional areas of interventions to increase uptake in facility delivery.
Social support including family and community contexts are mini universes of complex social, political, associational, economic, power and cultural dynamics, providing a different direction for encouraging the use of health services and facilitating behavior change than in a health facility (21)(22)(23)(24). Social support is considered a complex and dynamic process that often used in a broad sense, including social networks and social capital, which also involves individuals and their social networks, working to satisfy their needs, provide and complement the resources they have, and thus cope with new situations (7,17). One frequent criticism of research in this area is the lack of consensus about social support in terms of its de nition and how best to measure it (7,(17)(18)(19). However, many scholars and theorists conceptualized social support as a contextual phenomenon in the sense that it is both individual and a community characteristic re ecting the daily interaction between social relationships/neighbors that may bene t health through interpersonal trust and norms of mutual aid, promoting collective e cacy and neighborhood cohesion (1,25).
Further, studies have speci ed the common dimensions or forms of social support including functional and structural supports. Structural social support refers to the structure and quantity of social relationships, such as the size/density of networks and the frequency of interaction. Functional Social support, in contrast, refers to the function and quality of social relationships, such as perceived availability of help and the resources exchanged among individuals in a relationship such as informational support (advice/suggestions), instrumental support (aid or assistance), emotional support (empathy, care and trust), and accompaniment (7,10,26). Different studies including systematic reviews reported the association between social support and access to health care behaviors (7,16,17), and absence of social support was associated with increased maternal morbidity and complication (17,18,25).
Furthermore, studies in Ethiopia, Kenya and Bangladesh found that social support from a spouse or partner and a social network of family and friends has been found to in uence women's decisions regarding obtaining prenatal care, childbirth, and breastfeeding (17,18,(27)(28)(29), and other studies also found that female relatives and friends accompanying laboring women to maternity units (13,15,18), and the presence of a female relative during labor, is associated with improved labor outcomes (30). Various studies also acknowledged the diversity and uniqueness of each social support characteristics as a major source of strength, and identi ed that people who receive higher levels of social support tend to be less stressed and have better health status than more socially isolated people (1,25,31). Moreover, maternal social support can also be received in different contexts which leads to different healthcare seeking behaviors (32). For example, studies in SSA countries reported that pregnant women initiate ANC early and felt the importance of health facility childbirth, only when they are sick or experience any pregnancy-related health problems (32)(33)(34)(35). Early initiators emphasized that they go to the ANC and started to think health facility childbirth and need different supports from family and relatives if the pregnancy is under threat and or they are sick in the following comments. Moreover, if pregnant women perceived pregnancy to be a normal health condition and with previous successful home delivery, they felt seeking health care or giving birth at facilities was unnecessary; hence develop positive attitude towards home delivery and placed a low value on using antenatal care and health facility delivery (32)(33)(34). Thus, maternal social support plays a crucial role in improving maternal health outcomes, rst, in recognizing the need for health services, and, second, in facilitating or constraining the use of those services (10,30,36).
About one-third of currently married women in Ethiopia, however, have no nal say concerning their own healthcare decisions, with recent studies nding that 32.3% report that their husbands make the nal decision (15,27). According to the 2016 EDHS, more than 70% of women report social barriers to accessing MCH care, including getting money for advice or treatment (55%), long distance to a health facility (50%), lack of accompaniment (42%), and getting permission to go for treatment (32%) (6). Furthermore, in traditional and patriarchal societies like Ethiopia, where restrictions are placed on a woman's freedom of movement and contact with unrelated men, the in uence of the social environment through maternal social support factors may be important factors in determining whether MCH care, particularly health facility delivery is used or not (15,37).Therefore, assessing the distinctions among different dimensions and complex interactions of social support factors may provide a comprehensive view of how social support in uences health facility delivery, alongside other important factors.

Data source
The data for this study was obtained from a household survey conducted prior to delivering interventions (baseline survey) being evaluated in three districts in Jimma Zone, Ethiopia using a cluster randomized controlled trial (RCT) design. The intervention components under study in the cluster RCT were upgraded maternity waiting homes and local leader training that were hypothesized to improve levels of health facility delivery. Women reporting a pregnancy outcome (livebirth, stillbirth or abortion) up to 12 months prior to commencement of the baseline survey in January 2017 were eligible to be interviewed. A twostage sampling strategy was employed. First, 24 primary health care units (PHCUs) within the three districts were randomly selected for the trial, then, 160 women per PHCU were randomly selected from community-based lists of pregnant women generated as part of health post records. The questionnaire, administered in a face-to-face interview with women by trained research assistants in a private space at women's homes. Questionnaire contained modules on socio-demographics, maternal health service use, danger sign awareness, social support, decision-making practices and sources of information. At baseline, interviews were conducted with 3,784 (~99%) women of the 3,840 targeted sample size calculated to meet trial requirements. A protocol with detail methodology of the trial has been published previously and (14) and the trial identi er is NCT03299491.

Study setting
Jimma Zone is located 356 km from Addis Ababa in Southwest Ethiopia, Oromia National Regional State. Jimma Zone has 21 districts, and 42 urban and 513 rural kebeles. [1] The total population of Jimma Zone is estimated to be 3.2 million with the majority of the population living in rural areas (38). The three study districts selected for the cluster RCT were Gomma, Seka Chekorsa and Kersa, which had populations ranging from 180 000 to 270 000 in 2016 (39).

Design and analytic approach
A community based cross-sectional study design was used to assess association between health facility delivery and social support variables described in subsections below.The sample for this work included women of reproductive age group (15-49 years) who had uncomplicated pregnancy and delivery resulting in a live birth during the year preceding the survey. Women who faced pregnancy and delivery related complications such as bleeding, pre-eclampsia, eclampsia, mental health problems, infections, severe headache, water breaks without labour, persistent vomiting and related complications were more likely to receive immediate supports from family or relatives or close friends . Therefore, we excluded women who reported experiencing these complications during antepartum or intrapartum periods. Among the 3784 women surveyed at baseline, 480 reported experiencing antepartum or intrapartum complications and were excluded from the nal analysis. Thus, data from 3304 respondents were analyzed.
The primary dependent variable was the place of delivery of the most recent child, dichotomized as health facility or not health facility (home, in ambulance, on the road). Health facility delivery included those at the health center, hospitals and private hospital/clinic.

Variables of interest
Outcome variable The primary dependent variable was the place of delivery of the latest child, dichotomized as health facility or not health facility (home, in ambulance, on the road). Health facility deliveries included those at the health center, hospitals, and private hospital/clinic.

Independent variables
The main independent variables of interest included perceived social support, structural and functional social support characteristics. In addition to main variables we also included numerous independent variables that might be associated with place of delivery including age, education level, occupation, marital status, age at rst pregnancy, household income, availability of health facility, types of health facility, the time required to reach the nearest delivery facility, number of family members, social group participation, participated a programme that promote MCH cares, attitude towards health facility delivery, perception of health care services provision, planning to give birth at health facility, sources of information and decision-making autonomy (10,13,15,16,29,37). So, we developed social support questions on the basis of the result of the pre-intervention study which was conducted to explore major areas of maternal social supports, roles of community health actors and common sources of information in the same settings (15,33), and adopted social support questions from existing measures and review of the network literature (10,(36)(37)(38). So, we included key social support domains such as perceived social support, structural support and functional supports.
The perceived social support variable consisted of two items; rst, women were asked whether they had someone that they could depend on during their last pregnancy and childbirth when they were in need.
Responses were recorded as 1 = yes and 0 = no, and if the response was 'yes', it followed by asking the relationship of the women with the perceived individual (s) to provide supports. Responses were lists of options from 1 to 11 including husband/partner, parents, siblings, friends/neighbours, in-laws, religious groups, women developmental army [WDA], health workers, health extension workers (HEWs) and other to specify. Finally, these responses were collapsed into four categories: 1. Partner/family members (include partner, children or any family member in the household), 2. Relatives (like sibling, parent and in-law), 3.
Friends/neighbors, 4. Health workers (midwives, nurses, health extension workers). Functional social support had ve domains, including a) practical help with routine activities, b) material aid ( nancial or in-kind assistance), c) emotional support (empathy, care and trust), d) informational support (advice/suggestions), and e) accompaniment supports. For each section, subjects were rst asked whether they had received that type of functional support in the last pregnancy and childbirth. For example, to assess practical support, women were asked, "for your last pregnancy and childbirth, did you receive any practical help with routine activities?" (Example: help with child care, house chores, food preparation, cattle herding, etc). Responses were recorded as (1=yes and no = 0). If women answered yes, it was followed by "when did you receive this help?" Responses were recorded as 1-during pregnancy, 2during labour and childbirth, and 3-during postpartum. This question was also followed by structural support [types of network members] questions "from whom did you receive this help?" Responses were lists of options from 1 to 11 including husband/partner, parents, siblings, friends/neighbours, in-laws, religious groups, WDA, health workers, HEW and other to specify. These responses were collapsed into four categories:1. Partner/family members (include partner, children or any family member in the household), 2. Relatives (like sibling, parent and in-law), 3. Friends/neighbors, 4. Health workers (midwives, nurses, health extension workers). To provide an overall support size of each woman received during pregnancy, labour, and delivery, receipt of support was counted with 0 indicating no support received, 1 indicating received one type of support, 2 indicating received two types of support, and so on. Then, the total number of received supports was summed and divided by the number of maximum received supports, where 1 is the maximum score or women received highest number of different supports over the course of pregnancy and delivery. Density was conceptualized as the number of supporters/ ties that the women had during pregnancy and delivery. so, the density for each of the ve types of support was counted with 0 indicating do not receive any support from anyone, 1 indicating support received from one person, 2 indicating any support received from two persons, and so on. Therefore, the total number of ties/supporters was summed and divided by the number of maximum supporters, where 1 is the maximum score indicated that women have highest ties of social support over the course of pregnancy and delivery. For receiving continuous social support, we created a variable from the frequency of periods the women received all types of social supports from network members. A simple summative score from one open ended item, "when did you receive this help", with the responses of during pregnancy, during childbirth, and during postnatal period. Finally, those who reported they received supports during both pregnancy and delivery were considered as receiving continuous support/resulting in a nal continuous support score with values between 0 (not continuous) and 1 (received continuous support).
From socio-economic and healthcare variables, educational level was assessed based on women's responses on highest level of education completed, and categorized as; 1) no formal education, 2) primary school, 3) high school and 4) above high school. Women's responses on their primary occupation were collapsed into a nominal variable to re ect the main occupations listed; 1) housewife, 2) farmer, 3) traders, 4) others (private workers, unemployed, employee and student). Household income was created using information on the total household income of last year, which were then grouped into quirtiles; the rst quirtile ( lowest quirtile) corresponded to the lowest household income and the fourth quirtile corresponding to the highest household income. Travel time to closest health facility was assessed as continuous variable based on women's estimated time (one way walk in hours) required reaching the nearest health facility. Women's responses on their nal decision/ involvement in decision making on place of childbirth was assessed based a question whether they made decisions on their own, jointly with someone else or someone else. Then response was categorized as: 1= involved (if decision made by women/their own and jointly with someone else), 0 = not involved (if the decision made by someone else or not included the women). Attitude towards health facility delivery was measured through seven items using Likert scale [with response ranged from 1= strongly disagree to 5= strongly agree] resulted with nal score ranged from 7 to 35. Finally, the mean score was calculated with minimum score of 0.20 to maximum score of 1, in which the higher scores indicated having high attitude towards health facility delivery. Perception on healthcare provision was assessed based on evaluation scores that ranged from 1-5 to evaluate the service provided in the closest health facility; with response 5= very good and 1 = very poor. Finally, the mean score was calculated with minimum score of 0.32 to maximum score of 1, in which the higher scores indicated having highest perception on healthcare provision. Social group membership was assessed based on women's responses on membership of any social group, organization or association, and the responses were categorized as (1=yes, 0=no). Participation on programme that promote MCH cares was assessed based on women's responses on participation on any programme that promote MCH cares, and categorized as (1=yes, 0=no). Age and age at rst pregnancy were assessed based on interval age in years and number of household members was also assessed on interval numbers of household members. Whether the women planned to give birth at health facility was assessed by binary responses (1=yes, 0=no). Antenatal care use was assessed based on the number of visits during pregnancy as ANC visits and collapsed to binary responses as (1= ANC four or plus visits and 0= none/ less than four visits).

Analysis
First, descriptive analysis was carried out to summarize the socio-demographic and health service utilization characteristics of the respondents using frequencies and proportions. We also conducted bivariate and multivariate analyses of the association between each variable included in the study and health facility birth. We also ran separate logistic regression models to test two-way interactions, speci cally whether structural social support moderated the relationship between functional social support and facility birth. To identify variables associated with health facility childbirth, multivariable regression analysis was used. All independent variables and interaction terms that associated at bivariate levels (occupation, household income, number of household members, travel hour from closer health facility, participated on role that promotes maternal healthcare, involved in decisions on place of childbirth, attitude to facility delivery and perception on healthcare provision) were included in multivariable regression. From maternal social support variables perceived partner/family members to help women during labour and childbirth, women who received practical support, accompanied to health facility during last child birth, size of received supports, received continuous supports, received from partners/family members, from relatives/siblings and from friends/neighbors were also included in multivariate analysis. From two-way interactions or correlation between social support variables all tested interactions between each structural social support and functional social support variable, were not signi cantly associated with facility delivery . Age was not included because it is highly correlated with number of family members (r = .82), and from perceived social support, perceived relatives to help women during labour and childbirth was highly correlated with perceived partner/family members to help them (r = .89). All analyses were performed using SPSS software version 20.0 (SPSS Inc., 2008), and an alpha level of 0.05 was selected. with the average travel time (one way walk) from home to the closest health facility was around 1 hour. About 46% of women mentioned that they planned to give birth at the current place of delivery. Half of the women (50%) participated in a programme that promotes MCH cares and only 17.2% of women mentioned that they were a member of any social group. Almost half of the women (49%) mentioned that they involved in deciding place of delivery for their last childbirth.

Multivariate logistic regression analysis of social support and related factors with health facility delivery
After controlling for associated variables at bivariate levels, three socio-demographic and healthcare related factors including travel time from closest health facility, perception towards health facility use and decision in using health facility delivery resulted in statistically signi cant associations with health facility delivery. Therefore, the multivariate analysis showed that per unit increase in total score of travel time to health facilities the likelihood of health facility delivery also increased by 1.5 (AOR: 1.51, 95% CI 1.21 to 2.90), and women who were involved in the nal decision to identify their place of childbirth had twice the odds of using health facility delivery than women who did not deliver at a health facility (AOR: 2.12, 95% CI 1.73 to 2.58). Similarly, this study identi ed that with each unit increase in total score of perception towards the bene ts of health facility use the likelihood of health facility delivery also increased (AOR: 1.83, 95% CI 1.44 to 2.33).
Further, from social support variables our study revealed that women's perception that there were partner or family members to help them during labour and childbirth; women who received continuous support, women who were accompanied during childbirth, and women who received any support from their friends were all signi cantly associated with health facility delivery.
Women who perceived there were partner/family members to help them during labour and childbirth were more than three times more likely to have a health facility delivery than women who did not perceived there were partner/family members to help (AOR: 3.62, 95% CI 2.74 to 4.79). Women who described having companions to accompany them for health facility visits during labour and delivery had 1.6 times the odds of having used health facility delivery than women who did not this kind of support (AOR: 1.63, 95% CI 1.34 to 2.00).
Women who received continuous support during pregnancy, labour and delivery had almost twice the odds of having used a health facility for delivery than women who did not deliver at a health facility (AOR: 1.97, 95% CI 1.20 to 3.23), and women who received any type of social support from friends/neighbors during pregnancy and delivery were 1.6 times more likely to use health facility delivery than women who did not deliver at a health facility (AOR: 1.62, 95% CI 1.16 to 3.23) ( Table 4). We also ran separate logistic regression models to test two-way interactions, speci cally whether density of ties moderated the relationship between each network members (family members, relatives, friends and health workers) and health facility birth, as well as whether each network members moderated the relationship between each functional supports and health facility birth. At bivariate analysis, the result showed that except the interaction between emotional support and support received from relatives, all other interactions were not signi cant, but at multivariate analysis; all interactions were not signi cantly associated with health facility delivery (table 4).

Discussion
In order to understand culturally and socially relevant types of maternal social support, and to identify which measures may provide the best t for our speci c setting, we assessed the patterns of maternal social support and health facility delivery including the interactive effect of structural and functional social supports. In this study we found that around 47% of women delivered at health facility, consistent with recent studies in the country including data from the 2019 Mini EDHS and others which reported a proportion between 43%-52% (9,40,41). Previous reports in Ethiopia and in other low-income countries like Ghana, Kenya, and Bangladesh (7,10,16,42) showed much higher rates of facility deliveries (between 70%-74%), a discrepancy that could be explained by lower health facility delivery in rural areas, such as the study area of our analysis. The other explanation could be that our analysis considered women with uncomplicated pregnancy and childbirth, who may be less likely to delivery at a health facility because pregnancy and labour are seen as normal human events and not considered as a problem unless the women encountered complications and related health problems (15,42,43).
The current ndings showed that perceived support, types of support and the people providing this support were associated with place of delivery. Multivariable logistic regression analysis revealed that women who perceived that their partner /family members helped them during childbirth were more likely to deliver at health facilities, although this association should be interpreted with caution due to the possibility of reverse causality possible with cross-sectional data. Consistent with previous research (7,10,17) women who perceived that their husbands and family members were in favour of health facility deliveries were more likely to have given birth at one; receiving advice from family members during pregnancy further strengthened this likelihood.
In terms of functional social support, we found that women who were accompanied during labour and childbirth had a signi cantly higher odd of facility delivery than women without such support. The nding was supported by other studies which nd that women strongly prefer to have a birth companion (22,44), which is also associated with increased satisfaction with healthcare services. A Cochrane effectiveness review suggested that having a labour companion enhances utilization of MCH services, and improves outcomes for women (45,46). Further, in addition to enhancing health facility delivery, women who have companions to accompany them during labour and childbirth may also bene t from more rapid diffusion of health information, and through practical and affective support in accessing to local services and facilities (13,(45)(46)(47).
This study also supported the hypothesis that women who received continuous social support are more likely to use health facility delivery, which is in agreement with a previous study conducted in different LMICs (15,18,46,48). This suggests that women appreciate the continuous presence of maternal social supporters. Studies including systematic review recognizes the signi cance of continuous support of the woman's choice from her social network (such as her husband, partner, mother, or friend) during childbirth and WHO also encourages different organizations to issue practical guidelines promoting continuous support (4,49), as well as efforts to reduce the rate of dropout in the continuum of MCH care (1,25,50). A Cochrane effectiveness review on women's birth experience collected data from 11 RCTs comparing the impact of continuous and intermittent support; it reported that negative feelings about the childbirth experience and home delivery were signi cantly lower among women who received continuous support (44,46,48,51,52). The result of this study a rms these earlier ndings and recommendations of the positive role of continuous support during pregnancy and childbirth in enhancing health facility delivery, when the supporter is part of the woman's social network.
By companions of the woman's choice from her social network (such as her husband, partner, mother, or friend). Women who received continuous labour support were more likely to give birth 'spontaneously' From structural social support, the use of health facility delivery was positively and signi cantly associated when women received supports from close friends/neighbors. Although different studies report that friendship can have both negative and positive health impacts, many scholars report that the bene ts of friends are greater than those from relatives. For example, in terms of predicting health, friendship occasionally predicts health to an equivalent and, in some cases, larger degree compared to spousal and parent-child relationships (26,29,50). Different social support variables highlighted by different researchers as predictors of the health service utilizations, such as number of ties/density, emotional supports, interaction between functional and structural dimensions, participation in roles that promote healthcare, and member of any social groups, were not associated with health facility delivery in our study. Similarly, some studies argue that, instead of counting number of ties, amount of received supports and participation in activities that have different objectives than MCH care, it would be better to assess the quality of relationship and consistency of support from network members (24,50,53). Further, different studies on social support recommended caution during subgroup analyses and interpretation of social support variables (10,31,54), due to the fact that consistent patterns about the effectiveness of different social supports during pregnancy and childbirth may be enhanced or reduced by contexts in the birth setting, type of provider, and timing of onset of support (29,50,55). So, this result may imply about the main advantage of getting local context-based labour support which may target underserved pregnant women, provides service at no cost to the client, and seeks to employ supporters who are from the same local contexts in which they are living. In Ethiopia, receiving such in-kind social support during labour and childbirth is congruent with the social and cultural norms, where relationships and support from family and extended family, in-laws, close friends and customs are highly valued, including input and advice with regard to the decision about where and with whom to give birth.
Consistent with previous studies in in Ethiopia, Ghana, Kenya, and Nepal (13,15,17,43), our analysis identi ed a signi cant positive association between women who are involved in deciding their place of childbirth and health facility delivery. Different social support studies and theories have suggested that interactions between network structure and perceptions of advice from family members and close friends would help explain pregnant women's decision to utilize health facility birth (15,18,56). Further, this study indicated that women who had a positive perception towards the bene ts of health facility delivery were signi cantly more likely to seek health facility delivery. Several studies corroborated this nding, and the positive experiences of women in their childbirth helped them to have a positive perception, hence determining the childbirth place in their subsequent deliveries. Further, women whose families and close friends had positive perceptions about health facility birth were more likely to deliver in a health facility (57,58) .
Our ndings also revealed that women who lived further from the health facility were more likely to have used a health facility for delivery than those women who lived closer. In this regard, studies in Africa and Southeast Asia reported different ndings on distance from health facility as an in uencing factor for women's decisions to use MNCH care (13,57,59) . Qualitative and quantitative studies from the same setting in Ethiopia reported that women living a greater distance to the maternity facility were more likely to utilize MNCH services including maternity waiting homes (15,16,37,45). Within the study setting, there is a freely available ambulance transport service during labour and the availability of maternity waiting homes may help women from remote areas use childbirth care (13,15,16,37). The other reason could be providing women with supports and transportation funds before they go to a facility for delivery and managing transportation options, as well as using maternity waiting home may increase service utilization (3,13,28,42,60).
The ndings of this study, however, must be considered in light of various limitations. Women were only asked about current patterns of social supports, while the childbearing was done in the past, again leading to issues about the direction of the relationships as well as the issue of whether current patterns of maternal social support consistently re ect the future patterns. Again it is not clear whether participant's perceptions of network approval of facility delivery may have been in uenced by their actual experiences, rather than simple explanations of maternal support for facility delivery; mixed research methods including longitudinal studies could be important here in order to establish causality. Further, women could have responded with socially desirable answers such as reporting that they considered facility delivery in a positive light. In addition, ascertaining support for practical and advice regarding facility delivery does not provide the whole picture of social support. We therefore measured only some of the instrumental and informational support available to women. Despite these limitations, this study makes valuable contributions to the existing literature on social support in low resource settings. We also sampled women who had given birth 12 months prior to the survey which helped to limit recall bias written consent, participants were given time to read the consent forms and to ask questions before signing. For oral consent, the objectives of the research and ethical issues related to safety, con dentiality, and privacy were explained to uneducated participants in the local language of women's choice (Amharic or Afan Oromo) before asking if they agreed to the interview or focus group.

Consent to publish
Not applicable.

Availability of data and materials
Data used for this analysis (copy of the questionnaire used in the trial) will be provided by the authors upon reasonable request.

Competing Interests
The authors declare that they have no competing interests.