Menu
Menu
Menu
Menu
Similar to many sub-Saharan African countries, maternal mortality in Ghana ranks among the highest (39 th ) globally. Prior research has demonstrated the impact of social network characteristics on health facility delivery in sub-Saharan Africa. However, in-depth examination of the function of all members in a woman’s network, in providing various types of support for the woman’s pregnancy and related care, is limited. We qualitatively explore how women’s network social support influences facility delivery. Qualitative data came from a mixed methods evaluation of a Maternal and Newborn Health Referral project in Ghana. In 2015 we conducted in-depth interviews with mothers (n = 40) and husbands (n = 20), and 4 focus group interviews with mothers-in-law. Data were analyzed using narrative summaries and thematic coding procedures to first examine women’s network composition during their pregnancy and childbirth experiences. We then compared those who had homebirths versus facility births on how network social support influenced their place of childbirth. Various network members were involved in providing women with social support. We found differences in how informational and instrumental support impacted women’s place of childbirth. Network members of women who had facility delivery mobilized resources to support women’s facility delivery. Among women who had homebirth but their network members advocated for them to have facility delivery, members delayed making arrangements for the women’s facility delivery. Women who had homebirth, and their network members advocated homebirth, received support to give birth at home. Network support for women’s pregnancy-related care affects their place of childbirth. Hence, maternal health interventions must develop strategies to prioritize informational and instrumental support for facility-based pregnancy and delivery care.
This study was approved by the Ghana Health Service Ethical Review Committee and exempted from ethics review by University of North Carolina (UNC)–Chapel Hill’s Internal Review Board, as it was considered a program evaluation. Data for the study came from a mixed methods evaluation of an intervention to improve access to maternal and newborn health services and outcomes in rural Ghana by strengthening the referral process for pregnant women and sick newborns with complications. Between January and March 2015, household surveys were administered to 1260 women of reproductive age in 3 districts each in the NR and CR. We recruited a subsample of women who participated in the household surveys to participate in individual qualitative interviews (Fig 1). In-depth interviews (IDIs) were conducted with 20 women from across 2 randomly selected districts each in the NR and CR. Also, a sample of the women’s husbands (10 each in the NR and CR) was interviewed. Four focus group discussions (FGDs) were conducted with mothers-in-law (MILs) (8 to 12 per group) who were selected from a community in each of the district where the IDIs were conducted. We purposively sampled women who gave birth, and MILs (unrelated to the women selected for the IDIs) whose daughters-in-law gave birth, in the past 12 months prior to data collection. To enable comparison on use of facility delivery, we also sampled equal numbers of women who had facility births and homebirths. Two interviews each with NR and CR women, and one interview with a NR husband were lost, due to data corrupted audio files (Table 1). We used an egocentric network approach to collect qualitative data on women’s social network composition and function [36]. This approach enables examination of network characteristics from the women’s (focal individuals) perspectives, rather than interviewing all members of their networks. Women’s perceptions of their social networks are well correlated with the actual attributes of their network members, and thus serve as a reliable indicator of their network members’ characteristics [37]. To assess network composition, women were asked to first provide a comprehensive list of all network members involved in their pregnancy experiences. They then provided information on how they are related, and their frequency of contact and interactions with the network members; how long they have known the members; and how far network members live from them. Women also described the type of relationship they have with network members and which network members know and interact with each other. To assess network function, women were asked about how each network member was involved in their pregnancy experiences and the kinds of support and influence each member provided on their health service utilization during pregnancy and delivery. Husbands were asked to describe their wives’ pregnancy experiences, identify roles of individuals involved in health decisions about their pregnancy, and describe the kinds of support they and other family members provided their wives. In FGDs MILs were asked to identify roles of individuals involved in health decisions about women’s pregnancy, describe community perceptions about place of childbirth, and describe various supports MILs and family members provide women. Our objective was to understand important themes related to network support and women’s pregnancy and delivery experiences. During fieldwork the first author (LC) met with research assistants at the end of each day to review their summaries of the day’s interviews and to discuss emerging themes. Data saturation, the point at which additional data did not yield new insights on key themes relevant to the study, was reached before completing the interviews with women and their husbands [38]. However, the sample size was not modified because this study was part of a larger research study that covered other topics of relevance to the overall study (Table 1). Two additional FGDs with mothers-in-law were added in order to reach saturation for the FGDs. In each study region, male and female research assistants who were experienced in conducting qualitative data collection in the local languages of the study communities (Twi and Fanti in the CR and Dagbani and Lekpepkel in the NR) conducted the interviews and took detailed field notes. The first author trained them to collect qualitative social network data. Research assistants obtained informed verbal consent from all study participants, and also verbal consent from guardians of women under the age of 18 years. They read aloud informed consent forms to study participants, and guardians when appropriate. The research assistants then noted on the forms when verbal consents were received. Ethics review approval for the consent process was obtained from the Ghana Health Service and the UNC-Chapel Hill. All interviews were audio-recorded and transcribed into English. Our analysis began with close readings of all IDI and FGD transcripts. LEC wrote narrative summaries of pregnancy and delivery experiences, and of each FGD with MILs on their perceptions about childbirth experiences of women in their communities. LEC, with input from CB, then generated preliminary descriptive codes and memos on the various roles of network members in women’s pregnancy and delivery care experiences using these summaries. After another review of the transcripts and the research team’s discussion of emergent themes from the narrative summaries and field notes, LEC developed a core set of inductive codes in order to conduct thematic analysis of women’s social network characteristics and pregnancy/delivery care experiences. LEC applied these codes to the IDIs and FGDs using Atlas.ti software (version 7.0, Scientific Software Development GmbH, Eden Prairie, MN), and with input from CB modified and added new codes during the coding process [39]. After coding all transcripts, LEC conducted coding checks by reviewing all the transcripts to ensure that the coded data reflected codes defined in the codebook. Discrepancies identified were discussed with CB corrected to ensure coding accuracy and consistency. Subsequently, LEC worked with the research team in analyzing and interpreting that data. We reviewed the code outputs and developed code summaries and analytic matrices [40, 41]. We used both code summaries and narrative summaries, which captured women’s interactions with network members during pregnancy, to create a table summarizing women’s network composition, and described contextual information on the supportive roles of network members. In the matrices, we compared women who had facility versus homebirth and NR versus CR women, on themes of how network support in women’s receipt of pregnancy, labor, and delivery care affected their place of delivery.
N/A
