Barriers to institutional childbirth in rumbek north county, South Sudan: A qualitative study

Background South Sudan has one of the world’s poorest health indicators due to a fragile health system and a combination of socio-cultural, economic and political factors. This study was conducted to identify barriers to utilisation of institutional childbirth services in Rumbek North County. Methods Data were collected through 14 focus group discussions with 169 women and 45 men, and 18 key informant interviews with community leaders, staff working in health facilities, traditional birth attendants, and the staff of the County Health Department. Data were analysed using inductive content analysis. Results The barriers to institutional childbirth were categorised under four main themes: 1) Issues related to access and lack of resources: long distance to health facilities, lack of transportation means, referral problems, flooding and poor roads, and payments in health facilities; 2) Issues related to the socio-cultural context and conflict: insecurity, influence of the husband, lack of birth preparedness, domestic chores of women, influence of culture; 3) Perceptions about pregnancy and childbirth: perceived benefit of institutional childbirth, low childbirth risk perception, and medicalisation of childbirth including birth being perceived to be natural, undesirable birth practices, privacy concerns, and fear of caesarean section; and 4) Perceptions about the quality of care: inadequate health facility infrastructure and perceived neglect during admission. Conclusions Multiple factors hinder institutional childbirth in Rumbek North. Some of the factors such as insecurity and poor roads are outside the scope of the health sector and will require a multi-sectoral approach if childbirth services are to be made accessible to women. Detailed recommendations to increase utilisation of childbirth services in the county have been suggested. This study is reported according to the consolidated criteria for reporting qualitative research (COREQ) [13], and the COREQ checklist is presented in S1 Table. This study was conducted in Rumbek North County which in 2015 had a population of 59,740 people, based on the 2008 census projection [14], and was divided into 6 payams (sub-county units): Alor, Malueth, Mayen, Madol, Maper and Wunrieng. The county’s population is semi-nomadic and pastoralism is the main economic activity. As the population moves, it establishes temporary settlements called “cattle camps”. In 2015, the county had one Primary Health Care Centre (PHCC) located in Maper, and seven Primary Health Care Units (PHCUs). Each one of the PHCUs was run by one community health worker (CHW), one traditional birth attendant (TBA), and one drug dispenser. The PHCC had three professional health workers: a nurse, a midwife and a clinical officer (a holder of a three-year diploma in medicine); all expatriates. The nearest hospital from Maper PHCC is Rumbek State Hospital; located 100 Km away. South Sudan has a decentralised health system that is based on four levels of administrative structures: central, state, county, and community [10]. PHCUs and PHCCs are, respectively, the lowest and second lowest health facilities situated at the community (sub-county) level. PHCCs are mandated by the Ministry of Health to provide childbirth services; including emergency obstetric care, while PHCUs, depending on the availability of qualified staff, are supposed to attend to normal (uncomplicated) deliveries [15]. Each PHCU is supposed to be staffed by two CHWs and a community midwife while a PHCC is supposed to have one clinical officer, three professional nurses, two midwives, three CHWs, and lower cadre staff [15]. Routine data show that, in 2014, out of the expected 3278 childbirths in the county, only 36 (1.1%) were assisted by a skilled birth attendant and 56 (1.7%) took place in any health facility. This is a cross-sectional qualitative study that collected data utilising focus group discussions (FGDs) and key informant interviews (KIIs). FGDs were used to explore how different factors influence an individual’s decisions and perceptions related to care seeking for institutional childbirth. KIIs were used to gather in-depth information on the situation of maternal health care in the county and to triangulate some of the information gathered during FGDs. Villages in the country were stratified by payam and randomly selected as follows: two villages from each of Malueth and Mayen payams (the most populous) and one from each of Madol, Alor, Maper and Wunrieng. In each village, one FGD was conducted with women who had delivered in the past one year. Additionally, in a random sub-sample of half of the villages, husbands of women who delivered in the past one year were recruited to participate in men’s FGDs. Two extra FGDs with women were conducted in one cattle camp that was accessible during the study period. In each selected village, CHWs with the help of the village leaders invited (via face-to-face) 12 eligible participants to take part in the study. This was done one or two days before the respective FGD. The number of women who turned up for FGDs was often slightly higher than expected, and all were included. KIIs were conducted with the following categories of individuals: 1) CHWs in the PHCC and PHCUs, 2) TBAs (those working in health facilities and those in the community), 3) community leaders, and 4) staff at the County Health Department (CHD). Participants of KIIs were purposively selected in consultation with the project staff and were individuals perceived to be knowledgeable about the maternal health situation in Rumbek North. FGDs and KIIs were conducted in March 2015 utilising open-ended pretested question guides (S1 File). Each FGD was conducted by two Dinka (local language) speaking facilitators who were previously unknown to participants. The facilitators were of the same gender as participants, had at least high school level education, and were well versed with the local language and culture. One data collector was in charge of facilitating the sessions while the other one managed audio recordings and took field notes. The data collectors were trained for one day and were supervised by one of the co-authors (CW) who has experience in conducting qualitative studies and was present at all FGDs. The venues for FGDs were suggested by community leaders and often included local churches and under tree shades. KIIs took place at venues that were convenient to participants following prior arrangements with the study team. Based on time and logistical constraints, a total of 14 FGDs and 18 KIIs were conducted (Fig 1). Women’s FGDs had a median of 16 participants while men’s FGDs had a median of 9 participants. To avoid interruptions and to maintain privacy, other people who approached the groups were kindly requested to leave. There were no drop-outs during FGDs. All KIIs were conducted by one of the co-authors (CW) either directly in English (for CHWs and CHD staff) or through a translator (for the other type of informants). Both KIIs and FGDs were audio recorded. Each FGD session lasted for about one hour whilst each KII lasted for about 20 minutes. No repeat interviews were conducted. CHD: County Health Department; PHCC: Primary Health Care Centre; PHCU: Primary Health Care Unit; TBA: Traditional birth attendant. The audio recordings in Dinka were transcribed and translated into English by bilingual (Dinka and English) speakers while the recordings of KIIs conducted in English were transcribed by CW. It wasn’t logistically possible to return the transcripts to participants for review. The transcripts were edited and read through severally to obtain an overall picture and to identify emerging patterns. CW analysed the transcripts based on the inductive content analysis approach [16]. The analytic framework was adapted from a published systematic review of qualitative studies on determinants of delivery care use in low- and middle-income countries [17]. The framework contains four main themes: 1) Access and resource availability, 2) Influence of the socio-cultural context, 3) Perceptions of pregnancy and delivery, and 4) Perceptions of the quality of care [17]. Each transcript was entered into NVivo 10 (QSR International, Melbourne, Australia) where a coding frame had been set up using the key themes (and sub-themes) from the analytic framework. Each transcript was then read through and segments of the text that captured the predefined theme were coded. New emerging themes were also identified and coded. This was done is a similar way for all FGDs. Information from KIIs was used to triangulate findings from FGDs. The data for each theme and sub-themes were then pieced together to provide an overview of the content relating to that specific theme or sub-themes. Quotes were selected to represent a typical response or to illustrate a deviant opinion. Participants did not provide feedback on the findings. This study was approved by the Ministry of Health Ethics Committee and by Rumbek North CHD. Because the participants of FGDs, TBAs and community leaders were either illiterate or of low literacy levels, they provided verbal informed consent, which was audio recorded, after an explanation about the study. The ethics committee had approved this method of obtaining consent. CHWs and staff of the CHD provided written consent. Permission to conduct the study in selected villages was sought from village leaders. All collected information including audio recordings and transcripts were securely kept and were accessible only to the research team. Each participant of FGDs received a bar of soap to compensate for his/her time. No monetary incentives were provided.