Barriers and facilitators to health care seeking behaviours in pregnancy in rural communities of southern Mozambique

Background: In countries, such as Mozambique, where maternal mortality remains high, the greatest contribution of mortality comes from the poor and vulnerable communities, who frequently reside in remote and rural areas with limited access to health care services. This study aimed to understand women’s health care seeking practices during pregnancy, taking into account the underlying social, cultural and structural barriers to accessing timely appropriate care in Maputo and Gaza Provinces, southern Mozambique. Methods: This ethnographic study collected data through in-depth interviews and focus group discussions with women of reproductive age, including pregnant women, as well as household-level decision makers (partners, mothers and mothers-in-law), traditional healers, matrons, and primary health care providers. Data was analysed thematically using NVivo 10. Results: Antenatal care was sought at the heath facility for the purpose of opening the antenatal record. Women without antenatal cards feared mistreatment during labour. Antenatal care was also sought to resolve discomforts, such as headaches, flu-like symptoms, body pain and backache. However, partners and husbands considered lower abdominal pain as the only symptom requiring care and discouraged women from revealing their pregnancy early in gestation. Health care providers for pregnant women often included those at the health facility, matrons, elders, traditional birth attendants, and community health workers. Although seeking care from traditional healers was discouraged during the antenatal period, they did provide services during pregnancy and after delivery. Besides household-level decision-makers, matrons, community health workers, and neighbours were key actors in the referral of pregnant women. The decision-making process may be delayed and particularly complex if an emergency occurs in their absence. Limited access to transport and money makes the decision-making process to seek care at the health facility even more complex. Conclusions: Women do seek antenatal care at health facilities, despite the presence of other health care providers in the community. There are important factors that prevent timely care-seeking for obstetric emergencies and delivery. Unfamiliarity with warning signs, especially among partners, discouragement from revealing pregnancy early in gestation, complex and untimely decision-making processes, fear of mistreatment by health-care providers, lack of transport and financial constraints were the most commonly cited barriers. Women of reproductive age would benefit from community saving schemes for transport and medication, which in turn would improve their birth preparedness and emergency readiness; in addition, pregnancy follow-up should include key family members, and community-based health care providers should encourage prompt referrals to health facilities, when appropriate. This article is part of a larger formative research study conducted in Mozambique, India, Nigeria and Pakistan, in preparation to a cluster randomized controlled trial of a Community Level Intervention for Pre-eclampsia and Eclampsia (the CLIP trial) ({“type”:”clinical-trial”,”attrs”:{“text”:”NCT01911494″,”term_id”:”NCT01911494″}}NCT01911494) [16]. While the formative research was based on a mixed methods approach, the present article focuses on the qualitative component conducted within an ethnographic framework [17, 18]. A detailed description of these methods is presented elsewhere [19]. The study site consisted of five Administrative Posts (APs) within three districts in southern Mozambique: Xai-Xai and Bilene-Macia districts (in Gaza Province), and Manhiça district (in Maputo Province) (Fig. 1). These APs were purposely selected to reflect the diversity of socioeconomic and demographic characteristics in southern Mozambique, such as level of urbanization, population density, distance to a trading centre, presence of referral health facilities, and physical access to them. Each of the districts and respective APs included in this study is briefly described below. Map of the study area Xai-Xai, the capital town of Gaza Province, is located on the eastern coast, covering an area of 1,908 Km2, with a population estimated at 127,351 inhabitants [20]. The AP selected within Xai-Xai was Chongoene, which is a coastal region located 18 Km north-east of Xai-Xai. It has a population of 101,752 served by 8 primary health centres (PHC), with a total of 9 maternal-and-child health (MCH) nurses. In addition, there is access to the referral Provincial Hospital of Xai-Xai, a tertiary level facility. At the time the study was conducted, the area was also covered by 5 community health workers (CHW), locally named Agentes Polivalentes Elementares (APEs). Chongoene is the newly appointed head office of the district, and as result commerce, tourism, agriculture, and administrative services are thriving. Bilene-Macia is located in southern Gaza Province, with an area of 2.180 Km2 and a population of 151,548 [20]. Within this district, the AP of Messano was selected for this study. It has a population of 21,471 inhabitants, and is served by two PHC with two MCH nurses and four CHWs. The referral facility is Bilene-Macia Health Centre, a secondary level facility. The community infrastructure within Messano is weak, marked by poor access to the main road. The population is primarily employed in small-scale farming. Manhiça district is located in northern Maputo Province, 80 Km from Mozambique’s capital city. The district has an area of 2,689 Km2, a population of 245,829, and a mixture of urban and peri-urban communities [20]. The entire district population participates in the Health and Demographic Surveillance System (HDSS), in place since 1996 [21]. Due to the socioeconomic diversity found in this district, three APs, namely Três de Fevereiro, Ilha Josina Machel, and Calanga were selected for this study. Três de Fevereiro, located 31 Km north of Manhiça village, has 40,208 inhabitants. Four PHC with seven MCH nurses and three midwives serve the area, which had no CHWs at the time the study was conducted. Most residents are employed by the sugar and rice industry and engaged in informal trade and migrant labour in South Africa. This AP is intersected by the country’s 1st National Road. It has reasonable communication networks, roads, and public services. Ilha Josina is an island 50 Km north of Manhiça Village. It has a population of 9,346 inhabitants, mostly engaged in agriculture, served by one PHC with two MCH nurse and two CHWs. Calanga is a coastal AP, located 25 Km east of Manhiça Village, with a population of 9,524 inhabitants (mostly fishermen and small-scale farmers) served by one PHC with one MCH nurse and five CHWs. Both Ilha Josina Machel and Calanga are characterized by poor road infrastructure and transportation networks, severely affected by harsh weather conditions during the rainy season. Manhiça District Hospital is a referral facility for these three APs, although some patients from Ilha Josina Machel can also be sent to Xinavane Rural Hospital. Both hospitals are secondary level facilities. The study participants comprised of community members and health care providers. Community members consisted in women of reproductive age between 18 and 49 years (including pregnant women), male and female decision-makers (elders, husbands, partners, mothers, and mothers-in-law of women of reproductive age). Health care providers included formally-recognized cadres within the national health services (nurses, midwives, medical technicians) and traditional health care providers (TBAs, matrons and traditional healers). Although traditional health care providers in Mozambique often have interchangeable roles [22], it is worth mentioning some important differences among them. Traditional healers are mostly sought for the diagnosis, treatment and protection from illness, misfortune and other social concerns [23]; TBAs provide assistance to women during pregnancy, birth and the postpartum period; matrons are responsible for performing a variety of rituals including those for new-borns and adolescents [22, 24]. Data collection consisted of focus group discussions (FGD) and individual in-depth interviews (IDI) conducted with community members and health care providers (Tables 1 and ​and2).2). Focus groups and individual interviews were chosen to gain an understanding of the social norms and local contexts underlying care seeking practices, rather than the individual experiences and meanings assigned to them. Interviews were conducted when it was not practical to convene the required number of participants within a specific target group. Focus group discussions conducted In-depth interviews conducted The number of focus groups and individual interviews was pre-determined based on previous experiences of reaching data saturation regarding similar topics in different contexts [25, 26]. Both across and within-group saturation was assessed. For interviews with health care providers, the snowball sampling approach was used for recruitment, drawing initially on existing networks of local investigators and health professionals at the sites. Community members were identified through community leaders, who were provided with the required socio-demographic characteristics for inclusion. Focus groups were usually conducted out-doors at the community’s “circle” (location of community leaders’ office); individual interviews were conducted at the participants house. Interviews and focus groups were conducted by trained facilitators belonging to the social science research unit of the Manhiça Health Research Centre (CISM). Gender balance within the team members was ensured to cater for possible gender-sensitive issues, especially in one-on-one interviews. All interviews and focus groups, of which some were conducted in Portuguese and others in Changana (local dialect) according to participants’ preference, were audio recorded, and transcribed verbatim in Portuguese, preferably by the same team members who collected the data. Data quality checks were done by the social science team leader by reviewing the transcripts while listening to the audio recording. Data was analysed using NVivo version 10.0 (QSR International Pty. Ltd. 2012). Thematic data analysis was performed through the following steps: generating categories, coding text according to each category; annotating emerging themes and patterns and readjusting the categories and relationships between them; testing emergent themes through systematic searches of coded text; investigating alternative explanations through systematic searches of uncoded text. The social science team leader and the study’s senior social scientist conducted the coding of all Portuguese transcripts at CISM. Assistance was given from the CLIP social science co-ordinator from the University of British Columbia (UBC), who spent a significant amount of time onsite to support training, oversee data collection and perform the analysis quality control through repeat coding of one third of transcripts, which were translated into English, and evaluating coding agreement. All data collection was conducted after obtaining signed informed consent from each participant, as well as permission to record individual interviews and group discussions. Ethical approval for this study was obtained from the CISM Institutional Review Board (CIBS – CISM) and the UBC Review Board.