Implications of power imbalance in antenatal care seeking among pregnant adolescents in rural Tanzania: A qualitative study

Background Adolescent girls (10–19 years) are at increased risk of morbidity and mortality from pregnancy and childbirth complications, compared with older mothers. Low and middle-income countries, including Tanzania, bear the largest proportion of adolescent perinatal deaths. Few adolescent girls in Tanzania access antenatal care at health facilities, the reasons for which are poorly understood. Methods We conducted a qualitative thematic analysis study of the experiences of pregnant adolescents with accessing antenatal care in Misungwi district, Tanzania. We recruited 22 pregnant or parenting adolescent girls using purposive sampling, and conducted in-depth interviews (IDIs) about antenatal care experiences. IDI data were triangulated with data from eight focus group discussions (FGDs) involving young fathers and elder men/women, and nine key informant interviews (KIIs) conducted with local health care providers. FGDs, KIIs and IDIs were transcribed verbatim in Swahili. Transcripts were then translated to English and analysed using emergent thematic analysis. Results Four main themes emerged: 1) Lack of maternal personal autonomy, 2) Stigma and judgment, 3) Vulnerability to violence and abuse, and 4) Knowledge about antenatal care, and highlighted the complex power imbalance that underlies barriers and facilitators to care access at the individual, family/interpersonal, community, and health-systems levels, faced by pregnant adolescents in rural Tanzania. Conclusion Adolescent antenatal care-seeking is compromised by a complex power imbalance that involves financial dependence, lack of choice, lack of personal autonomy in decision making, experiences of social stigma, judgement, violence and abuse. Multi-level interventions are needed to empower adolescent girls, and to address policies and social constructs that may act as barriers, thereby, potentially reducing maternal morbidity and mortality in Tanzania. We conducted an in-depth qualitative thematic analysis study to explore the lived experiences of pregnant adolescents with accessing ANC in the rural Misungwi district, Tanzania. We used the social—ecological model as a conceptual framework, which categorizes barriers and enablers to seeking and obtaining health care on 4 levels: individual, interpersonal and family, community, and system levels [14]. Our study was nested as a sub-study within a larger longitudinal implementation and evaluation of the Mama na Mtoto (“Mother and Child”) intervention in Tanzania, which aimed to improve the delivery of essential health services to pregnant women, mothers, newborns and children under five; and to improve the health practices and utilization of essential health services by the same target groups [15]. This was conducted through implementation of district-led activities to improve the planning and delivery of high-quality facility-based maternal, newborn and child health (MNCH) services combined with strengthening the demand for these services at the community level while increasing linkages between the community and local health facilities through mobilization of a volunteer community health worker network. We conducted our study between July 2018 and September 2019 in the Misungwi district of Northern Tanzania. Misungwi has a population of approximately 400,000 people, over 90% of whom live in rural areas [16]. The district consists of over 724 small hamlets scattered throughout flatland terrain, where piped water, electricity and sanitation facilities are exceptionally rare [10]. Rural households are typically low-income and highly dependent on subsistence farming, pastoralism, petty trade, and fishing. Misungwi district is served by 49 health facilities (43 dispensaries, 4 health centres, 1 government hospital, and 1 private non-profit hospital). We recruited adolescent girls aged 10–19 years who were pregnant or parenting a child aged less than five years at the time of data collection, from four rural villages; Isesa, Buhunda, Nyamayinza and Kijima, using a maximum variation purposive sampling strategy, wherein we sought to include a variety of maternal ages, martial statuses, and levels of education. We obtained permission from the Misungwi District Medical Officer via the Misungwi District Executive Director to the village government, Village leaders, and through Health facilities administration, to conduct the study. Before recruiting adolescent girls, we held community engagement meetings with community leaders and community members to introduce the study, provide detailed information about the study, and explain its purpose. Community leaders and volunteer community health workers (CHWs), further engaged with community members to generate awareness and support for the study. Volunteer CHWs, who are seen as peers among villagers, were trained on study aims, protocol, and inclusion and exclusion criteria, and helped to identify potentially eligible adolescent participants. Trained research assistants met with prospective participants multiple times: initial visits involved meeting pregnant and parenting adolescents and their parents/partners at their household to provide detailed information about the study, to build trust and rapport, and to prepare for the interview process. If adolescents indicated interest in participating in the study, the research assistant invited them to engage in an in-depth interview (third visit) (IDI) in their preferred language (Swahili, or Sukuma) and at a location of their choice, to ensure comfort. Using the same method of community engagement, we also identified women and men who were parents, guardians, and in-laws of an adolescent girl who was pregnant or parenting a child under five at the time of data collection, as well as young husbands (aged less than 25 years), and invited them to participate in focus group discussions (FGDs). Finally, we invited local health care workers (nurses, midwives, physicians) to share their perspectives on the barriers and facilitators faced by pregnant adolescents in accessing ANC, in individual key informant interviews (KIIs). Participants who could not read or write were asked to select a trusted witness who could translate the information written in the consent form, and were asked to sign the consent with a thumbprint. All participants 18 years of age and above signed a written consent with a witness signing as well. Parents/guardians or husbands (18+) signed consent forms for adolescents less than 18 years of age, and the adolescents signed assent forms. Participation was voluntary and only those who fulfilled consent processes were interviewed. Confidentiality was observed and all information given by the participants was de-identified and assigned a pseudonym, or generic title for data analysis and reporting. We used a narrative data collection approach, supplemented by in-depth, semi-structured interviews for IDIs, and semi-structured interviews for FGD and KIIs. All data collection guides and questions were pilot tested [17]. Tanzanian members of the authorship team (WM, JK, PS) along with trained local research assistants who were fluent in the local languages, conducted all IDIs, FGDs and KIIs, and rotated through the roles of facilitator and note-taker. For IDIs we asked participants to tell us the story of their pregnancy and followed up with open-ended questions about their pregnancy circumstances and experiences, experiences with accessing ANC, decision-making around ANC, knowledge about ANC, and perceptions of their families, communities, and health workers’ attitudes about adolescent pregnancy. Our open-ended in-depth questions were based on the socio-ecological model and explored experiences on the personal, family, community, health-system and societal levels [18]. All interviews and focus groups were conducted by trained and experienced members of the research team. We conducted IDIs in the location that was selected as most comfortable by the participant. FGDs were held in schools, churches or local leaders’ offices, based on availability and accessibility for the participants. Interviewers explained the aim of the study and led all participants through the informed consent process before interviews or discussions were started. Interviews and FGDs were conducted and audiotaped in the preferred language of the participant (either Sukuma or Swahili), and field notes were taken. Participants were provided a transport allowance of 2000 Tanzanian shillings (equivalent to approximately 1 United States Dollar (USD)) and health care workers were given 5000 Tanzanian shillings (approximately 2 USD). All Sukuma interviews were transcribed verbatim in Swahili (as Sukuma is not commonly written) by research assistants who were fluent in both languages. Quality checks of the translations were conducted, wherein a second bilingual research assistant listened to the audio recording while reading the Swahili transcription, and added to or edited the transcript as needed, to ensure accuracy. A discussion among the research assistants was used to finalize the wording in cases where direct translations were challenging. Swahili recordings were transcribed verbatim in Swahili. All transcripts were then translated from Swahili to English by trained bilingual members of the research team for analysis. We then conducted a second set of quality checks by comparing the English translations to the original language recordings in a group that included native speakers of each language, to ensure that meaning nor content were lost in the translation process. We imported transcripts into NVIVO® 12 to conduct coding and emergent thematic analysis, using a constant comparison technique [19]. Constant comparison was used to ensure consistency in coding and analysis across multiple coders, and a large number of transcripts. The authors read the transcripts a minimum of twice each to familiarize themselves with and become immersed in the data. All members of the research team jointly coded the first three transcripts and used regular meetings to arrive at consensus on codes, and to create a common codebook. Subsequently we coded the remaining transcripts individually, with continuous sharing for consistency of codes. We grouped codes to form broader themes. An iterative process was employed, when needed, to re-categorize and revisit themes. FGD and KII data were triangulated with data obtained from IDIs. Themes and subthemes were organized according to the socio -ecological model (ref). This study was approved by the Catholic University of Health and Allied Sciences Research & Ethical Committee (CREC/201/2017), National Institute for Medical Research Lake Zone Institutional Review Board (MR/53/100/493), Mbarara University of Science and Technology Research Ethics Committee (MUREC 1/7), Uganda National Council for Science and Technology (SS 4386), and the University of Calgary Conjoint Health Research Ethics Board (REB17-1741). All informants 18 years of age and above signed a written consent with a witness signing as well. Parents/guardians or husbands (18+) signed consent forms for adolescents less than 18 years of age, and the adolescents signed assent forms. Only those adolescents with fulfilled consent requirements took part in the study. Consent for publication in peer reviewed journals was obtained from all participants for the use of anonymous quotes. All names appearing in this manuscript are pseudonyms, and do not reflect the true identities of the individual participants.