Background: Tanzania’s One Plan II health sector program aims to increase facility deliveries from 50 to 80% from 2015 to 2020. Success is uneven among certain Maasai pastoralist women in Northern Tanzania who robustly prefer home births to facility births even after completing 4+ ANC visits. Ebiotishu Oondomonok Ongera (EbOO) is a program in Nainokanoka ward to promote facility births through a care-group model using trained traditional birth attendants (TBAs) as facilitators. Results to date are promising but show a consistent gap between women completing ANC and those going to a facility for delivery. A qualitative study was conducted to understand psychosocial preferences, agency for decision-making, and access barriers that influence where a woman in the ward will deliver. Methods: In-depth interviews, focus group discussions and key-informant interviews were conducted with 24 pregnant and/or parous women, 24 TBAs, 3 nurse midwives at 3 health facilities, and 24 married men, living in Nainokanoka ward. Interviews and discussions were transcribed, translated, and analyzed thematically using a grounded theory approach. Results: Most women interviewed expressed preference for a home birth with a TBA and even those who expressed agency and preference for a facility birth usually had their last delivery at home attributed to unexpected labor. TBAs are engaged by husbands and play a significant influential role in deciding place of delivery. TBAs report support for facility deliveries but in practice use them as a last resort, and a significant trust gap was documented based on a bad experience at a facility where women in labor were turned away. Conclusions: EbOO project data and study results show a slow but steady change in norms around delivery preference in Nainokanoka ward. Gaps between expressed intention and practice, especially around ‘unexpected labor’ present opportunities to accelerate this process by promoting birth plans and perhaps constructing a maternity waiting house in the ward. Rebuilding trust between facility midwives, TBAs, and the community on the availability of health facility services, and increased sensitivity to women’s cultural preferences, could also close the gap between the number of women who are currently using facilities for ANC and those returning for delivery.
This was a qualitative study conducted in Nainokanoka ward in Ngorongoro district, Tanzania, the site of NDI’s EbOO project. Data collection was in the form of in-depth interviews (IDIs), focus group discussions (FGDs) and key-informant interviews, with individuals and groups identified as relevant stakeholders in decisions about where a woman will deliver in the ward. This included pregnant and/or parous women, TBAs, husbands and male elders, and ward health facility nurse midwives. The study employed a ‘grounded theory’ approach to analyze interview and focus group data that was collected. Grounded theory is an inductive analytical framework which begins with data collection. Although research questions identified in the study objectives provide a deductive framework that shape the creation of survey instruments, it is inductive analysis of IDIs and FGDs that identify emerging patterns in the data which are developed into a theory. The emergent theory can then suggest directions for further inquiry to strengthen and confirm it. Nainokanoka ward is one of 11 wards in the Ngorongoro Conservation Area (NCA). A ward is a third-tier administrative zone in Tanzania, with a population between 8000 and 15,000. Wards are subsumed into districts, which in turn are subsumed into regions. Nainokanoka ward is 1109 km2 and populated by Maasai pastoralists who are permitted to live in the NCA with certain restrictions on farming and land use [24]. The 2017 District Population and Livestock Census counted 14,166 people living in the ward (7419 females, 6747 males), approximately 50% of whom are under the age of 14 [24]. Under-five mortality in rural Ngorongoro district where the ward is located counted 39 deaths per 1000 live births [7], and MMR was estimated at 585 maternal deaths per 100,000 live births in the 2012 HPC [4]. Based on the crude birth rate for the region (43 live births per 1000 population [25]), there are an estimated 50 births per month in the ward. Nainokanoka ward is served by two dispensaries and one second tier health center located in 3 sub-villages: Irkeepusi, Nainokanoka, and Bulati (Fig. 2). The three sub-villages are approximately 15 km apart along the only road in the ward. All three facilities offer ANC and skilled delivery services with basic emergency obstetric care. Nainokanoka sub-village, in the center of the ward, is the location of the 2nd tier health center which does offer referral service (with ambulance) to a tertiary care center 80 km away to the South, in the town of Karatu — the nearest location for comprehensive emergency obstetric care. According to the ward medical officer, there are 4 clinical officers (at least 1 at each facility), 5 nurses, 2 medical attendants, 1 pharmacist and 1 lab technician who work in the 3 health facilities in the ward. There is no maternity waiting house at any of the health facilities. Nainokanoka Ward Health Facilities The decision criteria for sample size in a qualitative study is not standardized. Generally the strongest empirical justification for a qualitative sample is achieving a ‘saturation’ or redundancy of themes [26]. The assumption underlying saturation in this study is that all themes relevant to decisions and preferences around place of delivery will be presented through the 24 IDIs, 5 FGDs and 3 key-informant interviews, by 75 selected participants. Data collection, in the form of 24 IDIs, 5 FGDs and 3 key-informant interviews involving 75 community members were completed between December 10, 2018 and January 31, 2019. The study population was divided into four sub-groups identified as relevant stakeholders in delivery decisions in the ward: A purposive sampling technique was employed for all populations to best represent an equal cross-section of three sub-villages in the ward. Eight participants in each sub-population were chosen from each of the three sub-villages in the ward (Irkeepusi, Bulati, Nainokanoka). For women participating in IDIs there were also criteria for representation based on age, parity, and household distance from a delivery facility. Several multiparous women who had had both facility and traditional delivery were also recruited to provide insight into their experience with each. Purposive sampling was also used in FGDs to include TBAs from all sub-villages as well as several who had experience attending both facility and home deliveries. Married men with at least one child who were selected to participate in FGDs were selected proportionally to represent each sub-village. IDI and FGD questions were developed in collaboration with the project team who are themselves Maasai, living in the community, and familiar with the context. IDIs and FGDs with women were facilitated by an experienced female Maasai researcher (BL) who conducted them in the Maa language. Men’s FGDs were facilitated by a male Maasai researcher (KS) in the Maa language as well. All IDIs and focus groups were recorded then transcribed into Maa and translated into Kiswahili and English for coding and analysis. Translations were reviewed by two Maa speaking researchers for accuracy. Key-informant interview questions were developed by study authors based on initial analysis of data from IDIs and FGDs. Key-informant interviews were conducted in Swahili by a Swahili speaking researcher (BL), and transcribed and translated into English. Please refer to DECLARATIONS for details on Ethical Approval and Consent to Participate. Transcribed and translated data from de-identified key informant discussions, FGDs and IDIs, were analyzed and coded independently by two researchers—one, a Maasai from the community (KS), and one American (PM). Initial coding was done manually to identify major themes. Once major themes and subthemes were identified and compared, a common codebook of themes was created based on consensus of the researchers, and agreement from key EbOO project informants. The data was then recoded using NVIVO 12 analytical software for further analysis in preparation for use in publication. From the coded data, a theory was constructed, based on synthesis and deductive analysis by the two principle investigators. Results of the study were presented to key informants in the EbOO project as well as health professionals in the ward and professional colleagues involved in the project for feedback, as part of the project action-learning cycle. Feedback was used to further refine the emergent theory before final preparation of results for distribution and publication.
N/A
DIMA AI Care
