Experiences with and expectations of maternity waiting homes in Luapula Province, Zambia: A mixed-methods, cross-sectional study with women, community groups and stakeholders

Background: Luapula Province has the highest maternal mortality and one of the lowest facility-based births in Zambia. The distance to facilities limits facility-based births for women in rural areas. In 2013, the government incorporated maternity homes into the health system at the community level to increase facility-based births and reduce maternal mortality. To examine the experiences with maternity homes, formative research was undertaken in four districts of Luapula Province to assess women’s and community’s needs, use patterns, collaboration between maternity homes, facilities and communities, and promising practices and models in Central and Lusaka Provinces. Methods: A cross-sectional, mixed-methods design was used. In Luapula Province, qualitative data were collected through 21 focus group discussions with 210 pregnant women, mothers, elderly women, and Safe Motherhood Action Groups (SMAGs) and 79 interviews with health workers, traditional leaders, couples and partner agency staff. Health facility assessment tools, service abstraction forms and registers from 17 facilities supplied quantitative data. Additional qualitative data were collected from 26 SMAGs and 10 health workers in Central and Lusaka Provinces to contextualise findings. Qualitative transcripts were analysed thematically using Atlas-ti. Quantitative data were analysed descriptively using Stata. Results: Women who used maternity homes recognized the advantages of facility-based births. However, women and community groups requested better infrastructure, services, food, security, privacy, and transportation. SMAGs led the construction of maternity homes and advocated the benefits to women and communities in collaboration with health workers, but management responsibilities of the homes remained unassigned to SMAGs or staff. Community norms often influenced women’s decisions to use maternity homes. Successful maternity homes in Central Province also relied on SMAGs for financial support, but the sustainability of these models was not certain. Conclusions: Women and communities in the selected facilities accept and value maternity homes. However, interventions are needed to address women’s needs for better infrastructure, services, food, security, privacy and transportation. Strengthening relationships between the managers of the homes and their communities can serve as the foundation to meet the needs and expectations of pregnant women. Particular attention should be paid to ensuring that maternity homes meet quality standards and remain sustainable. This study focuses on Luapula Province in northern Zambia. Luapula Province is a sparsely populated (19.6 persons per square kilometre), primarily rural province with poor roads and expansive swamps [22]. It has the highest maternal mortality ratio in Zambia at 573 per 100,000 live births, compared to 483 per 100,000 live births nationally [22]. In 2013–2014, 68.4% of live births occurred in health facilities in Luapula Province [3]. The study also draws lessons from successful maternity homes in the Central and Lusaka Provinces. All study districts are similar in terms of population, road access, livelihood, [22] and poverty level [23]. At the time of this study, the health system comprised 68 health facilities, 21 of which had maternity homes in Luapula Province. There were 204 health facilities in Central Province and 294 health facilities in Lusaka Province [24]. In addition, in rural areas community groups, including chiefs or village headmen, Safe Motherhood Action Groups (SMAGs), Neighbourhood Health Committees, and traditional birth attendants promote reproductive, maternal, neonatal and child health (RMNCH), and HIV services [5]. This study employed a mixed–methods, cross-sectional research design [25, 26]. Data were collected between September and December 2013. Three data extraction tools described below were used to collect quantitative data. Qualitative data were collected through focus group discussions and key informant interviews in the four districts in Luapula Province. Additional qualitative data were collected in Serenje and Mkushi Districts in Central Province and Rufunsa District in Lusaka Province to contextualize findings from Luapula Province. Luapula Province was selected to be the focus of this study by the Ministry of Health, Provincial and District Medical Office Managers because it reported the highest maternal mortality ratio in Zambia [22]. In Luapula Province, only these four districts had maternity homes in 2013 [24]. The three districts in Central and Lusaka Provinces were selected because they had well-integrated and functioning maternity homes. Purposeful sampling was used to select the sites and recruit focus group discussion and key informant interview participants in all provinces [26]. In Luapula Province, qualitative data were collected from 17 of 21 facilities with maternity homes in 2013. The four maternity homes not included were incomplete at the time of data collection. In Central and Lusaka Provinces, qualitative data were collected from three CRHCs and a mission hospital that had maternity homes [26]. In the four districts of Luapula Province, 21 focus group discussions were conducted with 210 participants who attended antenatal care, postnatal care, family planning and children’s clinics in CRHCs and mission hospitals and members of the SMAGs. The participants were maternity home users, non-maternity home users, women who were pregnant for the first time, women who had delivered at home in 2012, elderly women and SMAG/Neighbourhood Health Committees members (Table 1). Each focus group discussion included eight to 12 participants. Interviews were conducted with 21 health facility in-charges, 10 couples from antenatal care clinics, 17 chiefs (four chiefs were unable to participate due to unforeseen circumstances), 12 village headmen, four District Community Health Officers, two District Community Nursing Officers and three staff members from partner agencies (Table 2). Number of focus group discussions and participants in Luapula, Central and Lusaka Provinces Number of key informant interview participants in Luapula, Central and Lusaka Provinces In the Central Province districts, four focus group discussions with six to 12 participants were conducted with 26 SMAG members responsible for mobilizing communities and organizing activities to support maternity waiting homes [26]. Rufunsa District had no SMAGs. In the three districts of Central and Lusaka Provinces, interviews were conducted with three District Community Medical officers, two District Nursing Officers and five health facility in-charges who supervised health staff and had developed systems to work with community groups on maternity homes [26]. Field guides were used by interviewers to guide focus group discussions and key informant interviews (Additional file 1). Focus group discussions focused on maternity home use, food availability, customs and traditions related to facility-based deliveries, maternity homes’ cost and length of stay, willingness to pay, transportation services for pregnant women, and general impressions of maternity homes. Key informant interviews investigated support for maternity waiting homes, mechanisms to sustain maternity homes and their operations. Focus group discussions and interviews lasted approximately 30 to 45, and 20 to 30 min, respectively. In Luapula Province, quantitative data was collected using three assessment tools. The Maternity Home Assessment Tool containing 42 items was used to collect data through direct observation and clinic staff interviews about the structures and amenities available in maternity homes (Additional file 2). In particular, this tool was created by the research team to collect data on maternity home ownership, funding, and building materials, as well as availability of water and electricity, rooms, beds, mattresses and cooking amenities. A Service Abstraction Form containing 23 items was used to extract annual deliveries from maternity registers accessed through facility in-charges (Additional file 3). A separate assessment tool created by Integrated Rural Development Initiative and Jhpiego was used to collect water sources and sanitation data. Nine trained research assistants collected all data in Mansa, Chembe, Samfya and Lunga districts. They received training on research ethics, the study’s protocol and data collection tools, and empirical content on collection and quality of qualitative and quantitative data through didactic and hands-on sessions. The first author supervised this team and collected data in Serenje, Mkushi and Rufunsa districts. Focus group discussions and key informant interviews were conducted in the local language, Bemba, until saturation was reached [27]. The data collection tools were piloted with 20 midwives at Levy Mwanawasa General Hospital and Bauleni, Chilenje and Kabwata Health Centres in Lusaka Province and revised before the study started. Qualitative data were transcribed from audio recordings in Bemba, translated into English, and back-translated into Bemba to ensure accuracy of the translation. The data were coded in Atlas-ti using codes derived from the field guide questions and emergent themes [28]. Working matrices were used to organize passages and themes by participant type and analysed each district’s responses in a framework [27]. The data coders, three based at Jhpiego in Baltimore and three based in Lusaka, refined the themes, found commonalities, wrote up findings, and returned iteratively to the raw data to find relationships between themes. Quantitative variables were entered into a Microsoft Access database and analysed with descriptive statistics by district using Stata [29]. Results were shared with stakeholders at a dissemination meeting in Lusaka. The study was approved by the Johns Hopkins University Institutional Review Board in Baltimore and the University of Zambia Research Ethics Committee. All focus group and interview participants were consented verbally prior to their participation in the study. A waiver of written consent was obtained as the study posed no more than minimal risk of harm.