Community’s experience and perceptions of maternal health services across the continuum of care in Ethiopia: A qualitative study

Background Continuum of care is an effective strategy to ensure that every woman receives a series of maternal health services continuously from early pregnancy to postpartum stages. The community perceptions regarding the use of maternal services across the continuum of care are essential for utilization of care in low-income settings but information in that regard is scanty. This study explored the community perceptions on the continuum of care for maternal health services in Ethiopia. Methods This study employed a phenomenological qualitative research approach. Four focus group discussions involving 26 participants and eight in-depth interviews were conducted with women who recently delivered, community health workers, and community leaders that were purposively selected for the study in West Gojjam zone, Amhara region. All the interviews and discussions were audio-taped; the records were transcribed verbatim. Data were coded and analyzed thematically using ATLAS.ti software. Results We identified three primary themes: practice of maternal health services; factors influencing the decision to use maternal health services; and reasons for discontinuation across the continuum of maternal health services. The study showed that women faced multiple challenges to continuously uptake maternal health services. Late antenatal care booking was the main reasons for discontinuation of maternal health services across the continuum at the antepartum stage. Women’s negative experiences during care including poor quality of care, incompetent and unfriendly health providers, disrespectful care, high opportunity costs, difficulties in getting transportation, and timely referrals at healthcare facilities, particularly at health centers affect utilization of maternal health services across the continuum of care. In addition to the reverberation effect of the intrapartum care factors, the major reasons mentioned for discontinuation at the postpartum stage were lack of awareness about postnatal care and service delivery modality where women are not scheduled for postpartum consultations. Conclusion This study showed that rural mothers still face multiple challenges to utilize maternal health services as recommended by the national guidelines. Negative experiences women encountered in health facilities, community perceptions about postnatal care services as well as challenges related to service access and opportunity costs remained fundamental to be reasons for discontinuation across the continuum pathways. The Ethiopian health system is structured into primary level care, secondary level care, and tertiary level care. The primary care structure found in a woreda (i.e., district) health system comprises a primary hospital, 4–5 health centers, and 20–25 health posts for the delivery of basic curative, preventive and promotive community and outreach services with a seamless continuum. Ethiopia has employed different community engagement approaches for decades through the use of voluntary community health workers with various names and scopes of practice including community health agents, community-based reproductive health workers, community health promoters or volunteers, and traditional birth attendants. Since 2004, the country has expanded community health services through the expansion of the HEP and actively engaging community volunteers to reach most communities and households [32]. In 2011, the community engagement has restructured and introduced the Women Development Army (WDA) strategy to further strengthen the HEP, and participation of individuals, families, and communities. Under the WDA strategy, women are organized and mobilized in groups to share actionable messages and influence each other to adopt and practice healthy behaviors [38]. Women development groups support Health Extension Workers (HEWs) in promoting key messages related to skilled maternal health care through social events such as coffee ceremonies, using peers during marketing, and other community events. They identify pregnant women and birth in their communities and link them to HEWs for early ANC and PNC care. The study was conducted in five rural woredas namely Burie Zuria, Dembecha, Jabi Tehnan, Quarit, and Womberima in West Gojjam Zone of Amhara region. The zone and districts were purposely selected based on the feasibility of establishing sampling frame as this zone has only five The Last Ten Kilometers (L10K) Project woredas while other L10K zones have 10 woredas, and availability of the established relationship with the zone and woredas which allowed practical feasibility to do the data collection. This helped the researchers to minimize difficulties and expenses involved in the planning and conduct of data collection. We utilized a phenomenological qualitative research approach. Based on the Andersen and Newman Framework for health services utilization model [12], we hypothesized that the continuous uptake of maternal care is influenced by the women’s and communities’ perceptions and previous experiences of care, women’s exposure to the health system at any stage of the antepartum, intrapartum, and postpartum stages, and health service and program-related factors. Accordingly, guided by this model, mothers’ and community’s experiences and perspectives regarding maternal health services were captured through in-depth interviews (IDI) and focus group discussions (FGDs) to answer the research questions in detail that is to gain individual and group/community perceptions and experiences regarding maternal health services. Three different groups of participants were selected to gain as much insight and understanding as possible about maternal health care services from the community’s perspectives. It is composed of women 15–49 years of age who have given birth in the last year before the date of data collection, community elders and community and religious leaders (all were males and hereafter referred to as community and religious leaders), and community volunteers (i.e. WDAs). Maximum variation sampling schemes were used to yield a wider perspective from various groups of stakeholders. We subdivided the woredas (i.e. districts) into better performing and low performing strata in terms of maternal health service utilization based on routine administrative data obtained from the zonal health department. Accordingly, two of the study woredas were classified as better-performing, and the other three as low-performing woredas. In each woreda, kebeles, the smallest administrative unit, were selected based on the feasibility of convening FGD participants and availability of HEWs in the Kebele. Lastly, from each kebele, study participants were recruited with the assistance of HEWs based on pre-set selection criteria that include having lived experiences of maternal health services, being recognized as influential and motivator of maternal health service uptake in the community, and being a community volunteer or WDA. Fig 1 presents the detailed sampling frame and respondents’ categories. The principal investigator and research assistants collected and analyzed informational redundancy, data saturation, after conducting three FGDs and six IDIs [39]. Data were collected in October-November 2019. Two research assistants who had experience in qualitative research along with the principal investigator collected the data. Interview and FGD guides were prepared with open-ended questions with probing questions (S1 and S2 Appendices) were used to capture the required information from study participants. The interview and discussion guides were prepared in English and translated to Amharic, the local language. The main topics of discussion included community perceptions about maternal health programs and health providers; the practice and experiences of antepartum care, facility delivery, and postpartum services; and reasons for discontinuation across the continuum of maternal health care. Then, follow-up questions were asked to help to explore their perceptions and experiences of receiving maternal health services in detail. Interviews and group discussions were conducted at a convenient place in the community. Focus group discussions were used to explore information about the social context and discussed the differences among participants. The size of the FGD ranged from 6–8 participants to elicit group-level perceptions by facilitating active interaction. The FGD sessions lasted between 49 and 99 minutes (discussions with WDAs lasted longer which might be due to their lived experience as a mother as well as their engagement in promoting maternal health messages as a community volunteer). Individual interviews were conducted with recently delivered mothers, WDAs, and community and religious leaders to investigate personal perspectives; some private issues, for instance, delivery experience; issues that were raised during FGD that needs further investigation; and to identify opportunities for improving maternal health services. Interview sessions lasted between 21 and 43 minutes. The principal investigator ensured the quality of the data by conducting regular reflective discussions with the research assistants. These discussions were held between the interviews or discussion sessions to discuss key findings, refine the FGD and IDI guides, and identify strategies that continually enhance the line of inquiry following the tradition of emergent design in qualitative research. Throughout the data collection, the study participants were probed to elaborate on or clarify what they have said during the interviews or group discussions to confirm the accuracy of the information captured and the meanings that the participants intended to ascribe to. Throughout the data collection and archive, we ensured confidentiality of the data. Individual personal identifiers were not collected and authors did not have access to information that could identify individual participants during or after data collection. All interviews and discussions were audio-taped with the consent of the study participants and the records were transcribed verbatim by the principal investigator and the research assistants. The principal investigator listened to the audio records and read through the transcripts several times to have an overall sense of the data and to organize the transcripts. The transcript texts were exported to ATLAS.ti software for analysis. A transcript analysis approach was employed which involved familiarization, coding and categorizing data, identifying themes and interpretation stages. Both deductive and inductive coding approach was applied. Guided by the conceptual framework and interview and discussion guides, pre-defined initial codes were developed (open coding) before data collection. Then, each code was further analyzed and disaggregated into categories and sub-themes (deductive axial coding). Iteratively, through reading the data, all data were subsequently classified into one of the codes. Additional codes were added while reading the data, categories, and sub-categories that had not been previously identified (inductive approach). Data were triangulated from responses obtained from IDIs and FGDs to compare them with responses from the different community groups. The categories and the concepts that emerged from interviews and discussions were verified by consistently linking the emerging categories with the data received from the other groups of informants to improve the trustworthiness of the qualitative data analysis. Quotes were used to enhance credibility and substantiate the narrative with participants’ own words. Reports of quotations for selected codes were generated in ATLAS.ti software. Themes and patterns of interrelationships between the themes were identified and reported. Ethical clearance was obtained from the Research and Community Service Office of the University of Gondar (reference number V/P/RCS/05/2505/2019; dated on 25 August 2019). The objectives, methodology, purpose of the study, and the benefits and risks of the study were explained to all study participants. Before data collection, participants were also informed of their right to voluntarily participate in the study. Verbal consent was sought and documented before conducting any interviews and discussions. Because the majority of the respondents were not expected to be able to read or write; written consent was not sought. If the respondent agreed to be interviewed after listening to the consent statement, the interviewer marked the consent form as consent given below the consent statement and signed below that. The interviewer continued with the interview only after receiving and documenting consent. The survey protocol submitted to the ethical review committee included the study questionnaire with the statement that described the consent-obtaining procedure. The primary investigator oriented data collectors on ethical and methodological issues and supervisors followed them throughout the data collection period to monitor any ethical breach. To ensure the privacy of study participants, respondents were interviewed in a conducive environment; and confidentiality of the data were guaranteed by preserving the anonymity of the study participants. Individual personal identifiers were not collected, to ensure the anonymity of data and the researcher would keep the information obtained from the research participant in private or will withhold information from others, to respect the confidentiality and privacy of study participants. Identifying information (names and addresses) was not included in the data collection instrument.