“We have been working overnight without sleeping”: Traditional birth attendants’ practices and perceptions of post-partum care services in rural Tanzania

Background: In many low-income countries, formal post-partum care utilization is much lower than that of skilled delivery and antenatal care. While Traditional Birth Attendants (TBAs) might play a role in post-partum care, research exploring their attitudes and practices during this period is scarce. Therefore, the aim of this study was to explore TBAs’ practices and perceptions in post-partum care in rural Tanzania. Methods: Qualitative in-depth interview data were collected from eight untrained and three trained TBAs. Additionally, five multiparous women who were clients of untrained TBAs were also interviewed. Interviews were conducted in February 2013. Data were digitally recorded and transcribed verbatim. Qualitative content analysis was used to analyze data. Results: Our study found that TBAs take care of women during post-partum with rituals appreciated by women. They report lacking formal post-partum care training, which makes them ill-equipped to detect and handle post-partum complications. Despite their lack of preparation, they try to provide care for some post-partum complications which could put the health of the woman at risk. TBAs perceive that utilization of hospital-based post-partum services among women was only important for the baby and for managing complications which they cannot handle. They are poorly linked with the health system. Conclusions: This study found that the TBAs conducted close follow-ups and some of their practices were appreciated by women. However, the fact that they were trying to manage certain post-partum complications can put women at risk. These findings point out the need to enhance the communication between TBAs and the formal health system and to increase the quality of the TBA services, especially in terms of prompt referral, through provision of training, mentoring, monitoring and supervision of the TBA services.

The study was conducted in Kongwa,one of the five districts in the Dodoma region. This study is part of a larger ongoing health system research project in the Dodoma region aimed at strengthening health systems for accessibility, equity and poverty alleviation in rural areas. The project explores both the supply side of human resources, including governance of the health system, equipment and supplies and the demand side including users and the community perspective on access to services. This study contributes to generation of knowledge on the demand side, through exploring the perspectives of different actors, namely women, their partners and community based health agents such as TBAs. Kongwa was selected because of its rural characteristics, since the focus of the entire project was to explore health systems’ in a rural setting, far from the capital city. Kongwa has a population of 248,656, with 90% living in rural areas. The area is agricultural and people mainly engage in cultivation and livestock keeping. Its residents are also engaged in other activities such as trade and mineral extractions. The area has a poor transport system with poor roads, causing difficulties for women from rural areas in seeking health services including referrals. There are a total of 46 health facilities in the district consisting of one district hospital, four health centers and 41 dispensaries. In the Dodoma region, 97.8% of pregnant women get at least one antenatal check-up, well short of the recommended four visits beneficial to all women [27]. There is no disaggregated data available in this region regarding the number of antenatal care visits that women receive. The available data is on national level indicating the percentages of mainland women utilizing ANC services in the rural areas:3.8% of women receive one visit, while 54.5% receive 2 – 3 visits, and 39.5% receive 4 or more visits. Institutional delivery in the Dodoma region drops to 45.9%, and only 33.8% of mothers receive post-natal check-ups, defined as check-ups of the mother and the child [22]. Kongwa district has engaged in a number of interventions aimed at improving maternal health, such as: 1) training service providers in Emergency Obstetric Care (EmOC), 2) implementing a waiting maternity home (chigonela) at the district hospital, where women who live far away or have complicated pregnancies can stay until delivery, 3) receiving equipment for facilities with help from developing partners such as delivery beds, and other equipment for conducting delivery such as gallipot and kidney dishes,4) motivating TBAs by paying to bring mothers to give birth at the facilities –the latter is currently stopped due to unavailability of funds from the government and 5) implementing community sensitizations on institutional delivery. The district has increased the number of health facilities from 42 in 2011 to 46 currently. The referral system has improved through the provision of mobile phones to all health facilities and securing of five ambulances for all four health centres and the district hospital. For this study, we conducted individual interviews with three different groups of participants. The first group comprised what we labelled as “untrained TBAs”. “Untrained TBAs” were those who are not linked in any way to the health-care system. They have not received any sensitization on health issues and they have not registered with the formal health care system. Eight individual interviews with “untrained TBAs” were conducted; they were approached by the researcher through the community leaders. The second group is comprised of what we called “trained TBAs”. “Trained TBAs” are registered and well known within the health-care system. They have received sensitizations on health issues from the formal health-care sector, have been advised to avoid attending deliveries and, in case they are faced with a delivery that cannot be referred, they have received information on aseptic techniques in order to conduct a clean delivery. Three trained TBAs participated, who were approached through those in charge of the health care facility. Trained TBAs, but not untrained TBAs, have been given incentives to bring mothers to deliver in health care facilities. However, such incentives had stopped due to shortage of funds when this study was conducted. Both the trained and untrained TBAs’ occupation was agriculture, and they were all local residents and of an old age. It was emphasized that the researcher was not representing the government or the Ministry of Health and the findings would not be used against TBAs in Kongwa. The third group comprised five multiparous women who have been clients of untrained TBAs. They were contacted through the untrained TBAs. Ages ranged from 27 to 54 years – mean age 40. Their parity ranged from 2 to 7 children (mean parity 5) and their main occupation was agriculture with a basic education of standard seven. The inclusion criteria for untrained TBAs were those who have been conducting deliveries and were referred by the village leaders. Untrained TBAs were located by village leaders because, since they are knowledgeable of the structure of the village, know the TBAs and they were also easy to be identified and approached by the researcher. For the trained TBAs, the inclusion criteria were those who have been conducting deliveries and were referred by the health-care providers. For untrained TBAs’ clients, the inclusion criteria were the clients who were referred by the interviewed TBAs. Data collection took place in February 2013. Participants were approached by the researcher, the aims of the study were explained and they were asked for their permission to conduct the interview and record it. The location of interviews for untrained TBAs and their clients were at their individual premises and for trained TBAs it was at the village office. The house where village meetings are conducted may be far away from the health services. All participants chose the location comfortable to them. The researcher made sure that no one else apart from the intended participants was there during the interview. Interviews continued until saturation was reached, meaning no new information relevant to answer our research question was emerging [37]. All individual interviews followed an open format, and several aspects were explored, such as practices, perceptions and post-partum care in general. Across the interviews, TBAs and women used different words to refer to the practices that they conduct during the post-partum period and their perceptions of formal post-partum care. Direct translations of these terms will be maintained in the quotations, but elsewhere we will use the term “post-partum care” when referring to a period after child birth of up to 42 days. The interviews were conducted in Kiswahili, which was the medium of communication of the researcher and all of the respondents. Transcriptions in Kiswahili were translated into English and entered into Open Code 3.4 for managing the data and facilitating the analysis [38]. The digitally recorded individual interviews were transcribed verbatim. Coding was conducted with the Kiswahili transcripts in order to be close to the text. Three of the authors are fluent in Swahili and were involved in this process. Codes were then translated into English and for developing categories and themes all authors were involved. Data were analyzed using qualitative content analysis, focusing on aspects related to postpartum care services [39]. After reading the interview transcriptions several times meaning units that referred to postpartum care were identified. From the meaning units – short summarized versions of the sentences – codes were developed. Codes were grouped together to build categories, which reflected the manifest content, i.e. what the interview transcripts conveyed regarding post-partum care (see the example of the coding process in Table 1). During the process of merging codes into categories, similarities and differences between the three different groups of participants were also highlighted. The field notes were used during analysis so as to capture the points that were taken during the interviews. Additionally, notes were used during the coding process to enrich and triangulate the information conveyed in the interviews. Example of the process of analysis from meaning unit to category *The category was later part of the theme “Caring rituals and being close to women”. In evaluating trustworthiness in qualitative research, criteria such as credibility, transferability, dependability and confirmability can be used [39,40]. In understanding the subjects’ reality and their complex situations we applied purposive sampling in our study. Purposive sampling was used, meaning that informants were selected due to their ability to contribute to answering our research questions [39,41]. They were selected because of their experiences of either being TBAs or being post-partum mothers and having received care from the TBAs. Village leaders assisted in selection of the informants after the researcher’s explanation on the objective of the study and the expected source of information. During the initial interviewing process, the researcher screened the informants based on the criteria and objective of the study and expected informants. The interviewing process continued after the researcher assured herself that the respondent met the criteria. We followed an emergent design: after each of the interviews, aspects that emerged and that we thought were relevant were included in the following interviews. Following an emergent design enhances dependability [37]. In order to enhance trustworthiness of the study the research team consisted of people with variety of professional background, including experts in sexual and reproductive health issues, public health and health system research. The detailed description of the study context, selection criteria, data collection and analysis process was complemented by quotes to allow readers to judge the transferability of the findings. Presentations of preliminary findings to the research team allowed critique from an outsiders’ perspective, especially alternative interpretations, which contributes to credibility. Interview guides underwent translation from English to Kiswahili, as the respondents were more comfortable in Kiswahili than in English. The original transcripts were analyzed in the Kiswahili language, which increased the credibility and dependability of our findings. During the data collection process, the researcher kept an open mind and tried to put her pre-understandings into brackets in order to ensure confirmability. Ethical approval to conduct this study was granted by the Senate Research and Publication Committee of Muhimbili University of Health and Allied Sciences (Ref. No. MU/ RP/AEC/Vol.XII/). Furthermore, permission to conduct this study was obtained from the Kongwa district executive director’s office, and lastly from the Kongwa district Medical Officer’s office. Informed consent was sought from each potential participant and names were not written anywhere in order to ensure confidentiality. They were all assured that information given would be treated with strict confidentiality.

N/A

• Dima Health