Reasons for the utilization of the services of traditional birth attendants during childbirth: A qualitative study in Northern Ghana

Background: Skilled delivery reduces maternal and neonatal mortality. Ghana has put in place measures to reduce geographical and financial access to skilled delivery. Despite this, about 30% of deliveries still occur either at home or are conducted by traditional birth attendants. We, therefore, conducted this study to explore the reasons for the utilization of the services of traditional birth attendants despite the availability of health facilities. Method: Using a phenomenology study design, we selected 31 women who delivered at facilities of four traditional birth attendants in the Northern region of Ghana. Purposive sampling was used to recruit only women who were resident at a place with a health facility for an in-depth interview. The interviews were recorded and transcribed into Microsoft word document. The transcripts were imported into NVivo 12 for thematic analyses. Results: The study found that quality of care was the main driver for traditional birth attendant delivery services. Poor attitude of midwives, maltreatment, and fear of caesarean section were barriers to skilled delivery. Community norms dictate that womanhood is linked to vaginal delivery and women who deliver through caesarean section do not receive the same level of respect. Traditional birth attendants were believed to be more experienced and understand the psychosocial needs of women during childbirth, unlike younger midwives. Furthermore, the inability of women to procure all items required for delivery at biomedical facilities emerged as push factors for traditional birth attendant delivery services. Preference for squatting position during childbirth and social support provided to mothers by traditional birth attendants are also an essential consideration for the use of their services. Conclusion: The study concludes that health managers should go beyond reducing financial and geographical access to improving quality of care and the birth experience of women. These are necessary to complement the efforts at increasing the availability of health facilities and free delivery services. This study adopted the phenomenology approach to qualitative enquiry.27 In phenomenological research, it is the participants’ perceptions, feelings, and lived experiences that are paramount and that are the object of study.28 This design was, therefore, deemed appropriate as the study aimed at documenting the lived experiences of women who delivered at TBA facility and the reason for their choice of facility. We adopted the social–ecological model. This model considers the complex interplay between individual, relationship, community, and societal factors in affecting the phenomenon of interest.29 The structures at each of the constructs in the model overlap and illustrate how factors at one level influence factors at another level. The individual constructs in the content of this study refer to the personal-level factors such as age, education, and income that influence individual health-seeking behaviour.30 The relationship which is the second level examines close relationships that may influence the likelihood of using TBAs for delivery. An individual’s closest social circle peers, partners, and family members influences their behaviour and contribute to their range of experience. The relationship factors also include previous experience with biomedical facilities or TBAs during childbirth. The third level (community) explores the settings, such as health facilities and neighbourhoods, in which social relationships occur and seeks to identify the characteristics of these settings that are associated with health-seeking behaviour31 during labour. The fourth and final level (societal) looks at the broad societal factors that help create a climate that drives people towards using the services of TBAs (Figure 1). Social–ecological model showing reasons for the utilization of the services of TBAs. The study was conducted in Tolon District and Yendi Municipality in the Northern region of Ghana. The Tolon district is divided into three sub-districts for the delivery of healthcare. There are health centres in each of the sub-districts and CHPS compounds in communities in the district. Access to health facilities has been reported to be higher than the regional average.32 The Yendi municipality has a government hospital located in Yendi and four health centres located at Yendi, Bunbonayili, Ngani, and Adibo. The municipality also has four7 CHPS compounds at Sunson, Kuni, Kamshegu, Oseido, Montondo, Yimahegu, and Kpasanado. There is also a clinic at Malzeri and a private clinic at the Church of Christ premises in Yendi.33 The selection of the district was based on information gathered from the Regional Health Directorate and available literature. These two districts are noted for a high number of TBA deliveries in the region despite the availability of health facilities. In Tolon, it has been reported that each community has more than two TBAs.34 Yendi was also selected because of the number of TBAs in the district and had served as a district for the training of TBAs in the region in the past. As a result, the district has more than 30 TBAs across various communities.35 The study population were women who had live birth in the TBA facilities in one selected district and a municipality. Health workers in the selected districts were used to identify the TBAs who have high attendance based on their district report. Their facilities were visited and women who had childbirth with TBA pending their discharge were recruited for the study. An initial screening sheet was used to select eligible women. To be eligible, the person should be residing in a community with a health facility and should be gainfully employed with a monthly income of more than the minimum wage of GHȻ310.00 ($53.13). Financial access and travel distance have been reported as known barriers to skilled delivery. This strategy was, therefore, employed to exclude people who had to use TBA services because of non-availability of a health facility. Higher costs associated with seeking supervised maternity services have been noted as very critical to the uptake of care for many women in Ghana and other developing nations.5,36 Free maternal health service was introduced in Ghana as a pro-poor strategy to reduce the financial barrier to healthcare during pregnancy and childbirth. The free maternal health policy was implemented in Ghana in July 2008 under the National Health Insurance Scheme (NHIS). The policy allows all pregnant women to have free registration with the NHIS after which they would be entitled to free services throughout pregnancy, childbirth, and three months postpartum. An in-depth interview (IDI) topic guide was used for the data collection. The topic guide was designed in English and translated into the local language (Dagbanli). The topic guide was designed according to the four constructs (individual, relationship, community, societal) in the social–ecological framework. For example, we asked questions about the reasons for utilizing the services of the TBA, the various actors in the decision-making process, and their views about the type of services provided by the TBA. For those who indicated they had previous experience with health facility delivery, we asked them to compare the services at the health facility with those rendered by the TBAs (see Supplementary file 1). These topic guides were pretested with five women who utilized the services of TBA in another suburb of Tamale. All the interviews were conducted by trained research assistants with previous experience in conducting qualitative interviews. We collected socio-demographic data such as age, education level, religion, and reproductive history at the end of the interview. The interviews for each day were transcribed before proceeding to conduct more interviews. The daily review and coding were useful in determining the point of saturation.37 Eligible women who refused to participate in the study were replaced. Four women who were eligible and recruited refused to participate for personal reasons. Interviews were conducted between March 2019 and June 2019. Interview sections lasted for 30–40 min. Reflexivity relates to the degree of influence that the researcher brings to bear on the research either intentionally or unintentionally.38 Reflexivity enhances the quality of research and also boosts understanding of how the researcher’s own interest could affect the research process.39 Bracketing, on the contrary, refers to an investigator’s identification of vested interests, personal experience, cultural factors, and assumptions that could influence how he or she views the study’s data.40 In qualitative research adopting phenomenology, it is important for researchers to disclose their personal biases and measures that were put in place to improve rigour, trustworthiness, and credibility of the research findings. Several strategies were adopted in reflexivity and bracketing. First, the research team did not have preconceived ideas and interest regarding the outcome of the study findings. The study was mainly informed by available literature that clearly shows that some women still have unskilled delivery despite the expansion of health facilities and introduction of free maternal healthcare policy. Furthermore, the research team and the research assistants were very open to study participants during data elicitation. There was no social or biological relationship between study participants and researchers. Although all the researchers have clinical training and practice, at the point of data collection and the research, none of the researchers was involved in clinical care. In addition, research assistants with experience in conducting qualitative interviews were recruited and trained by the lead investigator. They were informed on the need to have a neutral mind and behaviour towards study participants during interviews or data collection. They were further told that any biased behaviour, preconceived beliefs, or values could affect the data that would be collected and that could further have a negative effect on the outcome of the study findings. Recorded interviews were replayed to participants to make inputs and corrections. The interviews were transcribed verbatim. After the data collection and analysis, the findings were shared with some of the participants through a dissemination workshop. This enabled the participant to review and agree with the findings of the study as a form of member checking.41 In addition, a codebook was developed, reviewed, and accepted by the research team. Double coding of the data was done and compared. The coding trail was reviewed by an independent person for verification. Using the NVivo software, a coding comparison query showed a high level of agreement with a Kappa score of 0.92.42,43 All IDIs were recorded during the interview. The interviews were played to the interviewee after the interview for them to make the necessary corrections and addition. The recordings were transcribed verbatim. The transcripts were reviewed by an independent person who listened to the recordings and compared the content with the transcriptions. Daily interviews were shared with other authors to review and provide feedback on the process. This iterative approach strengthened the data elicitation process. Interviews continued until data saturation was achieved.44 Hybrid inductive and deductive framework45 were used in developing the codebook, coding of the transcripts, and developing the themes. Conceptual dimensions of the interview guides guided the preliminary development of the codebook. This was then revised to include the emerging themes from the data. This codebook was discussed and accepted by all authors. The transcripts were imported into QSR NVivo 12 for textual analysis. We used the case classification function in NVivo to identify each respondent and their attributes (socio-demographic and reproductive history). We first read through selected transcripts in NVivo and created nodes from the emerging issues in the data. Both free and free nodes were created during the coding until all the transcripts were coded. During coding, memos were written to key reflection from the data. The memos were linked to both the data sources and the nodes. Coded sections were regrouped into relevant categories and themes for presenting the results. Direct quotations were used, where appropriate, to support the themes. The main themes that depict reasons for patronizing the services of TBAs could basically be divided into biomedical health facility push factors and TBAs pull factors. These factors, which emerged from data, could be put into six sub-themes; good interpersonal relationship and practices by TBAs, post-delivery baby care and provision of special food by TBAs, requirements for labour in biomedical health facilities, preference for vaginal delivery and fear of caesarean section (C/S), perception about poor services in biomedical facilities and inexperienced midwives, and poor attitude of health workers during antenatal care (ANC) and facility delivery. The protocol for the study was reviewed and approved by the Ethics Review Committee of Ghana Health Service (GHS-ERC 18/02/2019). All participants signed an informed consent form before participation.