Involving traditional birth attendants in emergency obstetric care in Tanzania: Policy implications of a study of their knowledge and practices in Kigoma Rural District

Introduction. Access to quality maternal health services mainly depends on existing policies, regulations, skills, knowledge, perceptions, and economic power and motivation of service givers and target users. Critics question policy recommending involvement of traditional birth attendants (TBAs) in emergency obstetric care (EmoC) services in developing countries. Objectives. This paper reports about knowledge and practices of TBAs on EmoC in Kigoma Rural District, Tanzania and discusses policy implications on involving TBAs in maternal health services. Methods. 157 TBAs were identified from several villages in 2005, interviewed and observed on their knowledge and practice in relation to EmoC. Quantitative and qualitative techniques were used for data collection and analysis depending on the nature of the information required. Findings. Among all 157 TBAs approached, 57.3% were aged 50+ years while 50% had no formal education. Assisting mothers to deliver without taking their full pregnancy history was confessed by 11% of all respondents. Having been attending pregnant women with complications was experienced by 71.2% of all respondents. Only 58% expressed adequate knowledge on symptoms and signs of pregnancy complications. Lack of knowledge on possible risk of HIV infections while assisting childbirth without taking protective gears was claimed by 5.7% of the respondents. Sharing the same pair of gloves between successful deliveries was reported to be a common practice by 21.1% of the respondents. Use of unsafe delivery materials including local herbs and pieces of cloth for protecting themselves against HIV infections was reported as being commonly practiced among 27.6% of the respondents. Vaginal examination before and during delivery was done by only a few respondents. Conclusion: TBAs in Tanzania are still consulted by people living in underserved areas. Unfortunately, TBAs’ inadequate knowledge on EmOC issues seems to have contributed to the rising concerns about their competence to deliver the recommended maternal services. Thus, the authorities seeming to recognize and promote TBAs should provide support to TBAs in relation to necessary training and giving them essential working facilities, routine supportive supervision and rewarding those seeming to comply with the standard guidelines for delivering EmoC services. © 2013 Vyagusa et al.; licensee BioMed Central Ltd.

The study adopted a cross-sectional survey design. Being descriptive in nature, it involved the gathering of qualitative data by interviewing TBAs at the places which they preferred to be interviewed. This was supplemented with observations of TBAs’ practices. Quantitative data were also gathered as part of interviews done with TBAs. As mentioned above, the main study theme was about TBAs’ knowledge about EmoC aspects and their practices in delivering the desired services to the women in need. Other aspects investigated are as described below under the data collection section. The study covered 6 divisions, namely Kalinzi, Mahembe, Mwandiga, Nguruka, Buhingu and Ilagala, and all of them are located in Kigoma Rural District. This district is found between latitudes 4–6 degrees South and longitudes 29–30 degrees East. The district was selected based on the report indicating the district to have recorded a MMR of 757 deaths per 100,000 live births, the rate that was higher than the rates reported from elsewhere in Kigoma Region [49]. The latter Region is situated on the western border of Tanzania and has four administrative districts, namely Kigoma Urban (Kigoma Ujiji), Kibondo, Kigoma Rural (R) and Kasulu. The main portion of Kigoma (R) district lies along the shore of Lake Tanganyika whereby about 50% of her population lives. The Kigoma (R) district was found to be occupied with poor roads and other communication network systems including telephone infrastructure. According to the reports from the health authorities of this district and the experience of the first author in the present paper, communication was a challenge when it came to getting immediate information related to patients needing referral services. The majority of its inhabitants belonged to the ‘Ha’ tribe and Ha is the most popular indigenous language. Most of the residents were found being small-scale farmers, growing such cash crops as coffee and palm while potatoes, cassava, beans, maize and bananas were being grown mainly for food. Until 2005 when this study was being undertaken, there was neither any government (public) hospital nor any private hospital and therefore all referral and advanced services were being sought at the regional hospital known as Mawenzi while a few cases were referred to the private Baptist Hospital. Both of the latter two hospitals are situated in Kigoma Urban district. In terms of total number of health facilities and their types, the Kigoma (R) district possessed 5 public health centres (only two of these belonged to voluntary agencies while the rest were publicly (government) owned)). Until 23rd March 2013, records showed that the district had over 63 dispensaries among which 2 belong to parastatal organizations, 2 owned by Faith Based Organizations (Voluntary Agencies) and 2 belonging to private-for-profit entities (District Council Health Authorities, Kigoma Rural, per comm.). As some of these facilities were not easily accessible most of the deliveries were home-based. Originally, the plan was to recruit 160 trained TBAs among those who were registered by the District Council Health Authority. But, only 157 could be reached until end of the study. This depended on several factors especially the presence of the individuals targeted. Those who eventually were mobilized to participate have been selected using a convenience sampling approach. There was no baseline statistics on the total number of TBAs who were registered and those who were not registered so as to be used for calculating in advance of the study the sample representing different localities and gender mix for reasonable representation. The TBAs who were at last reached were found in all 6 divisions of the district and were identified through Consultation with local community leaders and community health workers. As highlighted before, qualitative and quantitative data collection techniques were adopted. This involved the gathering of the data using the data collectors after being oriented on the proper procedures. The data collectors were familiarized with both the use of research instruments, seeking informed consent from the study participants and carrying out field observations. The consent sought was either supported by each participant’s signature or thumb print as further elaborated later. Among these collectors were members of district health council management team (CHMT) and a few health facility based staff who were routinely working on maternal and child (MCH) services at the ANC and child health clinics. The staff concerned includes health officers, clinical officers and people with nursing profession. These were included in order to enable them see themselves the working environment of TBA in their delivery places, knowledge of TBAs on EmOC aspects, and their suggestions in relation to how the prevailing MCH service problems could be addressed if TBAs were to continue being involved effectively. This was believed to be an opportunity for equipping the staff concerned with the practical hands on answers about the working ability and environment of TBAs, thus enhancing their chance to recognize how better the experience gained from the field could be utilized for better actions. The individuals concerned were also expected to give the appropriate advice to any authorities that were much concerned with issues relating to TBAs’ involvement in maternal health services. This includes consideration on all possible or better practical ways of working together with TBAs in the respective communities. Careful measures were taken to ensure that the health service personnel involved in the data gathering process were those who were not popular or known. in the respective study communities. This was aimed at minimizing if not to avoiding obvious study biases or some respondents refraining from providing some key information. However, in a way, a few of the staff were found to be known in the study settings. Data were collected between April and May in 2005. In general, the information collected using different data collection tools (questionnaires, observational checklist, etc.) was based on questions intended to establish the respondents’ knowledge about EmoC issues and their practices in delivering the services to the needy pregnant women. A questionnaire used was translated in Kiswahili for easier use during the interviews with the identified TBAs. It comprised of a mixture of closed-ended and open ended questions. The former type of the questions was aimed at obtaining quantitative data that were supplemented with the data gathered using open-ended questions. The latter questions were mainly aimed at coming up with qualitative data that would supplement or help to explain some statistical data. The specific issues upon which the study questions were based addressed the respondents’ knowledge about the signs and symptoms of pregnant women who were at risk of facing maternal complications and risk of contracting HIV/AIDS infections. The respondents were also examined about their knowledge on the modes of HIV/AIDS transmission and prevention, the referral procedures, obstetric handling procedures for the pregnant women during labour as well as during and after childbirth (e.g. by observing aseptic techniques). Other issues investigated related to measures taken by TBAs after encountering a woman with haemorrhage and/or those who were facing other problems before, during and after childbirth. The interview schedules were conducted in one of the rooms, and this was either at the nearby study health facilities or in the hamlet leaders’ houses in the visited study villages. It was in such villages from where TBAs recruited for study were also selected. The places identified for interview have been selected after consultation with the TBAs themselves who felt that they could be more conveniently investigated while being there than and elsewhere. Their choice was respected instead of the previous plan of conducting the interviews and observations in the participants’ own premises. It was discovered that for various reasons the TBAs could not be comfortable if the study team saw the places where they were delivering their services. Moreover, field observations were conducted on TBAs’ delivery kits, interest being on the items contained in the kits, kit completeness and cleanness. Other planned issues for observation include TBAs’ service delivery environments, particularly state of the buildings and general surroundings in terms of sanitation and hygiene. Moreover, part of the observation include a review of TBAs’ note-books and the interest behind was to obtain the records showing whether or not the referrals were being made in correct way. To facilitate both the interviews and observations, TBAs were asked in advance to come with their delivery kits on the day they participated in the study. Each day after fieldwork, the research team was meeting to review and discuss the data collected by checking for their completeness and inconsistencies, making necessary corrections and compiling the data. Where applicable, the responses to open ended questions appearing on the questionnaire were coded to allow their frequency tabulation. The coded data were double entered in the database prepared using EPI-Info program. Analysis of these structured data was performed using SPSS software package. The one-way frequency tabulations were followed by chi-square (χ2) test technique so as to assess if there were any statistically significant difference in the observations made with respect to on the computed variable proportions. The difference observed was considered to be statistically significant at a P ≤ 0.05. To understand the level of knowledge among the respondents on signs, symptoms and danger signs of pregnant women at risk, a scale was used whereby respondents who mentioned at least three conditions were counted ‘highly or adequately knowledgeable’. Those who mentioned only one-to-two conditions were counted ‘moderately knowledgeable’ while those who could not mention any condition were categorized ‘poorly knowledgeable’. The analysis was thus limited in that no logistic regression analysis was performed. This is because it was found that the analysis done was still sufficient to help the study objectives to be achieved given the nature of the data collected. As shown under the results section, the denominators used in the calculation of the frequencies presented in the results section varied from one question to another and this was due to how the interviewees responded to specific questions. The unmentioned answer options or missing variables were excluded in the analysis. Official ethical clearance for the study was obtained from the Tumaini University through the Kilimanjaro Christian Medical Centre (KCMC) Ethics Committee. Permission to conduct the study in the district was also sought from Regional and District Medical and Administrative Offices. The study participants were informed about the study with assistance by local leaders. The individuals approached were also given the right explanations about the objectives and expected benefits of the study. They were also informed about the chance that anyone would be allowed to participate willingly and voluntarily. Those who were unable to read (and or write) were asked to put a thumb print on the informed consent form. The rest signed the consent form voluntarily. Participants were assured of the anonymity of their names and confidentiality of the information they wished to be treated so. All of the individual participants were informed of their freedom to drop out of the study any time they wished even after participating.

N/A

• vyagusa, dismas b.
• mubyazi, godfrey m.
• masatu, melkiory c.