Current PMTCT programmes do not reach many of the women who need them due to socio-cultural, economic, systemic and programmatic factors [38–40]. Home deliveries by traditional birth attendants are common in rural Zimbabwe and are increasing . Furthermore; PMTCT programmes require complementary approaches to prevent missed opportunities in this evolving context. This report is among the first attempts to evaluate the feasibility and acceptability of the participation of TBAs in PMTCT programmes .
TBAs in this rural context are elderly, married or widowed and have a minimum level of education. These socio-demographic characteristics are similar to those of TBAs in other settings . Untrained TBAs were younger, had less experience than trained TBAs and learned to attend births by themselves or by helping another TBA. The decrease in training programmes for TBAs in Zimbabwe could explain for the younger age amongst untrained TBAs.
The integration between community health workers such as TBAs and the formal health services can bring valuable benefits to community-based public health interventions and open ways for a number of activities related to prevention and care as has been shown in other reports [42, 43]. Complementary approaches, in which community-based interventions are paired with the strengthening and/or expansion of services at the health facility level, also have the potential to address a variety of other health challenges, such as uptake of HIV testing and compliance to PMTCT regimens [33, 44].
Recent studies have alluded to problems associated with home deliveries with reference to uptake of PMTCT interventions [45, 46]. The present study has demonstrated that beyond TBAs' current activities which include assistance during delivery and in the postpartum stage, they are willing to expand their scope of work in mother-and-child activities to include PMTCT with some limitations identified such as accompanying the child to the health centre for medication and assisting the health centre in the documentation of ANC services. This data suggests the need to strengthen the health care network system between the formal health services and the communities including TBAs. For this, integrating the services of TBAs into the mainstream health care delivery system is required. The existing health care system will have to build-up a partnership with TBAs who operate in the informal sector and help develop communication skills in the referral process. Furthermore, health authorities and health staff need to recognise the cultural and practical contribution of TBAs to the health system.
The success of community-based interventions that emphasize the participation of TBAs depends on the role and status of TBAs in a given community. In India, a 62% reduction in neonatal mortality was achieved through a community-based approach that included training of TBAs and local women to treat sick newborns at home . Furthermore, a trial in Pakistan reported substantial benefits in the reduction of perinatal mortality and maternal mortality by training and integrating TBAs in the health system .
Our results reveal that TBAs in this setting have limited knowledge on HIV/AIDS issues in general and PMTCT in particular. This can be improved through the training of these women who are already available and working in the community. This training will need to be tailored to the tasks that they are expected to perform, the knowledge and skills required and adapting the training curricula to the corresponding education level. A recent meta-analysis of 60 studies indicated that training TBAs was associated with significant improvements in their level of knowledge and the quality of advice they provided on family and child health .
The present study adds weight to the need of reinforcing TBA's knowledge on MTCT prevention measures before they can contribute to the provision of PMTCT services. At the moment, in this rural setting, TBAs' advice to women on HIV/AIDS issues (including PMTCT) is not frequent. At national level, there is a need to up-date the national TBA training manual by including basic concepts of HIV prevention . In Tanzania, it has been shown that TBAs, if given additional skills and motivated, can be used effectively in program implementation and contribute to reaching women who deliver outside health facilities with PMTCT interventions (from counselling to providing single-dose nevirapine [sdNVP]) . After careful selection, training and sustained and regular supervision, TBAs played an important role in supporting and referring pregnant mothers for facility-based PMTCT services in Uganda .
Performing a blood test was one of the activities TBAs refused to practice in the study. The application of this intervention by community workers such as TBAs is directly related to the national policy of each country and if ever promoted will need to include continuous supervision and follow-up. The first reported PMTCT programme to use TBAs to provide confidential HIV counselling and testing utilizing an oral fluid rapid test has been in Cameroon . TBAs also dispense NVP to HIV positive women and ensure that the newborn receives postpartum NVP prophylaxis as an integrated strategy in the PMTCT programme. This approach has been done through community participation, training and complemented with supervisory nurses who visit villages at monthly basis.
Involving TBAs in PMTCT interventions is supported by various individual reports that have shown improvement of TBAs' effectiveness in the provision of public health interventions [53, 54]. A recent systematic review showed that TBA training appears to increase antenatal care attendance rates by 38% .
In these two rural districts of Zimbabwe, the success of MTCT prevention programmes could be improved by strengthening community-based programmes, and include TBAs who could: establish links between the community and health services and, provide health education to encourage increased use of ANC services, hence accessibility to PMTCT; sensitize communities aiming at a family-centred PMTCT approach [24, 56] including information and communication of basic concepts on PMTCT and the importance of HIV testing for the pregnant women and their partners; provide community-based counselling and testing services for HIV [33, 57]; support of and adherence to follow-up of PMTCT regimens [45, 50]; provide mother-infant adequate referral [58, 59]; counsel women to bring newborns to a health centre for the infant dose of NVP; supervision of home administration of infant dose NVP through single-dose blister packs for newborns [60, 61]; support women to adopt safer infant feeding practices and promote and support of family planning and; provide care and support to affected families. The dispensing of medications to women in PMTCT programmes as has been reported elsewhere will strongly depend on the national policy of each country and the availability of an efficient training and supervision system [32, 51].
Even though doubts have been stated on the possible significance of health improvements that would be gained by attributing HIV prevention and care tasks to TBAs during home deliveries in developing countries [62, 63], our data allow the assumption that using these primary health workers for community health interventions including PMTCT may have a potential improvement of perinatal and maternal health indicators.
Another important finding was a positive association between women who delivered at home and the opportunity they had to choose the place of delivery. Decision-making power, gender inequalities and social pressure especially from spouses and other relatives has been reported to significantly influence the use of maternal and child health care [64, 65]. It has been well documented in Africa that women lack the power to make independent decisions with regard to their own health care and that of their children . Even though this study did not explore in detail gender and decision making health issues, we can suggest that measures aimed at encouraging women to deliver in health centres will have to involve men if they are to be successful. Furthermore, socio-economic factors can also explain this situation as educated women tended to be better economically empowered than non-educated women and could take decisions on their own .
In this study, many of the women who had delivered with a TBA and TBAs themselves cited cost fees as a major determinant of choice of place of delivery, which is consistent with other findings in similar settings [5, 68]. This factor, as well as the quality of care expressed in this study as negative experiences women encounter during their interactions with health workers in previous pregnancies, have been acknowledged as important reasons for the non-use of maternal services including PMTCT services and delivering outside the formal health service . In situations such as the one experienced here where women combine TBAs and professional care and where TBAs encourage women to use ANC service, strengthening existing basic antenatal service delivery in general and before introduction of additional interventions such as PMTCT programmes in particular is needed .
Women's cited fear of HIV testing and knowing one's HIV status was due to discrimination and stigmatization or breach of confidentiality. Fear of knowing one's HIV status has been described previously as an important reason for drop out from PMTCT services and low levels of HIV status disclosure . Songok et al raised concern that non-compliance to the intrapartum dose of sdNVP in a PMTCT programme in Kenya, was due to mothers giving birth at home and fear of TBAs knowing . Presence of TBAs discouraged them from taking the dose for fear of exposing their HIV status. It is thus important to extend and reinforce partnerships among different stakeholders at health centre and community level to support education and access to health information for all women in particular and the community in general in order to prevent stigma and discrimination. In addition, training, deployment and supervision of community health workers including TBAs, must emphasize the importance of confidentiality and the need to support women in the process of disclosure of their HIV-serostatus. HIV status disclosure may lead to improved access to HIV prevention and treatment programmes, increased opportunities for risk reduction and awareness of HIV risk to untested partners, which can result in greater uptake of voluntary HIV testing and counselling,  and adherence to the advise given to prevent postnatal and sexual HIV transmission .
Qualitative results shown in the FGDs confirm the willingness of TBAs to participate in PMTCT interventions. Nevertheless, as we found in the community survey, women who had delivered in a health centre or with a TBA agreed that for TBAs to be involved in PMTCT activities they need to be trained.
Several potential limitations should be addressed in conjunction with our findings. First, the sample was limited to two rural districts of Zimbabwe and as so, may not be representative of the whole of Zimbabwe. Nevertheless, the socioeconomic characteristics of women of reproductive age considered in this study reflect those of Zimbabwe and many other African settings. Second, respondents' accounts of subjective events around pregnancy and childbirth may have been prone to recall bias. However, studies have shown that the recollection of various factors related to pregnancy and delivery is accurate, even over long periods . A third consideration relates to possible inconsistencies especially where the respondent could not understand the agreed Shona terminologies as this might have affected the understanding of some questions by the respondents. This was controlled to some extent through training of data collectors on the terminologies to be used for the Shona translations to minimize errors and through piloting of the tools.