This realist evaluation sought to identify key mechanisms driving the implementation of an intervention of integrated childhood immunisation and FP services in BGRS, Ethiopia. It contributes to a growing body of literature that seeks to understand uptake of MCMs when FP services are integrated with other health services (7-9, 23-26). Recent studies from Rwanda, Zambia, and Ghana have looked specifically at the integration of FP with immunisations services and have found varying levels of success (7-9). Issues such as inconsistent training for HCWs, poor monitoring systems, and disjointed referral systems have all been cited as barriers to effective integration (7-9). Central to this evaluation was the exploration of if and how integration worked, for whom and what mechanisms drove MCM uptake. There is currently a lack of literature in which context and mechanisms may explain intervention implementation, and this evaluation offered a unique opportunity to explore this.
Integration
In terms of understanding how integrated the services were, and the level of service uptake over the study period, the monitoring and DHS data reviewed illustrate two important issues. First, the monitoring data demonstrated that uptake of MCMs among women who brought their children for at least one immunisation during the intervention period was approximately two-thirds (63%) with slight differences across Assosa and Bambasi. This indicates that FP services were quite well integrated and promoted, and that MCMs were appropriately offered during the provision of childhood immunisations. In this study context, FP counselling was offered by HEWs and HDA members predominantly at health and outreach posts, but also during household visists as these were all important locations where women received services. This is in line with evidence that suggests that flexibility about the point of delivery of integration is crucial for ensuring effective service delivery (9).
Second, the 2016 Ethiopia DHS data indicates a full immunisation rate in BGRS which, although higher than the national average, is well below the rate in several regions (14) and suggests that further community based healthcare approaches might help to increase access in BGRS. A study from the rural Dabat region of Ethiopia found that immunisation coverage rates were low when women had to travel long distances to health posts for immunisation services (27), which may also be the case for BGRS. Household visits by HEWs, in this context, were found to be a key component of the integration intervention. The benefits of using community based outreach approaches, such as household visits, have been shown to positively influence health seeking behaviours in rural and peri-urban areas, particularly in terms of immunisations (28).
Perceptions of key actors
Our findings suggest that acceptability of FP by religious leaders, and community members, including men, was a key factor that drove wider community acceptability and in turn, influenced a woman’s decision to use an MCM. The influence of religious leaders on the health seeking behaviours of communities is well documented. A recent study from Nigeria found that women’s attendance at ANC services increased after religious leaders in the community began promoting ANC services as an essential component of maternal and child health (29). In this context, religious leaders had a key role in delivering health messages to religious constituents. Similarly, Azmat et al (2011) determined that religious leaders in Pakistan had a strong influence within the community and that they should play a key role in informing the community about the benefits of FP. FP acceptance among religious leaders was influenced by exposure to messaging and information about FP from medical professionals (30). In the current evaluation, FP acceptance among religious leaders was influenced by their ability to see that FP aligned with their religious beliefs. Understanding how and why religious leaders accept FP is important when seeking to design and implement similar interventions in contexts where religious leaders have a strong influence over community members’ decision-making.
Similarly, this evaluation provides an understanding of how and why community members, particularly men, accept FP. Our findings suggest that male partners seeing religious leaders actively promote FP triggered them to accept FP themselves. Ethiopia’s 2016 DHS data indicate that men have more decision-making power than women within couples regarding FP in BGRS compared to almost every other region in the country (14). Studies from Nigeria and Malawi support the argument that men influence women’s decision-making about FP and that a key component of FP interventions should be male partner education to encourage their support for FP (31-33). Interventions that promote joint decision making may be successful. An intervention supported by the IRC in Ethiopia assessed the impacts of the use of a color-coded health calendar to increase uptake of immunisation services and found that the calendar promoted health discussions within households (34). DHS data also indicate that women in BGRS have lower rates of FP use than women in almost every other region in the country (35). They are also less likely to give birth with skilled birth attendants either in health facilities or at households (35). This indicates poor links with the formal health system and limited access to health services. Given this context, it is especially important to engage male partners and religious leaders to promote and protect women’s health and increase the women’s confidence as decision-makers. Hoyt et al demonstrate the influence of peers, male partners and the wider community on women’s decision making about MCM use in Benin, Ethiopia, Kenya, Malawi and Uganda (36).
Our findings also suggest that male partner education could be considered alongside gender transformative strategies that challenge traditional gender norms which may affect women and girls’ ability to access and use services. Research demonstrates that gender-transformative approaches that engage men in dialogue about FP and gender norms can increase FP access among women (37, 38) and have been successful at improving FP, antenatal care, maternal and child health behaviors and outcomes (39).
The willingness of the HEWs to attend training and to adhere to training guidelines was a key contributor to intervention outcomes, as these translated into HEWs feeling high levels of self-efficacy when delivering FP services. Understanding that interventions that increase feelings of self-efficacy among HCWs may yield positive intervention outcomes can help to improve intervention programming. However, HEWs were not trained in implant removal and this may have adversely affected uptake of implants by women, and self-efficacy of HEWs. National policy in Ethiopia does not task HEWs with implant removal (16). This has resulted in an unmet need of implant removal in the many rural and hard to reach areas (16, 40). To address this, the Integrated Family Health Programme (IFHP) has attempted to scale-up the availability of trained health professionals who can provide this service, but the programme has yet to be extensively rolled out. The Ethiopia Ministry of Health had begun to pilot the training of select HEWs in implant removal across the country, including in BGRS. Despite the lack of removal services by HEWs, the monitoring data indicated that implants were the most commonly accepted MCM among post partum mothers that brought a child for immunisation at health posts in Assosa during the study period. Similar findings have been documented in other studies within the context of Ethiopia (41). In this study, HEWs expressed concerns about women having to travel far distances and incur costs to access larger health facilies where implant removals are carried out. There is a wide body of literature that discusses the inverse relationship of distance and time spent travelling to a health facility and health care access and utilisation (42-44). This raises the issue of whether an intervention that encourages the adoption of a health service but that does not ensure an effective system for follow-up has a positive impact overall, or whether it contributes to health disparities.
HEWs perceived integration of immunisation and FP services to be advantageous. Relative advantages included reduced workloads, and a clear fit with their schedule, which focused on providing FP counselling during post-natal household visits and MCMs during the ’45 day immunisation’ visit. A recent Cochrane review of integrated interventions found that HCWs may become overloaded or deskilled in integration interventions leading to negative impacts on service provision and health outcomes (1). Our findings suggest that HEWs may perceive the integration of services differently when they are tasked with providing integrated services in households, compared to when they deliver such services to clients in health facilities. This may be due to the fact that health posts are usually staffed by two HEWs and are sometimes supported by additional HCWs. In our study, teamwork among HEWs,HCWs and HDA members was cited as a reason for manageable workloads. This evaluation identified mechanisms that included constructs of relative advantage and burden, indicating that in this context, in order for health workers to perceive the intervention positively, they needed to see how it would be advantageous to them or their clients, and how it would reduce their workload. Studies that have explored the training of community-based health workers have cited manageable workload, organisation of tasks, supportive supervision, adequate supplies and equipment, and respect from the community and the health system as key drivers of successful service delivery (9, 45). A recent study by Mayhew et al further supports our findings by concluding that structural factors at the health facility level, including issues of staffing and workload in integrated interventions can be mitigated and managed by HCWs themselves (46). The authors highlight that when HCWs felt agency or power over their own decision-making, they were able to overcome potential challenges of integration (46). These factors were mentioned by the HEWs interviewed in this evaluation and indicate that while the training they received was important, its effectiveness was dependent on having a supportive work environment that included workload sharing with colleagues which triggered constructs of self-efficacy.
Self-efficacy was a key construct within both the service delivery and community contexts. Self-efficacy has been described as ‘an individual's belief in his or her capacity to execute behaviors necessary to produce specific performance attainments’ (47). Among HEWs, feelings of self-efficacy were seen to drive motivation for and perceptions of the intervention. HEWs felt confident that they were sharing their workload with their co-workers. HEWs knew that they could carry out their work effectively, as the work was being shared, and this fostered a sense of teamwork among them. Studies that have assessed self-efficacy among health workers have found strong links between feelings of self-efficacy and motivation and have emphasized the links between team work, task-sharing, and self-efficacy (48-50). Among women, feelings of self-efficacy were triggered when there was community support for FP use, particularly from male partners. A study from Guatemala found that lack of knowledge about and availability of methods, fear of side effects and infertility, husbands being against family planning all negatively impacted feelings of self-efficacy (51). In the current evaluation, knowledge about the availability of FP methods, and support from husbands meant that women in the community felt confident in their ability to choose FP, thus leading to high levels of self efficacy. Karp et al describe a framework for women's and girls' empowerment in sexual and reproductive health with three key components: the existence of choice, the exercise of choice, and the achievement of choice (52). This has resonance with our study where the combination of the availability of FP services and the community support for FP meant that there was choice. This combination of factors allowed women to exercise their choice. Together, this led to the achievement of choice.
What does this study add?
This evaluation demonstrates the role that religious leaders, community members, and health workers playedin the implementation and uptake of an FP intervention in this srudy. By linking our empirical findings to published theories of acceptability, diffusion of innovations, and accessibility, we we were able to identify drivers of outcomes, that is mechanisms, with a higher level of abstraction than those identified in the empirical data alone. These constitute middle range theories that may be tested in other geographic sites. For example, we identified relative advantage and the related construct of self efficacy as an important mechanism in the acceptability and implementation of integrated FP and immunisation by health workers. More simply, we were able to provide theories identifying mechanisms that need to be triggered to facilitate the acceptability and implementation of FP when integrated with immunisation and improve the selection and implementation of effective interventions in the future.
Limitations
This study used data on the number of children immunised with Penta1 as a proxy measure for the total number of women who brought a child for immunisation at least once; however, this proxy i does not take into account the number of times a woman brought a child for an immunisation relative to whether she took up an MCM. These data do not account for women who were already using an MCM when they accessed the immunisation clinic, that is continuation, and therefore may underestimate use when defined as new acceptance and continuation.
Further, while SSIs were conducted with a wide range of stakeholders with selection based on the initial programme theory, it is possible that a larger sample size would have yielded data describing additional CIAMOs to those presented in this paper. Also, only one woman was interviewed specifically for her role as an FP user. HDA members and other community volunteers were sometimes also female FP users. While they provided their perspectives as female FP users, they were not interviewed specifically for that role. A larger sample of female FP users and non-users would have yielded more perspectives from these groups.