By Maryline Mireku and Lilian Otiso, LVCT Health
Conducting implementation research means adapting to, and coping with, changes in the external operating environment. In Kenya the REACHOUT team has had to conduct their research on improving the quality of close-to-community health programmes whilst the country implements the process of devolution. This has created new challenges and opportunities that we have needed to respond to. Our experiences provide lessons for other countries undergoing similar changes.
Devolution refers to the process of decentralization of government that Kenya underwent in 2014 from one centralized government to 47 counties with responsibility for hiring, budget and decision making leaving the national government with capacity building, standards and policy making
In 2013, LVCT Health conducted a context analysis on community health services in Kenya which identified lack of supportive supervision as a factor influencing performance of community health workers. The second phase of the study involved implementing a supportive supervision intervention in four community health units in Nairobi and Kitui counties, Kenya.
When we disseminated the findings of the first phase of our REACHOUT study in early 2014, health had not been fully devolved and there was still reliance on decision making at the Ministry of Health in central government.
After gaining ethical approval for the second phase of our study from the national Kenya Medical Research Institute (KEMRI) Ethical Review Committee (ERC) devolution meant that we had to proceed to County and Sub-County health offices to seek further approval before embarking on the implementation. As the study coordinator I was sure this process was not going to be difficult since these were offices we had worked with in other endeavours but also specifically during the first phase of the study. All I needed to do was send letters and the proposals to the offices and wait for feedback.
In the pre-devolution era most Counties priorities echoed those of central government. However, as we moved into the second phase of REACHOUT, we came to realize that County priorities had changed. Counties were now responsible for prioritizing and managing budgets in the face of major reduction in funds from the central government and donors. In addition, the central government appears to no longer have the resources to enforce implementation of national policies by the counties. The result of this was major changes to community health services implementation including changes in practice and in some cases the decision to not implement changes which had been recommended by the central government before devolution. For example, the Community Health Strategy (revised in 2014) called for changes in roles of community health providers such as the community health volunteers (CHVs) only being involved in mobilization of community members as opposed to the initial plan where they not only conducted mobilization but also collected data on household health status and in some cases provided basic curative services. But this has only kicked off in less than half of the Counties with the non-implementing ones sighting issues such as budgetary constraints and need for additional changes to the model to suit their needs as hindrances to adoption.
County Health Management Teams (CHMTs) are composed of heads of various divisions in the health department at county level headed by the county director of health (a new position). The constitution of CHMTs and the persons heading them (sometimes the County Executive Committee Member for Health) varied across the counties. New positions and functional groups such as County Research Technical Working Groups (TWGs) were created and formed a new layer we had to work with to conduct the study.
Due to the process of setting up, some of the offices created by the CHMT lacked an individual to head them due to lack of adequate staff or staff with the right qualifications or ongoing consultations. During the course of our work we noticed either prolonged vacancies in some of these positions or frequent changes in the people heading these positions. For example, in one of our study sites we had two offices headed by three different people in a span of six months.
This has had impacts on the research. We did not anticipate creation of new offices in the CHMTs with a responsibility of ethical appraisal of research studies. Before devolution the Ministry of Health structures (Provinces and Districts) accepted approvals from the existing ERCs and studies not need to undergo additional ethical approval. One of the Counties where we are implementing our study developed a County Research Office which was tasked with appraising all studies. Taking our study through another approval process meant we had to wait to begin the study and caused a delay that we had not planned for.
Unfortunately, even when new offices were created, there were several transfers with staff filling the positions for very short periods of time. In one of the offices we were working with we experienced three different people occupying an office in a span of three months. Upon raising my concern of the frequent changes one county official explained to me how these were some challenges with devolution. He concluded ‘at least the office still exists, we have had situations where posts were created only to be disbanded later’.
Similar changes were taking place at sub-county level. Some transferred staff would take some time off before reporting to their new posts. Expectations of transfers also made some sub county officials reluctant to be involved in activities especially because some of the transfers were happening sporadically. Often we would begin working with somebody only for them to be transferred midway. In one instance we were ‘accused’ of running activities at the Sub County level without the Sub County approval because the Medical Officer of Health who gave the approval and a good number of the Sub County Health Management Team (SCHMT) were transferred. We had to put our activities on hold and go back and re-sensitize the SCHMT.
Devolution has resulted in leaders being closer to the implementers. For researchers this has resulted in Key Informant Interviews with senior managers at County levels being more productive compared to pre-devolution where the provincial level managers would not know much about the implementers they were making key decisions for. Decision making by leaders has also taken a more practical approach and implementers have been able to successfully influence plans made by managers. In addition, issues which would take longer to be resolved due to the many bureaucratic levels that existed pre-devolution can now be managed faster. This is an advantage to researchers who hope to influence practice by giving decision makers feedback of good practices from interventions.
The experiences laid down here were from a time when Kenya had just devolved its health system. Researchers should anticipate similar issues where government systems and structures are undergoing change such as what we anticipate in the upcoming Kenyan general election or the recentBrexit vote for UK to exit from the EU. This has implications for research:
Devolution has increased interest in research by counties which may result in more buy in especially if the research is aligned to the county priorities. Delays may however occur due to longer approval processes.
Because our intervention focused on supervision we were interested in working with supervisors at the sub county level. However due to transfers we were forced to wait for individuals to report to their posts and sensitize them. These changes were necessary but in as much as they were anticipated on the ground the specifics were not known such as the timing and the regions/posts which would be affected. Researchers implementing in the Counties need to have plans which can be adapted to changes in the context they are working and avoid linear approaches and commonly held assumptions. We have had to be flexible and align our study to fit into these changing contexts although there is still uncertainty of what other changes will come up in the life of the study.
In such different and fluid contexts a researcher has to be ready to accept that at the end of your study, especially if long term, one may face difficulties in carrying out comparisons across sites, tracking changes and even making attributions. There is need therefore to embrace active observation to record possible confounders and their effect on the studies we are carrying out.
According to the new constitution, the role of the central government is development of policies and standards. From our experience, the central government has no mandate to enforce these policies. County governments are now amending or developing new policies contextualized to the local priorities. However, this process is affected by lack of structures and experience in utilization of evidence to make decisions to inform policies. For researchers in Kenya this means new players in the policy environment who have to be involved from the beginning to the end of the research for it to inform policy. Researchers will not only have to involve the Counties in advocacy but also work with some Counties to build capacity of some heads of department in this area.
Implementing research in uncertain contexts calls for researchers who not only anticipate change but also have the ability to adapt to the change at any time of the implementation. Our experiences from the Kenyan context offers lessons for others experiencing similar, rapid change.
Photo courtesy of Mkimemia
This project is funded by the European Union.