We started out this paper by presenting Tanzania’s commitment to promote decentralised health care, a system in which decision-making and priority setting processes have been devolved to the district level to ensure local level prioritisation, control and accountability . Whereas the decentralisation policy advocates for local planning processes that start at the grass-roots level and move up through the system, this ideal was hardly reflected in the findings of our study related to the PMTCT programme. Neither the district nor regional health plans were seen to incorporate the views and priorities of the categories of people most closely and strategically placed within the PMTCT programme. In spite of the existing structures and the clearly spelled out ideals of decentralisation, we found that in the case of the PMTCT programme the system remained heavily reliant on external funders who tend to guard their own globally generated priority agendas in a manner that disregards local experiences and priorities. Also the district’s dependency on basket funding with minimal internal revenues accruing from within the district has amplified the burden of non-inclusion of local priorities in the annual health plans.
In response to the changes, the donors who support the PMTCT interventions have made attempts at moving closer to the people by relocating their operational base from the national to the regional level. In other words they have shifted their arenas from the international, to the national and more recently to the regional levels. Although presently they work with the regional managers of the programme, the regional administrative level in Tanzania is still far removed from local communities and health facilities. According to the governmental structure the role of the region is to translate policy guidelines from the national level, advise and supervise districts, and review the district health plans to ensure conformity with the national guidelines before they are approved by the District Full Council . In this regard, the donor devolvement to the ‘local’ level appears to have failed to ease and facilitate the inclusion of the priorities identified by the districts in the PMTCT-related activities, when this level is at the heart of the grassroots-based priority setting and decision-making in the Tanzania’s decentralised system. The donors operating in the PMTCT field, according to this study seem to be fairly rigidly sticking to their own globally-defined set of priority areas with little or no willingness to include the district’s experienced demands. The dynamics at work related to priority setting in PMTCT thus seem to operate in a fashion that remain far-removed from the staff at the health facilities, the district administrators and, apparently, even from the programme managers at the regional level.
The PMTCT services in Tanzania have recently been fully integrated into routine RCH services , according to the 2011 PMTCT guidelines. It does however seem to remain extremely difficult to locate and coordinate where, when and in what ways the PMTCT activities should be integrated in the planning process within an extremely complex reproductive health package. The findings moreover add substance to the experience that the integration of the PMTCT services into the RCH services has revealed little success in terms of integrated planning, since health planners operate with the (very real) perception that the PMTCT programme is donor-driven and donor-funded. The implication is that the programme is left to others and consequently receives little attention within the complex priority setting processes that surrounds the production of the annual CCHP. An important assessment study of the performance of the health sector reforms and of decentralisation revealed a similar scenario: programmes dealing with diseases that were perceived to be located under ‘vertical programmes’ and thus under donor funding received far less attention during processes of prioritisation as they were perceived to be already catered for, the implications being that local priorities were often not incorporated . In the present study, little knowledge on how to plan properly using the ceiling allocated in a diverse set of priorities, communication gaps between representatives attending the planning sessions and PMTCT in-charges, and lack of interdepartmental collaboration have added to this difficulty and has made it challenging to carry out a meaningful prioritisation process in the PMTCT programme. It has previously been documented that challenges of conflicting personal interests coupled with poor interpretation and implementation of the guidelines are other aspects that may undermine the priority setting processes and overall performances of the district health planning teams .
Although the districts have been given the autonomy to prepare and implement their health plans, the Ministry of Health retains a central role in developing policies to be implemented at the local levels in the decentralised health care system . In practice, however, the MoHSW in this particular case seems to continue producing policy guidelines that are perceived as ‘must’ or ‘orders’ by the lower levels of administration. Although the basket funds are remitted to the districts, the planning teams are also in this case given guidelines to follow in their allocations implying: 5-10% (community initiatives), 15-20% (health centre), 10-15% (voluntary agency hospital), 25-35% (council hospital), and 15-20% (office of the DMO) . Thus the guidelines that may be helpful in guiding the process end up limiting the planning team’s capacity to plan as they would wish since the above distribution has to take precedence over all other additional items. Sometimes the ministry officials would ask for the inclusion of activities in the district health plans that are not of local priority. Members of the secretariat at the regional level, who review the CCHP for conformity with the national guidelines, would thus in this case e.g. cancel PMTCT-related activities that were initially planned for in order to accommodate national demands. These national demands are again often generated by global policy and funding bodies such as the UN system and become demands that nations find hard to refrain from. Daniels, the scholar behind the Accountability for Reasonableness Framework for priority setting, has argued that to improve fairness, the planning teams need to work like ‘a football team’ where all players work together for the common goal . Under this notion, activities perceived to be of important will receive attention in the prioritisation process regardless of whether they are from the vertical programme or not, or from lower level staff or not. In fact improved priority setting decisions improve the quality of service provision; they improve stakeholders’ satisfaction, and reduce complaints, thus enhancing trust and proper allocation of resources. The REACT project from which this sub-project emerged has reported positive results pertaining to stakeholder involvement in processes of priority setting, as documented by Maluka for example . However, more studies are deemed necessary to assess reasonably the status of priority setting processes within diverse health related programmes in this and other districts in order to draw broader conclusions.
The scenario that has emerged in the present study is challenging as it seems to question the legitimacy of the priority setting processes pertaining to the PMTCT programme as exercised in the country. Moreover, it asks questions on the manner in which the ideals behind decentralisation are fulfilled. The findings of the present study add to existing evidence of a continuation of top-down and external influence, whether donor or governmental documented in other studies [16, 32]. Johansson  has, for example, in his study on Tanzania revealed how priorities are set by international and national managers pertaining to the project related to the eligibility criteria for receiving antiretroviral drugs, a scenario which implies that lower level actors from where implementation takes place had no room to contribute. Other studies have focused on the challenges in priority setting processes at the district level [11, 16, 27]; indicating that health facility and community views are rarely taken into consideration, that no clear procedures are followed and that there are no clearly spelled out roles for the different committees at the district and health facility levels, rendering the health committees redundant and inactive. Weak health information systems, moreover, hinder the availability of credible and reliable evidence required by the District Health Plan team at the time of setting priorities [10, 33], a situation which in turn makes it difficult to make meaningful prioritisation. The involvement of lower levels in the planning process should be aimed at ensuring that resources are targeted to those in need. On this point, the Council on Health Research for Development (COHRED) asserts, ‘The ones who own the problem are the ones who can provide the solutions’ , and they need to actively participate in setting priorities .
In principle, one must be also careful when indicating that the study findings are relevant beyond the field we have focused on. However, we do wish to suggest that the findings of the present study together with similar findings from other studies [11, 16] indicate that the decentralisation policy seem to work to reinforce the established power structures rather than integrating priorities of lower levels stakeholders which was aimed at with the reform. It is also important to keep in mind that colonial legacies and customary power structures lead lower level staff to fear open disagreement, making it difficult for them to meaningfully execute the authority granted to them. This tendency creates a new form of dependency on the donors/ foreign experts and other higher ranking individuals, creating scenarios where local officers remain locked in a system over which they have little control.
In this particular case, the continued donor-dependence has contributed to the continuity of a top-down approach where PMTCT managers feel that they have little option but to adhere to the donor’s priorities, hence they experience a loss of autonomy. In the past few years the PMTCT related priority areas of the donors have been to improve access in terms of coverage of services . This has been a worthy contribution with tangible results in a vast numbers of African countries. The PMTCT services are today available in most health facilities with RCH services in Tanzania , and at least a single member of staff has been trained in each facility providing the services. These measures however, do not ensure the quality of the programme which continues to experience severe challenges [26, 36]. Moreover, the health systems challenges continue to hamper its successes and raise concerns over the quality of the services on offer. Similar findings were reported by Johansson from another rural district in Tanzania .
On the whole, effective implementation of sector-wide approaches where donors support the budget of the health sector through basket funding emerges as a useful way to enable districts to identify their own priorities . These approaches, when effectively implemented, can facilitate the smooth implementation of the decentralisation policy as it can allow for a shift from vertically-focused health programmes and centrally-controlled budgets to more comprehensive health planning and locally-controlled health budget structures at the district level. Such approaches, moreover, are aimed at reducing external power influences . In this study these aims do not seem to be fulfilled in the case of the PMTCT programme. The claim that funds were inadequate could result from a lack of readiness to allocate the budget items to the programmes perceived to be donor-driven or could stem from a lack of experience and capacity to prepare or implement the plan. A study by Semali  revealed that there might be opposition to the transfer of authority by district stakeholders by giving less priority to vertical programmes activities even when they are integrated into the horizontal services. Despite these challenges, it is important to note that improving the implementation of the decentralisation policy entails eliciting values and criteria for priority setting from lower level stakeholders in the health care system. In this effort, special attention should be paid to the methods to be used in bringing up local priorities. Only a continuous concern with the dynamics at work in the health systems will allow for continued pressures to be kept on local authorities in a manner likely to facilitate increasingly fair and inclusive priority setting processes whether in PMTCT or other health programmes. When priority setting processes are grounded in the local context, it is more likely that the decisions reached will be perceived as relevant by the stakeholders, and that decisions will ultimately be experienced as improving the quality of health services on offer, which should be the main outcome of a fair priority setting process in a decentralised health care system.
Study strengths and limitations
This study is based on a single rural district. As such it is hardly possible to generalise its findings to the rest of the Tanzanian districts. Nevertheless, the emphasis placed on donor funding of the PMTCT programmes and the integration of the activities in the CCHP is a national policy. Thus, it is likely that dilemmas of priority setting related to the continued central role of the donors and of MoHSW, and the challenges linked to the weaknesses inherent in the district planning team that were found in this study as hampering the effective implementation of the decentralisation policy can be of some relevance also in other districts in the country.