Introducing payment for performance in the health sector of Tanzania- the policy process

Background: Prompted by the need to achieve progress in health outcomes, payment for performance (P4P) schemes are becoming popular policy options in the health systems in many low income countries. This paper describes the policy process behind the introduction of a payment for performance scheme in the health sector of Tanzania illuminating in particular the interests of and roles played by the Government of Norway, the Government of Tanzania and the other development partners. Methods: The study employed a qualitative research design using in-depth interviews (IDIs), observations and document reviews. Thirteen IDIs with key-informants representing the views of ten donor agencies and government departments influential in the process of introducing the P4P scheme in Tanzania were conducted in Dar es Salaam, Tanzania and Oslo, Norway. Data was collected on the main trends and thematic priorities in development aid policy, countries and actors perceived to be proponents and opponents to the P4P scheme, and P4P agenda setting in Tanzania. Results: The initial introduction of P4P in the health sector of Tanzania was controversial. The actors involved including the bilateral donors in the Health Basket Fund, the World Bank, the Tanzanian Government and high level politicians outside the Health Basket Fund fought for their values and interests and formed alliances that shifted in the course of the process. The process was characterized by high political pressure, conflicts, changing alliances, and, as it evolved, consensus building. Conclusion: The P4P policy process was highly political with external actors playing a significant role in influencing the agenda in Tanzania, leaving less space for the Government of Tanzania to provide leadership in the process. Norway in particular, took a leading role in setting the agenda. The process of introducing P4P became long and frustrating causing mistrust among partners in the Health Basket Fund. The study was carried out in 2012 and 2013 in two locations: Dar es Salaam, Tanzania, and Oslo, Norway. A qualitative study design was adopted to explore narratives and perceptions surrounding the introduction of P4P in Tanzania. In-depth interviews were conducted, observational activities were carried out and reviews of policy document and other relevant secondary data were conducted. Data was collected by the first author in two phases, October-November 2012 in Oslo and January-June 2013 in Dar es Salaam. Below is an account of the method we used for data collection and how these methods are triangulated in the study. The first author participated in a number of meetings on P4P in Dar es Salaam in the period of January 2013 to June 2013 both as a participant and as an observer. Two of the meetings were particularly important. The first was a P4P stakeholders meeting, held in January 2013 which gave an overview of the status of P4P in health care in Tanzania and provided an opportunity to identify influential actors in the field. During this meeting initial contacts with potential informants in Dar es Salaam were made. The meeting also contributed to the identification of potential sources of secondary data for the study. The second meeting occurred when the first author was requested by the P4P joint assessment committee, which consisted of Norad, the World Bank and USAID, to assist as a resource person on literature on P4P in Tanzania and other contexts. This role was important for gaining access to and building rapport with central informants in the study. Overall, the participation in the P4P meetings were important for gathering background information, for refining the research questions, for the identification of potential informants, for the development of the interview questions [34], and for mapping of secondary data sources. Policy documents were of utmost importance to the study, and were used mainly from a realist perspective [34], that is, as a means to understanding the P4P policy and design in the Tanzanian context. Hence, policy documents were essential in providing background information to the study and in defining the questions and trajectories that were pursued in the in-depth interviews. Policy documents central to our study include: The Pwani region P4P pilot: design document [17], Health sector startegic plan III (July 2009-June 2015) Partnership for delivering the MDGs [35]. The national road map strategic plan to accelerate reduction of maternal, newborn and child deaths [36], Payment for performance strategy 2008–2015 [37], Implementation guidelines- payment for performance [38]. These policy documents have been instrumental in uncovering the political frames and in supplementing primary data collected from the representatives of the Ministry of Health and Social Welfare. Detailed information on a range of themes related to the introduction of P4P in the health sector in Tanzania was obtained through IDIs conducted with representatives of organizations and agencies identified during participant observation in the meetings and conferences. The questions that were asked in interviews were tailored to suit the perceived roles played by different actors, and this process was aided by the information obtained in policy documents, and during observations. Three interview guides were designed: one for Government officials in Tanzania; one for Norwegian informants, and one for other development partners and stakeholders. Although these three interview guides had different specific (for detailed interview questions refer to additional files 1, 2, and 3), they were all guided by the following general themes: trends and thematic priorities in development aid policy, countries and actors perceived to be proponents of the P4P scheme, and agenda setting in the Tanzanian Health Basket Fund and the introduction of P4P. The P4P agenda in Tanzania was first introduced into the Health Basket Fund. In choosing the informants for the study, we used purposive sampling following two criteria. To achieve the objective of the study we were interested in the views of members of the Health Basket Fund who were influential during the P4P introduction process by either supporting or questioning the P4P agenda. Through this criterion, we were able to identify the Ministry of Health and Social Welfare of Tanzania, the World Bank, the Norwegian Agency for Development Cooperation (Norad), the Danish International Development Assistance (Danida), the Swiss Agency for Development and Cooperation (SDC), German International Cooperation (GiZ) and Irish Aid. All these members (with the exception of GiZ) were formative members of the Health Basket Fund in 1999. They were selected based on the assumption that they therefore possessed more knowledge on the founding principles of the Health Basket Fund than members that joined at a later stage. Secondly, we were interested in organizations/agencies outside the Health Basket Fund that appeared to be important stakeholders in the P4P agenda setting and the subsequent P4P pilot. Based on this criterion we identified the Clinton Health Access Initiative (CHAI), an organization managing the P4P scheme in Tanzania on behalf of the Ministry of Health and Social Welfare. In addition, we included the Norwegian Embassy in Dar es Salaam, and the Norwegian Ministry of Foreign Affairs, who played an important role in introducing and funding the P4P programme in Tanzania. From these 10 organizations/agencies, a total number of 13 in-depth interviews with key-informants were conducted, 11 of these in Dar es Salaam and two in Oslo. Informant selection in the organizations focused on individuals knowledgeable of the P4P agenda setting and process in Tanzania, and the majority of our informants were representatives of their organizations in the Health Basket Fund. An overview of the interviews conducted is summarized in Table 1. Overview of IDIs All interviews were conducted in English and based on informed consent, and all except two were recorded. The two interviews were not recorded due to the preference of the informants. In addition to recording, rapid note taking was used in all interviews. The recorded IDIs were transcribed verbatim and error checked. The analysis of the material started with a review of transcripts which were later imported to NVivo 10 for data management purposes. Qualitative content analysis was undertaken, looking for both manifest and latent content [39]. Coding units were identified and condensed [39]. Sub-themes were developed from the codes and defined into themes that we used in presenting our results. Research clearance was granted in Norway through the Norwegian Social Science Data Services and in Tanzania through the Ifakara Institutional Review Board, and the National Institute for Medical Research (NIMR/HQ/R.8a/Vol.IX/1515). Individual consent was sought and obtained free of coercion.