THE INTERNATIONAL JOURNAL OF HEALTH PLANNING AND MANAGEMENT
Int J Health Plann Mgmt 2014; 29: e368–e382. Published online 7 November 2013 in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/hpm.2226
Implementation of Community Health Fund in Tanzania: why do some districts perform better than others? Stephen Oswald Maluka1* and Godfrey Bukagile2 1
Institute of Development Studies, University of Dar es Salaam, Dar es Salaam, United Republic of Tanzania 2 Development Studies, College of Business Education, Dar es Salaam, United Republic of Tanzania
SUMMARY In early 1990s, Tanzania, like other African countries, introduced user fees in public health systems. Although user fees were considered important in promoting health, they appear to reduce people’s access to health services. To counteract the detrimental effects of the user fees, various types of health insurances were introduced, including the Community Health Fund (CHF). Drawing from the review of minutes, health facility visits and key informant interviews, this study explored why implementation of the CHF in Tanzania has been more successful in some districts than in others. The findings indicate that in Lindi district, the enrolment rate for the CHF was very low. This was attributed to high premium rates, frequent drug stock-out, lack of trust by the community members to the health providers, low incentives and local politics. In contrast, in Iramba district, the performance was better. Availability of drugs in the health facilities, effective supervision, commitment of the top district-level officials and incentives to the health facility committees were the main factors that facilitated good performance of the fund in Iramba district. The focus of the implementation needs to be placed on the active engagement of the local-level leaders and politicians who are responsible for the implementation of the policy. Equally important is the availability of quality health services in the health facilities. Copyright © 2013 John Wiley & Sons, Ltd. KEY WORDS: Community Health Fund; health policy; policy implementation; health facility committees; Tanzania
INTRODUCTION In the early 1990s, Tanzania, like many other developing countries in Africa, was caught up in the web of reforms in the health and social services sectors that saw the adoption of the user fees for healthcare financing in public health systems. Although user fees were considered important in promoting health and economic *Correspondence to: S. O. Maluka, Institute of Development Studies, University of Dar es Salaam, Dar es Salaam, United Republic of Tanzania. E-mail: [email protected]
Copyright © 2013 John Wiley & Sons, Ltd.
IMPLEMENTATION OF COMMUNITY HEALTH FUND IN TANZANIA
e369
performance, they appear to reduce people’s access to health services especially the poor and the most vulnerable groups of the society (Gilson, 1997; Watkins, 1997; James et al., 2006) and sometimes lead to catastrophic health expenditures (McIntyre et al., 2006; Devadasan et al., 2007; WHO, 2010). To counteract the detrimental effects of the user fees, developing countries have vigorously been promoting the use of risk-sharing arrangements for healthcare financing, which makes healthcare payments predictable especially for the rural poor. Various types of health insurances were introduced. The first type of insurance is national or social health insurance, which is based on individual’s mandatory enrolment. The second is voluntary mechanisms, which include private health insurance and community-based health insurance. Community-based health insurance schemes are based on the premises of risk pooling and community solidarity to risks of falling sick and are conceptually designed to provide financial protection and reduce out-of-pocket payments for healthcare (Ouimet et al., 2007; Jacobs et al., 2008). In Tanzania, the government introduced the Community Health Fund (CHF) in 1995 as a mechanism for providing additional funds for financing health services and protecting households from out-of-pocket payments as a result of the implementation of user fees in the health sector (MoH, 1999). In 1996, the scheme was first piloted in Igunga district. In 1998, the Ministry of Health (MoH) decided to expand the CHF to nine additional districts. In 2001, the CHF Act was passed by the parliament, establishing the CHF as the official health plan at the local/community level (URT, 2001). The CHF is a district-level voluntary pre-payment scheme that targets the population living in rural areas and/or employed in the informal sector. The scheme is based on the concept of risk sharing whereby members pay a small contribution on a regular basis to offset the risk of needing to pay a much larger amount if they fall sick (Mtei and Mulligan, 2007). Table 1 provides details of the CHF in Tanzania. Low enrolment remains one of the biggest challenges to the CHF. By 2011, CHF had enrolled a total of 561 370 households, which was only 7.4% of the total population (MoHSW, 2012). Various factors have contributed to low enrolment, including perceived poor quality of services in public facilities, a widespread inability to pay membership contributions, the limited benefit package and lack of access to referral facilities (Kessy, 2008). These factors are compounded by weak management, lack of trust in management and a failure among communities to see the rationale for protecting themselves against the risk of illness (Kessy, 2008). However, Kamuzora and Gilson (2007) found that district health managers had a direct influence over the factors explaining low enrolment of the CHF. According to these authors, the district health managers did not allocate a budget for CHF administration activities and did not ensure supervision of health staff to support delivery of quality services (Kamuzora and Gilson, 2007). Although a body of evidence indicates that the CHF is ineffective in its implementation, there are wide variations between districts, suggesting that implementation has been more successful in some places than in others (Burns and Mantel, 2006). A recent review revealed numerous innovative refinements to enrolment and health service quality improvements financed by the CHF (Stoermer et al., 2012). Against this background, the study explored why the implementation of CHF in Tanzania has Copyright © 2013 John Wiley & Sons, Ltd.
Int J Health Plann Mgmt 2014; 29: e368–e382. DOI: 10.1002/hpm
e370
S. O. MALUKA AND G. BUKAGILE
Table 1. Details of the CHF in Tanzania (URT, 2001) Payment and benefits
Organisation
Membership based on household enrolment One membership card is issued per household and is valid for a period of 12 months The annual contribution from each household is defined by the districts The CHF schemes subsidised by the government in the form of matching grants, complementing the members’ contributions with an equal amount Membership allows the household access to medical services without further co-payments at the primary level Inclusion of secondary-level and tertiary-level healthcare is at the discretion of the district Households without a CHF membership must pay the predetermined user fees to access healthcare Poor households who are unable to pay may be issued CHF membership or an exemption letter upon recommendation by the village council At the district level, the CHF is overseen by the CHSB with representatives from the district authorities, public healthcare provider, private healthcare provider and the community There is usually a CHF coordinator responsible for tracking membership levels and funds contributed across the district At the ward level, the Ward Development Committee and the Health Facility Committee are responsible for mobilising people to join the CHF, tracking the membership base, overseeing premium collections, evaluating CHF operations and providing recommendations and granting exemptions
CHF, Community Health Fund; CHSB, Council Health Service Board.
been more successful in some places than in others. It is expected that this study will contribute to the small body of literature on health policy implementation in developing countries. The study adopted descriptive approach, seeking to provide indepth analysis of the factors that influence the performance of the CHF in Tanzania. The study focused on the experiences of various actors involved in the implementation of the CHF at the district and community levels in order to give a bottom-up perspective of the implementation process.
METHODOLOGY The study design and settings An exploratory case study design focusing on two districts—one relatively high performing (Iramba) and the other fairly low performing (Lindi)—was used (Yin, 2003). The case study approach was considered the appropriate method for this study because CHFs are inherently a complex intervention. They are dependent on the context in which they are implemented, and their implementation is not standardised. They also have a strong social component in that they are designed to promote and protect the health of the population and to reduce inequalities. The ‘high’ and ‘low’ performing categorisations were based on the enrolment of the Copyright © 2013 John Wiley & Sons, Ltd.
Int J Health Plann Mgmt 2014; 29: e368–e382. DOI: 10.1002/hpm
e371
IMPLEMENTATION OF COMMUNITY HEALTH FUND IN TANZANIA
Table 2. Key characteristics of the study settings
Population Community Health Fund enrolment rate* Hospitals Health centres Dispensaries Divisions Wards Villages Health workers available Shortage of health workers
Iramba district
Lindi district
236 282 people 28.1% 2 6 60 7 31 143 43% 57%
194 143 people 0.4% 1 6 30 9 30 134 50% 50%
*The national Community Health Fund coverage rate is about 7.9% (National Health Insurance Fund data of 2011).
Table 3. The sampled study sites
1 2 3 4 5 6 7 8 9 10 11 12 13
Name of the facility
Type of the facility
District
Kiomboi Ndago Kinyangiri Bomani Mampanta Misigiri Ulemo Kitomanga Rutamba Mchinga Kilangala Mahumbika Mnolela
Hospital Health centre Health centre Dispensary Dispensary Dispensary Dispensary Health centre Health centre Dispensary Dispensary Dispensary Dispensary
Iramba
Lindi
CHF.1 However, the districts were not selected on the basis of performance only; other factors such as accessibility were taken into account. Table 2 summarises key characteristics of the study settings. In each district, one hospital, two health centres and four dispensaries were purposively selected, in collaboration with the district medical office, to include only those at which the health facility committees and head of the health facility have been in operation for at least 1 year (Table 3). Data collection techniques This study was primarily based on the individual interviews with the key stakeholders. Documentary data and health facility visits were used to support, verify 1
CHF performance data were obtained from the Budget Speech of the Ministry of Health and Social Welfare 2011/2012. Copyright © 2013 John Wiley & Sons, Ltd.
Int J Health Plann Mgmt 2014; 29: e368–e382. DOI: 10.1002/hpm
e372
S. O. MALUKA AND G. BUKAGILE
and highlight the key issues that emerged. In Iramba, individual interviews were carried out in October 2012, whereas in Lindi, the study was conducted in January 2013. Each interview lasted for approximately 45 min and was carried out at the respondent’s workplace and/or home. In order to cover a wide range of views of different actors, a purposive sampling technique was used. The purpose of key informant interviews was to collect information from a wide range of people who have first-hand knowledge about the CHF operations. An interview guide was developed to assist the semi-structured interviews with key respondents. The use of open-ended questions permitted the participants to express their views and elaborate on issues that they felt were most relevant and important. At the district level, members of the Council Health Management Team (CHMT) and the Council Health Service Board (CHSB) were interviewed. At the health facility level, committee members at the district hospital and health centres as well as ward and village leaders were interviewed. In total, 83 interviews were carried out (51 interviews in Iramba and 32 interviews in Lindi) (Table 4). Furthermore, considerable documentary information was obtained, and our analysis was validated with observations at the health facilities. Data analysis Quantitative data that were recorded in numerical form were entered into an Excel worksheet for secondary analysis of totals, percentages and other such simple statistics as necessary. Thematic approach was adopted to analyse qualitative data (Ritchie et al., 2003). First, all the interviews were transcribed verbatim into Swahili. Second, the transcriptions were translated into English. Third, interview transcripts were entered into NVIVO 10 software (QSR International Pty Ltd, Australia) for storage, coding, text search and retrieval. Using NVIVO 10 software, data were coded to initial themes. Thereafter, data were sorted and grouped together under patterns that were more precise, complete and generalisable. As patterns of meaning emerged, the researchers searched for similarities and differences in performance between the two districts involved. Finally, the data were summarised and synthesised, retaining as much as possible key terms, phrases and expressions of the respondents. After this Table 4. Number of key informant interviews Category of respondents
Council Health Service Board Council Health Management Team Hospital governing committee Health centre and dispensary committees Councillors Village executive officers and village leaders Heads of health facility and health workers Social welfare officers Total Copyright © 2013 John Wiley & Sons, Ltd.
Number interviewed Iramba
Lindi
4 3 3 17 2 8 12 2 51
3 3 — 15 — 4 7 — 32
Int J Health Plann Mgmt 2014; 29: e368–e382. DOI: 10.1002/hpm
IMPLEMENTATION OF COMMUNITY HEALTH FUND IN TANZANIA
e373
analysis, data were triangulated to allow comparison across different categories of respondents for final interpretation and presentation of the results.
ETHICAL ISSUES The research was approved by the University of Dar es Salaam, which has been given the mandate by the Government of Tanzania and Tanzania Commission for Science and Technology to grant research clearance to its students and staff. The clearance was presented to the regional and district authorities who approved the study in their area. Verbal consent was sought from prospective informants after the researchers had explained the objectives of the study. Informants were informed of their right to withdraw from the study any time they wished. They were also assured of confidentiality of any information they would give. All interviews were tape recorded with the permission of participants, and the resulting recordings and transcripts were stored in a manner that protected their confidentiality. For example, no respondent’s name was recorded.
FINDINGS The following section presents the main findings of the study organised in three main themes: enrolment and the benefit package; how funds were collected, managed and used; and what works, how and under what conditions. Enrolment process and the benefit package It was found out that in Lindi district, households were required to contribute TSh 10 000 (equivalent of $6.5) annually. This premium covered a maximum of six members per household. Community members, who were neither CHF nor National Hospital Insurance Fund members, were offered health services through user fees, which were payable at the rate of TSh 1000 at the dispensary and TSh 2000 at the health centre level per visit. The CHF benefits were limited to one primary healthcare facility (dispensary or health centre) closest to the domicile of the member and were not transferable to other healthcare providers in the district. On the other hand, in Iramba district, it was found that households were required to contribute TSh 5000 (equivalent of $3.5) annually. This premium covered a maximum of six members of the household, a couple and their children below 18 years of age. All CHF members could obtain healthcare services in any public primary health facility in the district provided that they showed evidence that they were active members. The district had also extended CHF benefits to include secondary-level services. However, secondary-level services were limited to ambulance and the outpatient department. If the CHF member was admitted, she or he had to bear all the admission costs. When referral was made to the district hospital, a patient had to be given a referral letter from the primary healthcare facility. There was no reimbursement to the district hospital for services rendered to the CHF members. Copyright © 2013 John Wiley & Sons, Ltd.
Int J Health Plann Mgmt 2014; 29: e368–e382. DOI: 10.1002/hpm
e374
S. O. MALUKA AND G. BUKAGILE
How were funds collected, managed and used? In both study districts, funds were collected by the heads of the health facilities and submitted to the district accountant on a monthly basis. The funds collected from the memberships were used for health-related purposes and activities by the district, partly flowing back to the health facilities. At the district, funds were deposited in account number 6, managed by the District Medical Officer (DMO). In Lindi district, 40% of the funds flowed back to the health facilities. Similarly, in Iramba district, 20% of the funds were returned to the facilities on monthly basis. At the facilities, the CHF funds were used for minor rehabilitations and purchasing of supplementary drugs and medical supplies. In Iramba district, the CHF funds were also used to pay incentives to the members of the health facility committees, health workers and salaries for watchmen. Similarly, CHF funds kept at the district were used for purchasing supplementary drugs and medical supplies as well as paying incentives to the members of the boards. The majority of the CHMT members also reported that a good number of health facilities had been installed with solar power using CHF funds. The review of the minutes and Comprehensive Council Health Plans supports this point. It was revealed that out of 35 health facilities that had electricity, 28 used solar power, and seven had been connected to the national grid. In the 2012/2013 financial year, the CHMT had planned to install six more health facilities with solar power using CHF money. However, it was found out that in both study districts, heads of the health facilities and members of the committees felt that they had little control of the funds. They pointed out that, apart from the small percentage of funds that were returned to the facility, they did not know how the remaining percentage of CHF funds was used. According to the respondents, this reduced their morale. One respondent commented this way: ‘Although we collect CHF funds, we are not informed how the funds are used. I once raised this concern to the CHF coordinator. But I was told that CHF is a district scheme, and funds are managed at the district level. I told him that he was demoralising us. The household contributes in order to get better services at their facility. When you use our money to improve services in the facilities which have not contributed, you discourage people from renewing their membership’ (IDI with in-charge of health facility in Iramba). In contrast, the majority of the district health officials had different opinions on this point. First, they felt that the health facilities had no capacity to manage CHF money. In particular, health facilities had no trained accountants to enable them to manage a large amount of funds. Second, both health providers and the health facility committees were not permanently based at the facility. Membership for the health facility committees expired after 3 years. Similarly, health providers could be transferred to other facilities. As a result, respondents said that it would be difficult to control the management of funds. Third, in order to obtain matching funds from the central government, the district had to pool together contributions from all health facilities. The government does not pay matching grants to individual health facilities. According to the CHMT members, it is important to have a single account Copyright © 2013 John Wiley & Sons, Ltd.
Int J Health Plann Mgmt 2014; 29: e368–e382. DOI: 10.1002/hpm
IMPLEMENTATION OF COMMUNITY HEALTH FUND IN TANZANIA
e375
managed by the DMO. Furthermore, the vast majority of respondents in Iramba district reported that in order to allow members to obtain services in any public primary health facility in the district, it was important for the funds to be managed at the district level. Otherwise, health services would have been limited to one primary healthcare facility closest to the domicile of the member. What works and under what conditions?. In Lindi district, respondents across all levels of the district health system reported that CHF enrolment was extremely low. This matches with the analysis of minutes and the CHF registers found in the health facilities. It was evident that in almost all health facilities, the number of enrolment had been decreasing yearly (Table 5). Almost all respondents at the district and facility levels attributed low CHF enrolment rate to frequent drug stock-out in most of the facilities. This discouraged many households from joining the scheme and/or renewing their membership. ‘At this facility, there are very few people enrolled in CHF. When we started the scheme, many people joined. Many people dropout because of the frequent drug stock-out. When there are no drugs, we tell them to buy in the privately owned pharmacies. That is where the problem starts. They ask, ‘why do you tell us to buy drugs while we have already paid our money’? Consequently, many people have dropped out. Currently, there is only one person who has renewed membership’ (IDI with a health facility in-charge). Furthermore, few respondents pointed out that the community members do not trust the scheme managers and were not sure whether the funds collected would be managed properly. Some respondents remarked this way: ‘The main problem is the modality of collecting CHF contributions. In the current system, funds are collected by the in-charges of the health facilities. Community members have lost trust on them because of their behaviour’ (IDI with CHSB member). ‘Community members have lost trust on the health providers. They are not sure whether their money would be managed properly’ (IDI with health facility committee member). Our analysis further revealed that local politics affected CHF enrolment. In one health facility, it was reported that the opposition political party was very strong Table 5. CHF enrolment in the study facilities in Lindi district Facility name 1 2 3 4 5 6
Rutamba Health Centre Kitomanga Health Centre Kilangala Dispensary Mahumbika Dispensary Mchinga Dispensary Mnolela Dispensary
2009
2010
2011
2012
60 8 6 10 7 15
67 24 4 8 5 12
33 35 1 7 3 10
12 13 1 8 0 5
CHF, Community Health Fund. Copyright © 2013 John Wiley & Sons, Ltd.
Int J Health Plann Mgmt 2014; 29: e368–e382. DOI: 10.1002/hpm
e376
S. O. MALUKA AND G. BUKAGILE
and that the community members were sensitised not to contribute for the provision of the social services. The community members were informed that it is the responsibility of the government to provide free social services to the citizens. One respondent expressed this way: ‘Politics affects CHF enrolment. The entire village leadership is from the opposition political party. The villagers are sensitised not to pay for the health services. Instead, the government is supposed to provide free health services’ (IDI with incharge of a health facility). Conversely, in Iramba district, all respondents from the district to village levels reported that CHF enrolment had increased considerably in recent years. As a result, the number of out-of-pocket payments had noticeably gone down. Analysis of the minutes of the health facility committees and CHF registers confirmed this point. As Table 6 indicates, the CHF enrolment in the study facilities had been increasing over the past 3 years. A number of factors facilitated the uptake of the CHF in Iramba district. First, the CHMT had made efforts to ensure that drugs were available in all facilities and that patients were not required to purchase drugs. The most important approach used to ensure availability of drugs was frequent monitoring and auditing of drugs in the health facilities. According to CHMT members, this significantly helped to reduce irrational prescription and improper management of drugs by the health providers. One respondent remarked this way: ‘We have realised that the drugs we get from the Medical Stores Department are enough. There were problems of mismanagement of drugs by the health providers. Sometimes drugs were mismanaged by the health providers and even found in the privately owned shops. Through frequent auditing, we have managed to overcome this problem. Some health providers were given warning’ (IDI with a CHMT member). Frequent drug audit also made it possible for the CHMT members to identify drugs that were not needed at the facility. These were quickly sent to other facilities that were in need of such drugs.
Table 6. CHF enrolment in the study facilities in Iramba district Facility name
1 2 3 4 5 6 7
Ndago Health Centre Kinyangiri Health Centre Mampanta Dispensary Bomani Dispensary Misigiri Dispensary Ulemo Dispensary Total district enrolment
Number of households 2009
2010
2011
216 182 100 81 53 103 5053
163 190 132 141 79 192 3076
233 349 225 206 81 425 4890
CHF, Community Health Fund. Copyright © 2013 John Wiley & Sons, Ltd.
Int J Health Plann Mgmt 2014; 29: e368–e382. DOI: 10.1002/hpm
IMPLEMENTATION OF COMMUNITY HEALTH FUND IN TANZANIA
e377
In both study districts, respondents reported that drugs were delivered through the Integrated Logistics System. The health facilities sent orders to the district on a quarterly basis. The districts reviewed and approved orders, and sent them to the Medical Stores Department (MSD). The MSD processed the orders and debited the facility accounts. The districts could allocate supplementary funds from local sources. In Iramba district, the district health officials reported that they had managed to ensure that drugs were ordered in time. Further, the district had a small buffer stock of drugs at the district hospital. The health facilities that needed drugs could borrow from the district hospital. These had to be replaced when the health facility obtain its share from MSD. However, it was unanimously reported that of recent, especially from July 2012, the district experienced shortage of drugs mainly attributed to MSD. All district health managers reported that MSD had not delivered drugs to the district for almost 8 months. ‘Of recent, the district has been experiencing drug stock-out, particularly from July this year. This problem is mainly caused by the MSD. The MSD has not supplied us drugs for the past eight months. For almost three years we had no problem of drug stock-out’ (IDI with CHMT member) Another respondent added: ‘The main problem is the Medical Stores Department (MSD). You know, if you are a sole provider there is high possibility of becoming sluggish. For example, the district may send invoice to the MSD and you are told that drugs are available. But once you deposit funds to their account, you are told we have no drugs. The bad thing is that once funds have been deposited they cannot be returned. At the end of the day, you are forced to select other items which you did not order. In my opinion, there is need for another drug supplier who can compete with the MSD’ (IDI with a CHMT member). Second, the district council increased the amount of user fees from TSh 1000 to TSh 3000 at the dispensary level, and from TSh 1500 to TSh 4000 at the health centre. At the same time, the premium for the CHF remained TSh 5000 per household. The majority of respondents felt that the increase of the amounts for out-pocket payments, coupled with the availability of drugs at the health facility, significantly increased the enrolment of the CHF. They also reported that the number of outpocket payments had dropped. This approach began in January 2012 after being approved by the Full Council. The idea was initiated by the CHMT and forwarded to the Full Council for approval through the CHSB. By the time this study was conducted, the CHSB was planning to revise the amount of premium for CHF from TSh 5000 to TSh 10000. According to respondents, this would enable the district to expand CHF benefits to include in-patient secondary-level (hospital) services. Third, almost all respondents reported that there had been frequent sensitisation of the community members. Most often, sensitisations were done by the health facility committee members through village meetings. The CHSB also occasionally visited the health facilities. Analysis of all sources of data revealed that the health facility committees in Iramba district held meetings monthly to discuss various issues pertaining to the performance of the health facilities. It was evident in almost all facilities that implementation of the CHF was one of the permanent agenda in the Copyright © 2013 John Wiley & Sons, Ltd.
Int J Health Plann Mgmt 2014; 29: e368–e382. DOI: 10.1002/hpm
e378
S. O. MALUKA AND G. BUKAGILE
health facility committee meetings. Records of the meetings were available in all health facilities that the researchers visited. Two factors made it possible for the committees to hold their meetings as required. First, the district council managed to ensure that the health facility committee members were paid incentive to compensate their time. The chairperson and the secretary of the committee were paid TSh 3000 whereas other members were paid TSh 2000 after every meeting. Second, the CHMT had established supervision and monitoring system that required all health facilities in the district to submit minutes of the meetings along with other reports to the CHMT on monthly basis, from the first to the sixth date of every month. Closely linked to this, incentives were provided to the CHF coordinator. The CHF coordinator had been provided with a motorcycle and a laptop to enable him or her to better perform his or her responsibilities our respondents said. This had facilitated sensitisation and supervision of the CHF as well as keeping proper records related to the operations of the CHF. By contrast, in Lindi district, the committees held their meetings quarterly. The chairpersons and secretaries of the health facility committees were paid TSh 15 000 per meeting whereas other members were paid TSh 10 000. However, further analysis of the interviews across all respondents revealed that sitting allowances were hardly paid or were paid very late, sometimes even after three or four meetings (three to four quarters). Consequently, some members of the committees did not attend the meetings as required. Furthermore, in Iramba district, respondents at all levels of the district health system reported that in order to increase revenues, the district council decided to limit the number of exemptions by encouraging the village councils to pay for the elders who deserve exemptions. The elderlies who are 60 years and older were grouped by 10s to form households and were issued with CHF cards purchased by the village governments. The district council, therefore, benefited from the matching funds provided by the central government to all CHF beneficiaries.
DISCUSSION This paper has described the experiences of the district-level and local-level actors in implementing CHF in Tanzania and has illustrated a number of factors for the diverse implementation of the fund in the study districts. In this section, we expand on the most important issues raised by the respondents with reference to a wider literature on the implementation of health policy in developing countries. The findings indicated that in Lindi district, the uptake of the CHF was very low. On the contrary, in Iramba district, the performance was better. Availability of drugs in the health facilities, effective supervision, commitment of the district health management team and local government officials, and incentives to the health facility committees and boards were cited by respondents as important. Therefore, these are the factors that provide the most convincing explanation of why CHF perform better in some districts than in others. The finding indicated that Iramba district increased user fees as a mechanism to attract the population to join CHF instead of paying fees. A similar approach has Copyright © 2013 John Wiley & Sons, Ltd.
Int J Health Plann Mgmt 2014; 29: e368–e382. DOI: 10.1002/hpm
IMPLEMENTATION OF COMMUNITY HEALTH FUND IN TANZANIA
e379
been recently reported in other districts in Tanzania. For example, in Mpwapwa district, user fees were raised to TSh 4000 for hospitals, TSh 3000 for health centres and TSh 2500 for dispensaries (Stoermer and Macha, 2009). Although in Iramba district the evidence suggests that the CHF enrolment increased after the increase of user fees, this mechanism poses equity concerns. Those most at risk may be left without accessibility, not being able to pay either the enrolment premium or the user fees (Stoermer and Macha, 2009). As this approach seems to gain momentum across districts in Tanzania, equity implications need to be further examined. It would be important to examine trends in the utilisation of health services by the poor and vulnerable groups in districts where user fees have been increased. Effective supervision and commitments of the district health management teams were also mentioned by respondents as important factors for the good performance of CHF in Iramba district. Conversely, inadequate supervision and low commitment of the district health managers in Lindi district resulted into low enrolment. This finding is in line with the previous studies in Tanzania, which found that lack of commitment of some of the top managers at the district affected the implementation of the policy (Kamuzora and Gilson, 2007; Maluka, 2013). On the other hand, there are examples where the commitment and support of the district managers resulted into relative success of the CHF scheme. The Kongwa, Kondoa and Dodoma urban districts’ success has been seen to be due in part to the improved cooperation between the district health board, councillors and other leaders who together mobilise the community to join CHF, as well as through community groups (Stoermer et al., 2012). In Mpwapwa district, the personality of the DMO was reported as one of the factors leading to the good performance of CHF (Stoermer and Macha, 2009). The variation in policy implementation reported in this study suggests that the district managers and the health workers were able to exercise considerable discretion in the implementation of the CHF scheme, and it is largely their actions that have determined how the scheme took shape on the ground. The actions of the district-level actors may be seen as coping strategies to overcome the problems on them from above (Powell-Jackson et al., 2009). Bottom-up implementation theories have argued that in the face of resource uncertainties, policy implementers adopt coping behaviours to manage the high demands and time pressures that they face; and through these behaviours, they may re-interpret and reshape implementation process (Lipsky, 1980; Hill, 1997). It can reasonably be argued that district health officials and locallevel actors who attempted to deal with challenges were often driven by pressure to meet local needs. This finding reiterates the need of actively engaging the policy implementers in the design, management and implementation of policies (Kamuzora and Gilson, 2007). Policies need to be socially acceptable to the groups affected, and ideally, policy implementers, including the public, should be involved in the formulation process (Birt et al., 1997). Closely linked to this, our analysis suggests that financial incentive to the health facility committees and the board members facilitated good performance of CHF in Iramba district. The district council had in essence managed to ensure that the health facility committee members were paid incentives on time. This motivated members of the health facility committees to hold their monthly meetings as Copyright © 2013 John Wiley & Sons, Ltd.
Int J Health Plann Mgmt 2014; 29: e368–e382. DOI: 10.1002/hpm
e380
S. O. MALUKA AND G. BUKAGILE
required. The fact that the health facility committees held their meeting monthly made it possible to frequently discuss the implementation of the CHF in their respective health facilities. The district council also provided incentives to the CHF coordinator to enable him or her to better perform his or her responsibilities. On the other hand, in Lindi district, while the health facility committees were in place, meetings were held quarterly, and the majority of the members were demoralised because of the late and or non-payment of the sitting allowances. The district health managers need to design incentive mechanisms that would reinforce an effective implementation of the CHF. However, although financial incentives are important, the deciding factors are the will and ambition of the individuals and decision-making bodies within the district healthcare structures. In contrast, the finding indicated that CHF enrolment rate in Lindi district was extremely low. This was partly due to the high premium rates. The district health officials in Lindi, in collaboration with the health facility committees and boards, as well as the ward and village leaders, should put efforts in encouraging communities to join the CHF scheme. It is imperative to improve collaboration of the local-level leaders who are responsible for the implementation of the fund. The village and ward leaders need to be sensitised about the importance of the CHF in protecting the households against medical catastrophe and impoverishment. Additionally, the district council should consider the possibility of reducing the current CHF contribution rate of TSh 10 000 to TSh 5000. This would encourage many households to join the scheme, and the district would benefit from the matching grant provided by the central government. Evidence shows that in Mwanga district, membership steadily rose following the reduction of the premium rate after 6 months since the initial implementation from Tsh 10 000 to 5000 in 2006 (URT, 2007). Likewise, in Rombo district, in order to attract many households to join the CHF scheme, the board reduced CHF contribution from the former TSh 10 000 to affordable amount of TSh 5000 per household per year. This increased enrolment, reaching approximately 25% of the existing potential (Stoermer et al., 2012). Furthermore, in both study districts, frequent drug stock-out in the facilities constrained the implementation of the CHF. Drug stock-out was, for the most part, attributed to the inadequate and late supply from the MSD. Drug stock-out seems to be a persistent problem in Tanzania. Experience shows that despite the transition to an Integrated Logistics System from the one based on the distribution of standard essential medicine kits as of 2009, many health facilities in the country still face the problem of irregular supply of medicines, equipment and medical supplies from the MSD. It is evident that the Medical Store Department face challenges to efficiently and timely supply drugs to all facilities in the country. The researchers did not interview respondents from the MSD on this matter. However, evidence suggests that the main constraints facing the MSD include increased demand of drugs and medical supplies, limited financial and human resources, and long and bureaucratic procurement procedures (Mwananchi Cooperation, 2013). The Ministry of Health and Social Welfare should increase the capacity of the MSD and review the drug procurement procedures. In addition, the Ministry of Health and Social Welfare should consider the possibility of diversifying drug supply system in the country by opening opportunities for other government-subsidised drug suppliers. Copyright © 2013 John Wiley & Sons, Ltd.
Int J Health Plann Mgmt 2014; 29: e368–e382. DOI: 10.1002/hpm
IMPLEMENTATION OF COMMUNITY HEALTH FUND IN TANZANIA
e381
LIMITATIONS OF THE STUDY This study relied primarily on the review of minutes, health facility visits and key informant interviews with the community representatives and the district health managers. First, all data collection was completed before the analysis of data was undertaken. We recognise that a more iterative process of data collection and analysis would have improved the interview guides and experience of the interviewers. Second, the study did not interview the members of the CHF in order to assess their experiences in utilising health services and the factors constraining the enrolment of the fund. Future studies should focus on these aspects in order to understand better the actual practices of CHF scheme in Tanzania. Similarly, future studies should explore equity implications in the districts, which have recently increased user fees in order to motivate households to join CHF. Despite these limitations, the study provides a detailed account of the implementation of CHF in Tanzania.
CONCLUSION This study marked an important attempt in providing a deeper and more nuanced understanding of the implementation of CHF in Tanzania. The large variation in policy implementation reported in this study suggests that district managers, health workers and local leaders were able to exercise considerable discretion in the implementation of the CHF scheme, and it is largely their actions that have determined how the scheme took shape on the ground. The study confirms that the success of the CHF largely depends on the availability of health services, effective supervision mechanisms, commitment of the district health management team and local government officials, and incentives for the health facility committees and board members. The focus of implementation needs to be placed on the active engagement of the district health managers and local-level leaders such as counsellors, ward and village leaders, as well as health facility committees and boards who are responsible for the implementation of the policies. Equally important is the availability of quality health services in the health facilities. In particular, availability of drugs and medical supplies seems to be the most important way forward. The Ministry of Health and Social Welfare should increase the capacity of the MSD and review the drug procurement procedures. In addition, the Ministry of Health and Social Welfare should consider the possibility of diversifying drug supply system in the country by opening opportunities for other government-subsidised drug suppliers.
ACKNOWLEDGEMENTS We would like to express my gratitude to the Policy Research for Development, formerly Research on Poverty Alleviation (REPOA), for partially funding this study. We also thank local government officials, district health authorities and other stakeholders in the study districts for participating in the study. Further, we are thankful to our research assistants for assisting us in the data collection. The authors have no competing interests. Copyright © 2013 John Wiley & Sons, Ltd.
Int J Health Plann Mgmt 2014; 29: e368–e382. DOI: 10.1002/hpm
e382
S. O. MALUKA AND G. BUKAGILE
REFERENCES Birt CA, Gunning-Schepers L, Hayes A, et al. 1997. How should public health policy be developed? A case study in European Public Health. J Public Health Med 19(3): 262–267. Burns M, Mantel M. 2006. Tanzania review of exemptions and waivers, Report submitted to the Ministry of Health by the Euro Health Group. Devadasan N, Criel B, Van Damme W, Ranson K, Van der Stuyft P. 2007. Indian community health insurance schemes provide partial protection against catastrophic health expenditure. BMC Health Serv Res 7: 43. Gilson L. 1997. The lessons of user fee experience in Africa. Health Policy Plan 12(4): 273–285. Hill M. 1997. The Policy Process in the Modern State. Prentice Hall/Harvester Wheatsheaf: Harlow. Jacobs B, Bigdeli M, van Pelt M, Ir P, Salze C, Criel B. 2008. Bridging community-based health insurance and social protection for health care—a step in the direction of universal coverage? Trop Med Int Health 13(2): 140–143. James C, Hanson K, McPake B, et al. 2006. To retain or remove user fees?: reflections on the current debate in low and middle-income countries. Appl Health Econ Health Policy 5(3): 137–153. Kamuzora P, Gilson L. 2007. Factors influencing implementation of the Community Health Fund in Tanzania. Health Policy Plan 22: 95–102. Kessy F. 2008. Technical Review of Council Health Service Boards and Health Facility Committees in Tanzania. Report prepared for the MOHSW with financial support from DANIDA and SDC. Lipsky M. 1980. Street-Level Bureaucracy: Dilemmas of the Individual in Public Services. Russell Sage Foundation: New York. Maluka SO. 2013. Why are pro-poor exemption policies in Tanzania better implemented in some districts than in others? Int J Equity Health 12: 80. McIntyre D, Thiede M, Dahlgren G, Whitehead M. 2006. What are the economic consequences for households of illness and of paying for health care in low- and middle-income country contexts? Soc Sci Med 62: 858–865. Ministry of Health (MoH). 1999. Community Health Fund Operations Guidelines. Ministry of Health: Dar es Salaam.
Copyright © 2013 John Wiley & Sons, Ltd.
Ministry of Health and Social welfare (MoHSW). 2012. Health sector public expenditure review 2010/2011. Directorate of Policy and Planning—Ministry of Health and Social Welfare: Dar es Salaam. Mtei G, Mulligan J. 2007. Community Health Funds in Tanzania: a literature review. Report submitted to Ifakara Health Research and Development Centre. Mwananchi Cooperation. 2013. Challenges in the supply of drugs in Tanzania: who is to blame? Mwananchi Cooperation Friday 19 April. Ouimet MJ, Fournier P, Diop I, Haddad S. 2007. Solidarity or financial sustainability: an analysis of the values of community-based health insurance subscribers and promoters in Senegal. Can J Public Health 98(4): 341–346. Powell-Jackson T, Morrison J, Tiwari S, Neupane B, Costello AM. 2009. The experiences of districts in implementing a national incentive programme to promote safe delivery in Nepal. BMC Health Serv Res 9: 97. Ritchie J, Spencer L, O’Connor W. 2003. Carrying out qualitative analysis. In Qualitative Research Practice: A Guide for Social Science Students and Researchers, Ritchie J, Lewis J (eds). Sage Publications: London; 219–262. Stoermer M, Macha J. 2009. External Review of the Project “Health Insurance for the Rural Population (Tanzania)”. Report submitted to the Swiss Development Cooperation. Stoermer M, Hanlon P, Tawa M, Macha J, Mosha D. 2012. Community Health Funds in Tanzania: innovations study, Report submitted to GIZ. United Republic of Tanzania (URT). 2001. The Community Health Fund Act. United Republic of Tanzania: Dar es Salaam. United Republic of Tanzania (URT). 2007. Community Health Fund: Best Practices Workshop Report. Ministry of Health and Social welfare: Dar es Salaam. Watkins K. 1997. Cost-Recovery and Equity in the Health Sector: Issues for Developing Countries, Working paper, Oxfam. World Health Organization. 2010. World Health Report 2010—Health Systems Financing: The Path to Universal Coverage. World Health Organization: Geneva. Yin RK. 2003. Case Study Research: Design and Methods. Sage Publications Inc.: Thousand Oaks, CA.
Int J Health Plann Mgmt 2014; 29: e368–e382. DOI: 10.1002/hpm
Int J Health Plann Mgmt 2014; 29: e368–e382. Published online 7 November 2013 in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/hpm.2226
Implementation of Community Health Fund in Tanzania: why do some districts perform better than others? Stephen Oswald Maluka1* and Godfrey Bukagile2 1
Institute of Development Studies, University of Dar es Salaam, Dar es Salaam, United Republic of Tanzania 2 Development Studies, College of Business Education, Dar es Salaam, United Republic of Tanzania
SUMMARY In early 1990s, Tanzania, like other African countries, introduced user fees in public health systems. Although user fees were considered important in promoting health, they appear to reduce people’s access to health services. To counteract the detrimental effects of the user fees, various types of health insurances were introduced, including the Community Health Fund (CHF). Drawing from the review of minutes, health facility visits and key informant interviews, this study explored why implementation of the CHF in Tanzania has been more successful in some districts than in others. The findings indicate that in Lindi district, the enrolment rate for the CHF was very low. This was attributed to high premium rates, frequent drug stock-out, lack of trust by the community members to the health providers, low incentives and local politics. In contrast, in Iramba district, the performance was better. Availability of drugs in the health facilities, effective supervision, commitment of the top district-level officials and incentives to the health facility committees were the main factors that facilitated good performance of the fund in Iramba district. The focus of the implementation needs to be placed on the active engagement of the local-level leaders and politicians who are responsible for the implementation of the policy. Equally important is the availability of quality health services in the health facilities. Copyright © 2013 John Wiley & Sons, Ltd. KEY WORDS: Community Health Fund; health policy; policy implementation; health facility committees; Tanzania
INTRODUCTION In the early 1990s, Tanzania, like many other developing countries in Africa, was caught up in the web of reforms in the health and social services sectors that saw the adoption of the user fees for healthcare financing in public health systems. Although user fees were considered important in promoting health and economic *Correspondence to: S. O. Maluka, Institute of Development Studies, University of Dar es Salaam, Dar es Salaam, United Republic of Tanzania. E-mail: [email protected]
Copyright © 2013 John Wiley & Sons, Ltd.
IMPLEMENTATION OF COMMUNITY HEALTH FUND IN TANZANIA
e369
performance, they appear to reduce people’s access to health services especially the poor and the most vulnerable groups of the society (Gilson, 1997; Watkins, 1997; James et al., 2006) and sometimes lead to catastrophic health expenditures (McIntyre et al., 2006; Devadasan et al., 2007; WHO, 2010). To counteract the detrimental effects of the user fees, developing countries have vigorously been promoting the use of risk-sharing arrangements for healthcare financing, which makes healthcare payments predictable especially for the rural poor. Various types of health insurances were introduced. The first type of insurance is national or social health insurance, which is based on individual’s mandatory enrolment. The second is voluntary mechanisms, which include private health insurance and community-based health insurance. Community-based health insurance schemes are based on the premises of risk pooling and community solidarity to risks of falling sick and are conceptually designed to provide financial protection and reduce out-of-pocket payments for healthcare (Ouimet et al., 2007; Jacobs et al., 2008). In Tanzania, the government introduced the Community Health Fund (CHF) in 1995 as a mechanism for providing additional funds for financing health services and protecting households from out-of-pocket payments as a result of the implementation of user fees in the health sector (MoH, 1999). In 1996, the scheme was first piloted in Igunga district. In 1998, the Ministry of Health (MoH) decided to expand the CHF to nine additional districts. In 2001, the CHF Act was passed by the parliament, establishing the CHF as the official health plan at the local/community level (URT, 2001). The CHF is a district-level voluntary pre-payment scheme that targets the population living in rural areas and/or employed in the informal sector. The scheme is based on the concept of risk sharing whereby members pay a small contribution on a regular basis to offset the risk of needing to pay a much larger amount if they fall sick (Mtei and Mulligan, 2007). Table 1 provides details of the CHF in Tanzania. Low enrolment remains one of the biggest challenges to the CHF. By 2011, CHF had enrolled a total of 561 370 households, which was only 7.4% of the total population (MoHSW, 2012). Various factors have contributed to low enrolment, including perceived poor quality of services in public facilities, a widespread inability to pay membership contributions, the limited benefit package and lack of access to referral facilities (Kessy, 2008). These factors are compounded by weak management, lack of trust in management and a failure among communities to see the rationale for protecting themselves against the risk of illness (Kessy, 2008). However, Kamuzora and Gilson (2007) found that district health managers had a direct influence over the factors explaining low enrolment of the CHF. According to these authors, the district health managers did not allocate a budget for CHF administration activities and did not ensure supervision of health staff to support delivery of quality services (Kamuzora and Gilson, 2007). Although a body of evidence indicates that the CHF is ineffective in its implementation, there are wide variations between districts, suggesting that implementation has been more successful in some places than in others (Burns and Mantel, 2006). A recent review revealed numerous innovative refinements to enrolment and health service quality improvements financed by the CHF (Stoermer et al., 2012). Against this background, the study explored why the implementation of CHF in Tanzania has Copyright © 2013 John Wiley & Sons, Ltd.
Int J Health Plann Mgmt 2014; 29: e368–e382. DOI: 10.1002/hpm
e370
S. O. MALUKA AND G. BUKAGILE
Table 1. Details of the CHF in Tanzania (URT, 2001) Payment and benefits
Organisation
Membership based on household enrolment One membership card is issued per household and is valid for a period of 12 months The annual contribution from each household is defined by the districts The CHF schemes subsidised by the government in the form of matching grants, complementing the members’ contributions with an equal amount Membership allows the household access to medical services without further co-payments at the primary level Inclusion of secondary-level and tertiary-level healthcare is at the discretion of the district Households without a CHF membership must pay the predetermined user fees to access healthcare Poor households who are unable to pay may be issued CHF membership or an exemption letter upon recommendation by the village council At the district level, the CHF is overseen by the CHSB with representatives from the district authorities, public healthcare provider, private healthcare provider and the community There is usually a CHF coordinator responsible for tracking membership levels and funds contributed across the district At the ward level, the Ward Development Committee and the Health Facility Committee are responsible for mobilising people to join the CHF, tracking the membership base, overseeing premium collections, evaluating CHF operations and providing recommendations and granting exemptions
CHF, Community Health Fund; CHSB, Council Health Service Board.
been more successful in some places than in others. It is expected that this study will contribute to the small body of literature on health policy implementation in developing countries. The study adopted descriptive approach, seeking to provide indepth analysis of the factors that influence the performance of the CHF in Tanzania. The study focused on the experiences of various actors involved in the implementation of the CHF at the district and community levels in order to give a bottom-up perspective of the implementation process.
METHODOLOGY The study design and settings An exploratory case study design focusing on two districts—one relatively high performing (Iramba) and the other fairly low performing (Lindi)—was used (Yin, 2003). The case study approach was considered the appropriate method for this study because CHFs are inherently a complex intervention. They are dependent on the context in which they are implemented, and their implementation is not standardised. They also have a strong social component in that they are designed to promote and protect the health of the population and to reduce inequalities. The ‘high’ and ‘low’ performing categorisations were based on the enrolment of the Copyright © 2013 John Wiley & Sons, Ltd.
Int J Health Plann Mgmt 2014; 29: e368–e382. DOI: 10.1002/hpm
e371
IMPLEMENTATION OF COMMUNITY HEALTH FUND IN TANZANIA
Table 2. Key characteristics of the study settings
Population Community Health Fund enrolment rate* Hospitals Health centres Dispensaries Divisions Wards Villages Health workers available Shortage of health workers
Iramba district
Lindi district
236 282 people 28.1% 2 6 60 7 31 143 43% 57%
194 143 people 0.4% 1 6 30 9 30 134 50% 50%
*The national Community Health Fund coverage rate is about 7.9% (National Health Insurance Fund data of 2011).
Table 3. The sampled study sites
1 2 3 4 5 6 7 8 9 10 11 12 13
Name of the facility
Type of the facility
District
Kiomboi Ndago Kinyangiri Bomani Mampanta Misigiri Ulemo Kitomanga Rutamba Mchinga Kilangala Mahumbika Mnolela
Hospital Health centre Health centre Dispensary Dispensary Dispensary Dispensary Health centre Health centre Dispensary Dispensary Dispensary Dispensary
Iramba
Lindi
CHF.1 However, the districts were not selected on the basis of performance only; other factors such as accessibility were taken into account. Table 2 summarises key characteristics of the study settings. In each district, one hospital, two health centres and four dispensaries were purposively selected, in collaboration with the district medical office, to include only those at which the health facility committees and head of the health facility have been in operation for at least 1 year (Table 3). Data collection techniques This study was primarily based on the individual interviews with the key stakeholders. Documentary data and health facility visits were used to support, verify 1
CHF performance data were obtained from the Budget Speech of the Ministry of Health and Social Welfare 2011/2012. Copyright © 2013 John Wiley & Sons, Ltd.
Int J Health Plann Mgmt 2014; 29: e368–e382. DOI: 10.1002/hpm
e372
S. O. MALUKA AND G. BUKAGILE
and highlight the key issues that emerged. In Iramba, individual interviews were carried out in October 2012, whereas in Lindi, the study was conducted in January 2013. Each interview lasted for approximately 45 min and was carried out at the respondent’s workplace and/or home. In order to cover a wide range of views of different actors, a purposive sampling technique was used. The purpose of key informant interviews was to collect information from a wide range of people who have first-hand knowledge about the CHF operations. An interview guide was developed to assist the semi-structured interviews with key respondents. The use of open-ended questions permitted the participants to express their views and elaborate on issues that they felt were most relevant and important. At the district level, members of the Council Health Management Team (CHMT) and the Council Health Service Board (CHSB) were interviewed. At the health facility level, committee members at the district hospital and health centres as well as ward and village leaders were interviewed. In total, 83 interviews were carried out (51 interviews in Iramba and 32 interviews in Lindi) (Table 4). Furthermore, considerable documentary information was obtained, and our analysis was validated with observations at the health facilities. Data analysis Quantitative data that were recorded in numerical form were entered into an Excel worksheet for secondary analysis of totals, percentages and other such simple statistics as necessary. Thematic approach was adopted to analyse qualitative data (Ritchie et al., 2003). First, all the interviews were transcribed verbatim into Swahili. Second, the transcriptions were translated into English. Third, interview transcripts were entered into NVIVO 10 software (QSR International Pty Ltd, Australia) for storage, coding, text search and retrieval. Using NVIVO 10 software, data were coded to initial themes. Thereafter, data were sorted and grouped together under patterns that were more precise, complete and generalisable. As patterns of meaning emerged, the researchers searched for similarities and differences in performance between the two districts involved. Finally, the data were summarised and synthesised, retaining as much as possible key terms, phrases and expressions of the respondents. After this Table 4. Number of key informant interviews Category of respondents
Council Health Service Board Council Health Management Team Hospital governing committee Health centre and dispensary committees Councillors Village executive officers and village leaders Heads of health facility and health workers Social welfare officers Total Copyright © 2013 John Wiley & Sons, Ltd.
Number interviewed Iramba
Lindi
4 3 3 17 2 8 12 2 51
3 3 — 15 — 4 7 — 32
Int J Health Plann Mgmt 2014; 29: e368–e382. DOI: 10.1002/hpm
IMPLEMENTATION OF COMMUNITY HEALTH FUND IN TANZANIA
e373
analysis, data were triangulated to allow comparison across different categories of respondents for final interpretation and presentation of the results.
ETHICAL ISSUES The research was approved by the University of Dar es Salaam, which has been given the mandate by the Government of Tanzania and Tanzania Commission for Science and Technology to grant research clearance to its students and staff. The clearance was presented to the regional and district authorities who approved the study in their area. Verbal consent was sought from prospective informants after the researchers had explained the objectives of the study. Informants were informed of their right to withdraw from the study any time they wished. They were also assured of confidentiality of any information they would give. All interviews were tape recorded with the permission of participants, and the resulting recordings and transcripts were stored in a manner that protected their confidentiality. For example, no respondent’s name was recorded.
FINDINGS The following section presents the main findings of the study organised in three main themes: enrolment and the benefit package; how funds were collected, managed and used; and what works, how and under what conditions. Enrolment process and the benefit package It was found out that in Lindi district, households were required to contribute TSh 10 000 (equivalent of $6.5) annually. This premium covered a maximum of six members per household. Community members, who were neither CHF nor National Hospital Insurance Fund members, were offered health services through user fees, which were payable at the rate of TSh 1000 at the dispensary and TSh 2000 at the health centre level per visit. The CHF benefits were limited to one primary healthcare facility (dispensary or health centre) closest to the domicile of the member and were not transferable to other healthcare providers in the district. On the other hand, in Iramba district, it was found that households were required to contribute TSh 5000 (equivalent of $3.5) annually. This premium covered a maximum of six members of the household, a couple and their children below 18 years of age. All CHF members could obtain healthcare services in any public primary health facility in the district provided that they showed evidence that they were active members. The district had also extended CHF benefits to include secondary-level services. However, secondary-level services were limited to ambulance and the outpatient department. If the CHF member was admitted, she or he had to bear all the admission costs. When referral was made to the district hospital, a patient had to be given a referral letter from the primary healthcare facility. There was no reimbursement to the district hospital for services rendered to the CHF members. Copyright © 2013 John Wiley & Sons, Ltd.
Int J Health Plann Mgmt 2014; 29: e368–e382. DOI: 10.1002/hpm
e374
S. O. MALUKA AND G. BUKAGILE
How were funds collected, managed and used? In both study districts, funds were collected by the heads of the health facilities and submitted to the district accountant on a monthly basis. The funds collected from the memberships were used for health-related purposes and activities by the district, partly flowing back to the health facilities. At the district, funds were deposited in account number 6, managed by the District Medical Officer (DMO). In Lindi district, 40% of the funds flowed back to the health facilities. Similarly, in Iramba district, 20% of the funds were returned to the facilities on monthly basis. At the facilities, the CHF funds were used for minor rehabilitations and purchasing of supplementary drugs and medical supplies. In Iramba district, the CHF funds were also used to pay incentives to the members of the health facility committees, health workers and salaries for watchmen. Similarly, CHF funds kept at the district were used for purchasing supplementary drugs and medical supplies as well as paying incentives to the members of the boards. The majority of the CHMT members also reported that a good number of health facilities had been installed with solar power using CHF funds. The review of the minutes and Comprehensive Council Health Plans supports this point. It was revealed that out of 35 health facilities that had electricity, 28 used solar power, and seven had been connected to the national grid. In the 2012/2013 financial year, the CHMT had planned to install six more health facilities with solar power using CHF money. However, it was found out that in both study districts, heads of the health facilities and members of the committees felt that they had little control of the funds. They pointed out that, apart from the small percentage of funds that were returned to the facility, they did not know how the remaining percentage of CHF funds was used. According to the respondents, this reduced their morale. One respondent commented this way: ‘Although we collect CHF funds, we are not informed how the funds are used. I once raised this concern to the CHF coordinator. But I was told that CHF is a district scheme, and funds are managed at the district level. I told him that he was demoralising us. The household contributes in order to get better services at their facility. When you use our money to improve services in the facilities which have not contributed, you discourage people from renewing their membership’ (IDI with in-charge of health facility in Iramba). In contrast, the majority of the district health officials had different opinions on this point. First, they felt that the health facilities had no capacity to manage CHF money. In particular, health facilities had no trained accountants to enable them to manage a large amount of funds. Second, both health providers and the health facility committees were not permanently based at the facility. Membership for the health facility committees expired after 3 years. Similarly, health providers could be transferred to other facilities. As a result, respondents said that it would be difficult to control the management of funds. Third, in order to obtain matching funds from the central government, the district had to pool together contributions from all health facilities. The government does not pay matching grants to individual health facilities. According to the CHMT members, it is important to have a single account Copyright © 2013 John Wiley & Sons, Ltd.
Int J Health Plann Mgmt 2014; 29: e368–e382. DOI: 10.1002/hpm
IMPLEMENTATION OF COMMUNITY HEALTH FUND IN TANZANIA
e375
managed by the DMO. Furthermore, the vast majority of respondents in Iramba district reported that in order to allow members to obtain services in any public primary health facility in the district, it was important for the funds to be managed at the district level. Otherwise, health services would have been limited to one primary healthcare facility closest to the domicile of the member. What works and under what conditions?. In Lindi district, respondents across all levels of the district health system reported that CHF enrolment was extremely low. This matches with the analysis of minutes and the CHF registers found in the health facilities. It was evident that in almost all health facilities, the number of enrolment had been decreasing yearly (Table 5). Almost all respondents at the district and facility levels attributed low CHF enrolment rate to frequent drug stock-out in most of the facilities. This discouraged many households from joining the scheme and/or renewing their membership. ‘At this facility, there are very few people enrolled in CHF. When we started the scheme, many people joined. Many people dropout because of the frequent drug stock-out. When there are no drugs, we tell them to buy in the privately owned pharmacies. That is where the problem starts. They ask, ‘why do you tell us to buy drugs while we have already paid our money’? Consequently, many people have dropped out. Currently, there is only one person who has renewed membership’ (IDI with a health facility in-charge). Furthermore, few respondents pointed out that the community members do not trust the scheme managers and were not sure whether the funds collected would be managed properly. Some respondents remarked this way: ‘The main problem is the modality of collecting CHF contributions. In the current system, funds are collected by the in-charges of the health facilities. Community members have lost trust on them because of their behaviour’ (IDI with CHSB member). ‘Community members have lost trust on the health providers. They are not sure whether their money would be managed properly’ (IDI with health facility committee member). Our analysis further revealed that local politics affected CHF enrolment. In one health facility, it was reported that the opposition political party was very strong Table 5. CHF enrolment in the study facilities in Lindi district Facility name 1 2 3 4 5 6
Rutamba Health Centre Kitomanga Health Centre Kilangala Dispensary Mahumbika Dispensary Mchinga Dispensary Mnolela Dispensary
2009
2010
2011
2012
60 8 6 10 7 15
67 24 4 8 5 12
33 35 1 7 3 10
12 13 1 8 0 5
CHF, Community Health Fund. Copyright © 2013 John Wiley & Sons, Ltd.
Int J Health Plann Mgmt 2014; 29: e368–e382. DOI: 10.1002/hpm
e376
S. O. MALUKA AND G. BUKAGILE
and that the community members were sensitised not to contribute for the provision of the social services. The community members were informed that it is the responsibility of the government to provide free social services to the citizens. One respondent expressed this way: ‘Politics affects CHF enrolment. The entire village leadership is from the opposition political party. The villagers are sensitised not to pay for the health services. Instead, the government is supposed to provide free health services’ (IDI with incharge of a health facility). Conversely, in Iramba district, all respondents from the district to village levels reported that CHF enrolment had increased considerably in recent years. As a result, the number of out-of-pocket payments had noticeably gone down. Analysis of the minutes of the health facility committees and CHF registers confirmed this point. As Table 6 indicates, the CHF enrolment in the study facilities had been increasing over the past 3 years. A number of factors facilitated the uptake of the CHF in Iramba district. First, the CHMT had made efforts to ensure that drugs were available in all facilities and that patients were not required to purchase drugs. The most important approach used to ensure availability of drugs was frequent monitoring and auditing of drugs in the health facilities. According to CHMT members, this significantly helped to reduce irrational prescription and improper management of drugs by the health providers. One respondent remarked this way: ‘We have realised that the drugs we get from the Medical Stores Department are enough. There were problems of mismanagement of drugs by the health providers. Sometimes drugs were mismanaged by the health providers and even found in the privately owned shops. Through frequent auditing, we have managed to overcome this problem. Some health providers were given warning’ (IDI with a CHMT member). Frequent drug audit also made it possible for the CHMT members to identify drugs that were not needed at the facility. These were quickly sent to other facilities that were in need of such drugs.
Table 6. CHF enrolment in the study facilities in Iramba district Facility name
1 2 3 4 5 6 7
Ndago Health Centre Kinyangiri Health Centre Mampanta Dispensary Bomani Dispensary Misigiri Dispensary Ulemo Dispensary Total district enrolment
Number of households 2009
2010
2011
216 182 100 81 53 103 5053
163 190 132 141 79 192 3076
233 349 225 206 81 425 4890
CHF, Community Health Fund. Copyright © 2013 John Wiley & Sons, Ltd.
Int J Health Plann Mgmt 2014; 29: e368–e382. DOI: 10.1002/hpm
IMPLEMENTATION OF COMMUNITY HEALTH FUND IN TANZANIA
e377
In both study districts, respondents reported that drugs were delivered through the Integrated Logistics System. The health facilities sent orders to the district on a quarterly basis. The districts reviewed and approved orders, and sent them to the Medical Stores Department (MSD). The MSD processed the orders and debited the facility accounts. The districts could allocate supplementary funds from local sources. In Iramba district, the district health officials reported that they had managed to ensure that drugs were ordered in time. Further, the district had a small buffer stock of drugs at the district hospital. The health facilities that needed drugs could borrow from the district hospital. These had to be replaced when the health facility obtain its share from MSD. However, it was unanimously reported that of recent, especially from July 2012, the district experienced shortage of drugs mainly attributed to MSD. All district health managers reported that MSD had not delivered drugs to the district for almost 8 months. ‘Of recent, the district has been experiencing drug stock-out, particularly from July this year. This problem is mainly caused by the MSD. The MSD has not supplied us drugs for the past eight months. For almost three years we had no problem of drug stock-out’ (IDI with CHMT member) Another respondent added: ‘The main problem is the Medical Stores Department (MSD). You know, if you are a sole provider there is high possibility of becoming sluggish. For example, the district may send invoice to the MSD and you are told that drugs are available. But once you deposit funds to their account, you are told we have no drugs. The bad thing is that once funds have been deposited they cannot be returned. At the end of the day, you are forced to select other items which you did not order. In my opinion, there is need for another drug supplier who can compete with the MSD’ (IDI with a CHMT member). Second, the district council increased the amount of user fees from TSh 1000 to TSh 3000 at the dispensary level, and from TSh 1500 to TSh 4000 at the health centre. At the same time, the premium for the CHF remained TSh 5000 per household. The majority of respondents felt that the increase of the amounts for out-pocket payments, coupled with the availability of drugs at the health facility, significantly increased the enrolment of the CHF. They also reported that the number of outpocket payments had dropped. This approach began in January 2012 after being approved by the Full Council. The idea was initiated by the CHMT and forwarded to the Full Council for approval through the CHSB. By the time this study was conducted, the CHSB was planning to revise the amount of premium for CHF from TSh 5000 to TSh 10000. According to respondents, this would enable the district to expand CHF benefits to include in-patient secondary-level (hospital) services. Third, almost all respondents reported that there had been frequent sensitisation of the community members. Most often, sensitisations were done by the health facility committee members through village meetings. The CHSB also occasionally visited the health facilities. Analysis of all sources of data revealed that the health facility committees in Iramba district held meetings monthly to discuss various issues pertaining to the performance of the health facilities. It was evident in almost all facilities that implementation of the CHF was one of the permanent agenda in the Copyright © 2013 John Wiley & Sons, Ltd.
Int J Health Plann Mgmt 2014; 29: e368–e382. DOI: 10.1002/hpm
e378
S. O. MALUKA AND G. BUKAGILE
health facility committee meetings. Records of the meetings were available in all health facilities that the researchers visited. Two factors made it possible for the committees to hold their meetings as required. First, the district council managed to ensure that the health facility committee members were paid incentive to compensate their time. The chairperson and the secretary of the committee were paid TSh 3000 whereas other members were paid TSh 2000 after every meeting. Second, the CHMT had established supervision and monitoring system that required all health facilities in the district to submit minutes of the meetings along with other reports to the CHMT on monthly basis, from the first to the sixth date of every month. Closely linked to this, incentives were provided to the CHF coordinator. The CHF coordinator had been provided with a motorcycle and a laptop to enable him or her to better perform his or her responsibilities our respondents said. This had facilitated sensitisation and supervision of the CHF as well as keeping proper records related to the operations of the CHF. By contrast, in Lindi district, the committees held their meetings quarterly. The chairpersons and secretaries of the health facility committees were paid TSh 15 000 per meeting whereas other members were paid TSh 10 000. However, further analysis of the interviews across all respondents revealed that sitting allowances were hardly paid or were paid very late, sometimes even after three or four meetings (three to four quarters). Consequently, some members of the committees did not attend the meetings as required. Furthermore, in Iramba district, respondents at all levels of the district health system reported that in order to increase revenues, the district council decided to limit the number of exemptions by encouraging the village councils to pay for the elders who deserve exemptions. The elderlies who are 60 years and older were grouped by 10s to form households and were issued with CHF cards purchased by the village governments. The district council, therefore, benefited from the matching funds provided by the central government to all CHF beneficiaries.
DISCUSSION This paper has described the experiences of the district-level and local-level actors in implementing CHF in Tanzania and has illustrated a number of factors for the diverse implementation of the fund in the study districts. In this section, we expand on the most important issues raised by the respondents with reference to a wider literature on the implementation of health policy in developing countries. The findings indicated that in Lindi district, the uptake of the CHF was very low. On the contrary, in Iramba district, the performance was better. Availability of drugs in the health facilities, effective supervision, commitment of the district health management team and local government officials, and incentives to the health facility committees and boards were cited by respondents as important. Therefore, these are the factors that provide the most convincing explanation of why CHF perform better in some districts than in others. The finding indicated that Iramba district increased user fees as a mechanism to attract the population to join CHF instead of paying fees. A similar approach has Copyright © 2013 John Wiley & Sons, Ltd.
Int J Health Plann Mgmt 2014; 29: e368–e382. DOI: 10.1002/hpm
IMPLEMENTATION OF COMMUNITY HEALTH FUND IN TANZANIA
e379
been recently reported in other districts in Tanzania. For example, in Mpwapwa district, user fees were raised to TSh 4000 for hospitals, TSh 3000 for health centres and TSh 2500 for dispensaries (Stoermer and Macha, 2009). Although in Iramba district the evidence suggests that the CHF enrolment increased after the increase of user fees, this mechanism poses equity concerns. Those most at risk may be left without accessibility, not being able to pay either the enrolment premium or the user fees (Stoermer and Macha, 2009). As this approach seems to gain momentum across districts in Tanzania, equity implications need to be further examined. It would be important to examine trends in the utilisation of health services by the poor and vulnerable groups in districts where user fees have been increased. Effective supervision and commitments of the district health management teams were also mentioned by respondents as important factors for the good performance of CHF in Iramba district. Conversely, inadequate supervision and low commitment of the district health managers in Lindi district resulted into low enrolment. This finding is in line with the previous studies in Tanzania, which found that lack of commitment of some of the top managers at the district affected the implementation of the policy (Kamuzora and Gilson, 2007; Maluka, 2013). On the other hand, there are examples where the commitment and support of the district managers resulted into relative success of the CHF scheme. The Kongwa, Kondoa and Dodoma urban districts’ success has been seen to be due in part to the improved cooperation between the district health board, councillors and other leaders who together mobilise the community to join CHF, as well as through community groups (Stoermer et al., 2012). In Mpwapwa district, the personality of the DMO was reported as one of the factors leading to the good performance of CHF (Stoermer and Macha, 2009). The variation in policy implementation reported in this study suggests that the district managers and the health workers were able to exercise considerable discretion in the implementation of the CHF scheme, and it is largely their actions that have determined how the scheme took shape on the ground. The actions of the district-level actors may be seen as coping strategies to overcome the problems on them from above (Powell-Jackson et al., 2009). Bottom-up implementation theories have argued that in the face of resource uncertainties, policy implementers adopt coping behaviours to manage the high demands and time pressures that they face; and through these behaviours, they may re-interpret and reshape implementation process (Lipsky, 1980; Hill, 1997). It can reasonably be argued that district health officials and locallevel actors who attempted to deal with challenges were often driven by pressure to meet local needs. This finding reiterates the need of actively engaging the policy implementers in the design, management and implementation of policies (Kamuzora and Gilson, 2007). Policies need to be socially acceptable to the groups affected, and ideally, policy implementers, including the public, should be involved in the formulation process (Birt et al., 1997). Closely linked to this, our analysis suggests that financial incentive to the health facility committees and the board members facilitated good performance of CHF in Iramba district. The district council had in essence managed to ensure that the health facility committee members were paid incentives on time. This motivated members of the health facility committees to hold their monthly meetings as Copyright © 2013 John Wiley & Sons, Ltd.
Int J Health Plann Mgmt 2014; 29: e368–e382. DOI: 10.1002/hpm
e380
S. O. MALUKA AND G. BUKAGILE
required. The fact that the health facility committees held their meeting monthly made it possible to frequently discuss the implementation of the CHF in their respective health facilities. The district council also provided incentives to the CHF coordinator to enable him or her to better perform his or her responsibilities. On the other hand, in Lindi district, while the health facility committees were in place, meetings were held quarterly, and the majority of the members were demoralised because of the late and or non-payment of the sitting allowances. The district health managers need to design incentive mechanisms that would reinforce an effective implementation of the CHF. However, although financial incentives are important, the deciding factors are the will and ambition of the individuals and decision-making bodies within the district healthcare structures. In contrast, the finding indicated that CHF enrolment rate in Lindi district was extremely low. This was partly due to the high premium rates. The district health officials in Lindi, in collaboration with the health facility committees and boards, as well as the ward and village leaders, should put efforts in encouraging communities to join the CHF scheme. It is imperative to improve collaboration of the local-level leaders who are responsible for the implementation of the fund. The village and ward leaders need to be sensitised about the importance of the CHF in protecting the households against medical catastrophe and impoverishment. Additionally, the district council should consider the possibility of reducing the current CHF contribution rate of TSh 10 000 to TSh 5000. This would encourage many households to join the scheme, and the district would benefit from the matching grant provided by the central government. Evidence shows that in Mwanga district, membership steadily rose following the reduction of the premium rate after 6 months since the initial implementation from Tsh 10 000 to 5000 in 2006 (URT, 2007). Likewise, in Rombo district, in order to attract many households to join the CHF scheme, the board reduced CHF contribution from the former TSh 10 000 to affordable amount of TSh 5000 per household per year. This increased enrolment, reaching approximately 25% of the existing potential (Stoermer et al., 2012). Furthermore, in both study districts, frequent drug stock-out in the facilities constrained the implementation of the CHF. Drug stock-out was, for the most part, attributed to the inadequate and late supply from the MSD. Drug stock-out seems to be a persistent problem in Tanzania. Experience shows that despite the transition to an Integrated Logistics System from the one based on the distribution of standard essential medicine kits as of 2009, many health facilities in the country still face the problem of irregular supply of medicines, equipment and medical supplies from the MSD. It is evident that the Medical Store Department face challenges to efficiently and timely supply drugs to all facilities in the country. The researchers did not interview respondents from the MSD on this matter. However, evidence suggests that the main constraints facing the MSD include increased demand of drugs and medical supplies, limited financial and human resources, and long and bureaucratic procurement procedures (Mwananchi Cooperation, 2013). The Ministry of Health and Social Welfare should increase the capacity of the MSD and review the drug procurement procedures. In addition, the Ministry of Health and Social Welfare should consider the possibility of diversifying drug supply system in the country by opening opportunities for other government-subsidised drug suppliers. Copyright © 2013 John Wiley & Sons, Ltd.
Int J Health Plann Mgmt 2014; 29: e368–e382. DOI: 10.1002/hpm
IMPLEMENTATION OF COMMUNITY HEALTH FUND IN TANZANIA
e381
LIMITATIONS OF THE STUDY This study relied primarily on the review of minutes, health facility visits and key informant interviews with the community representatives and the district health managers. First, all data collection was completed before the analysis of data was undertaken. We recognise that a more iterative process of data collection and analysis would have improved the interview guides and experience of the interviewers. Second, the study did not interview the members of the CHF in order to assess their experiences in utilising health services and the factors constraining the enrolment of the fund. Future studies should focus on these aspects in order to understand better the actual practices of CHF scheme in Tanzania. Similarly, future studies should explore equity implications in the districts, which have recently increased user fees in order to motivate households to join CHF. Despite these limitations, the study provides a detailed account of the implementation of CHF in Tanzania.
CONCLUSION This study marked an important attempt in providing a deeper and more nuanced understanding of the implementation of CHF in Tanzania. The large variation in policy implementation reported in this study suggests that district managers, health workers and local leaders were able to exercise considerable discretion in the implementation of the CHF scheme, and it is largely their actions that have determined how the scheme took shape on the ground. The study confirms that the success of the CHF largely depends on the availability of health services, effective supervision mechanisms, commitment of the district health management team and local government officials, and incentives for the health facility committees and board members. The focus of implementation needs to be placed on the active engagement of the district health managers and local-level leaders such as counsellors, ward and village leaders, as well as health facility committees and boards who are responsible for the implementation of the policies. Equally important is the availability of quality health services in the health facilities. In particular, availability of drugs and medical supplies seems to be the most important way forward. The Ministry of Health and Social Welfare should increase the capacity of the MSD and review the drug procurement procedures. In addition, the Ministry of Health and Social Welfare should consider the possibility of diversifying drug supply system in the country by opening opportunities for other government-subsidised drug suppliers.
ACKNOWLEDGEMENTS We would like to express my gratitude to the Policy Research for Development, formerly Research on Poverty Alleviation (REPOA), for partially funding this study. We also thank local government officials, district health authorities and other stakeholders in the study districts for participating in the study. Further, we are thankful to our research assistants for assisting us in the data collection. The authors have no competing interests. Copyright © 2013 John Wiley & Sons, Ltd.
Int J Health Plann Mgmt 2014; 29: e368–e382. DOI: 10.1002/hpm
e382
S. O. MALUKA AND G. BUKAGILE
REFERENCES Birt CA, Gunning-Schepers L, Hayes A, et al. 1997. How should public health policy be developed? A case study in European Public Health. J Public Health Med 19(3): 262–267. Burns M, Mantel M. 2006. Tanzania review of exemptions and waivers, Report submitted to the Ministry of Health by the Euro Health Group. Devadasan N, Criel B, Van Damme W, Ranson K, Van der Stuyft P. 2007. Indian community health insurance schemes provide partial protection against catastrophic health expenditure. BMC Health Serv Res 7: 43. Gilson L. 1997. The lessons of user fee experience in Africa. Health Policy Plan 12(4): 273–285. Hill M. 1997. The Policy Process in the Modern State. Prentice Hall/Harvester Wheatsheaf: Harlow. Jacobs B, Bigdeli M, van Pelt M, Ir P, Salze C, Criel B. 2008. Bridging community-based health insurance and social protection for health care—a step in the direction of universal coverage? Trop Med Int Health 13(2): 140–143. James C, Hanson K, McPake B, et al. 2006. To retain or remove user fees?: reflections on the current debate in low and middle-income countries. Appl Health Econ Health Policy 5(3): 137–153. Kamuzora P, Gilson L. 2007. Factors influencing implementation of the Community Health Fund in Tanzania. Health Policy Plan 22: 95–102. Kessy F. 2008. Technical Review of Council Health Service Boards and Health Facility Committees in Tanzania. Report prepared for the MOHSW with financial support from DANIDA and SDC. Lipsky M. 1980. Street-Level Bureaucracy: Dilemmas of the Individual in Public Services. Russell Sage Foundation: New York. Maluka SO. 2013. Why are pro-poor exemption policies in Tanzania better implemented in some districts than in others? Int J Equity Health 12: 80. McIntyre D, Thiede M, Dahlgren G, Whitehead M. 2006. What are the economic consequences for households of illness and of paying for health care in low- and middle-income country contexts? Soc Sci Med 62: 858–865. Ministry of Health (MoH). 1999. Community Health Fund Operations Guidelines. Ministry of Health: Dar es Salaam.
Copyright © 2013 John Wiley & Sons, Ltd.
Ministry of Health and Social welfare (MoHSW). 2012. Health sector public expenditure review 2010/2011. Directorate of Policy and Planning—Ministry of Health and Social Welfare: Dar es Salaam. Mtei G, Mulligan J. 2007. Community Health Funds in Tanzania: a literature review. Report submitted to Ifakara Health Research and Development Centre. Mwananchi Cooperation. 2013. Challenges in the supply of drugs in Tanzania: who is to blame? Mwananchi Cooperation Friday 19 April. Ouimet MJ, Fournier P, Diop I, Haddad S. 2007. Solidarity or financial sustainability: an analysis of the values of community-based health insurance subscribers and promoters in Senegal. Can J Public Health 98(4): 341–346. Powell-Jackson T, Morrison J, Tiwari S, Neupane B, Costello AM. 2009. The experiences of districts in implementing a national incentive programme to promote safe delivery in Nepal. BMC Health Serv Res 9: 97. Ritchie J, Spencer L, O’Connor W. 2003. Carrying out qualitative analysis. In Qualitative Research Practice: A Guide for Social Science Students and Researchers, Ritchie J, Lewis J (eds). Sage Publications: London; 219–262. Stoermer M, Macha J. 2009. External Review of the Project “Health Insurance for the Rural Population (Tanzania)”. Report submitted to the Swiss Development Cooperation. Stoermer M, Hanlon P, Tawa M, Macha J, Mosha D. 2012. Community Health Funds in Tanzania: innovations study, Report submitted to GIZ. United Republic of Tanzania (URT). 2001. The Community Health Fund Act. United Republic of Tanzania: Dar es Salaam. United Republic of Tanzania (URT). 2007. Community Health Fund: Best Practices Workshop Report. Ministry of Health and Social welfare: Dar es Salaam. Watkins K. 1997. Cost-Recovery and Equity in the Health Sector: Issues for Developing Countries, Working paper, Oxfam. World Health Organization. 2010. World Health Report 2010—Health Systems Financing: The Path to Universal Coverage. World Health Organization: Geneva. Yin RK. 2003. Case Study Research: Design and Methods. Sage Publications Inc.: Thousand Oaks, CA.
Int J Health Plann Mgmt 2014; 29: e368–e382. DOI: 10.1002/hpm
