Poor People's Experiences of Health Services in Tanzania

Abstract

This literature review examines key findings on poor people's experiences of health services and the barriers to access among the very poor. It was first published in 2004 by the Women's Dignity Project/Utu Mwanamke, PO Box 79402, Dar es Salaam, Tanzania. E-mail: 〈[email protected]〉 and is reprinted here, almost in full, with their kind permission.

“…the poor must be brought from the margins into the mainstream. The process must be inclusive. The weakest economies and communities need special and differentiated help.” (President Benjamin W Mkapa, 24 February 2004)

TANZANIA faces serious challenges to improving the health and well-being of its people. The Ministry of Health and its partners in government, the donor community and civil society have responded with concerted action, in many cases achieving significant gains. These achievements are particularly notable given serious funding limitations in the health sector. The allocation to health has increased only slowly over recent years, from 7.5% in fiscal year (FY) 2000 to 8.7% in FY02, which is low in relation to projections in the Poverty Reduction Strategy (PRS) and to the Abuja commitment of 15%. Again, despite the PRS commitments, the absolute budgetary increase year-to-year has declined from a high of 41.12% in FY02 down to a 5.68% increase in FY04 (MoH 2004).

However, the low level of funding does not categorically preclude improvement of health services and ultimately health outcomes. “Differentiated” allocation of government funds, including in the health sector, can better prioritise the needs of the poor. The continuing disparities in health outcomes between the poorest and the richest Tanzanians and those in rural versus urban areas need to be addressed, along with the barriers to service experienced by the poor due to distance, formal and informal health charges, and other obstacles (R&AWG 2003). The new resource allocation formula that utilizes equity criteria to distribute funds across districts, and the increasing proportion of funds for preventive services, are both positive developments in reaching the poor. Nonetheless, additional actions are needed to mobilize meaningful change for Tanzanians living in poverty.

Tanzania is fortunate to have an extensive network of health facilities throughout the country, a mark of its commitment to ensure that people have access to essential health services.^* Decades of commitment to improving the health of the population, however, were compromised in the 1980s and '90s by volatile world commodity prices, worsening terms of trade, debt-servicing and structural adjustment policies–all of which combined to undermine the effectiveness and efficiency of the health sector. This is manifested in a dilapidated infrastructure, shortage of trained staff, de-motivated workforce, weak management systems and poor quality of care.

In an effort to stem the deterioration of the health system and address systemic financing, service delivery and management concerns, a health sector reform (HSR) program was initiated in 1995/96 including a move to decentralize authority and resources to the district level. The reforms centre on key issues that affect health service delivery: equity, efficiency, cost-effectiveness and quality of care. The 2003–2006 Health Sector Strategic Plan (HSSP) is the most recent elaboration of Ministry of Health objectives to implement services in the country. It states a particular commitment to reaching vulnerable populations–to increase access and utilization of health services, improve the quality of services, and to ensure equity in allocation of resources (MoH 2003).

One of the key components of the HSR is financial sustainability including through a policy of cost-recovery introduced in the early 1990s ^{22 15 26 35 9} (Mmbuji et al 1996, MoH 1995, Mubyazi 1998, Njau 2000, Hemed 2000). The rationale for introducing the policy was to generate additional revenues; improve the availability and quality of health services; strengthen the referral system and rationalize utilization of health services; and improve equity and access to health services (MoH 1994). Charges were introduced in four stages: at referral, regional and district hospitals in 1993, and by 1994 at the district level. Fees have also been introduced in some health centres and dispensaries in some districts and are slated to be rolled out formally in 2004. Other mechanisms such as the Community Health Fund (CHF), designed to serve the majority of the poor in a more equitable way, have also been introduced. The National Health Insurance scheme on the other hand is an insurance plan primarily for civil servants and some of their dependents.

Provisions for exemptions and waivers within the cost recovery programme were introduced with a view to protect vulnerable social groups and the very poor. An exemption is an automatic entitlement to free public health care services and is extended to children aged five years and under, for maternal and child health (MCH) services, to those with specific diseases, and for people with long term mental disorders. Exemptions, as such, are not necessarily confined to the poor. They are designed to protect vulnerable social groups. A waiver on the other hand is a conditional temporary entitlement that is provided after evaluations by the relevant authorities. It is to be granted to patients who do not automatically qualify for exemptions but are considered to be in need of such services and are “unable to pay.”

What have been poor people's experiences of health services following the introduction of the HSR and cost sharing? Are the goals of access, quality of care and equity being realized? What are some of the key barriers that poor people face in accessing services, and how have these impacted their health and long-term vulnerability?

This literature review examines key findings on poor people's experiences of health services and includes a particular focus on the barriers to access among the very poor due to cost sharing, an issue of special interest in the current era of “pro-poor” development. It highlights seven key issues for reflection and action: access, health care charges, participation and decision-making, governance and accountability, exemptions and waivers, how poor people cope, and health care-seeking behaviour and choice of providers.

Key barriers the poor face in accessing quality health care

Results of a baseline nationwide Service Delivery Survey (SDS) carried out by Tanzania Development Research Group (TADREG) indicates that even though three-quarters (75%) of the villagers were concerned about the costs of health care, a majority (71%) still expressed a readiness to pay more for health services, provided the quality of services improved. While people are willing to pay for care if quality is good, this pre-condition does not in general appear to exist. Public perceptions on accessibility to health care reveal that, overall, health services continue to fail the very poor. Deteriorating roads coupled with inadequate and unaffordable transport make it impossible for the poor to reach facilities. Poor women face particular obstacles as they have neither the time, money, nor necessarily the incentive to access distant and low quality care.

On the positive front, information obtained from exit interviews with in-patients and out-patients across eight public health facilities in the four regions of Mbeya, Kilimanjaro, Mtwara and Mwanza, shows that 90% or more of the 200 people interviewed were of the opinion that the cost sharing system had made a “remarkable” improvement in the quality of public health services. Improvements were seen in the quality of public health services and staff, availability of drugs and laboratory services, and shortened queuing time (Msambichaka 2003). This is consistent with evidence gathered from focus group discussions (FGDs) with civil society organizations and local politicians in the four regions…

Findings from an assessment of Hanang, which was one of the nine roll-out districts in the initial phase of expansion of the Community Health Fund (CHF) scheme, reveal that for all CHF members the benefits of the scheme outweighed the cost (Chee et al 2002).^*

Most members and non-members alike believed that it had led to improved services at CHF participating facilities. CHF resources had been used to improve the quality and range of services throughout the district by purchasing drugs and equipment and refurbishing health facilities; however, the majority of the funds (59%) had been used towards the construction of the district hospital. It is noteworthy that a large percentage of the members tended to be civil servants, and thus relatively well-off.

In general, however, quality of care and fees appear to be key factors in demotivating people to utilize facilities. As noted in an evaluation study of 13 hospitals, altogether, less than half of the 356 exit patients/escorts (201 patients and 155 patient escorts or relatives) interviewed from across 13 hospitals reported any improvement in quality of care following the introduction of user fees. More specifically they cited deficiencies in terms of drug availability (34% of the respondents), laboratory services (30%), and availability and cleanliness of beddings (42%) (Mmbuji et al 1996).

Findings from an Mbeya Region study on quality of care through the eyes of patients indicate that if able, the poor are willing to pay for better quality of care but services are not improving (Tibandebage & Mackintosh 2002). The government hospital in Mbeya Urban District, in particular the main maternity hospital that is a part of it, was reported to have particularly low levels of care. Stories about abuse, lack of care, lack of advice and lack of professionalism predominated. Altogether, of the 49 household interviews, dissatisfaction relating to health facilities in Mbeya Region was recounted by 30. Respondents described 41 incidents in all, with over half (22) of the incidents pertaining to the Government hospital (including the maternity hospital), and the rest referring to private/mission hospitals and the primary health care facilities. The government hospital management team on the other hand was of the opinion that the quality of care offered had on the whole improved: the quality and use of laboratory services had reportedly improved due to revenue generated from the cost sharing scheme.

The above study also illustrates problems in the referral system emerging from the interaction between government and private/religious facilities. Patients from mission facilities report facing harassment and mistreatment when referred to the government hospital “which they (patients) regard as dangerous, abusive and unwilling or unable to provide care and treatment.” Those people who can, refer themselves to mission hospitals and larger mission health centres and dispensaries. The problem is further compounded by divisions among the mission facilities that do not refer patients to hospitals that are not of their own denomination. Government facilities on the other hand generally referred patients to the most affordable option.

In a study carried out in Morogoro Rural and Kilombero region, the provision of services at health centres and dispensaries is noted to be of poor quality–they are small and understaffed, waiting times are very long and medicines are in short supply, especially in the fourth week of any month (SDC 2003). Results of a household survey carried out in three divisions of Lushoto District show that because of poor quality of service and poor availability of drugs, most patients by-pass the lower levels of health care to seek hospital treatment (Agyemang-Gyau & Mori 1999).

Availability of essential drugs

From the patients' perspective, a constant supply of essential drugs is a prerequisite to the credibility of health services and to the quality of health care provided… However findings from studies also reveal that even when the poor are able to find money for basic care, and even when essential drugs are available, their inability to purchase these medicines makes treatment actually impossible. Drugs are often found to be more affordable at government facilities but they run out quickly; they are more available at private and mission facilities but people generally cannot afford to buy them there (WDP 2003).

While ensuring the smooth flow of medicines to health centres and dispensaries, the standard pre-packed kits fail to take into account the varying morbidity patterns and therefore the diverse demand for health care across the country. This situation may change for the better with the introduction of the drug indent system.

The issue, however, is not only one of an inadequate supply of drugs, but as observed in Kondoa district, also of “unfair and inefficient” distribution of these drugs once they arrive at the dispensaries (Mujinja & Hausmann 1997, quoted in Mubyazi 1998). Even when essential medicines that are supposed to be free are “officially unavailable,” staff at one health centre often offered to arrange for these medicines if patients had the money. Generally, those people who are able to buy medicines from the drug stores do so, even if they are only able to buy partial doses (SDC 2003). Likewise, results from Dodoma indicate that for 70% of the respondents there have been no changes in the availability of essential drugs at the health facilities. Many people continue to buy their drugs from drug stores, though often they are not able to purchase the prescribed medication due to their meagre income (Rutaihwa 1997)…

Shortage of skilled providers

MoH has recently described the human resource situation as a state of crisis. Available evidence shows that public health facilities, in particular those located in the remote rural areas, face an acute shortage of skilled providers. Human resource development is–and will continue to be–a critical factor in raising the quality of care.

According to the MoH Health Sector Strategic Plan (2003), the majority of skilled health workers work in the large cities, and as a result, rural facilities are understaffed. In addition, substantial imbalances exist in the patient-provider ratio across districts, and between rural and urban areas. Approximately one-third of the existing labor force is unskilled, and it is estimated that only 55–60% of staff time is spent on productive activities…

A few examples illustrate the problem:

• 47% of vacant LGA level positions in the health sector for which permits have been issued are still not filled (JHSR 2004).

• An evaluation of maternal care in six districts in Tanzania confirmed the poor quality of care in most of the districts (Options 1998) plus inadequate human resources coupled with insufficient knowledge and skills.

• In Kondoa district the four rural health centres are largely understaffed-an average of two staff per rural health unit–because of staff retrenchments, staff transfers outside the district and deaths (Mujinja & Hausmann 1997, quoted in Mubyazi 1998)…

MoH is not the only agency responsible for the delays, however, nor for the dearth of qualified health workers. The Public Service Commission now controls hiring of civil servants and procedures are not clear. The education system in Tanzania is tragically weak, resulting in poorly educated personnel in all sectors including health…

“You are nobody if you do not have money.” (SDC 2003)

Access to services is viewed to be strongly dependent on “connections” and on ability to pay. Discrimination and lack of respect by health workers towards the very poor is a common theme emerging from a number of studies ^{47 49 52} (SDC 2003, Tibandebage & Mackintosh 2002, WDP 2003). Medical staff are often rude to the poor and dismissive.

Distance and transport

“Many [poor women] cannot afford transport costs so they sell their food, borrow, use herbs or just wait to die.” (Health worker, Mpwapwa, WDP 2003)

Evidence from a number of studies suggests that the cost of accessing care is a critical determinant of whether or not care is sought. This is particularly true where a properly equipped health facility is far away, infrastructure for transport is lacking, and in which communities are particularly poor, as evidenced by a large number of people living below the poverty line. Rural areas are particularly disadvantaged on many health and survival indicators, in terms of both health outcomes and service uptake. Many of these indicators are strongly related to poverty–and these differentials have sometimes increased during the 1990s (R&AWG 2003). The problem of access is especially problematic for pregnant women who may not be able to access an institutional delivery due to these obstacles.^* Only 44% of deliveries in Tanzania take place in a health facility and the trend has been downwards in the past decade (NBS November 2002).

Various studies confirm the major impediments of distance and transport, including the Tanzania Participatory Poverty Assessment (TzPPA 2003), a study of four districts in Kagera region (Mpembeni et al 2000), Kondoa district (Mujinja & Hausmann 1997, quoted in Mubyazi 1998), Morogoro Rural and Kilombero region (SDC 2003), and one in three districts of Mpwapwa, Korogwe and Tunduru (WDP 2003). Data from the '90s reveals that in terms of deliveries attended by a medically trained person–like other health problems–there exist great disparities in people's ability to get care between those living in rural and urban areas, and between rich and poor persons (^{40 8}R&AWG 2002, Gwatkin et al November 2003^† ).

In the TzPPA, distance was the second most frequently cited obstacle limiting people's capacity to treat illness, following the major obstacle of cost. According to the 2000/01 Household Budget Survey (HBS), nationally, nearly half a million households are estimated to live more than 20 kms from the nearest dispensary or health centre (NBS July 2002). Available information from Kagera, Kondoa, Mpwapwa, Korogwe and Tunduru reveals that on the average, patients have to walk between five and 10 kilometres, often across difficult terrain, to the nearest dispensary or health centre. Only 30% of the villages in Korogwe have a health facility.

Hospitals are especially out of reach and this has implications for emergency referral care. For example, findings from WDP (2003) reveal that distance to the nearest hospital can range from 10 to about 60 kms. Transport, whether public or private, is scarce and ambulances are severely lacking. In times of emergencies, people who can resort to hiring bicycles for Tshs 200 do so, but some people are not even able to afford this.

Generally, ambulance costs are prohibitive. In the SDC (2003) study, ambulance costs required an advance payment and at times amounted to Tshs 8,000. Many people in the study had been impoverished by serious illnesses requiring emergency referrals to the regional hospital, and many women in Morogoro Rural “have died in childbirth because they could not raise transport money” (SDC 2003). Ambulance costs in Mongo wa Mongo village in Mbulu District for example can come to as much as Tshs 80,000… As a result, women say that when obstetric emergencies arise, their only option is to “pray to God” (TzPPA 2003).

Health care charges, “unofficial payments” and bribes

Health charges have placed a particularly significant financial burden on the poorest households that have very little income flexibility. Official charges are not always affordable and “unofficial” charges are still in place… At times, fees have been an impossible barrier for the poor to overcome–denying them access to critical services. The proposed introduction of user fees at dispensaries and health centres “is likely to further raise the costs faced by users and may increase the incidence of informal charging…” (R&AWG 2003).

Between 1989 and 1991, Government conducted a comprehensive financing study of the potential of introducing user fees in public health facilities (MoH 1995). Findings indicated that… people incurred significant costs to purchase essential medicines and other small items that were often not in stock at the health facilities, to pay various “unauthorised” fees and for emergency transport, and to sustain the costs of waiting time, opportunity costs away from income earning, etc.

The Mbeya Region study reports a number of stories about children dying because of lack of funds for treatment, or of a mother who was refused MCH care because she was not able to pay a “fine” of Tshs 700 for not bringing the child back on time. The study also reports one woman who died in the maternity hospital because she was unable to pay for an emergency caesarean section, and of at least three other people that died because they did not have the required money and therefore were denied treatment (Tibandebage & Mackintosh 2002)…

Figure 1 Tanzania, 2003

Findings from the recently completed Policy and Service Satisfaction Survey (PSSS) show that cost of health care was cited to be one of the major problems faced by 54% of the respondents during the last 12 months, especially in the rural areas (REPOA 2003). Over half of all respondents said that the cost of treatment is the biggest health care problem they face, and overall cost of treatment is the third most acute household problem. Nearly three-fourths of all respondents thought that people's ability to pay for health services had decreased during the last five years. Around 40% of the respondents knew people who had been refused medical treatment, and over a quarter knew “a lot of people.” In 85% of cases, people were refused treatment because of lack of cash to pay for drugs and supplies. 57% were refused treatment because they could not pay the consultation fee, and 26% because of inability to pay unofficial payments to health workers. After cost, respondents cited the obstacles of drug availability, waiting time and distance to the health facility…

In addition to the obstacles placed by official charges, poor people also encounter the barrier of “unofficial charges” or bribes. According to TzPPA participants, corruption is the most significant obstacle they face to accessing social services, especially health care. Even in instances where government dispensaries and health centres do not charge officially, there are numerous reports of people having to pay for medicines and supplies that are supposed to be free, and bribes to get treatment ^{50 52 49 43} (TzPPA 2003, WDP 2003, Tibandebage & Mackintosh 2002, Rutaihwa 1997).

The rationale for introducing health care charges and related exemption and waiver procedures in the public sector was to generate additional revenues; to improve the availability and quality of health services; to strengthen the referral system and rationalize utilization of health services; and to improve equity and access to health services. The principle is that revenues are retained by, and used locally at, the facility level on items directly related to improving the availability and quality of health care (MoH 1995).

Revenue generated from user fees is also meant to supplement government budget allocations. However the 2003 Health Public Expenditure Review (PER) indicates that cost-sharing has “contributed relatively little to the sector resource envelope”–no more than 2% of recurrent costs (MoH February 2003). It is also not clear how much money has been generated by cost-sharing and what this money is used towards.

The question then becomes: does the potentially small financial gain from cost-sharing outweigh the potentially significant loss of health among people excluded from services due to fees? This is particularly true at the dispensary and health centre levels where poor people are more likely to access care.

In an earlier study carried out by the MoH, information obtained from 13 hospitals revealed that revenue generated from user fees was generally low compared to projections, and varied from one hospital to another (Mmbuji et al 1996). For 1993/94 and 1994/95, the revenue represented 1.0% and 2.46% of recurrent budgets, and 2.5% and 6.7% of the non-wage budget for the health sector, respectively. However, due to budgetary shortfalls for most hospitals, these revenues constituted a significant source of expenditure. Msambachika's study (2003) reports that between 1998/99 and 2001/02, user fees accounted for 20–65% of total hospital expenditure in hospitals studied. It is not clear from the report what constitutes the total available resources of the hospital, however, or if the resources examined included all sources or only discretionary monies.

In the studies reviewed, user fees had not yet been officially introduced at the dispensary level, though there were some health centres charging fees ranging from Tshs 100 to 1,000. However, even when the poor are able to pay the official cost of Tshs 100 at one of the government health centres, many are not able to purchase the required essential medicines, making it impossible for them to get effectively treated ^{47 49 43 52}(SDC 2003, Tibandebage & Mackintosh 2002, Rutaihwa 1997, Women's Dignity Project 2003).

The Mbeya study also illustrates that a number of patients failed to obtain referral care at the hospital because they ended up spending a significant portion of their money on transporting themselves to the distant government hospital. Subsequently, they were not able to pay for all the official fees or for essential operations. The costs of transferring a seriously ill patient can often be insurmountable for a poor family, totalling Tshs 8,000 to 12,000: 2,000 for transport, plus another 6,000 to 10,000 at the hospital, including an overnight stay (often with a relative). This does not include the costs of actual care.

Except when required to pay for an ambulance during an emergency, costs for drugs are often the largest of all the health costs. In the Lushoto district study, patients seeking hospital treatment spent on the average about Tshs 2,000 per episode of illness: about 80% on drugs and other fees, 10% on transport, food and accommodation and 10% on informal charges (Agyemang-Gyau & Mori 1999). A study carried out in Igunga District indicates households incurred an average total cost of Tshs 6,874 for health care during a three-month period in 2000 (Msuya 2003). Close to 50% (Tshs 3,164) was spent on drugs; the remainder was spent on transportation and communication charges, examination/prescription fees and “other” health care costs. A fifth of the 200 surveyed households had failed at least once in a year to get medical treatment due to lack of cash.

For many members, one of the greatest benefits of the CHF scheme in Hanang was that it ensured unlimited access to health care for the whole family (Chee et al 2002). Others considered it a positive form of savings for unpredictable illnesses in future. The average number of visits per CHF household to facilities was 32 (for the entire 11-month study period in 2001); the average per individual was four visits. According to the health workers in charge of these facilities, these figures do not reflect members' overuse of health services for minor conditions, but rather their willingness and ability to seek timely care when ill because they have prepaid for services.

However, in spite of these positive findings, the average membership rate for the seven CHF-participating facilities visited during the assessment was estimated to be fairly low, at around an average of 5% of all households in the catchment area in 2001 (ranging from 0.3% to 13% across all CHF facilities). Inability to pay a lump sum of Tshs 10,000 was the most common reason given for not joining or renewing CHF membership. Few were aware that the CHF membership fee could be paid in instalments. It was easier for people to pay user fees that were introduced around the same time as the CHF: Tshs 1,000 per person for a single visit to the health facility.^* For the same reason, the preferred mode of payment for the majority (70%) in Lushoto was fee-for-service; only a minority who reported relatively high household income favoured community health insurance (Agyemang-Gyau & Mori 1999).

Governance

“…the baby suffered from respiratory problems. At 1.00 am at night my husband and I had to rush the baby to the dispensary… When [Bwana Mganga] eventually opened the door, he looked at the child and said ’This child is not sick‘. When we insisted, he then said ‘So didn't you come with a lamp? How would I check him?…’ He then said, ‘Okay, give me a piece of paper on which I can write the prescription.’ In response to that we also explained that we had forgotten to come with any paper. He then gave us one paracetamol tablet and told us to go back home…On the same night, the baby's condition worsened… We were then compelled to go back to Bwana Mganga… The baby was still and not crying anymore. However, the Bwana Mganga just gave him an injection. When we arrived home the baby was already dead. We did not take up the issue to the Village Government Chairperson because it was too painful and we realized we still depended on the dispensary for further needs. In addition, the Village Government Chairperson also knows of the Bwana Mganga's arrogance yet he has never set aside time to listen to the people, especially women…” (TzPPA 2003)

Information from a number of studies reveals that overall, communities are not involved in decision-making processes regarding the provision of health care. Reliable mechanisms are not in place for airing grievances or lodging appeals, and for raising community issues to the district level for action. Similarly there are few–if any–mechanisms currently in place to provide information to people on health care expenditures at the local level or to engage people in monitoring the allocation and use of these funds. There is growing concern over the prospects of the Local Government Reform Programme (LGRP) in promoting self-governance in Tanzania. As aptly noted by Kawa (2003), even though the legal and institutional framework of village governance is in place, the current ongoing LGRP that ends in 2004 is not clear about mechanisms of empowering communities at grassroots level.

…The establishment of Community Health Service Boards (CHSBs) and Health Facility Committees^† should in this respect provide a role for community members in decision-making and oversight, assuming that CHSBs and Committees are representative of, and accountable to, their constituencies.

There are certainly some examples of village leaders and district authorities taking positive action in response to concerns expressed by communities. For example, some health workers are held accountable for quality of care provided, including poor or discriminatory practices; making information on cost-sharing available; assisting poor people to access care; mobilizing communities to put in place emergency transport systems and; managing community health problems (WDP 2003). However, this may be more the exception than the rule.

In general, and as clearly shown in the TzPPA (2003) and in the SDC (2003) study, knowledge of recent reforms and entitlements is weak. People, especially the poor, expressed a “sense of helplessness” in feeling they do not have the opportunity to raise concerns, particularly about rights, injustice and corruption. Very few of the households in the studies are participating actively in Village Government activities and meetings. This contrasts with findings of the Policy and Service Satisfaction Survey that reported more than a quarter of rural households have members who are or have been village, ward or district councillors and almost as many have been involved in preparing a village or ward plan (REPOA 2003).

People report not only being afraid of participating in “governance” structures, but do not think they have the right to do so. Furthermore, they believe such participation to be pointless. As such, despite substantial dissatisfaction with health care in the Mbeya Region study, in particular in the urban areas and especially with the government hospital, there was no evidence of anyone making a formal complaint to a facility.

The CHF in Hanang has a decentralized management structure that seeks to promote involvement of the communities, primarily through elected members of the District Health Board (DHB) and the Ward Health Committee (WHC). This has been limited, however, because the WHCs were not functioning in many of the wards. In addition, even though many CHF members and some non-members were aware that the WHC is responsible for the local management of the CHF, very few were adequately informed about the composition of the committee or its specific roles and responsibilities. Most CHF members had never been invited to or attended a meeting. They did not know how the CHF was managed or how decisions regarding the use of CHF funds were made. Information about the CHF and the fee was not widely disseminated at the outset.

The community-based health care approach in Kilombero demonstrates to some extent that the effectiveness of community-based initiatives can be strengthened through “community empowerment:” increasing people's awareness of their constitutional rights and the policies affecting their well-being ^{13 44}(Lwilla 2001, Saltiel & Tissafi 2002). A community-initiated, cost-sharing mechanism was piloted in two dispensaries in 1997 to address issues identified by the community as important to quality of care: shortage of medicines, inadequate furniture and lack of diagnostic equipment. In one village dispensary the community decided to pay a flat rate of Tshs 200 per visit and Tshs 100 as a consultation fee which would cater for drugs and maintenance respectively (the charges were subsequently revised); in the second village dispensary, they decided to pay according to the type and number of drugs prescribed (prices defined by the community) with an additional Tshs 100 to be paid once a month.

The task of identifying those eligible for waivers was given to hamlets and thereafter discussed in Village Councils and Ward Development Committees for approval. For appropriate collection and management of funds, communities employed their own cashier and selected their facility committee that was responsible to oversee the daily running of the facility including the mandate to approve all expenditures…

Governance and accountability

The most recent World Development Report (World Bank 2003) points to the need to strengthen accountability between–and among–three distinct constituencies: poor people, providers and policymakers…. Accountability in health is seen to ensure appropriate allocation, disbursement, utilization and monitoring of financial resources, including through publicly accessible information. It is also seen to assure compliance with procedures and standards of performance, and to respond to ongoing and emerging societal needs and concerns ^{2 30} (Brinkerhoff 2003, Mwangu 2001).

Overall in Tanzania, adequate management and information systems have not yet been put in place to ensure appropriate collection and utilization of fees, and to ensure health services are responsive to the priority needs of the beneficiaries. Diversion of funds at the district council level has been frequently reported to be a serious problem. A “Pro-Poor Expenditure Tracking Study” by REPOA (2001) that analyzed government expenditures on priority sectors indicates that there are widespread leakages of funds for other charges allocated to district councils. Council leaders often allocate funds to expenses other than those delivering community-level services. In addition, there is a lack of transparency with regards to receipt of funds and their allocation from Treasury. Heads of dispensaries and health centres are often not informed by their District Executive Directors on disbursements and the allocation to departments. As a result, a reduced amount of funding is actually used in the delivery of health services. The situation has been further aggravated by continual delays in disbursements of basket funds to the districts (MoH 2004).

One of the key findings from the Hanang CHF assessment was that the current administration and management procedures do not allow for an accurate, ongoing self-monitoring of the CHF performance (Chee et al 2002). Findings revealed many inconsistencies in membership, utilization, and financial data between the national and district level, district and facility level, and in records within the facility. In fact, the ward CHF health committee that is supposed to be responsible for management, administration, and promotion of the CHF at the local level was only functioning in one of the seven wards. As a result, during 2000 for example, most wards expended less than 20% of CHF funds collected in 1998 and 1999, and had significant unused balances in their district CHF sub-accounts by January 2001.

Access to information is a prerequisite to ensuring effective community participation in monitoring the provision of health services. This includes information on budget allocations, actual expenditures, use of medical supplies, who qualifies for waivers and exemptions, etc. down to the community and/or facility level. Aggregate data on allocations to districts provides some useful information, but is too generalized to enable effective monitoring and tracking (NGO Policy Forum 2003b).

Exemptions and waivers

An assessment of existing practices related to exemptions and waivers is not possible because hospitals do not generally keep accurate records of exemptions and waivers granted. For example in one study none of the 10 hospitals surveyed (a mix of government, private and missionary) was using the relevant forms to determine eligibility for waivers; most were not using any documentation at all ^{34 22}(Newbrander and Sacca 1996, Mmbuji et al 1996). In another study of eight public health facilities in Mbeya, Kilimanjaro, Mtwara and Mwanza, public hospitals did start off by maintaining a proper information system. However according to the researcher, since the government failed to compensate them for exemptions and waivers granted, there was no reason to maintain what health workers and authorities essentially believed to be a costly and time-consuming procedure (Msambichaka 2003).

Exemption, and in particular waiver, procedures are generally thought to be cumbersome and inefficient and have loopholes that allow for the misuse and sometimes abuse of the system. Although incomplete and insufficient, available information shows that government hospitals especially in rural areas do exempt vulnerable groups from formal charges though the practice appears to be sporadic. There is limited evidence of systematic implementation of the waiver policy. The lack of consistency is due in part to health service providers not following procedures that in themselves are not clear; but also due to the low uptake and lack of insistence on free services by the poor. This is due, primarily, to people not being aware of their rights. Even if aware of their rights, they are not sure that these rights will be respected.

There is no doubt that the key to the success of an exemption and waiver system is its financing. Systems that have compensated providers for the revenue foregone from granting exemptions (Thailand, Indonesia and Cambodia) have been more successful than those who have expected the provider to absorb the costs of exemptions (Kenya) (Bitran & Giedion 2003). However, according to hospital management teams interviewed in Msambichaka's (2003) study, the government has failed to adhere to its original commitment of compensating public sector hospitals for exemptions and waivers, thus minimizing resources at those facilities. This not only impacts on a health facility's capacities, but it may also deter providers from effectively exempting those who qualify in order that the facility does not appear to be under-performing financially.

According to Msambachika (2003), a crude analysis of available data from the eight public hospitals surveyed in Mbeya, Kilimanjaro, Mtwara and Mwanza, waivers constituted only a small proportion of the total exemptions–less than 5%. According to the hospital management teams, civil society organizations and local politicians, exemptions are generally granted to those who do in fact qualify for them, though the exemptions may not necessarily target the very poor, or emergency cases. This is supported by evidence from available records: in all the eight hospitals visited, most of the recorded exemptions were granted to under-fives and for MCH services. There were of course exceptions: frequent requests for exemptions from ex-civil servants who are relatively well-off, and granting of exemptions to people presented by staff as their relatives.

However, preliminary findings of the Lindi Rural study indicate that in fact, only 49% of acute cases and 20% of admitted cases of under-five-year-olds were exempted (SCF 2003). Findings also suggest that the least poor tend to benefit more from the scheme, in particular for chronic illness and admission: 23% of the least poor and 20% of the moderate were exempted, compared to only 12% of the poorest. In the 1998 Service Delivery Survey, there were frequent complaints that services that were formally free could now only be obtained at a cost–including delivery, basic drugs and family planning services–all of which are supposed to be exempted (TADREG 1998).

Available evidence from several studies cited in this report illustrates that even if official fees are exempted or waived, the poor and vulnerable still end up having to pay at times for other costs. These include drugs and some small charges (e.g. cards, materials, first consultations, etc), not to mention the non-fee costs of care such as food, transport and bribes. In Kagera region for example, pregnant women may officially be exempted from charges, but they were still required to make contributions set by the local government councils to provide for deficits or delayed supplies (Mpembeni et al 2000):

“Every pregnant woman has to pay Tshs 500 for kerosene and also needs to purchase the maternity card. Most of us cannot afford these costs.”

“If you have not paid Tshs 700 you do not get TT vaccine or a maternity card so if you have no money you don't need to waste time.”

In Nzanza village, Meatu District, pregnant women end up paying a minimum of Tshs 7,700 for childbirth: 500 for an ANC card, 200 for a syringe, 2,000 for gloves and 5,000 for a “thank-you” to attendants (TzPPA 2003). As a result, many women end up delivering their babies at home without assistance from a trained attendant. In Mpwapwa, pregnant women will only be assisted with deliveries at the health centre if they “carry things like gloves, razor blades, etc. If you have no money to buy them you will not be attended. They will harass you and ask you to deliver on your own.” (37-year-old woman, Mpwapwa, WDP 2003)…

Why implementation of the exemption and waiver scheme is poor

Exemptions are generally easier to implement because the vulnerable groups are easily identifiable on medical grounds or by age. However, immense difficulties are encountered in implementing the waiver procedures ^{10 50} (Hutton 2003, TzPPA 2003). At present, there is no standardized procedure in place. Rather, there is a great deal of personalised negotiation around payment in all facilities charging fees. The task, then, becomes open to interpretation and uneven application, and possible abuse. Regular patients and those who are known to staff are often more likely to gain deferment or waivers.

In the 2002 Mbeya study (Tibandebage & Mackintosh), the government hospital had a formal system of waivers from official fees for inability to pay. A total of 375 people were registered in 1997 though it is not clear how many of these waivers were eventually granted. The management board had the power to exempt patients who were seriously ill and unable to pay, but no such cases were cited. The hospital in effect had no procedure in place of individually scrutinizing such waivers, which in theory involved an appeal to a social worker. No one in the patient or household interviews had used, or even attempted to use, the waiver system or knew of anyone else who had done so.

The CHF design in Hanang specified that communities would identify families eligible for free CHF cards (Chee et al 2002). A list of names had been submitted to the district council more than 18 months ago (from when the assessment was carried out) but nobody from this list had been given waivers. The assessment concludes that policies in Hanang district are insufficient to protect the poor from the burden of health costs.

Undoubtedly, clear criteria for granting waivers are a pre-requisite to reducing confusion and ambiguity among those responsible for managing the system, and among potential recipients. A decision needs to be made at what level the threshold must be drawn and how to identify eligible persons. Several other decisions need to be taken as well regarding the waiver procedures. For example, should waivers be granted to households or on the basis of individual targeting? Should they be permanent or temporary? How frequently should eligibility be reassessed? Should waiver eligibility be pre-determined following clear guidelines, or when individuals seek care in the facility? Should they be based at the household (as in the CHF) or at the individual level?…

With the exception of one study in Lindi Rural, available evidence indicates that people are relatively better informed of exemption categories. FGDs with civil society organizations and local politicians in Msambichaka's (2003) study show that most people understand the policy of exemptions. This is supported by information obtained from exit interviews with a total of 200 in-patients and outpatients across all eight facilities that indicates that 84% of the patients are satisfied with the exemption system. Even though 74% believed the procedures to be cumbersome, bureaucratic, inefficient and not user-friendly, many are willing to put up with it because people are generally satisfied with the quality of available services.

In Lindi Rural, public awareness about the exemption scheme was noted to be very low among health workers, village officials and the community (SCF 2003). Communities do not view it as their right, but rather as a special favour granted to them by service providers.

Coping mechanisms

A number of studies have shown that poor households with limited assets resort to a number of short-term survival strategies to pay for health care, especially in times of emergencies. This further impoverishes them and contributes to their long-term vulnerability ^{1 25 52 47 45 50}(Agyemang-Gyau & Mori 1999, Msuya 2003, WDP 2003, SDC 2003, SCF 2003, TzPPA 2003). Coping strategies include using their own savings (if they have any); possible contributions from relatives or others; engaging in petty trade; selling critical assets such as crops, animals, land and their labour; borrowing money; taking a loan; and bonding their assets.

At times the poor are forced to take their children out of school. Often they face food shortages and cope by reducing the number of meals taken in a day. They send children away to live with relatives or to eat with neighbours.

Generally, borrowing money is a problem for poor women who “are the least trusted group as they are not capable of earning enough money to cater for their families and pay debts” (WDP 2003). The situation is exceptionally difficult for people living on the margins of society–street boys, migrant workers and sex workers (SDC 2003). They do not have the required support structure to access money and other necessities, and no one from whom to borrow money. Sex workers are ridiculed and abused by their neighbours and health service providers, and in effect have to continue working to pay for their treatment, even when sick.

Health care-seeking behaviour and provider choice

Findings from a baseline household survey in southern Tanzania reveal that even in a very poor area that might easily be assumed to be uniformly poor, care-seeking behaviour is worse among poorer families than among the relatively rich (Schellenberg et al 2003)…. Carers of children from wealthier families had better knowledge about danger signs, were more likely to bring their children to a health facility when ill, and were more likely to have had a shorter journey (less than 90 minutes) to the health facility than poorer families. Their children were more likely to have received anti-malarial treatment and antibiotics for pneumonia, and were more frequently admitted to a hospital. The rate of hospital admission in the lowest socio-economic status quintile was almost half that of the highest.

More often than not, poor people's incomes are sufficient for subsistence only. In Kilombero and Ulanga, for example, about 75% of the monthly household consumption and expenditure in 1997 was for food (Schellenberg et al 2003). Medicines, when available, were expensive, resulting in non-compliance, incomplete treatment or irrational regimens… When costs were prohibitive, many of the poor in Ifakara either opted not to seek treatment at all or resorted to the traditional healer… (SDC May 2003, p.33). In Mwakidila community in Tanga District…the poor often resorted to use of traditional herbs and practiced environmental cleanliness for treating malaria, even though these measures were not perceived to be the most effective (Mutalemwa 2002).

By and large, if they can afford it, the rural poor usually opt for government health facilities because they are cheaper compared to private and religious facilities, and they tend to be near by. PSSS results indicate that over four-fifths of households used government rather than private or faith-based health facilities (REPOA 2003)… All the respondents in the Tanga study agreed that when home treatment fails, government health facilities are the most affordable source of care for the very poor and vulnerable, even if they are compelled to pay for consultation or registration fees, and even if they have to buy the prescribed drugs (Mutalemwa 2002). It is however, important to note that in this instance government services allowed them to obtain timely treatment with an option of paying later to the village office…

Mission facilities were generally known for staff commitment and positive attitude towards patients, for availability of drugs and tests, cleanliness, few hassles and no bribery practices. However perhaps the most important factor was the willingness of mission facilities to defer payment, grant a partial waiver and start treatment if necessary. Private facilities were known mainly for short waiting times and availability of drugs.

Conclusion

The issues documented in this literature review are a call to all actors–government, policymakers, donors, non- governmental organizations, faith based organizations, health workers and others–to make quality health care available to people, whether rich or poor.

In order for the goals of the PRS and health sector to be realized, however, particular commitments must be made to those who are impoverished, marginalized and otherwise vulnerable ^{40 10}(R&AWG 2002, Hutton 2003). Action must reach beyond policies and guidelines to meaningful changes in service delivery and health outcomes. A major challenge for central and local governments is to deliver more and better services with the additional, though not necessarily sufficient, resources mobilised under the PRS.

At the same time, the health sector is under-funded^* and health expenditures have not always been in favour of the poor ^{40 7}(R&AWG 2002, Gwatkin, 2003), despite the PRS commitments. Government health services have traditionally benefited the better-off more than the disadvantaged, especially for secondary and tertiary care, which accounts for most government health care expenditures (Castro-Leal et al 2000, quoted in Gwatkin 2003). More recently though, the share of spending on primary and preventive services has increased and funding is shifting more towards “other charges” (e.g. medicines and other supplies) indicating a move to bring services closer to the people. This is a very positive development.

Although health spending as a proportion of the total government budget has increased from 7.5% in FY00 to 8.7% in FY03, these figures are still low and have not reached the goal for per capita expenditure needed to achieve basic health outcomes for the population (NGO Policy Forum 2003a). There also exist high and unexplained inequalities in per capita resource allocation between regions and between districts. The new resource allocation formula instituted to distribute health funds according to more equitable criteria is an important achievement for pro-poor development. More actions to this effect are seriously needed.

Under-skilled and de-motivated personnel, deficiencies in quality of care, weak and confusing management systems, lack of information provided to health consumers, and lack of access by the very poor to treatment characterize much of the current situation. The health care system therefore requires not only a massive investment of funds but also a renewed commitment and vision among all actors to generate fundamental change. This is a particular imperative for Tanzanians living in poverty, for whom treatment is becoming increasingly unavailable, and for whom expensive private care is not an option. Several themes emerge from this review for reflection and action:

Access

• Poor quality of care, health care charges (official and “unofficial”), long distances coupled with poor roads and inadequate and unaffordable transport facilities, and poor governance and accountability mechanisms–all limit poor people's access to health care.

• Lack of essential drugs and supplies, of “skilled providers,” discrimination against those who are not able to pay, and poor referral systems result in poor quality of care.

Health care charges

• Health care charges have placed an impossible financial burden on the poorest households, who are often excluded from using health facilities when they most need them.

• Cost-sharing revenue generated has not necessarily impacted positively on quality of health care, nor on access to health care by the poorest.

• User fees are not the only charges the poor have to pay; other costs include travel time, transport costs, other “unofficial” costs including bribes, and for drugs and supplies.

• The CHF may have improved the quality and range of available services, but the scheme is not necessarily benefiting the very poor in a more equitable way.

Participation and decision-making

• Adequate management and information systems have not been put in place to ensure appropriate collection and utilization of fees.

• Communities are generally not involved in planning and financial management of health services to ensure that health services focus on meeting their priority needs.

• Ordinary people at the community level do not have access to information about budgets, incomes, expenditures, use of medical supplies, etc. and are therefore not in a position to effectively monitor their use.

Governance and accountability

• Community participation is limited in part due to a general lack of knowledge about recent reforms, but also because poor people do not know their rights or feel they can exercise them.

• Reliable mechanisms for raising concerns and for channelling these to the district level for action are not in place.

• Studies point to dissatisfaction regarding a range of health system issues: cost, quality assurance, access, availability and equitable distribution of basic services, abuses of power, financial mismanagement, corruption and lack of responsiveness.

Exemptions and waivers

• Exemptions, and waivers in particular, are not effective as a means of protecting vulnerable social groups and the poorest of the poor…

• A lack of clear criteria and policy guidelines for establishing people who are eligible for waivers results in individual ad hoc decisions, with no clear records or follow-up by management. Poor people themselves are not routinely informed of the procedures for getting exemptions and/or waivers.

• Lack of funding to health facilities to compensate for loss in revenue due to exemptions and waivers has a negative impact on the facilities, performance and discourages facilities from granting them.

The dilemma, then, is how to make quality care available to all–including the poor–in an environment of limited and insufficient financial resources and severely constrained human and material resources… With the second Poverty Reduction Strategy currently being developed, this is a perfect time to meaningfully address the issues raised in this review and to strengthen health care provision for the poor.

Notes

* The country has an extensive network of 5031 health facilities (4412 dispensaries, 402 health centres, 217 hospitals and 100 training institutions). 60% are government-owned; the remainder are voluntary, parastatal and privately owned.

* The CHF was first introduced in December 1995 on a pilot basis in Igunga district (MoH 2003). By the end of 2003, the CHF was operating in 37 districts, and another 90 districts had been sensitized and were ready to implement the scheme. A household (husband, one wife and all children under 18 years old) joins the CHF by paying an annual membership fee, usually between Tshs. 5,000 and 10,000, which provides unlimited access throughout the year to outpatient services at CHF-participating facilities. The CHF is essentially a district-level pre-payment scheme for primary care services targeted at the rural population and those working in the informal sector. It is designed to serve the majority of the poor in a more equitable way.

* In Kondoa district, for example, more than three-fourths (81%) of the expected 16,469 deliveries in 1996 were delivered at home and only a handful of these with the assistance of a trained TBA and/or a health staff.

† DHS 1996 and 1999 data analysed by income quintile.

* The CHF accounted for 10% of the total district health budget in 2001; it contributed to less than 20% of the total fees collected at health facilities for health care services (Chee et al 2002). The majority of the contribution of the preceding two years was from user fees and this continued to grow, from 20% of fees collected in 1999 to 77% in 2001.

† Community Health Service Boards (CHBs) are one of the tools of devolving authority and ownership to the LGAs and taking over responsibility for management decisions for district health services. They are being rolled out to all LGAs and the district hospitals (MoH 2003). At lower levels in the health system, Health Facility Committees are being established.

* The absolute per capita amounts of allocations for health services were US$ 3.40 in 1998/99 and US$ 6.60 per capita in 2003/04 (MoH September 2003). The target for the MoH has been to reach US$9 per capita to ensure delivery of the essential health package, though US$12 has been agreed as the international threshold.