Bulletin of the World Health Organization

Beyond fragmentation and towards universal coverage: insights from Ghana, South Africa and the United Republic of Tanzania

Diane McIntyre a, Bertha Garshong b, Gemini Mtei c, Filip Meheus d, Michael Thiede a, James Akazili e, Mariam Ally f, Moses Aikins g, Jo-Ann Mulligan h, Jane Goudge i


Over the past few decades, health sector reforms in many African and other low- and middle-income countries have increased inequities in access to affordable health care. A growing reliance on out-of-pocket payments and privately organized care has resulted in health care provided on the basis of ability-to-pay, which has disadvantaged lower-income socioeconomic groups.

The 2005 World Health Assembly called for universal coverage in health systems.1 WHO defined this as securing “access to adequate health care for all at an affordable price”.2 This definition allows for a high level of fragmentation in health-care provision and financing. Fragmentation refers to the existence of a large number of separate funding mechanisms (e.g. many small insurance schemes) and a wide range of health-care providers paid from different funding pools. Different socioeconomic groups are often covered by different funding pools and served by different providers. Fragmentation reduces the possibilities for income and risk cross-subsidies in the overall health system.3 Although WHO adds that a “crucial concept in health financing policy towards universal coverage is that of society risk pooling”,2 this aspect appears to be ignored in many of the policy prescriptions directed at low- and middle-income countries in recent times.46

To achieve universal coverage through pooling risk to the greatest extent possible, and an equitable health system where ability-to-pay determines financing contributions and the use of services is on the basis of need for care, user fees and other out-of-pocket payments must be reduced and the level of prepayment should be increased7 in a way that maximizes the size of risk pool(s). This can be done by increased tax funding and/or by introducing mandatory (i.e. social or national) health insurance.

The aim of this analysis is to explore the extent of fragmentation within the health systems of three African countries (Ghana, South Africa and the United Republic of Tanzania); how this developed; how each country has attempted to address the equity challenges arising from this fragmentation and what remains to be done to promote universal coverage. This paper draws on the results of the first phase of a three-year project analysing equity in the finance and delivery of health care in Ghana, South Africa and United Republic of Tanzania. More information is available at: http://www.heu.uct.ac.za/shield. A brief overview of the health systems of these countries, using a framework for analysing health-care financing in terms of its key functions,8,9 is provided in Table 1.

From colonialism to fragmentation

Under colonial rule, many African countries, including Ghana, South Africa and the United Republic of Tanzania, organized their health systems primarily to benefit a small elite group of colonials and their workers.10,11 Health-care provision occurred mainly through hospitals in urban areas, with direct payment at the point of use. The rest of the population relied on services from a range of providers such as traditional healers and missionary health centres.

After independence, the governments of Ghana and the United Republic of Tanzania provided medical care free of charge to their populations at public health facilities. Health care was financed through general taxes and external donor support,12 user fees were removed and attention was directed to developing a wide range of primary health-care facilities across the country. At that time private practice was limited, and even prohibited by law in the United Republic of Tanzania in 1977. Post-colonial South Africa, in contrast, did not usher in democratic elections, and apartheid policies reinforced inequities in the distribution of health services between the urban and rural population as well as along racial lines.13 User fees remained in place, albeit relatively token. In addition, private voluntary insurance organizations, called medical schemes, were established by mining and other companies as a way to provide for the health-care needs of their “white” employees (classified as such under the former South African Population Registration Act).14

However, because of worsening economic conditions in the 1980s and the inability to sustain recurrent government expenditure to provide free health care to their populations, Ghana and the United Republic of Tanzania initiated health sector reforms as part of broader structural adjustment programmes under the guidance of The World Bank and the International Monetary Fund. These macroeconomic policies, embedded in neoliberal ideology, aimed mainly at reducing government spending to address budgetary deficits, introducing cost recovery mechanisms through user fees15 and liberalizing health services to allow private sector involvement.16,17 Although not under similar pressure from international financial institutions, the South African government subscribed to many of the prevailing neoliberal macroeconomic policies of the time and introduced similar reforms.18 In particular, South Africa increased the level of user fees substantially and vigorously promoted the growth of the private health sector.19

The reforms in all three countries had a profound impact on the financing and organization of the health sector. The liberalization of the health sector led to a rapid increase in the number of private health providers, many of them informal and unregistered.20 In general, these health sector reforms undermined the potential for cross-subsidies in the overall health system and resulted in increased inequalities in access and utilization of health services. By the end of the 1990s, public resources for the health sector had declined sharply and health system funding relied heavily on cost recovery policies and voluntary health insurance. Following the re-introduction of user fees, the utilization of health services decreased significantly in Ghana12,21 and the United Republic of Tanzania,22 particularly among people on low incomes. As well as the decline in utilization, user fees were also associated with delays in seeking treatment and increased reliance on self medication.23

An additional component of financing reforms during this period was the introduction of risk sharing strategies through community based health insurance (CBHI) in Ghana and the United Republic of Tanzania and the dramatic increase of private voluntary health insurance in South Africa. These voluntary insurance schemes have fuelled health system fragmentation, with over a hundred individual medical schemes in South Africa and, similarly for CBHI, in Ghana24 even though they cover a small proportion of the population (less than 14% in South Africa, and less than 1% in Ghana and the United Republic of Tanzania).

The current fragmentation of the health system into large numbers of small insurance risk pools, especially in South Africa and until recently in Ghana, and the relatively high share of out-of-pocket expenditures as a percentage of total expenditure on health (47% in the United Republic of Tanzania, 45% in Ghana and 11% in South Africa)25 severely limit the potential for universal coverage. Out-of-pocket payments represent the most extreme form of fragmentation as they place the burden of health-care funding on an individual and translate into health service use, and hence benefits, being distributed according to ability-to-pay rather than need for health care.

Health services for different socioeconomic groups and groups with varying health-care needs are not financed in the same way. For example, although CBHI schemes often cover relatively poor communities, they exclude the poorest.26,27 Similarly, private voluntary insurance schemes in South Africa cover the wealthiest groups and have sought to exclude those with the greatest health risks28 who are then dependent on publicly funded health care for which they are generally required to pay user fees (except at the primary care level). This effectively prohibits risk-related and income-related cross-subsidies between groups of different socioeconomic status and health-care needs.

What are some of the effects of this fragmentation? First, some households face a “catastrophic” burden of health-care payments, with expenditure that exceeds 10% of total household income29 or 40% of non-food household expenditure.30 For example, it was estimated that 1.3% of households in Ghana experience “catastrophic” payments (which is above average from a study of 59 countries).30 Second, poorer groups are not able to benefit from publicly funded health services to the extent that their relative burden of ill-health would suggest, as their utilization is deterred by user fees. For example, while the poorest quintile of the population in Ghana in the 1990s received 12% of the benefit of using public health services, the richest quintiles received 33%; the comparable figures for the United Republic of Tanzania are 17% and 29%.31 Finally, an effect of fragmentation in South Africa, which is not evident in the other two countries, is an uncontrolled cost spiral within the private sector. This is largely due to the inability of the many separate medical schemes to negotiate effectively with powerful collectives of private sector providers.13 Thus, fragmentation is not only of concern from an equity perspective, but also in relation to health system efficiency and affordability.

Promoting cross-subsidies

It is worthwhile considering whether and how these three African countries have set about addressing the equity problems of their highly fragmented health systems. User fee exemptions and waivers have been implemented as partial remedies for the lack of a comprehensive system of cross-subsidies in all three countries, in an effort to reduce the economic burden of ill health on poor and vulnerable households and improve access to health care. The current South African health system features free health care for vulnerable groups (particularly pregnant women, children aged less than 6 years, the disabled and the elderly), waivers for the poor and free primary health services for all. In Ghana, exemptions focus mainly on diseases regarded as being of public health importance (e.g. leprosy, tuberculosis), specific services for children and pregnant women (e.g. immunizations, antenatal care) and people with extremely low incomes. The situation is similar in the United Republic of Tanzania with exemptions for priority groups and selected health conditions and waivers on grounds of poverty.

In all three countries, exemptions for specific demographic groups and diseases have been implemented relatively successfully. However, waivers directed at protecting the poorest people have proven to be ineffective,32,33 largely due to the perennial problem of identifying them, as well as a lack of awareness on eligibility criteria and the deterrent of excessive “red-tape”. In addition, the issue of whether user-fee revenue lost from waivers is reimbursed influences the extent to which they are granted at facility level. For example, all exemptions and waivers in Ghana are meant to be reimbursed to individual facilities out of pooled government and donor funds. Inadequate budgeting for exemption and waiver reimbursements and long delays in paying reimbursements have led to some facilities refusing to grant them.33

Developing effective mechanisms for identifying and protecting people with very low incomes is critical in all three countries. Even if user fees were completely abolished, as is happening in a growing number of African countries, it would still be necessary to identify people with the lowest incomes to protect them in relation to other financing mechanisms (e.g. to partly or fully subsidize their health insurance contributions). In addition, if universal coverage is to be achieved, it is necessary to explore ways of achieving funding pools that are as large and integrated as possible, to maximize income and risk cross-subsidies and to allocate pooled resources in an equitable way.

The key pooled funding mechanisms for health care are tax (and donor) funding and health insurance schemes. Although African heads of state, through the 2001 Abuja Declaration, committed themselves to allocating 15% of government budgets to the health sector,34 there has been progress towards this goal only in Ghana, where the health sector’s share of the budget increased from 8.2% in 2004 to 15% in 2006.35,36 A significant component of this growth results from increases in salaries and allowances in the health sector. In contrast, in South Africa the health sector’s share of the government budget has in fact declined from 11.5% of the total government budget in 2000/2001 to 10.9% in 2007/2008.13 In the United Republic of Tanzania, public spending has increased negligibly from just under 10% in the early 2000s to 10.2% of total public spending in 2005/2006.37

In relation to health insurance schemes, there has been little progress in expanding insurance coverage within South Africa. The uncontrolled spiral in medical scheme expenditure and contributions has in fact contributed to a decline in the proportion of the population covered from 17% in the 1990s to about 14% currently. The benefit package has also declined, with many schemes only covering inpatient care and chronic illnesses specified in the Prescribed Minimum Benefits regulation. In contrast, Ghana and the United Republic of Tanzania have made significant progress in expanding insurance coverage. In both countries, until the recent introduction of mandatory insurance, community-based health insurance had been the predominant form of health insurance and it had achieved very limited coverage. These schemes generally only cover outpatient care at primary health-care level. In 1999, the United Republic of Tanzania introduced the National Health Insurance (NHI) fund for civil servants, which now covers 5% of the population. More recently, the National Social Security Fund has introduced a Social Health Insurance Benefit to cover formal sector workers in private firms and some employees in the public sector. Registration to date has been relatively low and accounts for less than 1% of the population. Ghana has taken the boldest steps towards universal coverage by introducing an NHI scheme in 2003, which will ultimately cover all Ghanaians. By December 2007, 55% of the population had registered with the NHI and 44% had received their membership cards.38 The mandatory health insurance schemes in both countries cover quite comprehensive outpatient and inpatient services at public sector and accredited nongovernment facilities.

It is not only expansion of population coverage by pooled funding that is important from a universal coverage perspective but also the degree to which different funding pools are integrated. In South Africa, there has been some consolidation of insurance coverage, with the number of medical schemes declining over the past few years. Nevertheless, there remain over 120 schemes, each with several benefit options that operate as separate pools,39 severely fragmenting the pooling of risk across the insured population. There is an intention to introduce a risk-equalization mechanism between these separate schemes but this is yet to be implemented. In Ghana, each of the district mutual health insurance schemes that comprise the NHI effectively constitutes a separate risk pool. The NHI fund could assume risk-equalization responsibilities, but this has not been done explicitly to date. Instead, it simply transfers certain funds to individual district mutual health organizations. These include the payroll-based health contributions of formal sector employees and government funds used to subsidize the contributions of informal sector workers and the poor. In the United Republic of Tanzania, the decision to introduce the Social Health Insurance Benefit as a mandatory scheme separate from the NHI fund appears to have been largely influenced by the preferences of private sector employers and employees; it is of concern from the perspective of fragmenting risk pools that there is no mechanism for risk-equalization between these mandatory schemes.

Risk-equalization is a mechanism for allocating resources that are pooled via health insurance. Mechanisms are also required to ensure the equitable allocation of funds pooled via tax revenue. Both mechanisms for risk-equalization between insurance schemes and for the allocation of general tax resources ensure that the relative risk of ill-health or likely health-care needs of the population served are taken into account. All three countries use some form of needs-based formula for guiding the allocation of tax resources between different geographic areas. For example, Ghana uses a formula including the regional population size, the population below the poverty line and rates of under-5 mortality to determine the allocation of tax and donor-pooled funds for non-salary budgets in the health sector. The same variables, with the addition of a variable reflecting the transport needs within each district, are used in the formula used to allocate donor “health basket” funds and “block grants” (tax funds and donors’ general budget support) to districts in the United Republic of Tanzania. In South Africa, budgets for the full range of services provided by provinces are allocated to provinces on the basis of a formula which includes estimates of the relative need for these services, with the health component being based largely on the size of the population not covered by private health insurance.

However, the equity promoting effect of these needs-based resource allocation formulae is to some extent offset by the allocation of other streams of tax and donor funds through separate channels. For example, in the United Republic of Tanzania districts are allocated a matching grant equal to member contributions to each district community health fund. While this provides an incentive for districts to generate community health fund revenue, it cannot be described as an equitable allocation mechanism as relatively poor districts are less able to generate these contributions in the first place. In Ghana, there is a separate funding channel for reimbursing fee revenue lost through granting exemptions. Thus, resources are not allocated according to the relative need for fee exemptions (e.g. based on the poverty levels in that district) but on the basis of the number of exemptions actually granted. As indicated previously, there are considerable problems in the implementation of exemptions, which are likely to be more severe in areas of lowest-income, which have the lowest staffing levels and weakest service delivery and exemption implementation capacity.


There is growing international consensus that out-of-pocket payments are contrary to the goal of universal coverage, particularly given the ineffectiveness of fee waivers in providing financial protection to the poor. There is also consensus that universal coverage can only be achieved through prepayment funding mechanisms. However, it is of concern that financing strategies (such as CBHI and private voluntary health insurance) that inevitably further fragment health systems are still being promoted as useful financing mechanisms for low- and middle-income countries.46

The analysis presented indicates that South Africa has made the least progress in addressing fragmentation, while Ghana appears to be pursuing a universal coverage policy in a more coherent way. To achieve universal coverage, the size of risk pools must be maximized. Further, resource allocation mechanisms must be put in place, whether these are to equalize risks between individual insurance schemes or to equitably allocate general tax (and donor) funds. Ultimately, there is a need to achieve as much integration of financing mechanisms as possible to promote universal cover with strong income and risk cross-subsidies in the overall health system. ■


We thank our colleagues from SHIELD (Strategies for Health Insurance for Equity in Less Developed countries)who contributed to the initial country specific reports: Irene Agyepong, John Gyapong and Frank Nyonator in the Ghana report; Marianela Castillo-Riquelme, Ermin Erasmus, Lucy Gilson, Vimbayi Mutyambizi and Moremi Nkosi in the South African report; and Anne Mills and Natasha Palmer in the Tanzanian report. We would also like to thank Bronwyn Harris for her helpful comments on an earlier version of the manuscript.

Funding: This research was funded by IDRC (Grant number 103457) and the European Commission (Sixth Framework Programme; Specific Targeted Research Project no: 32289). Diane McIntyre is supported by the South African Research Chairs Initiative of the Department of Science and Technology and National Research Foundation.

Competing interests: None declared.



  • Health Economics Unit, University of Cape Town, Anzio Road, Observatory, 7925, South Africa.
  • Health Research Unit, Ghana Health Service, Accra, Ghana.
  • Ifakara Health Research and Development Centre, Dar es Salaam, the United Republic of Tanzania.
  • Development, Policy and Practice, Royal Tropical Institute, Amsterdam, the Netherlands.
  • Navrongo Health Research Centre, Ghana Health Service, Upper East Region, Ghana.
  • Ministry of Health and Social Welfare, Dar es Salaam, the United Republic of Tanzania.
  • College of Health Sciences, University of Ghana, Accra, Ghana.
  • Health Economics and Financing Programme, London School of Hygiene and Tropical Medicine, London, England.
  • Centre for Health Policy, University of the Witwatersrand, Johannesburg, South Africa.