The World Health Assembly of 2005 called for all health systems to move towards universal coverage, defined as "access to adequate health care for all at an affordable price". A crucial aspect in achieving universal coverage is the extent to which there are income and risk cross-subsidies in health systems. Yet this aspect appears to be ignored in many of the policy prescriptions directed at low-and middle-income countries, often resulting in high degrees of health system fragmentation. The aim of this paper is to explore the extent of fragmentation within the health systems of three African countries (Ghana, South Africa and the United Republic of Tanzania). Using a framework for analysing health-care financing in terms of its key functions, we describe how fragmentation has developed, how each country has attempted to address the arising equity challenges and what remains to be done to promote universal coverage.The analysis suggests 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, health systems must reduce their reliance on out-of-pocket payments, maximize the size of risk pools, and resource allocation mechanisms must be put in place to either equalize risks between individual insurance schemes or equitably allocate general tax (and donor) funds. Ultimately, there needs to be greater integration of financing mechanisms to promote universal cover with strong income and risk cross-subsidies in the overall health system.

Little is known about health system equity in Tanzania, whether in terms of distribution of the health care financing burden or distribution of health care benefits. This study undertook a combined analysis of both financing and benefit incidence to explore the distribution of health care benefits and financing burden across socio-economic groups. A system-wide analysis of benefits was undertaken, including benefits from all providers irrespective of ownership. The analysis used the household budget survey (HBS) from 2001, the most recent nationally representative survey data publicly available at the time, to analyse the distribution of health care payments through user fees, health insurance contributions [from the National Health Insurance Fund (NHIF) for the formal sector and the Community Health Fund (CHF), for the rural informal sector] and taxation. Due to lack of information on NHIF and CHF contributions in the HBS, a primary survey was administered to estimate CHF enrollment and contributions; assumptions were used to estimate NHIF contributions within the HBS. Data from the same household survey, administered to 2224 households in seven districts/councils, was used to analyse the distribution of health care benefits across socio-economic groups. The health financing system was mildly progressive overall, with income taxes and NHIF contributions being the most progressive financing sources. Out-of-pocket payments and contributions to the CHF were regressive. The health benefit distribution was fairly even but the poorest received a lower share of benefits relative to their share of need for health care. Public primary care facility use was pro-poor, whereas higher level and higher cost facility use was generally pro-rich. We conclude that health financing reforms can improve equity, so long as integration of health insurance schemes is promoted along with cross-subsidization and greater reliance on general taxation to finance health care for the poorest.