Brief historical overview

The early history of private health insurance saw the unregulated formation of company-based health care funds, which were subsequently formalized under the Friendly Societies Act (1956) and later under the Medical Schemes Act (1967) (see Box 12.2). Medical schemes were initially subject to community rating and minimum benefits requirements, and different schemes provided very similar benefits. Reforms implemented during the 1980s and early 1990s led to increasing use of private insurance principles (Reference McLeodMcLeod, 2005). The 1986 Browne Commission supported the application of risk and experience rating of contributions, as well as more individualized benefits (Department of Health, 2002). It also argued that there would be significant cost savings if members paid small claims themselves and only claimed from pooled funds after paying an initial amount. These measures would substantially reduce the size of risk pools and the degree of cross-subsidy between young, old, healthy and sick people.

Before 1989, a medical scheme could only vary its contribution rates based on income and number of dependants. After 1989, contributions could also be based on age, geographic area, extent of cover provided, length of membership, size of group and actual claims experience. Schemes were therefore able to reduce existing cross-subsidies in line with the view prevailing among insurance companies that cross-subsidies within health insurance were unfair. Hence, in 1994, the first democratic government inherited a medical scheme system characterized by risk rating. The same year, a leading scheme introduced personal medical savings accounts (MSAs), with no intervention from the regulator, and other schemes rapidly followed suit (see Chapter 13 in this volume).

Aggressive competition among medical schemes for low-risk members (risk selection) in the 1990s had adverse consequences in terms of health care equity and access, with older people and those with chronic diseases particularly affected. Throughout the 1990s, benefits declined and vulnerable groups were increasingly excluded from coverage. By 1999, no open schemeFootnote ⁴ permitted anyone over the age of 55 to join as an individual member and almost all open schemes applied lifetime exclusions for pre-existing conditions and used experience rating without restriction (Department of Health, 2002). Since 1994, policy formulation and legislation have turned back to solidarity principles, although medical schemes still operate in a voluntary membership environment. The substantially revised Medical Schemes Act No. 131 (1998) (the Act) has applied since 1 January 2000.

The 1998 Medical Schemes Act

Since 1998, medical schemes have been clearly demarcated as the only vehicles providing for the reimbursement of actual spending on health (known as indemnity cover). Hence, any funding arrangement that is intended to assist in meeting the actual costs of medical services must satisfy the requirements of the Medical Schemes Act. These requirements include being registered under the jurisdiction of the Council for Medical Schemes (CMS), having financial guarantees, maintaining prescribed solvency levels, having at least 6000 principal members,Footnote ⁵ adhering to product design requirements and regular reporting to the Registrar of Medical Schemes, who is the chief executive of the CMS.

Crucially, the Act re-introduced three key principles to enhance risk pooling within schemes (Reference McLeodDoherty & McLeod, 2003):

• open enrolment: open schemes have to accept anyone who wants to become a member at standard rates;

• community rating: everyone must be charged the same standard rate, regardless of age or state of health. Community rating currently applies to each benefits option in each medical scheme rather than for the industry as a whole;

• prescribed minimum benefits: a minimum benefits package as regulated by the CMS must be offered by all schemes. Beneficiaries must be covered in full for the specified conditionsFootnote ⁶ with no financial limits or cost sharing. Schemes may insist on the use of a contracted network of preferred providers and drug formularies to manage care.

As the Act came into force, some schemes attempted to create hybrid products combining a medical scheme plan with high deductibles (no underwriting, community rating, low commission) with a health insurance policy that exactly met the deductible (underwriting, risk rating, very high commission). The net effect was essentially to continue to offer risk-rated cover. Another challenge to the new environment included the excessive use of reinsurance contracts to transfer all of the risk from the non-profit medical scheme to a registered long-term insurer in the same financial group. These activities required decisive action by the new regulator. Despite considerable opposition over several years, the regulator prevailed and the industry settled down to the new legislative environment.

The Medical Schemes Act of 1998 provided for much stronger governance of medical schemes and the industry as a whole. Each scheme is now governed by a board of trustees, of which half are elected by members, and their duties are codified by the Act. The Act also tasked the CMS with the protection of beneficiaries, rather than with the protection of the industry, as was previously the case. Its executive arm, the office of the Registrar of Medical Schemes, has grown substantially from being a small deputy directorate within the Department of Health to being a statutory body with expanded regulatory powers. Funded by a levy on medical scheme members, the office is staffed by a professional team including accountants, lawyers, health economists and health professionals. In addition to regulation, the CMS is responsible for accrediting organizations that provide services to medical schemes.

Trends in private health insurance

Medical schemes have steadily declined in number from 305 in 1974 to 83 schemes in 2015 (CMS, 2016), and further consolidation is expected. In 1994, open and restricted schemes had approximately equal numbers of members. Open medical schemes, under the direction of aggressive insurance companies, began to use brokers to attract business with favourable risk profiles and gradually the entire open scheme market followed. Brokers were only legally recognized from 2000, by which time nearly 70% of all medical schemes members belonged to open schemes. Brokers’ costs incurred by open schemes constituted 1.2% of gross contribution income and 14% of non-health care spending by these schemes in 2015 (CMS, 2016). This cost is of concern because it mostly relates to members being ‘churned’ from one open scheme to another, rather than to acquisition of new members. By 2007 there were more individual health brokers accredited with the CMS (over 9700)Footnote ⁷ than there were GPs (approximately 7000; Econex, 2009).

Member numbers have increased steadily from 1.4 million in 1974 to 3.95 million in 2015, and the number of beneficiaries grew from 3.5 million to 8.8 million over the same period (CMS, 2016). In the mid-2000s, there were an estimated 1.5 million people without cover who live in households in which somebody has cover (Reference McLeodMcLeod, 2007). Although these numbers are somewhat dated, they highlight that some families do not register all their children with schemes. While there are age restrictions whereby children cannot be covered when they reach 21years, a key issue influencing coverage is that contributions must be paid for each dependant (even though the contribution rate per child is lower than that for adults), suggesting that some families struggle to afford medical scheme cover.

Based on the available information, the distribution of health coverage is as follows. Over 84% of the population is entirely dependent on tax-funded hospital-based services (inpatient care and specialist outpatient care). About 20% of this population uses private sector primary care services (mainly GPs and retail pharmacies) on an out-of-pocket basis while the remaining 80% use public sector primary care services. Just over 16% of the population is covered by medical schemes. About 45% of medical scheme beneficiaries belong to schemes with an MSA component.

The problem of medical scheme affordability is considered to be the greatest obstacle to growth in the industry. At higher income levels, some 75–85% of the employed are already members. This share falls rapidly for lower-income groups so that, among those with earnings just above the tax threshold, only one third are members (Reference McLeodMcLeod, 2010). Less than 10% of workers with earnings below the tax threshold can afford medical scheme cover. This drop in affordability is compounded by the inability of many lower-income families to cover all of their family members, even with an employer subsidy.

Changes in relation to key actors Footnote ⁸

Although medical schemes are non-profit entities, they are surrounded by a number of for-profit entities that provide administration, marketing, managed care, consulting and advisory services. Consumers can therefore confuse the non-profit medical schemes and the high-profile listed companies that act as third-party administrators. Since 2000, scheme administrators and managed-care companies have been required to be accredited by the CMS. There has been some consolidation among the largest administrators, but many small administrators have been set up to profit from this lucrative business. In 2015, there were 16 third-party administrations, down from 23 in 2009 (CMS, 2016). The market is highly concentrated and the largest three administrators provide services for 80% of beneficiaries. Less than 7% of beneficiaries belong to schemes that are self-administered. In 2015, there were 25 accredited managed-care organizations; there are some concerns about the risk-taking role of some of these organizations. As indicated earlier, the increasing and substantive involvement of brokers over the past decade and a half is also of concern.

The role of employers in medical schemes has been declining for some time. Over the past decade, a growing number of employers have chosen to limit their involvement with medical schemes to paying contributions to open medical schemes. This is partly due to the changing nature of employment, with people no longer spending their entire working life with one employer. In the 1970s, it was typical for an employer to provide medical cover and other employee benefits over and above a cash salary. In the 1980s and 1990s there was a growing move to so-called cost-to-company remuneration; as employers became concerned about the total cost of an employee (cash salary plus benefits), they fixed the total remuneration level, but also gave employees more flexibility in choosing cash or benefits. Medical scheme membership is seldom offered as an additional benefit but included as part of an employee’s total remuneration. Moreover, the presence of brokers and the incentives offered to them to direct members to open schemes has exacerbated the declining role of employers.

Trends in relation to income and risk cross-subsidies

Medical schemes are not permitted to differentiate contributions by age, gender or state of health, but they can use income-related contributions as a factor to build in a deliberate cross-subsidy from high-income to low-income workers. The proportion of options within restricted schemes making use of income-related contributions has declined significantly over time, from 83.9% in 2004 to 61.0% in 2006. Nevertheless, restricted schemes make greater use of income-related contributions than open schemes. Open schemes struggle to obtain reliable income information and have largely moved away from any internal income cross-subsidies, not least because they fear attracting low-income older people instead of more desirable low-income, young workers.

Since 1994, surveys have documented a significant movement towards excluding pensioners from company funding for health care, which reduces risk cross-subsidies within medical schemes. The full effect of transferring investment risk and medical inflation risk to older people and future pensioners will take some time to unfold. Reference McLeodMcLeod et al. (2003) described the issue as “a future affordability time bomb”, warning that it will affect the industry when those joining companies from around 2000 onwards reach retirement age. The issue has been receiving some attention from the Department of Social Development as part of the evaluation of retirement reform proposals (Reference McLeodMcLeod, 2007) that postulate setting up a mandatory social security system for retirement.

Developments over the last decade

A very positive development over the last decade has been the government’s re-establishment of a restricted scheme for public sector workers. In the apartheid years, the public sector had a restricted scheme for each racial group, but during the 1990s public sector workers were allocated an employer subsidy for medical schemes and could use this to join any open medical scheme of their choice. This approach fuelled the growth of the role of brokers in the 1990s. A review of remuneration for public sector employees identified major shortcomings in relation to medical scheme cover, including inequality in access to cover, affordability concerns, lack of value for money, spending inefficiencies and little integration with public sector health care (Reference McLeod, Ramjee, Harrison, Bhana and NtuliMcLeod & Ramjee, 2007). Despite a relatively generous medical scheme subsidy, less than half of the 1 million state employees were using the subsidy, and those that did so were using open schemes.

In 2002, the cabinet approved a framework policy for a new restricted medical scheme for public sector employees only, centred on the principles of equity (equal access to basic benefits), efficiency (with respect to costs and delivery of benefits) and choice of benefits (employees could choose more expensive cover and pay for it themselves). The Government Employees Medical Scheme was registered in January 2005 and became operational in January 2006. Government attracted potential members by offering a higher medical scheme subsidy within the Government Employees Medical Scheme and making it the only option available for new employees. The scheme grew rapidly and by December 2009 it had over 400000 principal members or over 1.1 million beneficiaries. It has since grown more slowly, reaching a peak of 1.85 million beneficiaries in 2013 and declining to 1.77 million beneficiaries in 2015 (CMS, 2016). (The evolution of the number of beneficiaries of open and restricted schemes is shown in Fig. 13.2 in Chapter 13.) It is the largest restricted scheme, accounting for over 45% of all restricted scheme beneficiaries, and the second-largest medical scheme in South Africa. The high subsidies, income-related contributions and strong purchasing power of the new scheme proved successful in many respects. In particular, a low-wage-earning civil servant and his or her family are able to join the lowest cost option without contributing (that is, the government pays the full contribution). The implementation of the Government Employees Medical Scheme sets an example to other employers by demonstrating that it is possible to develop benefits packages that are affordable to all employees.

There were also efforts to promote risk pooling across individual schemes and across benefits options within schemes. A proposed amendment to the Medical Schemes Act in 2008 would have given the CMS responsibility for the operation of a Risk Equalization Fund. However, this amendment was never passed by parliament; efforts to introduce a range of new regulations of medical schemes were overtaken by the ruling party’s decision to rather pursue a tax-funded, single purchaser system (see below). The purpose of the Risk Equalization Fund that was proposed was to protect open enrolment and community rating by ensuring that risk selection (deliberately attracting the young and healthy) was penalized. The Risk Equalization Fund would have created an industry-wide risk pool for those covered by medical schemes, and ensured an industry-wide community rate for the PMB, regardless of the age and disease profile of each option.

Policy-makers also signalled a desire to reduce the number of benefit options. In 2004, an International Review Panel argued that only a limited number of supplementary benefits options should be allowed above the PMB (Armstrong et al., Reference Armstrong2004). The 2008 Medical Schemes Amendment Bill took a different line, suggesting that all hospital events should be covered in a single basic pool and that benefits above that level must also be pooled across all members in the scheme that choose an extension of cover. In essence, this would have changed benefit design from a series of vertical silos, each with its own community rate, to a series of horizontal pools and non-overlapping benefits pools, each with a community rate. Members would then have been able to choose which benefits they wanted and the contribution rate would have been determined by summing the prices of components of the benefits package. There was a strong negative reaction to this proposal from the schemes and the bill was not passed. Benefit package reform has largely stalled; the only change, introduced in 2008, was to allow efficiency-discounted options (EDOs) (CMS, 2016: p.27). There is no difference in benefit coverage between EDOs and non-EDOs; EDOs are simply more restrictive in the network of service providers that members can use. By 2016, only eight open schemes had introduced an EDO (CMS, 2016).

Financial protection

The extent to which medical schemes have contributed to ensuring universal protection against the financial risks of ill health has changed considerably over time. In colonial times, South Africa had a strong public health sector that covered almost all citizens. Mission hospitals served some rural populations where there were no public facilities, but over time these hospitals were integrated into the public health sector. During the apartheid era, separate public health services were provided for white and black people (for example, there were always at least two public hospitals in large towns). Token user fees were charged at public facilities and as there was no incentive for facilities to generate user fee revenue, given that all revenue was transferred to the Treasury, fee exemptions were applied liberally. Generous exemptions constituted substantial protection from direct health care costs, although many patients, particularly in rural areas, had to incur considerable transport costs due to the relatively limited provision of services in rural areas.

Medical schemes initially developed to cover the costs of privately provided primary care (particularly GP visits, retail pharmacy services and some specialist care), essentially a response to the rising expectations of wealthier, particularly white, people. In this way, they provided financial protection for those who wished to spend more on health care. However, the costs that were covered, mainly for routine primary care, were relatively small, and medical scheme members are now expected to cover these costs on an out-of-pocket basis or through MSAs.

In the 1980s, there was an explicit government policy of promoting the growth of the private sector (Working Group on Privatization and Deregulation, 1986; South Africa (Republic) 1987). As noted by a policy-maker at that time: “Health authorities must not be seen as an infinite source of health facilities and medical care. More people should be able to make use of private health facilities” (Reference RossRoss, 1982). As a result, by 1990, 62% of general doctors and 66% of specialists worked in private practice (Reference Rispel and BehrRispel & Behr, 1992), compared with about 40% of doctors in the early 1980s (Reference NaylorNaylor, 1988), and the number of for-profit general hospital beds nearly doubled from around 9900 beds in 1988 to over 18400 in 1993 (McIntyre et al., Reference McIntyre1995). Public facilities were racially desegregated in 1988 and the use of public hospitals by medical scheme members began to decline very rapidly, with an equally rapid growth in the use of private hospitals. During this period, medical scheme membership among the black population grew particularly fast, largely as a result of trade union demands for employers to provide medical scheme cover to black staff and as part of the government’s efforts aimed at “the co-option of certain groups of urban blacks by improving their immediate living environment and welfare services, and by offering them a stake in the capitalist system” (Reference 413PricePrice, 1989).

These developments made out-of-pocket payments for privately provided health care unaffordable, particularly in relation to the growing use of private specialist and hospital care. Consequently, medical schemes are now an important mechanism for providing protection from potentially catastrophic costs for those who choose to use private facilities. However, medical schemes also provide less financial protection than previously because of rapidly rising health care costs and reductions in benefits (increasing cost sharing and growing exclusion of services from cover). Many scheme members cover these costs out of pocket. By the late 1990s, two thirds of all out-of-pocket payments for health care were made by medical scheme members (Cornell et al., Reference Cornell2001). The most significant development in recent years has been the introduction of PMBs in 2000. These have been critical to providing financial protection for those choosing to use privately provided health care.

Equity in financing

Contributions to medical schemes are community rated and so not explicitly linked to age or health status. However, given that each scheme has a number of different benefit options, it is possible that there is an element of self-selection into more comprehensive and higher-cost benefits options by higher-risk members, if they have sufficiently high incomes to afford these options. The most recent financing incidence study found that the incidence of medical scheme contributions is regressive across members. Contributions by the lowest-income 40% of members are equivalent to 14% of household consumption expenditure, whereas contributions by the richest quintile of members are less than 6% (Reference McIntyreMcIntyre, 2010). This is largely due to most contributions being charged as a flat monetary amount. Higher-income individuals tend to choose more comprehensive benefits packages and thus, in general, pay higher contributions than lower-income groups, but the regressivity of contributions indicates that the differentiation of the flat contribution rate across benefits packages is not as great as the differences in income levels across members.

When considering the overall incidence of health financing across all South Africans, medical scheme contributions are strongly progressive for two main reasons (Reference Ataguba and McIntyreAtaguba & McIntyre, 2012). First, the highest-income groups belong to medical schemes and, second, the contributions these groups make to schemes are very high. However, the fact that the magnitude of the contributions is related to inefficiencies and poor cost containment within this environment (see below) detracts from this so-called progressivity: a far greater proportion of the population could benefit from the funds in medical schemes if the schemes operated more efficiently. Also, although the higher-income groups that belong to medical schemes bear the greatest burden for financing health care, they also derive all the benefits provided by the schemes.

Evaluation of the impact of medical schemes on equity in financing also needs to take into account the substantial subsidy provided by the government in the form of tax relief for contributions. According to the most recent estimates, these tax credits amounted to 16 billion South African Rand in the 2014/2015 financial year (Department of Health, 2015). In addition, the government devotes a considerable amount of general tax revenue to pay medical scheme contributions on behalf of civil servants, estimated to exceed 20 billion South African Rand in 2014/2015 for the largest civil servant schemes. These tax credits and contributions to medical schemes reduce the amount of tax resources available for those who rely on publicly funded health services. The magnitude of this subsidy has been repeatedly criticized by health sector analysts, especially since the tax deductions particularly benefit those with the highest incomes and do not benefit employees who fall below the income tax threshold. Partly as a result of these criticisms, the Treasury revised tax concessions for medical schemes in 2005, spreading the subsidy more equitably across tax payers; nevertheless substantial public funds are directed to supporting medical schemes.

Equity in provision and use

Private providers are heavily concentrated in urban areas; this does not create too many barriers to accessing services for medical scheme members because most of them live and work in urban areas, although as service provision is relatively biased towards the wealthiest localities within urban areas, it is the lowest income scheme members who face limitations in access. Nevertheless, given the overall magnitude of private provision relative to the population served, there are very few barriers to service access for medical scheme members.

In terms of use of health services, an analysis of data from the only household survey that permits the calculation of utilization rates indicate greater utilization rates among medical scheme members (an average of 5.5 outpatient visits per beneficiary per year and 144 inpatient admissions per 1000 beneficiaries per year) in 2008 than among those who are not members (4.1 outpatient visits per person and 92 inpatient admissions per 1000 people per year) (Harris et al., Reference Harris2011). Use of services by medical scheme members occurs largely in the private sector (85% of outpatient and 82% of inpatient care), with very limited use of private providers by those not covered by medical schemes (17% of outpatient and 4% of inpatient care).

Medical schemes have played an important role in removing financial barriers to using health care when needed for those with cover. However, in terms of the overall health system, they have entrenched inequities in access to health care between those with private cover and those relying on publicly funded health care. Developments since 1995 may have changed this pattern. Severe reductions in medical scheme cover of routine health care spending may have led to differences in rates of service use by socioeconomic statusFootnote ⁹ among medical scheme members. The introduction of free primary care at public facilities in 1996 has also reduced financial barriers to health care for those without medical scheme cover. Unfortunately, recent data on utilization rates that could demonstrate these likely changes are not available.

Quality of care and efficiency in the delivery of services

There are almost no data that would allow evaluation of the quality of care or efficiency of service delivery. There is no routine assessment of the quality of care purchased by the medical schemes, and patients have limited recourse when they receive poor quality of care. They may report providers to the Health Professions’ Council of South Africa, but experience has shown that this institution seldom takes action. A Hospital Rating Index developed by the largest medical scheme administrator (Discovery), which took into account cost, quality and value for money, was scuttled due to vociferous opposition from private hospitals. The reality is that scheme members often have no real choice of hospital. Almost all private specialist practices are based in private hospitals and thus the specialist to whom one is referred automatically determines the hospital of admission. In addition, choice of hospital for specific procedures is usually limited outside the largest metropolitan areas.

Although data on efficiency are limited, it is evident that medical schemes have few incentives to operate efficiently. Evidence on the rate of spending increases, particularly in the periods 1981–1993 and 1996–2004 (McIntyre et al., Reference McIntyre1995, Reference McIntyre2007), indicates that schemes have paid very little attention to ensuring that services are provided at the lowest possible cost. This appears to have continued over the last decade. Medical schemes use the standard set of managed-care practices, such as pre-admission certification and drug utilization review, but it is not clear that these have been effective. One initiative aimed at enhancing technical efficiency is the promotion of generic medicine use through user charges for prescriptions, but this is a relatively passive approach on the part of the schemes. Government legislation and regulation, such as enabling generic substitution by pharmacists and medicine-pricing regulations, have been far more effective in controlling the costs of medicines than actions by medical schemes.

Medical schemes have very limited purchasing power and are unable to effectively negotiate reimbursement rates with providers, either in terms of the form of payment mechanism or the level of the reimbursement rate. This is particularly true in relation to private hospitals (which account for the largest portion of medical scheme spending), given how concentrated the market has become, with three groups owning over 80% of all private hospital beds (Reference van den Heevervan den Heever, 2007).

A few of the largest medical schemes have succeeded in negotiating a per diem rate with certain private hospital groups, and a few schemes serving low-income employees pay GPs on a capitation basis, but fee-for-service is by far the dominant payment mechanism. As indicated in Box 12.2, former statutory mechanisms for establishing provider reimbursement rates are no longer in place. There has been discussion since 2007 about establishing a central bargaining chamber, and in 2014 the Competition Commission initiated a Market Inquiry into the Private Healthcare Sector to address growing concerns about the high costs of private health care.

In terms of allocative efficiency, there is potential for medical schemes to change their benefits packages as health care priorities change, but there is no evidence that they have adapted benefits on this basis. Once again, it is regulatory intervention that has promoted allocative efficiency goals. For example, the PMBs have focused on chronic diseases and hospital care to address the major burden of disease among scheme members, including adding AIDS to the list of chronic diseases, a clear demonstration of allocative efficiency concerns. The PMBs have also been involved in the development of very detailed treatment protocols, to enhance cost-effectiveness. However, medical schemes have undermined allocative efficiency by reducing cover of primary care, which must be paid for via MSAs or out of pocket, whereas specialist visits are still usually covered. This has promoted a relative reduction in the use of GPs and an increase in the use of specialists.

The existence of 83 schemes and some 300 benefit options does not contribute to administrative efficiency. The rate of increase in non-health care spending within medical schemes, particularly in the late 1990s and early 2000s, was dramatic. This spending, which includes administrative costs, managed-care management services, reinsurance costs and broker fees, accounted for only 8.4% of total scheme spending in 1997 but had increased to 16.2% by 2006 (CMS, 2008). This was of considerable concern, especially given that managed-care efforts do not seem to be particularly effective and brokers are not introducing new members but simply juggling existing members between different schemes. The Council for Medical Schemes consistently interrogated these costs in their annual reports. By 2015, non-health care spending had been reduced to 8.6% of medical scheme spending (CMS, 2016).

Impact on the health system

The existence of voluntary private health insurance has had a profound and largely negative impact on the overall health system. First, it has seriously limited the potential for income and risk cross-subsidies, separating 16% of the population (the highest earners) from the rest of the population. At present, there is no risk pooling between schemes, or between different benefit options within schemes. Of more concern is the fact that the wealthy minority belong to schemes that dispose of 42% of the financial resources in the health sector. This disproportionate spending highlights the system’s deficiency in relation to income cross-subsidies.

Second, voluntary private health insurance has been the main contributor to high and rapidly increasing health care spending, which is not translated into commensurate improvement in health status. Public spending on health care flatlined in real per person terms from the mid-1990s to 2003, whereas the spending spiral in medical schemes continued unabated, with real contribution rates per beneficiary increasing by over 6% per year between 1996 and 2006. However, real public spending on health care has been increasing over the past decade, while medical scheme contributions have continued to grow at rates well above the consumer price index. Real per person spending on medical scheme beneficiaries was 3.6 times greater than the government’s per person spending on nonbeneficiaries in 1996. This differential rose rapidly to 7.6 times by 2003, but has since declined to 4.8 times in the early 2010s.

Third, private provision could not have grown as rapidly in the absence of voluntary private health insurance. Prepayment to schemes (as opposed to out-of-pocket payments to private providers) and the fee-for-service provider payment mechanism have induced a migration of health professionals from the public to the private sector. This has created a vicious cycle of health professional incomes becoming far higher in the private sector (with the possible exception of nurses), resulting in a further exodus of staff to the private sector; a growing number of professionals serving a stagnant pool of medical scheme beneficiaries, leading to increases in the fees charged by these professionals as well as supplier-induced demand and uncontrolled growth in medical scheme spending; and medical scheme cover becoming increasingly unaffordable at the same time as schemes being unable to expand cover to lower-income workers. All of this leaves the public sector with a limited number of health professionals attempting to serve the needs of the vast majority of the population in the context of growing health needs, particularly those related to the HIV/AIDS epidemic, contributing substantially to a decline in the quality of publicly provided care.