ARTICLE
          WOULD SOCIAL HEALTH INSURANCE
     IMPROVE SOUTH AFRICAN HEALTH CARE?
    WHAT OTHER MIDDLE INCOME COUNTRIES
                                   CAN TEACH US
                                                                       Max Bachmann
  Introduction
   Government finance for health care is unlikely to increase markedly in South
 Africa in the near future, despite major deficiencies in basic services. While the
 state's ability to raise taxes is constrained, there are competing demands from
 other neglected social sectors such as education and housing. Health care
 expenditure is however already substantial, accounting for about 6% of Gross
 Domestic Product, half being financed through private health insurance (Mc-
 Intyre, 1993). Social health insurance, usually referred to as National Health
 Insurance (NHI), has been advocated as a mechanism for expanding public
 control of health resources and/or for raising additional revenue for health care,
 c f ^ T 1 ^ * 6 P 6 1 ^ 1 ™ 5 of the advocates (Broomberg and De Beer, 1990;
                      Can N a t i o n a l
    Thfw M                                 Congress, 1992; Heard, 1993).
    ine world Bank has promoted social health insurance for middle-income
 developing countries internationally. The 1993 World Development Report's
 main rationale for promoting health insurance is that by providing a payment
 mechanism for financing 'discretionary' care for those who can afford it, NHI
 auows^govenuneiils to focus expenditure on ensuring universal access to
 J ^ T T T ^ f * 6 ' i n t e r v e n t ions which have the greatest impact on death
                        d B a n k 199
                                    '      3:119-123). The Reportalso argues that by
                        m S U r a n C e C 0 v e r to
                                                     **•*» a *& package of essential
                              1 ta 8et            e
                          S! ' * • * *Penditure on the uninsured poor. Other
                                ^ i n c r e a s e fi"^*31 resources for public health care
                             Cttati0n
                          ?                ' ****** PeoPIe are more willing to pay in-
                          i n retUm for a
                                                   ^gibtebenefit, and because a health
                                                         b
                                   " * **"*»*              * governments to finance non-
ma^unS^SSmtr^ 80 " 6 ^                    mi
comDuLsorv mnt!rtl »• ^                     <ldle-income countnes. It is usuaUy based on
compulsory contributions by employers and employees in the form of a payroll
26
                                                              TRANSFORMATION 24 (1994)


BACHMANN                                                                   ARTICLE
tax. Individuals unable to afford compulsory insurance would usually be ex-
cluded from benefits, but could potentially be subsidised by the premiums of
wealthier workers and/or by government contributions. NHI may be conceived
narrowly as a financing mechanism, or more broadly as an integral part of a
national health care programme. While the feasibility of NHI financing in South
Africa needs detailed investigation, of equal concern is the potential impact of
NHI on the nature and distribution of health care provision.
    South Africa has poorer health indicators than other countries with similar per
capita incomes (Table I). A large proportion of the population has poor access to
basic effective services. Distribution of health care is grossly unequal, whether
 one compares geographical areas, racial categories, or public and private sectors.
 Doctors appear to be the most unequally distributed resource. Doctonpopulation
 ratios are 11 times higher, and bed:population ratios 1.9 times higher, in the
 country as a whole than in 'homelands', (Naylor, 1988), and doctonpopulation
 ratios are 6.8 times higher in the private than in the public sector (Rispel and
 Behr, 1992). Health care expenditure per capita in 1987 was 4.3 times higher for
 whites, and 2.5 times higher for Asians, than for Africans (Mclntyre and
 Dorrington, 1990). Although less than 20% of the population has private in-
 surance coverage, 59% of doctors are in the private sector, while public services
 for the uninsured 80% employ only 41% of doctors (Rispel and Behr, 1992).
 Under a democratic government, popular pressure for greater equity in health
 care is likely to increase.
     Fee-for-service payment for private health care - in which doctors, hospitals
  and others are paid proportionally to the quantity of services they provide - has
  impaired efficiency by encouraging excessive medication, surgery, hospitalisa-
  tion, and doctor consultations in South Africa (Broomberg and Price, 1990a,
  1990b; Price and Broomberg, 1990). Costs to private insurance of medication
  and hospitalisation increased by 36% and 49%, respectively, between 1989 and
   1990 (Mclntyre, 1990), reflecting increasingly technical care. The public health
  service is also excessively hospital-oriented, with about 30% of the total health
  budget devoted to 13 teaching hospitals (Mclntyre, 1993).
    As this paper will show, there are many options within a social health insurance
  programme, with corresponding policy implications. It will be argued that a well
  managed public programme in South Africa could improve the efficiency of
   health care by, for example, promoting primary health care, and could improve
   equity by extending better health care to a greater proportion of the population.
   However, as universal insurance coverage is improbable given high unemploy-
   ment rates, NHI would probably have to begin with people in formal employment
   and their families, with the state continuing to provide services directly to the
   rest of the population. Thus publicly managed NHI is likely to involve the new
   TRANSFORMATION 24 (1994)                                                       27


ARTICLE                                                                BACHMANN
government in formalising inequality between the insured and uninsured, in-
stitutionalising the social divide between the employed and unemployed. The
government faces a dilemma: either to allow the private health sector to continue
to lavish health resources on the minority that can afford them (possibly adding
some regulations), or to take control of the finances and accept responsibility for
what will probably be a two-tier national health system. The central policy issue
is control of health care through control of finance. In any case, protracted
negotiations with employers, unions, and health professionals will be necessary
to accommodate conflicting interests. It is thus unsurprising that the African
National Congress's National Health Plan for South Africa (African National
Congress, 1994) has deferred a decision on this question, proposing only that a
commission of enquiry be established after elections, following a strategy similar
to the Clinton administration's national health plan in the United States.
  This paper will draw heavily on the experience of middle-income countries of
Asia and Latin America. These countries are more economically and socially
comparable to South Africa man are developed nations, or the poorer countries
of sub-Saharan Africa. The former countries have similar per capita incomes to
South Africa, indicating that similar levels of health care are affordable (Table
I). They also have similarly unequal income distributions, making equity con-
siderations important Many Latin American countries have more than SO years
of experience of social health insurance, and some Asian countries have, more
recently, attained near-universal insurance coverage.
Insurance coverage
  The proportion of the population insured (coverage) is crucial to the effects of
insurance on equity (World Bank, 1993:121). Less than 20% of South Africans
have private health insurance, the rest depending largely on public services, or
out-of-pocket payments for private services (Mclntyre and Chetty, 1992). Where
employees are insured through their work, their dependents tend not to be
insured, especially among lower earners. Coverage is low compared to countries
of similar income (Table I). Could NHI significantly increase this proportion?
  Coverage has increased steadily in most Latin American countries, starting in
the 1930's and accelerating in the last two decades (Table I). Virtually the whole
populations of Brazil (Roemer, 1991), Costa Rica (Roemer, 1991) and (South)
Korea (Yang, 1991) are, at least nominally, entitled to insurance-funded care. In
Brazil universal entitlement was simply decreed by government during the
late-1980s, although a large proportion of the population did not contribute
(Roemer, 1991). In most countries, compulsory insurance began with urban
workers employed by large enterprises in a few economic sectors, and was
gradually extended to other sectors and smaller enterprises.
28                                                    TRANSFORMATION 24 (1994)


 BACHMANN                                                                         ARTICLE
                                         TABLE 1
 ECONOMIC, DEMOGRAPHIC AND HEALTH INDICES (1991), AND HEALTH
    INSURANCE COVERAGE (BETWEEN 1985 AND 1990), IN SELECTED
                                        COUNTRIES
COUNTRY         GNPPER            POP.      %POP.             LIFE       INFANT    HEALTH
                 CAPITA      (MILLION)      URBAN         EXPECT. MORTALITY            INS.
                   (U8$)                                    YEARS          RATE     COVER
                                                                     (PER 1000)   (%POP.)
LOW INCOME
CHINA                 370          1150          60             69           38          40
ALL LOW
INCOME                350         3127           39             62           71
LOWER-MIDDLE INCOME
BOLIVIA               650             7           52            59           83          30
PHILIPPINES           730            63          43             65           41          37
PERU                 1070            22           70            64           53          19
THAILAND             1570            57           23            69           27          39
TURKEY              1780             57           63            67           58          35
COSTARICA           1850              3           48            76           14         100
CHILE               2160             13           86            72            17         13
ALL LOWER-
MIDDLE INCOME       1590            774           54            68           42
UPPER4IIDDLE INCOME
S.AFRICA            2560             39          60             63           54          20
VENEZUELA           2730             20           85            70           34          50
ARGENTINA           2790             33          87             71           25          75
URUGUAY             2790              3          86             73           21          67
BRAZIL              2940            151           76            66           58         100
MEXICO              3030             83           73            70           36          60
S. KOREA            6330             43           73            70            16        100
ALL UPPER-
MIDDLE INCOME       3510            627           73            69           34
ALL HIGH
INCOME             21050            822           77            77             8
SOURCES: All data from World Bank (1993), except health insurance data from Roemer
(1991) (Argentina, Brazil, China, Costa Rica, Phillipines, Turkey), Kutzin and Bamum (1992)
(China, Korea), Grosch (1990) (Bolivia, Costa Rica), Mesa Lago (1989) (Peru, Mexico),
Mclntyre and Chatty (1992) (South Africa).
 TRANSFORMATION 24 (1994)                                                                 29


 ARTICLE                                                               BACHMANN
  Extension of insurance to rural populations has usually been restricted by their
inability to afford premiums, or by shortages of rural services. Rural Koreans
were initially reluctant to pay insurance contributions, as the lack of rural
services made benefits unlikely (Yeon, 1985). In Brazil and Mexico, special
funds subsidised by government and social security were established to extend
coverage to rural areas (Roemer, 1991; Knight and Mahar, 1979; Leal de Aranjo,
 1973). Brazil also financed insurance for farm workers with a tax on certain
agricultural products (Mills, 1983). Extension of rural health care has been
especially successful in Costa Rica, but this was largely financed through the
government, assisted by foreign aid (Roemer, 1991).
   Brazil, Korea and Costa Rica were only able to take the last step to universal
insurance coverage once the majority of their populations were already insured.
As the World Development Report cautions, 'only a few middle income countries
that have adequate financial resources, political resolve, and administrative
capacity will be able to achieve such universal insurance coverage' (World Bank,
 1993:161).
   For South Africa the feasibility of increasing coverage depends crucially on
the numbers of people in the urban formal economy. With estimates of un-
employment among the economically active population exceeding 30% (Cooper
et al, 1992), and the precariousness of employment for much of the rest, it is
doubtful whether sufficient additional payroll taxes could be raised to cover the
majority of the population, without severely decreasing benefits currently en-
joyed by the insured few through redistribution. It is also questionable whether
a democratic government will have the political power and will to use NHI for
bridging inequalities between employed and unemployed. It would probably be
most feasible for the government to follow the example of most other countries,
concentrating initially on expanding and regulating insurance for workers in
formal sector employment, with extension to dependents and workers in less
formal employment as economic growth allows.
Equity
   In middle-income countries with substantial NHI coverage, numerous factors
contribute to the persistence or amelioration of health care inequalities (Table
II). Doctors are usually the most unequally distributed resource. In Brazil the
numbers of doctors per 10 000 population were 17.S in the southeast, 6.9 in the
northeast, and 6.6 in the north (Roemer, 1991). In Korea the corresponding
figures were 7 in urban areas and \2 in rural areas (Yang, 1991).
  Insured people usually have more and better health care than do the uninsured,
and coverage is invariably higher in urban than in rural regions. Poor people
living in peri-urban shack settlements are also likely to have very poor access to
30                                                    TRANSFORMATION 24 (1994)


BACHMANN                                                                     ARTICLE
                                    TABLE II.
RANGES OF RESOURCE SUPPLY AND HEALTH INSURANCE COVERAGE
     BETWEEN LEAST SERVED AND BEST SERVED REGIONS IN EACH
                                    COUNTRY.
COUNTRY       DOCTORS:10000POP.           BEDS: 1000 POP.             COVERAGE(%)
                   LOW-HIGH                  LOW-HIGH                   LOW-HIGH
Argentina           8.1-47                    4.3-8.4                     6-100
Bolivia                                       0.2-0.9                     11-33
Chile                2.1-5                    3.2-4.8                     39-95
Costa Rica           1.9-12                    1.1-5.7                   54-100
Ecuador             4.0-14                     1.6-2.8                    3- 20
Mexico               2.4-21                   0.4-3.3                    17-100
Panama               4.0-11                   2.7-55                      11-75
Peru                 0.3-19                   0.6-3.0                      3-27
Uruguay              6.5-35                    1.8-3.8                    17-68
South Africa        0.3"-10"                  4.0*-7.0"
SOURCE: McGreevey 1990, [27] except South African data from Mclntyre and Chetty [8].
LEGEND: * homeland, ** metropolitan, # non-metropolitan
 health care and other services (Harpham et al, 1988). In most countries, the
 wealthy are most likely to be insured. In Brazil (Isuani, 1985) and Dominican
 Republic (McGreevey, 1990), for example, coverage increased with increasing
 income level and job status. Over time, expansion of coverage has proceeded in
 order of decreasing wealth. Unequal coverage may exist even within households,
 when workers' dependents are not covered, as in Peru (Suarez-Berenguela,
  1988), or pay higher fees, as in China (Kutzin and Barnum, 1992).
   Inequalities also exist among the insured. Entitlement to care does not guaran-
 tee that care is provided. Local services may not exist. User fees, time and
 transport costs, and lack of knowledge of the benefits of care may deter utilisa-
 tion, especially for the rural poor. In Mexico City social security clinics were
 located far from insured residents of poor suburbs, who preferred to pay to use
 private practitioners closer to home, demonstrating the inequalities that exist
 even among insured people living in cities (Ward, 1987). In Brazil, social security
 spent twice as much per capita in the richer southeast of the country as in the
 poorer northeast (McGreevey, 1990). Utilisation rates among insured Brazilians
 were twice as high in the southeast as in the northeast, four times as high in urban
 as in rural areas, and ten times as high in the urban southeast as in the rural
 TRANSFORMATION 24 (1994)                                                          31


 ARTICLE                                                                BACHMANN
 northeast, reflecting unequal access to services (Roemer, 1991). Insured
Koreans' utilisation rates increased with increasing income, especially for hospi-
 tals (Yeon, 1985). These examples show that insurance coverage is not in itself
sufficient to ensure access to care.
   There is a danger that NHI may drain scarce resources, such as finance and
professionals, from government health services. In most Latin American
countries, social security consumed shares of public health finance that were
disproportionate to numbers covered. As coverage increases, government expen-
diture on the uninsured, and on preventive services, may decline. Roemer claims
that government health spending has generally been maintained during expan-
sion of social security (Roemer, 1987). In Costa Rica, expansion of social
security was accompanied by a marked increase in the government supply of
primary care (World Bank, 1993:70). In Uruguay, by contrast, government
expenditure, and the quality of its care, declined as insurance increased (Tollman
et al, 1990). Similarly, increased social security spending in Brazil was also
accompanied by a decrease in government spending (McGreevey, 1990).
   Of particular concern is the tendency of doctors to move from public service
into private practice. NHI has typically increased demand and financing for
private care. Thus Abel-Smith argues that an ample supply of doctors is a
prerequisite for NHI (Abel-Smith, 1991). In Latin America, a profusion of
private medical schools has helped increase supply to meet the demand (Roemer,
 1991). However, despite relatively high doctonpopulation ratios there, doctors
are still largely unavailable in rural areas. Compulsory community service for
medical graduates has succeeded in supplying doctors to areas of greatest need
in many countries, and is an important option for South Africa (Roemer, 1991).
However, such coercion frequently results in poor quality public work, and
unofficial private work (moonlighting) (Roemer, 1991). Improved staffing,
management and amenities of rural facilities, as well as higher pay, are also
necessary to attract professionals to where they are most needed.
   NHI has the potential to redistribute resources and wealth from the richer to
the poorer, but it may do the opposite. While premiums are ostensibly shared
between employers and workers, it may be impossible to determine who ul-
timately pays, as employers can shift their costs onto workers (through lower
wage increases), or onto consumers (through higher prices), depending on labour
and commodity markets (Musgrove, 1985). A key objective of NHI is the
subsidisation of poorer and sicker people by the wealthier and healthier, as
premiums do not depend on individual health risk, but are usually based on
income. Whether NHI does redistribute wealth depends on the distributions of
payments and benefits in relation to income. Like most income taxes, NHI
contributions are also progressive in Costa Rica, as in France and Germany, but
32                                                     TRANSFORMATION 24 (1994)


BACHMANN                                                                    ARTICLE
are regressive in Korea, where a low ceiling on maximum contributions means
that higher earners contribute a lower proportion of their income (Yeon, 1985).
Picard has proposed that NHI in South Africa be financed from a progressive
earmarked income tax (Picard, 1993), but it is doubtful whether people earning
more than R70 000 per annum would be prepared to pay more than 10% of their
income towards NHI, as he suggests would be necessary, at the same time as
their benefits were being reduced.
   NHI may also be regressive if government subsidies transfer tax revenue,
drawn partly from consumption taxes paid by the poor, to health care for the
wealthier. Forms of government subsidy of health insurance include i) employer
contributions for government workers, ii) tax-deductible premiums, iii)
provision of public sector care to insured patients at below-cost prices, and iv)
professional training. All four mechanisms operate in South Africa. In China,
government subsidies of health insurance were markedly greater for wealthier
urban workers than for the rural poor (Kutzin and Barnum, 1992). Government
subsidies may, however, be targeted at the poor, as in Dominican Republic, where
subsidies increased with decreasing income of beneficiaries (McGreevey 1990).
  The direction of redistribution depends not only on who pays, but also on the
ratio of benefits to contributions. Despite the inequalities of Brazilian health care,
social health insurance transfers wealth from richer to poorer regions. While the
poorer northeast contributed 9% of revenue, it received 17% of expenditure. The
wealthier southeast contributed 63% of revenue, but only received 53% of
expenditure (Roemer, 1991). The city of Sao Paulo contributed 42% of revenue,
but drew 24% of spending (McGreevey, 1990).
  Korean health insurance, by contrast, appears to aggravate inequality, for two
reasons (Yeon, 1985). Firstly, high co-payments and obligatory 'gifts' to doctors
deter the insured poor from using health services, especially hospitals and
specialists. Secondly, higher earners have more dependents, many of whom are
elderly. Yeon showed that benefit to contribution ratios were 25 - 30% higher for
wealthier than for poorer beneficiaries, indicating that the wealthy gained more
from health insurance than did the poor.
   These examples demonstrate the complex influences of health insurance on
equity. If health insurance in South Africa, whether private for-profit, or com-
pulsory NHI, remains confined to a minority of the population and continues to
drain financial and professional resources away from the public service, it will
aggravate inequality. If, however, NHI markedly expands coverage, and if
services to the rural and peri-urban poor are extended, with subsidisation through
tax or insurance of the poorer by the wealthier, equality will be enhanced. The
former outcome is more likely without significant political support for NHI.
TRANSFORMATION 24 (1994)                                                            33


ARTICLE                                                               BACHMANN
Efficiency
    Efficiency of health care refers to health effects achieved in relation to
resources used. It has become increasingly apparent that greater expenditures on
highly technical, specialised and hospital-based care have not led to commen-
surate improvements in health. By contrast, marked health improvements have
resulted from measures ensuring universal access to essential curative and
preventive services through the primary health care approach (World Bank,
1994). While available data does not allow comparison between countries of
health benefits per unit of expenditure, the efficiency of various health systems
can be indirectly inferred from the prevailing patterns of care.
   Health care under NHI in Latin America and Asia has typically been exces-
sively costly, hospital-oriented, specialised, highly technical, and urban based.
The greatest inefficiencies do not necessarily arise directly from NHI itself, but
rather follow from the 'perverse incentive' of fee-for-service reimbursement of
private care providers, which rewards excessive hospitalisation, investigation
and treatment Levels of care are typically poorly defined, with the hospital as
the usual first point of contact in countries such as Brazil (McGreevey, 1982)
and Korea (Kutzin and Barnum, 1992; Yeon, 1985). In Brazil, the proportion of
the social security budget directed to hospitals increased from 40% in the 1960s
to 50% in the 1980's (Kutzin and Barnum, 1992). Brazilians receiving private
outpatient care are eight times as likely to be admitted to hospital as are those
attending public facilities (Roemer, 1991). In Brazil, the proportion of hospital
inpatients receiving X rays and laboratory tests increased by two to three times
from 1970 to 1981 (McGreevey, 1982). Korea has a surplus of highly expensive
equipment of doubtful value: for a population of 43 million, it has 28 hospitals
equipped for heart transplants, 26 lithotrypsy machines, and computerised
tomography in every hospital with over 200 beds (Yang, 1991). Over-medication
is also encouraged, severely so in Korea (Roemer, 1991) and China (Kutzin and
Barnum, 1992), where drugs account for 57% and 50% of their respective health
insurance budgets.
  In the countries examined curative care has been emphasised, frequently at the
expense of preventive and promotive care. While social security health spending
increased in Brazil between 1978 and 1982, government spending on com-
municable disease control decreased by 41%, including a drop of 57% for
malaria, and 80% for bilharzia (McGreevey, 1982).
  Primary care has usually been neglected in NH-funded systems. In most cases,
primary care through NHI comprises payments to urban private practitioners.
Fee-for-service remuneration has stimulated over-servicing, as in Korea, where
the number of doctor consultations per illness episode rose from three in 1980
to five in 1990 (Yang, 1991). There are, however, important exceptions to this
34                                                    TRANSFORMATION 24 (1994)


BACHMANN                                                                  ARTICLE
trend, in which primary care is provided through health centres. Urban
polyclinics were developed by social security institutes in a number of Latin
American countries, with sufficient scale to efficiently provide a mixture of
general and specialist care (Roemer, 1987). Brazil has about 7 000 polyclinics,
in addition to the 10 000 (mostly public) health posts and centres (Roemer, 1991).
Costa Rica greatly increased the distribution of health centres in rural areas, but
this was largely undertaken by the government who delegated all hospital
management to social security institutes (World Bank, 1993:70).
   The inefficiencies described have led to steeply escalating costs in most of the
countries reviewed, to which they have responded by increasing premiums and
co-payments, and cutting benefits (Ron et al, 1990; Roemer, 1991; Grosch,
 1990). Some have attempted to change methods of paying providers but have at
times met resistance. In Korea, for example, doctors successfully resisted limits
to excessive remuneration for prescribing drugs, which is a major source of
income (Roemer, 1991). Cost control is, however, possible in industrialised
countries with near-universal NHI coverage, as has been shown by European
countries during the 1980's, but has only been achieved through sustained
government and insurance institute intervention in health care (World Bank,
 1993:125). Central control of health care funding through NHI could enable
redirection of resources to the most efficient forms of care (Broomberg and De
Beer, 1990), but engagement of insurance institutes and government in health
care organisation is crucial (Kutzin and Bamum, 1992; World Bank, 1993).
Organisation
  Forms of public sector control of NHI vary widely. In Korea the health ministry
has a minimal role, such as restricting the establishment of new hospitals (Yang,
 1991). In Brazil, by contrast, the entire insurance system is based within the
health ministry (Roemer, 1991). The degree of real control may, however, be
 unclear. In Brazil, the social security institute retained much of the staff and
structure it had before incorporation into the ministry, and wielded more
economic and institutional power than did the rest of the ministry. In none of the
countries reviewed has NHI been an integral part of a unified government-
provided national health service.
  Amajor choice is whether aNHI institute should provide care itself- the 'direct'
 method - or whether it should purchase services from private or public providers
 - the 'indirect' method (Roemer, 1987; Roemer, 1991). The early decades of
Latin American NHI were characterised by direct provision, in which salaried
professionals served beneficiaries in hospitals and clinics owned by social
 security institutes (Roemer, 1987). Through the 1970s and 1980s the indirect
 method has become predominant in Latin American and Asian NHI, although
                                                                                 3S
 TRANSFORMATION 24 (1994)


ARTICLE                                                                   BACHMANN
the direct method is still dominant in India and Turkey (Ron et al, Roemer 1991).
Both the World Bank (World Bank 1994,131) and, in South Africa, the Centre
for Health Policy (Picard, 1993) argue that the state would improve efficiency
by increasing competition between independent providers for public contracts.
The Bank warns, however, that 'managed competition requires a high degree of
government administrative capacity to set the rules and monitor provider
performance'.
   Health maintenance organisations (HMOs), in which pre-payment entitles
members to a range of services, are a form of insurance. However, as the
providers of care bear the costs of care themselves, they have an incentive not
to overtreat HMOs providing urban primary care have expanded in Brazil and
cover half of Uruguay's population (Tollman et al, 1990; Briscoe, 1990). Limita-
tions of HMOs identified elsewhere include exclusion of people at higher risk
of illness, incentives to enroll too many people and under-treat them, and their
failure in rural areas (Tollman et al, 1990). Managed care can take many forms,
however, and some of these problems could be avoided through careful design,
including a judicious combination of incentives and regulation.
    HMOs have expanded in South Africa, and have been shown to be more
efficient than insurance-funded fee-for-service care (Broomberg and Price,
 1990b). The Chamber of Mines is currently taking steps to get greater control
over the provision of employees' health care, rather than leaving it to private
insurers and professionals (Fourie, 1993). There is also scope for greater worker
control of HMOs, as shown by the participatory management of the ISCOR
medical benefit scheme (Kruger, 1993). The latter scheme however also il-
lustrates the limitations of HMOs: only a quarter of low earners could afford to
join the most basic scheme, and most workers opted exclusion of their depend-
ents.
  These examples show that NHI in South Africa could be organised in a number
of ways. In any case, the state will need to take an active role in organising
insurance institutes, and in directing finances to areas of greatest need. It is,
however, likely that the first task facing a future health ministry, of integrating
the highly-fragmented public health authorities into a unitary service organised
at national, regional and district levels, will be so great that it will not have the
capacity to simultaneously manage the reorganisation of private care through
NHI.
Political influences
   Equity and efficiency are most likely where insurers, providers, the state and
the public have similar objectives, and where programmes are coordinated.
Conflicts of interest are inevitable, however, and are played out through overt or
36                                                     TRANSFORMATION 24 (1994)


BACHMANN                                                                      ARTICLE
latent political processes.
   Mesa Lago (1989) and Borzutsky (1989) have described how social security
in Latin America developed under multi-class alliances of industrialists, the
middle class (especially government employees) and the organised working
class, and was a focus of intense bargaining. The state used social security to
'co-opt, neutralise and control' pressure groups. Much decision-making occurred
outside formal structures through elite networks, including government officials,
professionals, hospital owners, and social security administrators. The power of
social security institutes and labour ministries was often greater than that of
health ministries. The privatisation drives of military regimes changed health
care from direct provision of services by social security institutes and public
services, to fee-for-service payment to private providers. While the poor have
generally been excluded from these processes, more recently popular pressures
have prompted the expansion of coverage in countries like Brazil and Korea.
These examples show how powerful interest groups can influence the direction
of health care under NHL
   In South Africa, the greatest resistance to a redistributive and regulated NHI
system is likely to come from the current beneficiaries of private insurance, and
from the medical profession which would resent any restrictions of professional
autonomy and income. Also, if NHI covers only part of the population, there may
be substantial popular opposition. Much will depend on the ways in which trade
unions and employer organisations see their constituencies' interests being
furthered, and how NHI proposals fit in with other social security developments
aimed at income maintenance.
Conclusion
  The evidence of countries economically comparable to South Africa shows that
NHI can increase funding for health care, but also poses risks. If NHI is used to
expand the present system of fee-for-service payment for private care, the
already critical inefficiencies and inequities will be aggravated. An additional
danger is that government health resources could be drawn into expensive
ineffective care for the few. Efficiency can however be increased under NHI by
i) directing resources to those services which will have the greatest impact on
health, by developing a health system based on primary care, ii) continual
evaluation of processes and outcomes of care, and critical assessment and
regulation of new technology, and iii) creating financial incentives for providers
of care, especially doctors, to practise efficiently. Equity can be increased by i)
ensuring that the majority of the population is entitled to, and has access to,
adequate care, regardless of their ability to pay for it, and ii) subsidisation by the
wealthy of the poor.
                                                                                    37
TRANSFORMATION 24 (1994)


ARTICLE                               -                                                 BACHMANN
   To promote health for all, NHI needs to be linked to an integrated national
system of health care. Particularly needed are programs to ensure access for the
poor in rural and peri-urban areas, including die unemployed and the elderly, as
have been attempted in Brazil, Mexico and Costa Rica NHI has the potential to
raise and direct large sums of money equitably and efficiently, but it is not a
complete solution. Political and organisational complications may distract the
state from focusing on the leading priorities - extending basic services to the
poor, and putting the fragmented public health system together again.
ACKNOWLEDGEMENTS
    The author is grateful to Jonathan Broomberg, Diane Mclntyre and Menick Zwarenstein for
    their critical comments on earlier drafts of this paper.
REFERENCES
Abel Smith. B (1991) - 'Financing health for all'. World Health Forum. 12.
African National Congress (1992) - 'ANC policy guidelines on health'. South African Medical
         Journal, 82.
African National Congress (1994) - A national health plan for South Africa (Johannesburg: ANC).
Borzutsky, S (1985) - 'Politics and social security reform', in Mesa Lago (1985).
Briscoe, J (1990) - Brazil: the new challenge of adult health (Washington: World Bank).
Bloomberg, J and C De Beer (1990) - 'Financing health care for all - is national health insurance
         the first step?'. South African Medical Journal. 78.
Bloomberg, J and MR Price (1990a) - 'The impact of the fee-for-service reimbursement system
         on me utilisation of health services: Part 1. A review of the determinants of doctors'
         practice patterns'. South African Medical Journal, 78.
Broomberg, I and MR Price (1990b) - 'The impact of the fee-for-service reimbursement system
         on the utilisation of health services: Put 0 . Comparison of utilisation patterns in medical
         aid schemes and a local health maintenance organisation'. South African Medical
         Journal, 78.
Cooper, C, J Schindler, C McCauL et al (1992) (eds) - Race relations survey 199112,
         (Johannesburg: South African Institute of Race Relations).
Fourie, I (1993) - 'Analysing the specific needs of your work force' (paper presented at
         Affordable Primary HealmCue conference, Midrand).              ^ ^
Grosch, ME (1990) - 'Social spending in Latin America: the story of the 1980V, World Bank
         Discussion Papers, 106 (Washington: World Bank).
Harpham, T, T Lusty and P Vaoghan (1987) - In the shadow of the city: community health and the
         urban poor (Oxford: Oxford University Press).
Isuani, EA (1985) - 'Social security and public assistance', in Mesa Lago (1985).
Knight, PL D Mahar and R M o r a (1979) - Brazil: human resources special report, Volume II
Kroger, HJ (1993) - 'Effective decision making and communication structures' (paper presented at
         Affordable Primary Health Careconference, Midrand).                        ^ ^
Kutzin, J and H Bamum (1992) - 'How health insurance affects delivery of health care in
         developing countries', Population, Health and Nutrition Working Paper (Washington:
         World Bank).
Leal de Aranjo, L (1973) - 'Extension of social security to rural workers in Mexico', International
         Labour Review. 108.
McGreevey, W (1982) - 'Brazilian health care financing and policy: an international perspective'.
         Population, Health and Nutrition Notes (Waiffington: World Bank).
McGreevey. W (1990) - 'Social Security in Latin America', World Bank Discussion Paper, 110
38                                                                 TRANSFORMATION 24 (1994)


BACHMANN                                                                                    ARTICLE
Mclntyre, D (1993) - 'Health sector financing and expenditure in South Africa', Health
         Economics Unit Working Paper, 4 (Cape Town: Department of Community Health).
Mclntyre, D and K Chetty (1992) - Health Economics Unitfactsheet: health sector resources in
         South Africa (Cape Town: Department of Community Health).
Mclntyre, D and RE Dorrington (1990) - 'Trends in the distribution of South African health care
         expenditure'. South African Medical Journal. 78.
Mesa Lago, C (1985) (ed) - The crisis of social security and health care: Latin American
         experiences and lessons (Pittsburgh: University of Pittsburgh Press).
Mesa Lago. C (1989) - Ascent to bankruptcy: financing social security in Latin America
         (Pittsburgh: University of Pittsburgh'Press).
Mills, A (1983) - 'Economic aspects of health insurance', in K Lee K and A Mills (eds) - The
         economics of health in developing countries (Oxford: Oxford University Press).
Musgrove, P (1985) - "The impact of social security on income distribution', in Mesa Lago (1985).
Naylor, C D (1988) - 'Private medicine and the privatization of health care in South Africa',
         Social Science and Medicine, 27(11).
Picard, J (1993) - A National Health Service for South Africa. Part II: a proposal for change
         (Johannesburg: Department of Community Health).
Price, MR and J Bloomberg (1990) - 'The impact of the fee-for-service reimbursement system on
         the utilisation of health services: Part III. A comparison of Caesarian section rates in
         white nulliparous women in the private and public health sectors'. South African Medical
         Journal.liT
RispeL L and G Behr, (1992) - Health indicators - policy implications (Johannesburg: Centre for
         Health Policy, University of Witwatersrand).
Roemer, MI (1987) - 'Health service financing by social security', International Journal of
         Health Punning and Management, 2.
Roemer, MI (1991) - National health systems of the world. Volume I (Oxford: Oxford University
         Press).
Slabber, CF (1992) - 'A new South Africa - a new health care strategy'. South African Medical
         Journal, 82.
Suarez-Berenguela, RM (1988) - 'Financing the health sector in Peru', Living standards
         measurement study, Working paper 31 (Washington: World Bank).
Tollman, S, D Schopper and A Torres (1990) - 'Health maintenance organizations in developing
         countries: what can we expect?'. Health Policy and Planning, 5(2).
Ward, P (1987) - 'Reproduction of social inequality: access to health services in Mexico City',
         Health Policy and Planning, 2.
World Bank (1993) - World Development Report 1993 (Washington: Oxford University Press).
Yang, BM (1991) - 'Health insurance in Korea: opportunities and challenges'. Health Policy and
         Planning, 6(2).
Yeon, H C (1985) - Who gains and who loses? An overview of equity and efficiency in Korea's
         health insurance system (Boston: Takemi Program in International Health, Harvard
          School of Public Health).
TRANSFORMATION 24 (1994)