Journal of Public Health Policy

, Volume 32, Supplement 1, pp S102–S123 | Cite as

Inequities in access to health care in South Africa

  • Bronwyn Harris
  • Jane Goudge
  • John E Ataguba
  • Diane McIntyre
  • Nonhlanhla Nxumalo
  • Siyabonga Jikwana
  • Matthew Chersich
Original Article

Abstract

Achieving equitable universal health coverage requires the provision of accessible, necessary services for the entire population without imposing an unaffordable burden on individuals or households. In South Africa, little is known about access barriers to health care for the general population. We explore affordability, availability, and acceptability of services through a nationally representative household survey (n=4668), covering utilization, health status, reasons for delaying care, perceptions and experiences of services, and health-care expenditure. Socio-economic status, race, insurance status, and urban-rural location were associated with access to care, with black Africans, poor, uninsured and rural respondents, experiencing greatest barriers. Understanding access barriers from the user perspective is important for expanding health-care coverage, both in South Africa and in other low- and middle-income countries.

Keywords

out-of-pocket payments access health-care utilization inequities household survey South Africa 

Introduction

More than a billion people, mainly in low- and middle-income countries (LMICs), are unable to access needed health services as these are unaffordable.1 In South Africa, health-care access for all is constitutionally enshrined; yet, considerable inequities remain, largely due to distortions in resource allocation.2,3,4 Access barriers also include vast distances and high travel costs, especially in rural areas; high out-of-pocket (OOP) payments for care;5 long queues;6 and disempowered patients.7 These barriers, created by uneven social-power relationships, resonate with access hurdles experienced elsewhere in LMICs.1,8 Globally policy attention has turned to universal health coverage (UHC) as a remedy for inaccessible, unaffordable health services.

Achieving equitable UHC requires the provision of accessible, necessary services (‘depth’) for the entire population (‘breadth’), and accommodating the ‘differential needs’ and financial constraints of disadvantaged groups (‘height’).8 Access is therefore the opportunity and freedom to use services,9 and encompasses the circumstances that allow for appropriate service utilization, plus a sufficiently informed individual or household (demand-side) empowered to exercise choice within the health system (supply-side).9,10 The ‘degree of fit’ between demand- and supply-sides, rather than each in isolation, determines the degree of access achieved.9

South Africa's apartheid past still shapes health, service, and resource inequities.2 Racial, socio-economic, and rural-urban differentials in health outcomes, and between the public and private health sectors remain challenging.2,3,11 In 2005, spending per private medical scheme member was ninefold higher than public sector expenditure, and one specialist doctor served fewer than 500 people in the private sector but around 11 000 in the public sector.11 Large information gaps remain about health access in the general population in South Africa, especially around utilization rates and OOP payments for health care.12 Documenting demand-side perspectives of users, too-long neglected, could inform future policies.8,9 We conducted a national household survey to fill these gaps and to examine access barriers.

Methods

In 2008, we conducted a household survey in South Africa, with households selected using multi-stage sampling, detailed elsewhere and in this edition.13,14 The team selected five randomly-chosen households within 960 enumerator areas. Within each household, we administered the questionnaire to an adult responsible for household health decisions. If the health-head declined or was ineligible, the household was substituted by the one to the immediate left. We verified 20 per cent of questionnaires telephonically and double entered the data. The Universities of Cape Town and the Witwatersrand provided ethics approval for the study and all respondents provided informed consent.

Measurement of access and need

We collected information on: health status; utilization of outpatient (annualized number of visits/person in the last month), and inpatient (number of admissions per 1000 people/year) services; health-insurance status; and reasons for delaying care when someone was ill and then the illness worsened in the previous year. We examined the access dimensions9 of availability (distances and travel mode to facilities), acceptability (reasons for provider choice, user satisfaction and health system perceptions, including reasons for delayed care), and affordability. For affordability, we calculated household ‘burdens’ of OOP payments (0–4 per cent of total household expenditure is low-moderate, 5–9 per cent high, and above 10 per cent catastrophic).5 We calculated the transportation burden by dividing transportation costs for outpatient visits by total household monthly expenditure.

The team assessed equity by considering whether access was equal among those with equal need for health care, by contrasting levels of need with service use in different population groups, such as socio-economic quintiles. We used two indicators of need: self-assessed health status as poor or very poor in main respondents;12 and, among the total population, any household member experiencing recent illness or injury.

Data analysis

We analyzed the data using STATA® 11 and weighted it for differential probability of participant inclusion. To detect differences among categorical variables we used the Rao–Scott F statistic to determine P-values.15 Given the large sample size, virtually all associations were significant. Using principal component analysis, we developed a composite index of socio-economic status based on variables including access to water and sanitation, housing characteristics, and age and gender of household heads, which are strongly related to socio-economic status in South Africa.16 We then categorized respondents into five socio-economic quintiles, from poorest (quintile 1) to wealthiest (quintile 5).

Results

The health decision-maker declined participation or was ineligible in 238 households (5 per cent), 223 of which were substituted. The 4668 households sampled contained 21 159 individuals. Four-fifths were black Africans, almost half had only primary schooling or less, and 39.9 per cent inhabited rural areas (Table 1). A quarter were employed, a similar proportion unemployed, and the remainder either pensioners (8.3 per cent), or children and students (44.3 per cent). Median per capita expenditure (for everything, including, but not restricted to, health services) was US$26.7/month (IQR=$13.3–$53.3; $1=R7.5), with 28.9 per cent spending below $15/month. Conversely, 5.4 per cent spent $250/month or more. Most did not have health insurance (88.4 per cent).
Table 1

Population characteristics and health service utilization in the general population in South Africa

Variable (% of study population)

Outpatient visits (per person/year)

Total outpatient

Inpatient admissions (per 1000 people/year)

 

Clinic/CHC

Public hospital

Private

 

Public

Private

Age

 <18 (37.9)

1.8

0.4

0.7

2.9

30.5

5.8

 18–24 (15.1)

1.3

0.6

0.7

2.7

84.4

8.6

 25–49 (31.1)

2.1

1.1

1.5

4.7

113.1

20.1

 50–64 (10.4)

4.1

1.7

2.2

8.0

109.6

51.5

 65+ (5.4)

4.6

2.1

2.0

8.7

176.2

27.5

Sex

 Female (55.7)

2.7

1.1

1.2

5.1

101.6

13.8

 Male (44.3)

1.5

0.6

1.1

3.3

53.8

20.0

Race

 Black African (82.2)

2.4

1.0

0.9

4.2

76.3

9.2

 Colored (9.5)

2.1

0.7

1.2

4.0

97.4

28.9

 Indian or Asian (2.2)

1.2

1.2

3.8

6.2

59.3

50.7

 White (6.1)

0.5

0.4

3.5

4.4

116.6

84.7

Area type

 Rural (39.9)

2.8

1.0

0.7

4.6

58.3

4.0

 Informal-urban (17.2)

2.2

0.9

0.9

4.0

100.1

3.5

 Formal-urban (42.9)

1.6

0.8

1.7

4.1

93.9

33.6

Education

 Non-primary (47.1)

2.7

0.9

0.7

4.4

61.2

5.9

 Some secondary (31.7)

1.9

1.1

1.0

4.0

101.2

14.8

 Complete secondary (16.7)

1.5

0.8

1.7

4.o

84.7

35.9

 Tertiary (4.6)

0.8

0.4

4.2

5.3

153.1

71.0

Employment

 Employed (23.9)

1.3

0.8

2.0

4.1

105.1

41.3

 Unemployed (23.5)

3.0

1.4

1.1

5.5

112.5

11.0

 Pensioner (8.3)

5.2

2.4

1.8

9.4

163.9

20.5

 Student/child (44.3)

1.7

0.5

0.6

7.9

34.7

5.6

Health insurance

 None (88.4)

2.4

1.0

0.7

4.1

88.2

3.3

 Insured (11.6)

0.4

0.4

4.7

5.5

25.5

118.6

Health statusa

 Excellent (22.7)

1.5

0.7

1.5

3.7

136.3

32.5

 Good (32.9)

2.5

1.3

2.0

5.8

121.2

21.7

 Average (26.5)

5.5

2.3

2.7

10.4

190.2

44.1

 Poor (14.7)

7.8

3.6

2.6

14.0

291.4

23.7

 Very poor (3.1)

5.6

6.7

2.7

14.9

406.8

36.8

aRestricted to main respondent (n=4668).

Need and utilization

We assumed that the 17.8 per cent of main respondents who reported poor health and the similar proportion of the total population who experienced illness or injury in the preceding month were in need of care. Need was unevenly distributed, although patterns varied between these measures (Table 2). For main respondents, over 20 per cent of those in the poorest three quintiles needed care compared to just 5.6 per cent of the richest; yet, the socio-economic status of those ill or injured was fairly evenly distributed. Similarly, need was higher among main respondents with only primary education or less (31.7 per cent) than those with tertiary qualifications (3.5 per cent), but education was not associated with recent illness or injury. Almost 20 per cent of women in both groups needed care: 1.3–1.6-fold more than males. Relative to other groups, more black Africans (20.2 per cent) reported poor health. In contrast, a third of Indians or Asians, decreasing to 16.6 per cent of black Africans, were recently ill or injured among the total population. While almost 20 per cent of the uninsured in both groups needed care, 6.2 per cent of insured main respondents reported poor health, and 23.6 per cent of the total insured were recently ill or injured. Within both groups a third of those above 65 years needed care.
Table 2

Differentials in need, utilization, and health system access between socio-economic quintiles

Variable

Category %a

Quintile 1

Quintile 2

Quintile 3

Quintile 4

Quintile 5

Total

Need

 

Poor or very poor healthb

21.3

25.0

20.8

16.2

5.6

17.8

 

Ill or injured

19.7

16.2

15.9

16.9

19.6

17.6

Utilization

 Outpatient total population

Public clinic/CHC

68.8

64.3

61.4

51.4

21.6

54.7

 

District hospital

17.9

15.8

9.8

8.1

3.9

11.6

 

Regional hospital

8.1

9.4

9.7

9.7

5.3

8.5

 

Tertiary hospital

1.7

2.5

7.2

7.6

5.6

4.8

 

Private non-hospital facility

15.7

19.8

22.8

31.3

60.8

28.8

 

Private hospital

3.1

1.7

1.6

4.5

12.2

4.4

 

Public sector visits (mean/year)

3.7

3.2

3.3

3.0

2.3

3.1

 

Public sector visits if ill/injured (mean/year)c

15.9

12.6

11.2

10.2

6.2

10.9

 

Private sector visits (mean/year)

0.4

0.5

0.7

1.3

2.9

1.2

 

Private sector visits if ill/injured (mean/year)c

2.1

2.1

2.7

5.1

8.7

4.4

 Inpatient total population

District hospital

51.1

48.5

35.1

17.7

5.7

31.0

 

Regional hospital

31.5

34.5

33.8

33.2

15.4

29.5

 

Tertiary hospital

12.3

16.4

28.2

30.5

17.1

21.1

 

Private hospital

5.2

0.7

2.9

18.6

61.9

18.4

 

Public sector visits (mean)d

74.2

74.3

105.7

70.8

76.2

80.5

 

Public sector visits if ill/injured (mean)d

224.0

206.4

264.5

248.7

200.2

227.7

 

Private sector visits (mean)d

5.1

0.5

3.4

13.5

74.2

16.6

 

Private sector visits if ill/injured (mean)d

13.7

3.4

9.9

32.3

192.8

44.7

Availability

 Transport to facility

Walked

41.1

38.6

45.5

41.0

13.9

37.0

 

Public transport

54.1

55.6

45.7

42.1

22.2

45.2

 

Private vehicle

2.5

3.6

6.2

15.2

63.6

15.9

 

Other

2.4

2.3

2.6

1.7

0.3

2.0

 Travel time

Mean minutes

38.2

34.2

30.6

26.5

20.2

30.7

 Chose facility as it's closest

Outpatient

57.0

59.9

54.8

55.2

41.2

54.0

 

Inpatient

51.4

62.6

42.7

48.0

34.6

47.5

Affordability

 Chose facility as don’t have to pay

Outpatient

37.7

41.6

39.9

35.8

14.7

34.5

 

Inpatient

23.2

38.0

29.4

31.7

9.0

26.2

 Delayed care as

Transport unaffordablee

21.1

17.0

11.4

9.9

1.1

12.2

 

Unable to get time off worke

1.7

3.8

7.4

7.3

15.9

7.1

 OOP household burdenf

Transport 5–9%

26.2

18.1

12.2

8.4

2.3

13.9

 

⩾10%

19.0

11.7

10.3

6.2

2.5

10.3

 

Outpatient public service 5–9%

2.8

1.4

1.8

1.4

0.9

1.8

 

⩾10%

2.3

2.5

0.5

1.0

0.9

1.5

 

Outpatient private service 5–9%

13.3

17.6

9.6

15.2

5.2

10.2

 

⩾10%

59.5

46.7

40.2

21.2

5.2

23.7

 

Inpatient public services 5–9%

10.2

4.8

3.4

5.5

4.3

5.6

 

⩾10%

7.1

5.2

4.7

7.2

7.7

6.2

 

Inpatient private services 5–9%

7.2

0.0

6.3

0.0

1.7

1.9

 

⩾10%

21.4

50.0

0.0

5.4

8.6

8.4

Acceptability

 Chose facility as respectful service

Outpatient

6.4

6.2

8.7

13.2

19.4

10.4

 

Inpatient

8.8

6.0

7.0

11.6

15.4

9.7

 Delayed care as

Queues too longe

8.1

6.7

7.7

11.8

9.2

8.5

 

Care likely to be ineffectivee

9.6

6.1

7.0

3.0

4.5

6.1

 

Won’t be treated respectfullye

5.3

0.4

3.2

3.2

3.1

2.9

 

Illness not seriouse

56.1

62.0

71.7

75.6

79.9

68.9

aAmong whole population unless stated otherwise.

bMain respondent only.

cThose ill or injured in the last month (of whole population).

dMean admissions/1000 people/year.

eDid not seek care when ill, then illness worsened (in past year).

f0–4% burden is low to moderate, 5–9% high, above 10% catastrophic.

Utilization

Outpatient care

While total utilization was similar across socio-economic groups (2.3–3.7 visits/year), the poorest mainly visited primary health care (PHC) facilities (in quintile 1, 68.8 per cent attended clinics, and 17.9 per cent district hospitals), while the richest were thrice as likely to use tertiary hospitals (including national central, academic, and specialist hospitals) (Table 2). Patients with only primary education or less, made 3.4 times as many visits to public clinics as those with tertiary qualifications (Table 1).

Outpatient visits in the private sector were concentrated in the richer quintiles. For ambulatory private care (including general practitioners, private dentists, and pharmacies), utilization rose steadily from 15.7 per cent of the poorest to 60.8 per cent of the wealthiest; a group that also used private-hospital outpatients four times more than those in quintile 1 (Table 2). In addition, use of private-outpatient services was high among those with tertiary education (4.2 visits), Indians or Asians (3.8 visits), Whites (3.5 visits), and the insured (4.7 visits).

Inpatient care

For the total population, the mean days admitted per 1.000 people/year was 80.5 in the public sector and 16.6 in the private sector, rising to 227.7 public and 44.7 private admissions for the ill or injured. Most inpatient care took place in public, rather than private facilities for all but the richest, with 61.9 per cent admitted privately compared to just 5.2 per cent in quintile 1, and 0.7 per cent in quintile 2 (Table 2). For public sector admissions, people in the lowest quintile mostly used district facilities (53.8 per cent), with only 13.0 per cent admitted to a tertiary facility. Conversely, public sector admissions among the richest quintile were predominantly at tertiary (44.8 per cent) or regional (40.3 per cent) hospitals. Main respondents in very poor health experienced threefold as many public admission-days as those reporting excellent health (406.6 versus 136.3) – almost double that of those who were ill or injured.

Insurance status was associated with differential utilization of inpatient care, especially in the private sector, with a mean 118.6 admission-days for the insured versus only 3.3 for those without insurance. For rural-dwellers and those living in informal-urban areas, total utilization of inpatient private facilities was just a tenth of urban-formal residents (33.6 admissions). Similarly large differentials occurred in private admissions between those in the poorer, more rural provinces of Limpopo (2.3 admissions), Mpumalanga (6.7 admissions), and Eastern Cape (6.9 admissions) versus the urban, better-resourced provinces of Gauteng (32.5 admissions), and the Western Cape (32.8 admissions). While women's total inpatient utilization was higher than men (115.4 versus 73.8), men were 1.5-fold more likely to be admitted to private hospitals.

Availability

The majority used public transport (45.2 per cent) or walked to outpatient health services (37.0 per cent), although two-thirds (63.6 per cent) of the richest used private means. People in formal-urban areas were sevenfold more likely to use private transport than rural residents, and only 6.6 per cent of whites used public means. Average travel time to a facility was 30.7 min, but almost twice as long for the poorest (38.2 minutes) than the richest (20.2 min) (Table 2). Similarly, travel times in rural areas were long (38.2 min). Travel was shortest for whites (17.5 min), followed by Indians or Asians (22.4 min), coloreds (25.8 min), and black Africans (32.5 min).

Two-thirds using public sector outpatient primary care and 53.5 per cent using public sector hospital outpatients chose the facility because it was close. The wealthiest quintile appeared more willing to travel, with only 30.4 per cent of users selecting private-outpatient services for their proximity. ‘Closest service’ was also important for half (49.8 per cent) of those using public-inpatient facilities, while this influenced just a quarter of private inpatients. Referral was the commonest reason for selecting private hospitals (38.4 per cent), compared to 28.5 per cent in the public sector. Of inpatients, 28.9 per cent in the public sector and 14.3 per cent in the private, were taken there in an emergency.

Affordability

In the public sector, ‘not having to pay’ informed the choice of over half using primary care, 30.4 per cent using hospitals as outpatients, and 29.0 per cent of inpatients. Less than 5 per cent gave this reason for using private services.

Of household members who delayed seeking care, 21.1 per cent of the poorest versus 1.1 per cent of the richest said this was due to unaffordable transport costs (Table 2). Transport costs were similarly a problem for 42.5 per cent living in Eastern Cape and 19.5 per cent in Limpopo. Unaffordable transport also obstructed immediate care for 18.2 per cent of children under 6, and 13.8 per cent of the uninsured, but only 1.0 per cent of insured. Inability to leave work prevented immediate care-seeking for 10.6 per cent of the insured, as well as 15.9 per cent of the richest, declining step-wise across quintiles to 1.7 per cent of the poorest.

For those who sought outpatient care, transport costs were catastrophic (⩾10 per cent of household expenditure) for 19.0 per cent of the poorest, falling to 2.5 per cent of the wealthiest (Table 2). Financially catastrophic transport costs occurred in 15.3 per cent of those living in rural areas, 14.7 per cent of the unemployed, and 12.0 per cent of those uninsured (Table 3). This also affected more black Africans (11.8 per cent) than other groups, and only few whites (1.6 per cent).
Table 3

Factors associated with high and catastrophic out-of-pocket payments for health care as a percentage of household expenditure

Variable

OOP transport to outpatient care

OOP outpatienta

OOP inpatientb

   

Public

Private

Public

Private

Burden on householdc

5–9

⩾10

5–9

⩾10

5–9

⩾10

5–9

⩾10

5–9

⩾10

Age

 <18

13.6

8.7

1.7

0.6

9.4

20.5

10.8

6.1

0.0

3.9

 18–24

13.5

13.7

3.5

2.5

11.7

31.1

4.9

4.6

0.0

10.5

 25–49

14.9

11.2

3.1

1.9

10.3

23.9

7.0

8.7

2.4

7.6

 50–64

13.8

9.1

1.8

1.8

11.0

22.5

5.7

6.3

3.0

7.7

 65+

13.1

12.0

0.2

1.4

8.9

41.5

2.3

3.2

0.0

16.2

Sex

 Female

15.0

10.6

1.9

1.4

10.4

25.9

6.5

5.3

2.5

8.6

 Male

12.1

10.4

2.6

1.6

10.0

23.7

6.8

9.8

0.9

7.2

Race

 Black African

16.5

11.8

2.4

1.7

12.2

36.3

7.3

6.8

1.8

7.0

 Colored

3.7

7.9

0.4

0.2

6.5

8.5

3.7

5.1

1.7

7.7

 Indian or Asian

0.5

2.7

1.2

0.0

10.4

3.3

5.1

9.1

4.5

9.8

 White

0.3

1.6

0.0

0.0

6.2

5.4

0.0

4.7

1.1

8.9

Area type

 Rural

22.4

15.3

2.9

2.1

11.5

54.1

10.1

8.2

9.7

17.9

 Informal-urban

8.6

10.6

1.7

1.3

16.6

30.7

4.1

5.2

0.0

1.2

 Formal-urban

6.7

5.1

1.4

0.7

8.6

14.4

4.8

5.2

1.2

7.5

Education

 Non-primary

18.8

11.9

1.6

1.7

11.7

34.5

7.8

7.2

0.0

15.6

 Some secondary

10.9

10.7

2.6

1.0

13.4

27.2

5.9

6.2

1.6

7.6

 Complete secondary

9.6

10.3

3.6

2.2

8.3

20.4

6.1

6.5

1.8

7.0

 Tertiary

6.2

2.2

1.7

0.0

5.7

8.2

6.3

8.4

3.4

5.3

Employment

 Employed

11.3

7.6

4.0

2.1

9.0

19.4

7.1

12.0

1.3

6.6

 Unemployed

16.0

14.7

2.5

2.4

14.4

33.7

5.9

5.2

6.6

12.8

 Pensioner

13.9

9.6

0.5

0.7

9.7

35.1

2.8

3.7

0.0

14.2

 Student/child

14.0

8.9

1.9

0.8

9.4

22.7

9.5

5.8

0.0

3.5

Health insurance

 None

16.0

12.0

2.1

1.5

17.4

43.0

6.6

6.6

2.7

25.2

 Insured

4.1

2.4

2.5

0.0

2.0

4.0

9.5

7.7

1.6

4.6

Health status

 Excellent

15.2

5.4

3.9

1.4

3.7

22.7

4.8

6.3

2.4

6.1

 Good

12.2

9.3

2.3

1.7

11.7

15.8

4.9

6.4

0.0

7.1

 Average

13.9

14.7

2.9

1.6

14.4

28.1

6.5

6.4

5.7

14.6

 Poor

19.9

13.7

2.0

2.2

15.6

46.9

5.7

8.9

5.5

6.1

 Very poor

16.6

11.6

0.0

1.6

12.9

25.3

6.5

0.0

17.1

0.0

aMost recent visit, excludes transportation.

bMost recent admission, excludes transportation.

c0–4% burden is low to moderate, 5–9% high, ⩾10% catastrophic.

OOP: Out-of-pocket payments.

OOP payments for outpatient care in the public sector were low-to-moderate (0–4 per cent) for most households across the different variables, catastrophic only in a very small minority (except 5.3 per cent in Limpopo Province). In contrast, these levels were 23.7 per cent for households that sought private-outpatient care, varying markedly by race (from 36.3 per cent for black Africans to 3.3 per cent for Indians/Asians); area type (54.1 per cent for rural, through 30.7 per cent for informal urban to 14.4 per cent for formal urban); and socio-economic status (59.5 per cent for the poorest, compared to 5.2 per cent for the wealthiest; Table 2). Over two-fifths of those above 65, and 35.1 per cent of pensioners faced financially catastrophic costs following private-outpatient visits.

Unsurprisingly, insurance status was strongly linked with financial catastrophe, experienced by 43.0 per cent of the uninsured versus just 4.0 per cent of the insured utilizing private-outpatient care (Table 3). Similarly, for private inpatients, five times as many uninsured respondents faced catastrophic costs than those with insurance. OOP payments were also catastrophic for 14.2 per cent of pensioner private inpatients and 16.2 per cent of those aged above 65, compared to 3.9 per cent below 18. A fifth of the poorest, and half in quintile 2, experienced catastrophic costs as private inpatients, while for the upper three quintiles this burden was low-to-moderate.

Most (88.2 per cent) encountered low-to-moderate OOP burdens as public sector inpatients. However, almost twice as many men (9.8 per cent) than women (5.3 per cent) experienced catastrophic costs, as did rural (10.1 per cent) relative to urban-formal dwellers (4.8 per cent). Catastrophic payments for public inpatients were also borne by 9.1 per cent of Indians or Asians, falling to 4.7 per cent of Whites, as well as 12.0 per cent of the employed, 10.4 per cent of those living in the largely rural Limpopo Province, and 8.2 per cent of rural-dwellers.

Despite free PHC services and hospital user fee exemptions for uninsured children under 6,2,17 OOP were made by 17.0 per cent of children under 6 as public sector inpatients and 7.7 per cent of uninsured patients attending a PHC facility.

Acceptability

Long queues (8.5 per cent), perceived ineffective care (6.1 per cent), and anticipated disrespectful treatment (2.9 per cent) partly accounted for delayed care-seeking. Most commonly, delays were due to a belief that the illness was not serious enough to warrant immediate care (68.8 per cent), highest among the richest and insured (Table 2).

Desire for respectful treatment influenced the health-seeking behavior of almost a quarter (22.3 per cent) attending private-outpatient services, but only 4.1 per cent accessing public PHC services, and 5.7 per cent using public hospitals. For inpatients, anticipation of respectful treatment was twofold as important for private (17.1 per cent) than public patients (7.5 per cent). Around four-fifths of main respondents who used public-inpatient services in the past year reported being treated respectfully by health providers, compared to 92.9 per cent of private inpatients (Table 4). Over half of all respondents (54.7 per cent) felt that patients at public hospitals are ‘rarely treated with respect and dignity’. Perceptions, however, varied by source, with 46.3 per cent of those actually admitted to a public hospital in the past year holding this view, compared to 54.7 per cent who had never been admitted, and 54.2 per cent basing their views on media reports.
Table 4

Service dissatisfaction among main respondents

Variable (% dissatisfied with service used in past month)

Outpatient service

Inpatient service

 

Public

Private

Public

Private

Clean facility

10.0

6.2

10.9

4.5

Consultation in private

13.8

8.2

15.3

4.5

Health problems kept confidential

10.3

6.8

15.9

3.3

Treated with respect and dignity

19.6

9.7

21.2

7.1

Drugs received improved their health

17.5

9.1

13.3

3.9

Timely medical attention

37.5

17.7

25.5

5.1

Overall quality of care

22.4

9.0

19.3

5.1

Dissatisfaction levels were high regarding the time taken to receive services: 37.5 per cent of public outpatients, 17.7 per cent of private outpatients, and 25.5 per cent of public inpatients (Table 4). Other acceptability factors that evoked dissatisfaction included cleanliness, privacy, and confidentiality.

Finally, confidence in the effectiveness of care received influenced outpatient facility choice for almost half (43.5 per cent) of those in the private sector, but only 9.6 per cent using PHC, and 13.7 per cent using public hospitals, although slightly higher in public inpatients (18.8 per cent). Among public sector patients, between a third and a quarter were dissatisfied with the overall quality of care received, while fewer were dissatisfied with overall quality of private care (9.0 per cent outpatients, 5.1 per cent inpatients).

Discussion

In 1971, the ‘inverse-care law’ was coined, because the availability of good medical care varied inversely with population health needs.18 Forty years on, many poor or disadvantaged social groups are denied equal access to good-quality services, despite their greater need.1,8 Delineating access barriers is a first-step towards reversing inequities and is a prerequisite for achieving UHC.1,8

This study is strengthened by the use of two need measures: self-reported health status among main respondents; and recent illness or injury within all household members. When gauged by health status, need varied predictably by socio-economic status, gender, and residence. However, need assessed by the second measure was less clearly differentiated across such variables. It is well documented that low-income groups cannot ‘afford’ to be ill, and therefore under-report or ‘ignore’ illness.12,19,20 Further, remembering recent illness in other household members may incur recall bias, or recent illnesses of other members may be conflated with service utilization.

Need for health care is difficult to measure as it is embedded within social norms and constructions of illness and perceptions of health.20 In our results, a perception that illness was not serious enough to warrant medical attention was the commonest reason for delayed treatment. This perception was especially prominent among the rich and insured, suggesting that they might have had less serious illnesses than the poor and uninsured. Unaffordable transport, anticipation of disrespectful providers, and a belief that care would be ineffective were more prominent access barriers for these latter groups.

For care seekers, total utilization was similar across socio-demographic groups. Marked disparities were, however, noted between the type of care accessed, both between private and public sectors, and within the public sector itself. Utilization of higher-level public facilities was greatest among richer, urban, and insured. Because tertiary hospitals are concentrated in the largely urban, wealthier provinces of Gauteng and Western Cape, and are better resourced and specialized than district facilities,2 this finding raises equity concerns. As elsewhere, access to specialized, doctor-led curative services within the public sector illustrates the inverse-care law.8 This emphasizes the need for considering ‘depth’ dimensions of UHC (type of services offered) alongside the ‘breadth’ (coverage for all). The finding also raises questions around referral systems that may unfairly privilege certain groups, and why some groups of people appear to ‘by-pass’ the district health system – a cornerstone of efforts to address access inequities.2

Costs of accessing services can be crippling for poor households.21 Our results suggest that the poorest bear disproportionate cost burdens. OOP burdens of outpatient care also fall on uninsured members, largely from their use of private providers. Although the poorest quintiles make more use of public PHC services, around a fifth of quintiles 1 and 2 also used private-outpatient care. Considerable private sector use across all socio-economic quintiles is not unique to South Africa, and accounted for over 20 per cent of outpatient visits for the poorest groups in 39 LMICs.22 However, in South Africa this burden on the poor bears vivid testimony to the country's distinctive private-public sector split, which severely limits cross-subsidization from wealthy to poor, and from healthy to sick. It emphasizes the need for reforming the private health sector in South Africa.12,23

As in many other LMIC contexts,1,22 transportation costs and travel distance emerged as key access barriers, especially for black Africans, poor, and rural residents. Although the Clinic Upgrading and Building Programme has improved service availability,3 we found that access barriers relate to the geographic inaccessibility of health facilities, particularly in largely rural and poorly resourced provinces. However, within the same geographical setting, different households cope differently with illness.5 This suggests a need for holistic and inter-sectoral approaches to support worse-off households, including mobile services, grants, and user fee exemptions.1,5

We found that a considerable portion of the groups exempted from user fees still pay for services. This undermines the equity-objectives of the government's exemption policies2,17 and risks undoing this important financial protection for poor households and vulnerable groups.5 It also illustrates the ‘discretionary power’ of providers and bureaucrats who determine who ultimately qualifies for exemptions.24

Understanding how frontline staff shape acceptability of health care is crucial.25 Respectful treatment, especially in the private sector with financial incentives to influence user choice, attracts users to certain facilities. Fewer public service users felt they were treated with respect and dignity. Provider respect engenders trusting patient-provider interactions, which sustain access, particularly for socially disadvantaged groups who generally bear the brunt of unacceptable care.25,26 Strengthening interventions to change organizational culture and management practices,25 and ensuring compliance with the Patients Rights Charter,24 are important for addressing the differential acceptability needs of disadvantaged groups (the ‘height’ of UHC).8

Finally, our results show that perceptions about health care vary according to whether respondents had recently used public sector services (more positive) or not (more negative); a reminder that the acceptability of health care is socially ingrained,25 and shaped by the media, and experiences of family and friends. Policy-makers therefore need to challenge negative perceptions and stereotypes, while simultaneously addressing legitimate concerns about the quality of care on offer.13 Improved acceptability, stimulating a shift from private to public services, would diminish the adverse financial burdens incurred through private providers.

Limitations

Poor recall might account for the total OOP payments in this survey being approximately $66 million below other recent estimates, triangulated from the Medical Schemes Council, Treasury and National Health Accounts.11 Monthly premiums paid to medical schemes were not considered in our OOP calculations, yet these pose substantial cost burdens. Although we enquired about preventative and related non-curative services, responses were restricted largely to curative care. Further, methodologically, temporal relationships cannot be established between variables within cross-sectional surveys,27 where, for example, high OOP payments might account for present socio-economic status. Future cohort and qualitative analyses might define the order of such events.

Conclusion

To achieve equitable UHC, the right to access health must be realized across society so that those who need care are able to access it regardless of who or where they are, or their ability to pay.1,28 Our findings concur with previous South African studies, confirming that poor, uninsured, black Africans, and rural groups have inequitable access.2,3,5,7 These inequities mirror the South African context, signaling the limited power of vulnerable social groups to ‘claim and use entitlements and opportunities’,8 and resonating strongly with studies in other LMICs.1,8,21,22 Undoing this status quo requires a comprehensive approach to UHC, which considers the ‘breadth’, ‘depth’, and ‘height’ of access, rather than financing reform alone.1,8

A financing-centered approach to National Health Insurance may reduce some of the affordability barriers, but will not deal with other access barriers found in this survey. Indeed, given the choice, many prefer using the private sector, even if it incurs catastrophic payments. This results in greater resources flowing to private facilities, thus worsening the public sector. Similarly, efforts to revitalize PHC and district systems – which might enhance affordability and availability – need to consider acceptability, whether intended users will actually access these services. Improving public sector service quality and perceptions thereof, and creating equitable access to different levels of public care, could reduce use of private providers and thus minimize financially catastrophic charges. These steps would create a closer ‘fit’ between the equity-seeking objectives of present policies and the inequitable, unhealthy realities that many continue to face.

Notes

Acknowledgements

For their highly valued contribution to the data collection, management and analysis, we would like to thank our colleagues, including Vanessa Daries, Veloshnee Govender, Okore Okorafor, Robert Moeti, Adelaide Maja, Natasha Palmer, Anne Mills, and Olufunke Alaba. For conceptual guidance, we would like to thank Duane Blaauw and Laetitia Rispel. SACBIA survey was a collaborative initiative between Health Economics Unit, University of Cape Town; Centre for Health Policy, University of the Witwatersrand; South African National Department of Health (NDoH); and the London School of Hygiene and Tropical Medicine. NDoH funded the survey through a European Union grant. The Community Agency for Social Enquiry collected the data. Diane McIntyre is supported by the South African Research Chairs Initiative of the Department of Science and Technology and National Research Foundation.

References

  1. World Health Organization. (2010) The World Health Report – Health Systems Financing: The Path to Universal Coverage. Geneva: World Health Organisation.Google Scholar
  2. Coovadia, H., Jewkes, R., Barron, P., Sanders, D. and McIntyre, D. (2009) The health and health system of South Africa: Historical roots of current public health challenges. Lancet 374 (9692): 817.CrossRefGoogle Scholar
  3. Gilson, L. and McIntyre, D. (2007) Post-apartheid challenges: Household access and use of health care in South Africa. International Journal of Health Services 37 (4): 673–691.CrossRefGoogle Scholar
  4. Constitution of the Republic of South Africa. (No. 108 of 1996). Statutes of the Republic of South Africa – Constitutional Law 38: 1241–1331, http://www.info.gov.za/documents/constitution/1996/a108-96.pdf, accessed 4 May 2011.
  5. Goudge, J., Russell, S., Gilson, L., Gumede, T., Tollman, S. and Mills, A. (2009) Illness-related impoverishment in rural South Africa: Why does social protection work for some households but not others? Journal of International Development 21: 231–251.CrossRefGoogle Scholar
  6. Nteta, T.P., Mokgatle-Nthabu, M. and Oguntibeju, O.O. (2010) Utilization of the primary health care services in the Tshwane region of Gauteng province, South Africa. PLoS ONE 5 (11): e13909.CrossRefGoogle Scholar
  7. Schneider, H., le Marcis, F., Grard, J., Penn-Kekana, L., Blaauw, D. and Fassin, D. (2010) Negotiating care: Patient tactics at an urban South African hospital. Journal of Health Services Research & Policy 15 (3): 137–142.CrossRefGoogle Scholar
  8. Frenz, P. and Vega, J. (2010) Universal Health Coverage with Equity: What We Know, Don’t Know and Need to Know. Background paper for the global symposium on health systems research. Montreux, Switzerland: First Global Symposium on Health Systems Research, http://www.hsr-symposium.org/images/stories/9coverage_with_equity.pdf, accessed 4 May 2011.Google Scholar
  9. Thiede, M., Akweongo, P. and McIntyre, D. (2007) Exploring the dimensions of access. In: D. McIntyre and G. Mooney (eds.) The Economics of Health Equity. Cambridge: Cambridge University Press, pp. 103–123.CrossRefGoogle Scholar
  10. Penchansky, R. (1977) The Concept of Access: A Definition. Hyattsville, MD: National Health Planning Information Centre.Google Scholar
  11. McIntyre, D. et al (2007) A Critical Analysis of the Current South African Health System. Cape Town, South Africa: Health Economics Unit, University of Cape Town and Centre for Health Policy, University of the Witwatersrand.Google Scholar
  12. Ataguba, J. and McIntyre, D. (2009) Financing and Benefit Incidence in the South African Health System: Preliminary Results. Cape Town: Health Economics Unit. Report No. 09-1.Google Scholar
  13. McIntyre, D., Goudge, J., Harris, B., Nxumalo, N. and Nkosi, M. (2009) Prerequisites for national health insurance in South Africa: Results of a national household survey. South African Medical Journal 99 (10): 725–729.Google Scholar
  14. Nxumalo, N., Alaba, O., Harris, B., Chersich, M. and Goudge, J. (2011) Utilization of traditional healers in South Africa and costs to patients: Findings from a national household survey. Journal of Public Health Policy 32 (Supp 1): S124–S136.CrossRefGoogle Scholar
  15. Rao, J.N.K. and Scott, A.J. (1984) On chi-squared tests for multi-way contingency tables with cell proportions estimated from survey data. Annals of Statistics 12: 46–60.CrossRefGoogle Scholar
  16. McIntyre, D., Muirhead, D. and Gilson, L. (2002) Geographic patterns of deprivation in South Africa: Informing health equity analyses and public resource allocation strategies. Health Policy and Planning 17 (1): 30–39.CrossRefGoogle Scholar
  17. McIntyre, D. and Gilson, L. (2002) Putting equity in health back on the social policy agenda: Experience from South Africa. Social Science and Medicine 54 (11): 1637–1656.CrossRefGoogle Scholar
  18. Tudor Hart, J. (1971) The inverse care law. Lancet 297 (7696): 405–412.CrossRefGoogle Scholar
  19. McIntyre, D., Gilson, L., Valentine, N. and Soderlund, N. (1998) Equity of Health Sector Revenue Generation and Allocation: A South African Case Study. Washington DC: Partnerships for Health Reform.Google Scholar
  20. Sauerborn, R., Adams, A. and Hien, M. (1996) Household strategies to cope with the economic costs of illness. Social Science and Medicine 43 (3): 291–301.CrossRefGoogle Scholar
  21. Xu, K., Evans, D.B., Kawabata, K., Riadh, Z., Klavus, J. and Murray, C.J. (2003) Household catastrophic health expenditure: A multi-country analysis. Lancet 362 (9378): 111–117.CrossRefGoogle Scholar
  22. Saksena, P., Xu, K., Elovaino, R. and Perrot, J. (2010) Health Services Utilization and Out-of-Pocket Expenditure at Public and Private Facilities in Low-income Countries. Geneva: World Health Organisation. Background paper no. 20.Google Scholar
  23. African National Congress. National Health Insurance. (2010) ANC National General Council, Additional Discussion Documents. 24–28 September; Durban, South Africa: African National Congress. pp. 1–48.Google Scholar
  24. Nkosi, M., Govender, V., Erasmus, E. and Gilson, L. (2008) Investigation the Role of Power and Institutions in Hospital-level Implementation of Equity-oriented Policies. Johannesburg, South Africa: Centre for Health Policy, University of the Witwatersrand; Health Economics Unit, University of Cape Town and London School of Hygiene and Tropical Medicine.Google Scholar
  25. Gilson, L. (2007) Acceptability, trust and equity. In: D. McIntyre and G. Mooney (eds.) The Economics of Health Equity. Cambridge: Cambridge University Press, pp. 124–147.CrossRefGoogle Scholar
  26. Gilson, L. and Erasmus, E. (2005) Supporting the Retention of HRH: SADC Policy Context. Johannesburg: Centre for Health Policy. Equinet Discussion Paper No. 26.Google Scholar
  27. Mann, C.J. (2003) Observational research methods. Research design II: Cohort, cross sectional, and case-control studies. Emergency Medicine Journal 20 (1): 54–60.CrossRefGoogle Scholar
  28. Hunt, P. (2006) The human right to the highest attainable standard of health: New opportunities and challenges. Transactions of the Royal Society of Tropical Medicine and Hygiene 100 (7): 603–607.CrossRefGoogle Scholar

Copyright information

© Palgrave Macmillan, a division of Macmillan Publishers Ltd 2011

Authors and Affiliations

  • Bronwyn Harris
    • 1
  • Jane Goudge
    • 1
  • John E Ataguba
    • 2
  • Diane McIntyre
    • 2
  • Nonhlanhla Nxumalo
    • 1
  • Siyabonga Jikwana
    • 3
  • Matthew Chersich
    • 1
    • 4
  1. 1.Centre for Health Policy & Medical Research Council Health Policy Research Group, School of Public Health, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, Private Bag X3, WitsJohannesburgSouth Africa
  2. 2.Health Economics Unit, School of Public Health and Family Medicine, University of Cape TownCape TownSouth Africa
  3. 3.National Department of HealthPretoriaSouth Africa
  4. 4.Department of Obstetrics and GynecologyInternational Centre for Reproductive Health, Faculty of Medicine, Ghent UniversityGhentBelgium

Personalised recommendations