Patient satisfaction with health care providers in South Africa: the influences of race and socioeconomic status

Abstract

Objectives. The first democratic government elected in South Africa in 1994 inherited huge inequalities in health status and health provision across all sections of the population. This study set out to assess, 4 years later, the influence of race and socioeconomic status (SES) on perceived quality of care from health care providers.

Design. A 1998 countrywide survey of 3820 households assessed many aspects of health care delivery, including levels of satisfaction with health care providers among different segments of South African society.

Results. Fifty-one percent (n = 1953) of the respondents had attended a primary care facility in the year preceding the interview and were retained in the analysis. Both race and SES were significant predictors of levels of satisfaction with the services of the health care provider, after adjusting for gender, age, and type of facility visited. White and high SES respondents were about 1.5 times more likely to report excellent service compared with Black and low SES respondents, respectively.

Conclusion. In South Africa, race and SES are not synonymous and can no longer be considered reliable proxy indicators of one another. Each has distinct and significant but different degrees of association with client satisfaction. Any assessment of equity-driven health policy in South Africa should consider the impacts of both race and SES on client satisfaction as one of the indicators of success.

South Africa’s first democratic government, elected in 1994, inherited huge inequalities in health status and health care provision across racial, socioeconomic, and urban/rural boundaries [1,2]. The racist policies of the apartheid era ensured that before 1994 a strong correlation between race and socioeconomic status (SES) could be assumed, or demonstrated, for almost any given indicator of development. This meant that these categories could serve as proxy indicators of social inequality with respect to health status and services.

The new government made a commitment to address these inequalities and improve public services, stating that ‘a transformed South African public service will be judged by one criterion above all: its effectiveness in delivering services which meet the basic needs of all South African citizens’ [3]. A cornerstone of this transformation process was many policies aimed at improving the health of the population in general and inequalities in health care in particular. The new policies placed a strong emphasis on the development of primary health care within the country and a particular emphasis on improving access to health care [4]. South Africa is not alone in proclaiming the rhetoric of equality in its efforts at health reform, and it has been a call repeated around the globe in the past decade [5,6]. The real test of the recently transformed health system in South Africa is whether it delivers quality care equitably across the barriers of race and SES, or whether certain inequalities remain entrenched. Client satisfaction is a fundamental indicator of success in any form of service delivery and is therefore a key component of such a test.

Literature on the racial, SES, and ethnic differences in health has focused on health status, care-seeking behaviour, and health care provision differences [7–9], illustrating significant disparities in health status and provision of care by race, SES, and ethnicity. Research mainly emanating from the United States highlights significant differences in perceived quality of care by race, SES, and ethnicity [10–16]. For example, levels of satisfaction were significantly lower in Hispanics and Asians compared with Blacks and Whites [10]. In a recent report, Farmer and Ferraro [17] confirm the significant associations of race and SES with a variety of health variables and further show them each to have a distinct and separate influence on health.

Research examining patient satisfaction with health care provision in South Africa and, more specifically, the perceived quality of care given by the health care providers is limited. Moreover, the influence of race and SES on levels of patient satisfaction is relatively unexplored. National surveys in South Africa have highlighted the levels of satisfaction by race of patients attending public and private sector health care providers [18,19]. A 1994 national survey showed that 48% of White (during the apartheid period, all people in South Africa were classified Black, Indian, Coloured, or White according to the Population Registration Act of 1950) respondents reported receiving excellent services compared with 26% of African and 24% of Coloured respondents [18]. A fifth (22%) of respondents below the minimum living level reported excellent services compared with 35% above this minimum living level. Focusing solely on race, a survey in 1998 found that in the public sector, 26% of Indian respondents were dissatisfied with services provided at clinics/hospitals compared with 12% of Black respondents [19]. In the private sector, 7 and 8% of Black and Coloured respondents, respectively, were dissatisfied with services rendered by private doctors, compared with 2% of Indian and 4% of White respondents, respectively.

Using data from the second countrywide Kaiser Household Survey of 1998 [20], this research set out to determine the extent to which race and/or SES continue to influence the level of satisfaction with health care providers 4 years after the election of the new democratic government. It also tries to determine whether their influence is separate and distinct or inseparable and synonymous as currently assumed.

Methods

The aim of the 1998 Kaiser National Household Survey on health inequalities in South Africa was to document the South African public’s awareness, perceptions, and attitudes towards health policy, health status, health care utilisation, access and barriers to health care as well as quality of health care services [20]. This article reports on respondents’ satisfaction with health care providers and how this was influenced by race and/or SES.

The process of assessing client satisfaction within a medical setting is complex and has led to the development of several different measurement tools [21,22]. Any instrument designed to measure client satisfaction must be based on an understanding of what the client means when they express an opinion on the nature of the service they received. Research has also shown that ‘expectations are seen as dependent on the context of the clinical encounter and the past experience and knowledge of the patient’ [23]. To address these issues, we validated the client satisfaction tool used in this study to ensure that the research team had a meaningful understanding of the concerns of the clients and in line with studies conducted elsewhere that have demonstrated how to ensure reliable instruments [24,25].

To ensure construct validity of the survey instrument, we prepared a draft questionnaire and submitted for review to an expert group using a Delphi technique. The revised version was tested through 14 focus groups across the country in both rural and urban areas, and then the final draft was piloted through a survey amongst 100 households drawn from the sample frame. Variables with the potential to influence client satisfaction recorded in the original survey included waiting time, the state of physical facilities, and travel costs. However, the literature confirms that the interpersonal dynamic between patient and provider plays a prominent role in shaping race and class differences in client satisfaction [16,26,27]. The focus of this article is therefore the client satisfaction with the health care provider and how this is influenced by race and SES.

Data for the survey was collected through a national face-to-face survey of almost 4000 randomly selected households, with 4 households selected in 1000 selected enumerator areas [20]. A final sample of 3820 was attained. The selected enumerator areas were stratified by province, race, and urban or rural area type. The household survey adopted the urban and rural definitions used in the national census [28]. Cities, towns, townships, and suburbs were classified as urban settlements. Enumerator areas comprising informal settlements, hostels, institutions, industrial and recreational areas, and smallholdings within or adjacent to any formal urban settlement were also classified as urban. Any area that was not classified urban was considered to be rural. In each enumerator area, the stands to be visited were identified by the fieldworker supervisor after the selection of a random starting point.

Respondents’ satisfaction with the health care providers was measured in the survey by asking participants whether the way health care providers treated them was ‘excellent’, ‘good’, ‘fair’, or ‘bad’. No test–retest estimates of this item were performed and is a weakness of the study. Responses to the level of satisfaction with the provider were dichotomized to excellent and good/fair/bad (less than excellent) because of the low number of respondents who reported fair and bad services. The analysis was restricted to those who had visited a doctor or clinic in the 12 months preceding the interview.

The level of satisfaction with the health care provider was compared by race and SES, adjusting for gender, age of the respondent, and type of facility visited (private or public). A SES indicator was created based on (i) the basic services that the household accesses, (ii) the difficulty a household experiences in paying for a range of basic goods and services, (iii) an estimate of the number of consumer durables in the household, (iv) the highest educational level in the household, (v) the reported monthly income of the household, and (vi) the number of people per room in the household [20]. Three SES categories (high, middle, and low) were created by dividing the SES scores into thirds. The indicator is similar to that currently being used by Statistics South Africa [29] and has been used by others within the South African context [30].

The independent impact of race and SES on the level of satisfaction with the care provider was assessed by carrying out binary logistic regression analyses incorporating level of satisfaction as the dependant variable, age, gender, and type of facility as control variables, and race and SES as the explanatory variables. The results are presented in the form of odds ratios (OR) and 95% confidence intervals (CI). An OR where the 95% CI excluded one was considered statistically significant.

Results

Of the 3820 households surveyed, 1953 (51%) respondents had sought care from a primary care provider in the year preceding the interview. The percentage of missing data was generally less than 1%, and the final model included 1928 records of a possible 1953 records (98%). Table 1 highlights the demographic characteristics of respondents who attended a primary care facility in the 12 months preceding the interview compared with those who did not. Substantially, more females, older respondents, White respondents, and those from the high SES group attended a primary care facility in the preceding year than those who did not.

The level of satisfaction with health care provider by race and SES is summarized in Table 2. Almost 40% of respondents reported receiving excellent treatment from the health care provider, whereas 49% reported a good service, 9% a fair service, and only 2% a bad service. Table 2 also highlights the significant association between race and SES and levels of satisfaction with the services of the health care provider. Compared with 61% of White respondents, only 31% of Black and 38% of Coloured and Indian respondents rated the services of the health care provider as excellent. More than half of respondents from the high SES group felt that they received an excellent service compared with a quarter and third of low and middle SES respondents, respectively.

In the unadjusted logistic regression analysis, both SES and race were significantly related to levels of satisfaction with the health care provider (Table 3). White respondents were 3.38 times more likely and Coloured and Indian respondents 1.35 times more likely than Black respondents to have reported the treatment provided by the health care provider as being excellent. Similarly, high SES respondents were 3.4 times more likely and middle SES respondents 1.53 times more likely than low SES respondents to have reported the treatment provided by the care provider as being excellent. After adjusting for gender, age, and type of facility visited, the influences of race and SES are less obvious, but significant relationships remain. White and high SES respondents were 1.55 and 1.41 times more likely to report excellent service than Black and low SES respondents, respectively.

Discussion

The study revealed high levels of satisfaction with health care providers in South Africa. Forty percent of respondents rated the care provider as excellent, and a further 49% rated the provider as good. However, this high overall level of satisfaction with the care provider may mask some of the underlying differences in levels of satisfaction across the different strata of South Africa society. Comparing the 1994 patient satisfaction survey results (18) with the 1998 results (Table 2), it appears that the percent of White respondents who rate their health care experience as ‘excellent’ increased by 13% (48–61%) over that time. Among Coloured/Indian respondents, the increase was 14% (24–38%), and 5% among Blacks (26–31%). The 1994 survey, however, assessed patient satisfaction in general, and the analysis in this article particularly analysed the levels of satisfaction with the patient–health care provider interaction.

This article set out to specifically examine the extent to which satisfaction with health care providers in South Africa is influenced by race and SES. Without controlling for other factors, there were substantial and significant differences in the levels of satisfaction with the health care providers across race and SES divides. For example, White and high SES respondents were 3.5 times more likely to report receiving excellent services from the health care providers compared with Black and low SES respondents, respectively.

This research further set out to explore whether the influence of race and SES on the level of satisfaction with health care providers was separate and distinct or inseparable and synonymous. The findings from the multivariate analysis suggest that although there is a close relationship between the two, race and SES influence the level of satisfaction separately and independently of each other. After adjusting for gender, age, type of facility, and SES, White respondents were 1.55 times more likely to report excellent services than Black respondents. Similarly in the adjusted analysis, high SES respondents were 1.41 times more likely to report excellent services than low SES respondents even after controlling for race.

One possible explanation for these differences in client satisfaction with health care providers by race and SES may be because of differences in client values, including how they expect to be attended to by the health care provider. An alternate explanation could be that the actual treatment provided might have been different because of the race or class influence on the patient–provider dynamic. This would amount to inequitable or discriminatory health care practice, something the newly elected democratic government made a firm commitment to address by implementing a range of policies to eradicate the inequities from the Apartheid era [3,4]. Assuming the latter, the persistence of these differences might be because of the relatively short time horizon for policy realisation and/or a reflection of a failure of policy efforts within the health system reform process to secure equality, at least in how health care providers treat their clients.

Further research may resolve whether differences in patient values or differences in health care provider behaviour are behind the observed differences in satisfaction with the provider–patient interaction. Such research would strengthen our understanding of the causal relationships that shape the provider–patient interaction in the South African context and help to map client satisfaction trends over time. Research in South Africa has generally regarded race and SES as inseparable and synonymous [31]. However, the findings of this study point to the value of including SES into future research and policy analysis.

This study found that 4 years after the election of a new democratic government, race and SES continue to significantly influence levels of satisfaction with health care providers. Further, although there remains a strong link between race and SES, they exert an influence separate and independent of each other. Any assessment of equity-driven health policy in South Africa should therefore consider the impacts of both race and SES on client satisfaction as one of the indicators of success.

References

Author notes

¹Community Oral Health, University of the Western Cape, Cape Town, Western Cape, South Africa, ²Fifth Quadrant Actuaries and Consultants, Cape Town, Western Cape, South Africa, and ³Strategy and Tactics, Cape Town, Western Cape, South Africa