Background
In November 2003, the South African Cabinet announced the
Operational Plan for Comprehensive HIV and AIDS Care, Management and Treatment for South Africa [
1]. This policy document set out a coherent, comprehensive strategy to deal with human immunodeficiency virus (HIV) and acquired immunodeficiency syndrome (AIDS) and explicitly included the use of anti-retroviral medicines for the treatment of HIV and AIDS (anti-retroviral treatment or ART) for HIV/AIDS patients with CD4 count ≤ 200 cells/μl or World Health Organization (WHO) stage IV disease [
2]. In 2007, the
HIV & AIDS and STI National Strategic Plan (NSP) 2007-2011 re-affirmed this emphasis on treatment by adopting as its two primary aims to reduce new HIV infections by 50%, particularly through the prevention of mother-to-child transmission (PMTCT), and to reduce the impact of the epidemic by extending access to appropriate treatment, care, and support to 80% of people diagnosed with HIV [
3].
Preliminary research results showed that ART provided by the South African government health services is as effective as that provided in high-income countries [
4]. The magnitude of the improvements in health and quality of life attributable to ART is similar to the estimates for non-governmental organisations (NGOs) and research-supported ART services in developing countries and for routine treatments in Europe and the United States [
4‐
10]. Studies of the ART programme outcomes have also indicated that South Africa ranks among the top African countries in patient survival and retention rates [
11,
12]. Finally, despite severe human resource shortages, health personnel provide high-quality care, which results in high levels of patient satisfaction with ART-related services [
13‐
15].
These favourable treatment outcomes should not encourage complacency. Despite having the largest public-sector ART programme in the world, the South African ART roll-out is hampered by limited financial and human resources [
16‐
18]. The health-care system is facing a crisis and the government is falling behind in its efforts to meet the
NSP targets, including the treatment of 80% of people who require ART by 2011 [
3]. By the middle of 2008, 568,000 HIV-infected patients were receiving ART in South Africa, with the public health sector accounting for 79% of this total. Based on the current Department of Health criteria for defining anti-retroviral eligibility (CD4
+ count < 200/μL or World Health Organization [WHO] stage IV), anti-retroviral coverage in adults was only 40.2% in 2008 [
19]. Moreover, coverage varied significantly between the provinces, with the Free State, the subject of this study, ranking last (25.8%)[
19]. This creates a risk of ART facilities becoming "islands of excellence in a sea of under provision" [
20‐
22]. ART facilities have a positive impact on the lives of patients enrolled in the programme, but after five years of treatment roll-out, their effects are still inequitable, serving less than half of the people who currently require ART in South Africa [
23]. The large gap between the number of people eligible for ART and the number of people actually accessing care creates a nationwide problem, with long waiting lists. Recent studies by Jacobs et al. (2008) and Rosen et al. (2005) have shown that the large patient numbers and resource constraints mean that ART in South Africa will be rationed for many years to come [
24,
25].
Bennet and Chanfreau (2005) have stressed that it is crucial to monitor who has access to treatment when anti-retroviral programmes are implemented in regions in which they are most needed [
26]. Knowing the characteristics of the enrolled patients can potentially facilitate the successful implementation of the
NSP in two ways. First, knowledge of the characteristics of the enrolled patients can assist policy makers in developing measures to retain these patients in care, a necessary condition for every ART programme. Second, information about these patients' characteristics can identify explicit and implicit rationing systems when the over- and under-represented groups of patients are evaluated [
27]. In this respect, Rosen et al. (2005) identified different types of rationing systems, using socio-economic, geographic, and medical criteria [
25]. The information produced could then be usefully applied by policy makers to target patient groups that are not reached in the implementation of the current ART programme.
Longitudinal studies of who is actually accessing the South African public-sector ART programme are scarce [
28]. This study has therefore one main research question: who is accessing public-sector ART in the Free State Province of South Africa? The study aimed to extend the current literature by investigating, in a quantitative manner and using a longitudinal study design, the participants enrolled in the public-sector ART programme in the period 2004-2006 in the Free State Province of South Africa. Interview data for 912 ART patients were used to assess the differences in the demographic (age, sex, population group, marital status), socio-economic (education, income, neo-material indicators), geographic (travel costs, relocation to obtain ART), and medical (CD4
+ count, viral load, time since first diagnosis, treatment status) characteristics between patients enrolled at various points in time. Furthermore, by identifying differences between the patients who enrolled at different time points, the study assessed whether certain patient groups were benefitting disproportionately from the public-sector ART programme, so that the distribution of public health benefits was uneven across the population. This information could be usefully applied in the development of policy measures that target groups of eligible patients who are not currently accessing ART.
Discussion
During the introductory years of the South African public-sector ART programme (2004-2006), anti-retroviral coverage rose from 4.9% in 2004 to 19.1% in 2006 (and to 40.2% in 2008). The preliminary patient outcomes of the programme were positive [
4,
49], indicating that South Africa ranks among the top African countries for patient survival and retention rates, approaching those reported in industrialised countries [
11,
12]. Despite significant progress in the provision of ART in South Africa, human and financial resource shortages create a large gap between the supply and demand for ART. Currently, only four of every 10 South Africans requiring treatment are actually receiving it. The situation is most acute in the Free State, where ART coverage is only 25.8% [
19]. Against the backdrop of this gap between supply and demand, Stewart et al. (2006) raised concerns that the provision of ART will create and/or exaggerate inequities in service provision [
51]. Therefore, it is crucial to monitor who gains access to these life-saving drugs when anti-retroviral programmes are implemented in regions with high HIV prevalence and limited resources [
26]. Using Rosen's typology of implicit rationing systems [
25], we measured the demographic, socio-economic, geographic, and medical characteristics of the patients enrolling in the Free State public-sector ART programme between 2004 and 2006.
The study results revealed significant differences in all four domains. Demographically, the results showed a significant increase over time in the proportion of females entering the programme, resulting in an under-representation of male patients in the public-sector ART programme. A similar trend was found in an Ethiopian study by Kloos et al. (2007) and a Thai study by Le Cœur et al. [
52,
53]. Because HIV infection rates are higher among women, it is to be expected that more women than men are receiving ART. However, our findings show that the proportion of women undergoing treatment was significantly higher than would be expected given the demographic trends [
54]. In 2006, three of every four (75.6%) patients enrolling in the ART programme were female, whereas the expected percentage based on national reports was only 56% [
55]. A WHO report on the global access to anti-retroviral therapy [
55] speculated that the relative over-representation of women in ART programmes could be because women are better integrated into community networks, and thus have better health-care information and are better able to access public health facilities than are men, especially where these women access PMTCT programmes. Closely related, recent studies suggest that antenatal clinics are often the first point of entry into HIV/AIDS prevention, care and treatment [
56,
57]. Nattrass (2008) hypothesised that South African men with AIDS-related illnesses are less likely to access ART than their female counterparts because masculine norms encourage them to deny weakness and to avoid seeking treatment for any ailment for as long as possible [
54]. Another possible explanation for this difference is the fact that the South African employment rate is substantially higher among males (50.5%) than among females (34.0%) [
58]. Because each visit to the clinic can be assumed to require a full day off work as a result of the protracted waiting times at the facilities [
14,
59], the cost of losing a day of work for health care may still be too high [
60,
61], resulting in the reported under-representation of employed (predominantly male) patients. However, further research is required to fully disentangle the complex relationship between sex and public-sector ART access. Our study also revealed significant but very small differences in the ages of the patients enrolling at different time points, with more recent patients being slightly younger. However, this very small reduction in the mean age of treatment initiation does not warrant any policy measures. Finally, the proportion of coloured patients accessing public-sector ART increased over time, although black patients still constituted the largest population group in the programme. These findings confirm those of previous studies performed by Natrass (2006), Tladi (2006), and Hudspeth (2004) [
62‐
64].
As we would expect in a public-sector health-care programme, the ART patients can be described as relatively poor in certain respects: only one of every five respondents was employed and this percentage remained stable over time. Analysis of the association between socio-economic status and access to public-sector ART revealed a growing trend in labour-market participation among ART patients over time. In other words, patients enrolling in 2006 were more likely to be either employed or looking for work than were patients enrolling during the first two years of the programme. However, our analysis of employment trends displayed no significant direction, indicating that the proportion of employed patients remained low. As suggested above, one potential explanation is that employed (predominantly male) patients are under-represented in the ART population because of the protracted waiting times at the facilities, warranting further research on this topic. Our study found no significant association between the remaining classical indicators of socio-economic status (monthly income, disability grant, employment status) and entry into the ART programme. The analysis of neo-material indices of health inequality showed a clear trend in the housing conditions of the enrolled patients over time: the toilet facilities (both their place and type) improved significantly during the three-year study period. Previous studies by Anderson et al. (2006) and Booysen et al. (2006 and 2007) have already shown that that the initial group of ART patients was not the poorest of the poor [
29,
65,
66]. Our finding that the majority of patients had access to a flush toilet, either in their dwelling or their yard, confirmed these conclusions and also suggests that these inequities worsened over time.
In previous studies, the distance from home to the health facility was often cited as one of the most important barriers to the initial uptake of ART and to patient adherence [
42,
67]. Our study showed that the mean transport cost of a one-way trip to the facility increased significantly (from ZAR 6.8 to ZAR 10.1) over the first three years of the Free State public-sector ART programme. A potential explanation would be that the initial (most urgent) treatment patients moved to live close to the ART facility, but the study results showed no significant association between the year of enrolment and whether the patient had relocated to access ART. It is also noteworthy that the intention of the Free State was that the programme implementation would occur in a phased or staggered manner, and as resources and experience became available [
44]. During phase I of the ART roll-out in the province (May-December 2004), four treatment sites, 13 assessment sites, and three combined sites were established. In phase II (2005/2006), a further three treatment, eight assessment, and seven combined sites were introduced [
68]. Because of this gradual establishment of more ART facilities, it would be logical to see an accompanying gradual reduction in the mean distance (and transport cost) to the nearest ART facility. The evidence suggesting that cost is an important barrier to ART uptake and adherence indicates that further research is required to investigate this worrying trend.
Finally, our study results reveal that eligible patients (CD4 < 200 cells/μL and/or WHO stage IV AIDS) are gradually enrolling at a slightly earlier stage of HIV infection. The interval between the first HIV-positive diagnosis and ART initiation is steadily decreasing over time. This is also reflected in the increasing baseline CD4 cell counts at ART initiation. This trend towards earlier enrolment and initiation of treatment is potentially important because many studies have indicated that delayed ART initiation (low baseline CD4 and high viral load) significantly reduces the chance of ART success [
21,
69]. These results for the initial years of public-sector ART are also consistent with the recommendations for ART scale-up made by Walensky et al. (2008) [
23]. Their simulation model showed that the prioritisation of the sickest patients (i.e., those with the lowest CD4 counts) would lead to significantly fewer deaths compared with the number of deaths in non-prioritised scenarios. However, this positive trend does not justify complacency. The overall low baseline CD4 cell count among the ART patients (120.8 cells/μL) in our sample is similar to that observed in previous studies in South Africa [
35,
70,
71]. These results once again show that increasing the recommended CD4 count at which patients begin ART from 200 to 350 cells/μL blood, as stated in the new WHO protocol for ART, will require extensive efforts to reach these currently untreated patients [
72]. This will prove to be a difficult task in the context of crippling human and financial resource shortages for health services [
17]. Finally, the study showed a clear trend of more ART-naïve patients being enrolled in the public-sector ART programme over time, whereas earlier patients had received ART from other sources before transitioning into the programme. The introduction of free public-sector ART clearly attracted patients from alternative ART programmes, while gradually eligible patients increasingly directly enrolled into the public-sector programme.
The strengths of this study include its longitudinal character and the availability of information on an understudied population. To the best of our knowledge, this is one of the first studies to assess, in a quantitative manner, the demographic, socio-economic, geographic, and medical characteristics of a representative sample of HIV-infected patients enrolling in South Africa's public-sector ART programme [
52]. However, there are some limitations to our study. First, our data set only includes eligible patients enrolling before 2007. Further research is required to monitor the more recent trends in patient characteristics. Second, the study only included patients eligible for ART who actually presented for and accessed treatment. As stated before, the Free State's anti-retroviral coverage was only 25.8% in 2008 [
19]. Consequently, further research is required to investigate the characteristics of the most vulnerable HIV/AIDS patients, those eligible for treatment but who do not access it. Third, a comparatively small number of variables, based on the available survey items, was used to measure the demographic and medical characteristics of the patients, potentially restricting the reliability and validity of these measures. Finally, the reported associations between the patient characteristics and their entry into the ART programme may not be generalisable to alternative settings. We can only ascribe the findings to patients enrolled in a public-sector ART programme and, more specifically, to patients enrolled in South Africa's public ART programme, as implemented in the Free State Province.
Conclusions
Our analysis of 912 HIV/AIDS patients starting ART in the Free State public-sector ART programme showed significant changes in the demographic, socio-economic, geographic, and medical characteristics of the patients during the first three years of the programme. The patients accessing treatment tended to be female and unemployed. Although relatively poor, those who were most likely to access ART services were not the most impoverished social group. The proportion of female clients increased and their socio-economic situation improved between 2004 and 2006. The increasing mean transport cost for a visit to the facility is worrying, because this cost is an important barrier to ART uptake and adherence. Eligible patients are accessing ART at an increasingly earlier stage of HIV infection, which is a positive trend because delayed ART initiation significantly reduces the chance of ART success. Although our knowledge of the characteristics of those patients actually accessing public-sector ART in a resource-poor setting is still in its infancy, this analysis has both practical and theoretical implications.
Theoretically, this study contributes to our understanding of the impact of a large-scale public ART programme on the population of patients eligible for treatment. The study results show that as long as the demand for ART exceeds the supply, implicit rationing of anti-retroviral therapy for HIV/AIDS is inevitable. Despite national policies that recognise the importance of equitable access to ART, the early evidence in the Free State is that these processes can have unwanted inequitable effects. The research community should devote more attention to the potential unintended consequences of ART scale-up in the context of large patient numbers and limited human and financial resources.
In practical terms, the study results show the importance of monitoring the population with access to ART programmes in the regions where they are most needed [
26]. A knowledge of the characteristics of the public-sector ART clients can potentially facilitate the successful implementation of the
NSP. To reach the
NSP's ambitious goal of providing treatment to 80% of HIV/AIDS patients who require it by 2011 [
3], policy makers need data from research studies about the eligible patients who are accessing ART and those who are not, to maximise the policy outcomes in a context of limited resources. A knowledge of the characteristics of the enrolled patients can assist policy makers in developing measures to retain these patients in care, a
conditio sine qua non for every ART programme. Moreover, the information provided can be used to identify and target patient groups that are not currently benefitting from the implementation of the ART programme.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
EW participated in the design of the study, performed the statistical analysis, and wrote the manuscript. KP and FB gave advice on interpreting the results. CH, FVL, DvR and HM were involved in revising the article for important intellectual content. All authors read and approved the final manuscript.