South Africa is estimated to have 6.6 million HIV-infected individuals, more than any other country, of whom 1.2 million will require antiretroviral therapy (ART) in the next 5 years [1,2]. A national HIV/AIDS treatment programme was launched in November 2003 . By December 2005, only 150 000 had been started on antiretroviral drugs, well below the World Health Organization (WHO) ‘3 by 5’ goal for South Africa [4,5]. Industry is a major healthcare provider for employees, and therefore has an important role to play in ART provision [6,7]. The Anglo American group South African operations in 2002 had approximately 140 000 employees with an estimated HIV prevalence of 24%. Of the estimated 33 400 HIV-infected employees, based on limited published data, which assume that individuals with CD4 cell counts below 200 cells/μl require ART , 10% (3400) were thought to require ART immediately. In November 2002, Anglo American implemented a policy decision to make ART available as part of their programme of HIV care to all employees [9,10]. The programme provides HIV treatment and care for employees throughout the country in both rural and urban areas that range from well-equipped hospitals to small occupational health clinics with a doctor visiting 1–2 days a week.
With the rapid implementation of new ART programmes, monitoring and evaluation is critical to ensure systems function effectively. Evaluation of treatment outcomes using cohort analysis, similar to that used by tuberculosis treatment programmes, has been recommended for ART programmes [11,12]. Cohort analyses provide information on the number of patients starting ART, retention rates and treatment outcomes over time. Cohort analyses provide a useful tool to monitor changes in programme performance over time. This paper describes the treatment outcomes of a large workplace ART programme using a cohort approach, and related average costs. We also review results from studies conducted within the programme, and lessons learnt while implementing the programme.
HIV treatment programme design
The objectives of the workplace HIV programme are to provide antiretroviral drugs to employees using a standardized system for ART delivery, to build the capacity of healthcare workers to provide HIV treatment and care, and to provide a framework for operational research.
The programme is largely nurse based, thus reducing the reliance on medical doctors who are a scarce resource. The centrally managed programme includes HIV specialist support, training, data management, laboratory and pharmaceutical services. A standardized, centrally managed, nurse-based ART programme has enabled a rapid scale-up of the programme to 70 ART delivery sites in all nine provinces in both rural and urban settings. The treatment is provided free of charge to all employees of the company; the programme is described in more detail elsewhere .
The criteria for starting ART are WHO stage 4, irrespective of CD4 cell count; CD4 between 250 and 350 and WHO stage 3; and CD4 cell count less than 250 cells/μl, irrespective of WHO stage. The first-line regimen consists of two nucleoside reverse transcriptase inhibitors (zidovudine and lamivudine) and one non-nucleoside reverse transcriptase inhibitor (efavirenz). In women of child-bearing potential, nevirapine is used in place of efavirenz. The second-line regimen is abacavir, didanosine and lopinavir/ritonavir. After starting ART patients are seen every 2 weeks until 8 weeks and then every 3 months. Blood monitoring for haematological and hepatic toxicity is performed at 0, 2, 6 and 12 weeks and then 3-monthly. CD4 cell counts and viral loads are measured at baseline, 6 weeks and 6 months, and then every 6 months thereafter.
Data are collected on standardized forms that are faxed or couriered to the central data management centre. The data are double entered onto an Access database (Microsoft Corporation, Seattle, USA). To ensure confidentiality, databases are password protected and patients are identified using unique clinic numbers; names are not captured on the database. Patient data is integrated with laboratory and pharmacy electronic records using flat file transfers.
For pharmaceutical supply, all but two companies use a pharmacy courier system by which treatment is dispensed centrally on a named-patient basis, and a courier service distributes monthly supplies of medication to the sites. In the other two companies, pharmacies are located on site and supply medication directly to patients.
When patients left employment we attempted to refer them to another ART programme for continuing treatment, but information on transfers out was not routinely collected during this period.
Monitoring and reporting
Sites are monitored regularly to ensure adherence to guidelines and accurate and complete data, initially on a monthly basis and then every second month once the sites are established. Sites are inspected at the monitoring visits for storage of medication and correct storage of patient files to ensure confidentiality is being maintained. After each visit, a monitoring report is made available to the site.
Monthly and quarterly progress reports are provided to sites and companies. Data are imported from the access database into STATA (Stata Corporation, College Station, Texas, USA) for analyses. Outcomes reported include the numbers enrolled into the HIV programme and starting ART, side effects and serious adverse events, adherence to treatment and discontinuation of treatment.
For the estimation of ART coverage, we used updated workforce and HIV prevalence figures for 2005. We estimated that 25% of HIV-infected employees required antiretroviral drugs, which differs from published estimates  because: (i) company criteria for starting treatment are less stringent than used in published estimates ; and (ii) ART availability will prolong survival and thus increase the prevalence of ART-requiring individuals. In one participating company in which more than 90% of the workforce have undergone voluntary counselling and testing, approximately 25% of the total estimated number of employees with HIV infection have started ART (unpublished data).
An economic evaluation is being conducted prospectively alongside implementation, which includes the collection and analysis of economic costs of ART provision. The cost analysis is incremental to existing health facilities and adopts an employer–provider perspective. Economic costs are measured using an ingredients costing approach, whereby resource quantities are measured and multiplied by respective prices. We cost resources consumed across the central support organisation, Anglo company, ART site and patient levels. Costs included in the analysis are comprehensive, including senior management time, infrastructure required and monitoring and administration, as well as more commonly included costs such as drugs, laboratory and patient visits. Cost per patient virally suppressed included costs for all those who had the opportunity to have 6, 12, 18 or 24 months on treatment whether or not a ‘successful’ outcome was obtained. This included employees who left the workforce up until the termination of employment through ill health retirement, death or other reasons. A discount rate of 10% was used to annualize capital costs. Staff time was measured through time allocation surveys and interviews and was valued at company wage rates per relevant job grade, drugs at prevailing African access prices and laboratory tests at private sector prices. Costs are presented in 2005 US dollars (US$1 = ZAR6.5).
Treatment outcomes were determined for sequential cohorts of patients starting ART in each 6-month period between October 2002 and December 2005. Outcomes are reported using all individuals who were evaluated at the specified timepoint as the denominator. Follow-up was censored at 31 December 2005. For the purposes of this analysis, patients were classified as having ‘stopped treatment’ if a treatment stop was recorded as being a result of patient request, failure to return for clinic visits or to collect medication, or non-adherence.
Approval for this study was obtained from the Anglo Gold Health Services Medical Research and London School of Hygiene and Tropical Medicine ethics committees. All patients gave written informed consent to use their routine information for research purposes.
Between October 2002 and December 2005, 4064 patients were started on ART. The six larger sites had 200–700 patients per site compared with 10–200 at the 63 smaller sites. A total of 2262 patients who started ART, and from whom informed consent was formally obtained to participate in the programme evaluation, were included in the analysis. Of these patients, 94.7% were men and the median age was 41 years (interquartile range 36–47). At the start of ART the median CD4 cell count was 158 cells/μl, the median viral load was 50 366 copies/ml, and the majority of the patients had either WHO HIV stage 3 (45.4%) or 4 (27.4%) disease. By December 2005, the estimated ART coverage was 83% based on 104 528 employees, an average HIV prevalence of 23%, and assuming that 25% of all HIV-infected employees (diagnosed or undiagnosed) fulfilled the company criteria for starting ART.
A total of 93.6% of patients at 14 752 clinic visits reported not missing any tablets over the previous 3 days. Almost half the patients (46.8%, 1058/2262) experienced one or more adverse events (any clinical event whether drug related or not). Of those patients who had an adverse event, 22.0% (225) experienced one or more severe or life-threatening (grade 3 or 4) adverse events. Of the 374 grade 3 or 4 adverse events the majority were caused by haematological (46.3%) or liver function test (19.8%) abnormalities. By the end of December 2005, 30% (680/2262) of patients were no longer on ART because of death (112, 5.0%), termination of employment (185, 8.2%), defaulting or stopped treatment (261, 11.5%), or other reasons (122, 5.4%).
Treatment outcomes are shown in Table 1. Of the 2262 evaluable patients, the median CD4 cell count increased by 90, 113 and 164 cells/μl at 6, 12 and 24 months, respectively. At 6 weeks, a greater than 1 log reduction of viral load was observed in 85%. Some 75% of patients had viral suppression below 400 copies/ml at 6 months, and 72% at 12, 18 and 24 months, respectively. The mean weight increased by 2.8, 3.8 and 4.6 kg by 6, 12 and 24 months, respectively.
The monthly cost per patient on treatment across sites averaged over 12 months of programme implementation ranged from US$160–345, decreasing to between US$145 and 229 for 24 months of programme implementation. The average cost per patient virally suppressed across sites with sample sizes sufficient for calculation to December 2005 (14 sites covering 88% of patients on programme) was cumulatively US$1752 [95% confidence interval (CI) US$1500–2004] at 6 months, US$3887 (95% CI US$3259–4514) at 12 months, and US$6950 (95% CI US$5710–8191) at 24 months.
Table 2 shows treatment outcomes stratified by 6-month periods of enrolment from October 2002 to September 2005. Early treatment outcomes (proportion with a > 1 log reduction in viral load at 6 weeks) improved over time whereas later treatment outcomes (proportion with a viral load < 400 copies/ml at 6, 12 and 24 months) have remained relatively stable. The proportion lost were lower in the early compared with the later cohorts.
Treatment outcomes in this workplace programme have been consistently good. Self-reported adherence is high and clinical events, although common, are mostly mild. Of concern is the fact that mortality remains high and a quarter of patients are lost to follow-up largely through termination of employment and defaulting or stopping treatment.
The virological outcomes of this workplace ART programme are comparable with that reported from other ART programmes in resource-poor and resource-rich countries. In a Ugandan ART pilot project, in which patients were required to pay for drugs at reduced prices, the proportion who achieved viral load suppression below 400 copies/ml one year after starting treatment ranged from 50 to 65% . Similar results have been reported in Senegal  and Haiti . In Baltimore, USA, 68% of patients attending an HIV clinic over a 7-year period achieved viral suppression to less than 400 copies/ml . Similar results have been reported in routine programmes in Europe . Costs are higher than previously reported costs from public sector trials or programmes [19,20] as a result of both the inclusion of more comprehensive costs and all organizational levels, as well as higher input prices faced in a private sector programme.
The cohort method of evaluation allows a programme's performance to be monitored over time. We have previously shown that the clinic site was independently associated with virological failure at 6 and 12 months [21,22]. The cohort analysis included all companies, which may have masked a variation in programme performance within individual companies. The cohort analysis of losses by 12 months indicates that this has remained stable overall. The earliest cohort seems to show lower losses, which could be a chance finding or differences in practice during the pilot phase of the programme.
Higher than expected losses from the programme were experienced. Death, loss of employment and defaulting treatment were the major reasons for loss. Mortality on ART remains high. In a substudy, mortality was compared with a historic cohort before ART was available. Mortality was not reduced in the first 6 months of therapy, but declined significantly by 71% after 6–12 months and 89% after more than 12 months of ART, respectively . Continuity of care for those leaving employment must be ensured. Of concern is that patients leaving the workforce often did not always attend the HIV clinic before returning to their homes in rural areas, thereby compromising continuity of care. In order to achieve continuity of care companies provide bridging treatment until patients are transferred onto another ART programme. Patients who had left the programme but were still in employment were interviewed to determine the reason for leaving the programme. The use of traditional medicines, non-disclosure, side effects, unwillingness to have blood tests, and being unconvinced of the benefits of treatment were cited as reasons for leaving (unpublished data). The losses from the programme, as a result of deaths, left employment and stopping treatment, seem to be higher in the later cohorts and this may have been a result of clinicians' initial reluctance to start ART in very sick individuals at the start of the programme.
Adherence to treatment is a major determinant of virological outcomes. Although self-reported adherence in this study is high, treatment outcomes suggest that adherence is less good than reported. A qualitative study was conducted to identify factors associated with poor adherence. A lower level of education, dissatisfaction with clinic services and a poor understanding of how ART works were associated with poor adherence .
This workplace programme uses zidovudine, lamivudine and efavirenz as its first-line regimen. Although clinical events were common, they were mostly mild. Anaemia and hepatotoxicity accounted for most of the severe adverse events.
A centrally supported workplace ART programme that uses a standardized protocol enables a systematic evaluation of the costs and savings that result in changes in healthcare utilization, labour turnover, absenteeism and productivity. As the programme has matured, the costs of programme implementation have reduced by 10–34% over the 24-month period. The workplace ART programme has had beneficial outcomes with respect to absenteeism and healthcare utilization . The incidence of tuberculosis has also declined, which among miners on ART is half that of a historic cohort of ART-eligible individuals attending clinic before ART was available .
A number of important lessons have been learnt from managing this workplace antiretroviral programme. Retention and treatment adherence are key concerns in the programme, which are being addressed by continued emphasis on the counselling of patients, and ensuring adequate staffing levels, with sufficient training and support. On the basis of our experience, the following recommendations can be made: (i) A dedicated HIV programme manager is required within each company to take overall responsibility for the HIV prevention and treatment programme; (ii) At each treatment site a dedicated HIV coordinator is required to administer the programme locally (staff, training and follow-up of patients); (iii) The focus on counselling should always be maintained. Sufficient staff should be in place to ensure there is adequate time for counselling. Training on counselling should be ongoing and should include counselling on disclosure and stigma. Counsellors should have regular debriefing sessions; (iv) The Aurum centralized ART management programme recommends one doctor, 2.5 nurses and 2.5 counsellors for every 250 patients on ART. This staffing ratio is sufficient to accommodate an additional 250 patients not on ART but requiring HIV care; (v) The integration of clinical and pharmacy services is essential. When these are separate, a system needs to be in place to notify clinic staff of patients who do not return for their medication; (vi) Treatment collection points should be close to where an individual works in order to minimize absenteeism, which may in turn lead to stigma from fellow workers; and (vii) Training is essential to maintain a competent cadre of staff working in the HIV treatment programme.
Monitoring treatment outcomes in sequential cohorts is a useful way of monitoring a programme's performance. A company can in many ways be considered a microcosm of a country. At a country level, key factors determining a successful response to the HIV epidemic include: strong leadership; a culture of openness and non-discrimination; a ‘multisectoral approach’ involving all sectors; and a ‘continuum of care’ with integration of HIV services. Challenges in implementing a workplace ART programme are similar to public sector programmes.
The authors would like to take this opportunity to acknowledge the contribution of all the company leaders who were instrumental in ensuring the programme was successful: thanks to the doctors, nurses and counsellors at all the ART delivery sites for their dedication to the programme. To all the members of the Aurum ART consortium, thanks for your continued dedication in the project. They would also like to acknowledge other staff in the Aurum ART and economics teams for their contributions to this programme.
Sponsorship: This work was funded by Anglo American Group companies. A.G. is supported by a Public Health Career Scientist award from the UK Department of Health.
Conflicts of interest: G.J.C. was employed by an AngloGroup company (AngloGold Ashanti) during the execution of this work.
2. World Health Organisation. HIV/AIDS epidemiological surveillance report for the WHO African region 2005 update
. Geneva: WHO; 2005.
3. South African Department of Health. Operational plan for comprehensive HIV and AIDS care, management and treatment for South Africa
. 19 November 2003. Available at: www.doh.gov.za
. Accessed: April 2007.
5. International Treatment Preparedness Coalition (ITPC). Missing the target. A report on HIV/AIDS treatment access from the frontlines
. AIDS Law Project, Cape Town, South Africa 28 November 2005.
7. Plumley B, Bery P, Dadd C. Discussion paper: beyond the workplace: business participation in the multi-sectoral response to HIV/AIDS.
In: XIVth International Conference on AIDS
. Barcelona, Spain, 7–12 July 2002. Abstract number-ThOrG1503.
8. Auvert B, Males S, Puren A, Taljaard D, Carael M, Williams B. Can highly active antiretroviral therapy reduce the spread of HIV?: a study in a township of South Africa. J Acquir Immune Defic Syndr 2004; 36:613–621.
9. Wall Street Journal, New York, USA. Giant Anglo American will supply free Anti-AIDS drugs to African workers, 7 August 2002.
10. Charalambous S, Grant AD, Day JH, Chaisson RE, Hayes RE, Churchyard GJ. Feasibility and acceptability of a specialist clinical service for HIV-infected mineworkers in South Africa. AIDS Care 2004; 6:47–56.
11. Harries AD, Gomani P, Teck R, de Teck OA, Bakali E, Zachariah R, et al
. Monitoring the response to antiretroviral therapy in resource-poor settings: the Malawi model. Trans R Soc Trop Med Hyg 2004; 98:695–701.
12. Libamba E, Makombe S, Mhango E, de Ascurra Teck O, Limbambala E, Schouten EJ, Harries AD. Supervision, monitoring and evaluation of nationwide scale-up of antiretroviral therapy in Malawi. Bull WHO 2006; 84:320–326.
13. Charalambous S, Grant AD, Day JH, Pemba L, Chaisson RE, Kruger P, et al
. Establishing a workplace antiretroviral therapy programme in South Africa. AIDS Care 2007; 19:34–41.
14. Weidle PJ, Malamba S, Mwebaze R, Sozi C, Rukundo G, Downing R, et al
. Assessment of a pilot antiretroviral drug therapy programme in Uganda: patients’ response, survival and drug resistance. Lancet 2002; 360:34–40.
15. Laurent C, Diakhate N, Gueye NFD, Toure MA, Sow PS, Faye MA, et al
. The Senegalese government's highly active antiretroviral therapy initiative: an 18-month follow-up study. AIDS 2002; 16:1363–1370.
16. Severe P, Leger P, Macarthur C, Noel F, Bonhomme G, Bois G, et al
. Antiretroviral therapy in a thousand patients with AIDS in Haiti. N Engl J Med 2005; 353:2325–2334.
17. Moore R, Keruly JC, Gebo KA, Lucas GM. Improvement in virologic response to highly active antiretroviral therapy in clinical practice from 1996 through 2002. J Acquir Immune Defic Syndr 2005; 39:195–198.
18. Grabar S, Le Moing V, Goujard C, Leport C, Kazatchkine MD, Costagliola D, Weiss L. Clinical outcome of patients with HIV-1 infection according to immunologic and virologic response after 6 months of highly active antiretroviral therapy. Ann Intern Med 2000; 133:401–410.
19. Badri M, Maartens G, Mandalia S, Bekker LG, Perrod JR, Platt RW, et al
. Cost-effectiveness of highly active antiretroviral therapy in South Africa. PloS Med 2006; 3:e4.
20. Boulle A, Kenyon C, Skordis J, Wood R. Exploring the costs of a limited public sector antiretroviral treatment programme in South Africa. S Afr Med J 2002; 92:811–817.
21. Charalambous S, Stenson A, Fielding KL, Pemba L, Churchyard GJ, Grant AD. 6-Month outcomes among HIV-infected individuals starting antiretroviral therapy in South Africa.
In: 3rd International AIDS Society Conference on HIV Pathogenesis and Treatment
. Rio de Janiero, Brazil, 24–27 July 2005 [Abstract MoPe11.2C41].
22. Charalambous S, Fielding K, Stenson A, Pemba L, Innes C, Senoge S, et al
. 12-Month outcomes among HIV-infected individuals starting antiretroviral therapy in South Africa.
In: XVIth International AIDS Conference.
Toronto, Canada, 13–17 August 2006 [Abstract WEPE0123].
23. Grant AD, Fielding KL, Charalambous S, Stenson A, Whetham J, Puso T, et al
. Risk factors for early mortality among HIV-infected individuals starting antiretroviral therapy in South Africa.
In: 3rd International AIDS Society Conference on HIV Pathogenesis and Treatment
. Rio de Janiero, Brazil, 24–27 July 2005 [Abstract MoPe11.6C25].
24. Dahab M, Hamilton R, Charalambous S, Dube N, Tsimele J, Govender N, et al
. A quantitative study of barriers & facilitators to adherence in an antiretroviral therapy programme in South Africa.
In: XVIth International AIDS Conference.
Toronto, Canada, 13–17 August 2006 [Abstract TUPE0151].
25. Muirhead D, Kumaranayake L, Pozo-Martin F, Charalambous S, Pemba L, Grant A, et al
. Early savings resulting from employer-sponsored ART in a large South African workforce setting.
In: 2nd South African AIDS Conference.
Durban, South Africa, 7–10 June 2005 [Abstract 652, oral presentation].
26. Churchyard GJ, Fielding KL, Charalambous S, Whetham J, Pemba L, Day J, Grant AD. Antiretroviral therapy reduces risk of TB among Southern African gold miners.
In: XVIth International AIDS Conference.
Toronto, Canada, 13–17 August 2006 [Abstract MOPE0172].