Managing the costs of HIV/AIDS: A case study of a South African contract cleaning company

Abstract

This paper, based on a case study of a South African contract cleaning company in Durban, KwaZulu-Natal, adds to the recent literature on the management of the financial impact of the HIV/AIDS pandemic. After situating the research alongside recent contributions that have examined large enterprises, and small, medium and micro enterprises, the paper provides a profile of the company and its predominantly female workforce. The company's management of costs incurred due to HIV/AIDS is critically assessed from the perspective of financial sustainability, using an AIDS Projection Model developed by Matthews (2007). It was found that while continued employment of this workforce is economically sustainable, both from the perspective of the business and the associated provident fund, the costs to employees are far from equitable. The paper therefore recommends the implementation of a holistic HIV/AIDS management programme, including treatment and prevention activities.

Keywords:

• HIV/AIDS
• labour economics
• social impact
• South Africa

1. Introduction

In their recent review article of the state of research into the HIV/AIDS pandemic in sub-Saharan Africa, Conrad & Ross (2008:1–3) observe that the main effort to understand and formulate policy has continued to emanate from a public health perspective (see, for example, UN, 2003; NDoH, 2008). They further note that, while a range of social sciences, including anthropology, sociology and demography, have become increasingly engaged in HIV/AIDS research, economic work on HIV/AIDS has not been salient in the literature despite some notable exceptions (see for example Quattek, 2003).

This seems peculiar since, as the same authors point out, HIV/AIDS presents immense challenges for the productivity of any economy. While macro-economic analyses often attempt to incorporate the economic effects of HIV/AIDS, as Johnston (2008) has pointed out, a host of problems reside in these attempts. Moreover, these problems affect a range of methodologies, including those grounded in cross-country statistical regression; standard neoclassical growth models based on labour inputs, capital stock and technological change; and computable general equilibrium models – to the extent that Johnston (2008:97) declares that the debate about the limitations of macro-modelling is ‘bifurcated’. To add to these difficulties, studies conducted at the level of the household, while offering insight into particular situations, perennially run the risk of not being sufficiently generalisable to lend themselves to policy development. We are therefore left with the problem of finding a level of analysis that is both relevant for urgent policy formulation and able to incorporate insights from a critical perspective on the economic phenomenon of HIV/AIDS.

This problem has at least been partially circumvented by literature that has focused on business-level studies of HIV/AIDS impacts. With the publication of The Impact of HIV and AIDS on Selected Business Sectors in South Africa (BER, 2005), the South African Government recognised the seriousness of the impact of HIV/AIDS on the private sector. In addition, some case studies have been done of micro enterprises (see, for example, Chao et al., 2007), small and medium enterprises (Connelly & Rosen, 2004) and large South African companies (Dickinson & Stevens, 2005), and a global baseline report was recently commissioned by the Global Business Coalition on HIV/AIDS (GBC, 2006). The present study adds to this area of HIV/AIDS research by focusing on a medium-sized contract cleaning company in Durban, KwaZulu-Natal. It investigated the extent to which the prevalence of HIV/AIDS threatens the financial sustainability of the business in the medium term and, for the longer term, how the potential costs may be defrayed so that both the profitability of the enterprise and the well-being of its employees may be sustained.

The following five sections are organised as follows. Section 2 provides an account of the contract cleaning industry in South Africa and its employees and how it currently manages HIV/AIDS. Section 3 explains the theoretical framework, derived from Rosen et al. (2000), that informs the investigation and describes the approach that was used to gather sero-prevalence data. Section 4 presents the results, explaining how these data, and data from company files on past employees, were used to produce a future HIV/AIDS cost-structure scenario. Section 5 specifies the key finding of the research, that this company's future financial sustainability is not threatened by the incidence of HIV/AIDS in its workforce. The concluding section finds that, while prima facie it may appear that the provident fund associated with the company is facing financial difficulties, this is in fact not the case: a number of options are available to both the company and the fund to mitigate any perceived threat, but these options do come at a cost – principally to employees of the company.

2. Managing HIV/AIDS: The industry, the company and its employees

2.1 Contract cleaning in South Africa

The contract cleaning industry in South Africa employs approximately 80 000 people in labour-intensive work. Most are unskilled female workers (cleaners) who earn relatively low wages (R1500 per month at the time of writing) and have very limited health and pension benefits. There are approximately 340 companies in the industry. The employer organisation, the National Contract Cleaning Association, represents over 240 companies. In contrast, less than 30% of the employees belong to trade unions, although several unions operate in this industry. The Bargaining Council of the Contract Cleaning Industry (BCCCI) in KwaZulu-Natal was created to maintain fair and equitable industrial relations and conditions of employment agreements for the industry (BCCCI, 2005). It is made up of an equal number of employer and employee (union) members and covers approximately 11 000 employees (15%) in the industry. Moreover, the South African Business Coalition on HIV/AIDS (SABCOHA) was formed in 2000 in an effort to encourage the private sector to acknowledge the real threat of the epidemic to businesses and to respond appropriately (SABCOHA, n.d.). It is with this background of the institutional players and governance structures in mind that we now turn to examine the case study.

2.2 The case study company and its employees

The medium-sized contract cleaning company that was the object of this study is located and operates in Durban, KwaZulu-Natal. At the time of the study it had 498 employees, 443 of whom were working as carers and cleaners in six aged care facilities. Some were engaged in bathing elderly people while others cleaned the premises. Some were also engaged in cleaning in four charitable and corporate organisations. There were also 30 registered nurses employed on a full-time basis, in the aged care facilities, as well as 10 supervisors for the overall operations and 15 head office administrative staff. Of the 443, there were 419 black women employees, 15 black men and nine coloured women. Of these, 75% were aged 35 years or more and had been working for the company on a long-term basis. The employee profile was typical of contract cleaning companies: mainly women (the female to male ratio was 97:3), predominantly black in terms of South Africa's current population categorisation model, mostly middle-aged and mostly having worked for long periods for the company.

The average wage for a cleaner working for this company was R14 244 per annum. Regulated casual wage rates were R7.71 per hour. These wages were also typical for the industry, but on the low side, computing to R80 a day for casual employees and R1490 a month for full-time employees. Other conditions of employment fell within industry-level agreements. Full-time employees were entitled to four weeks paid annual leave in December. All employees belonged to the BCCCI provident fund, administered by Negotiated Benefits Consultants. Under the fund scheme, employees contributed 6% of their salary and the company contributed another 6%. Benefits in the case of death included one year's salary and R3000 to assist with funeral costs. However, contrasting with the universal membership of the provident fund, there was no medical aid scheme or employer contribution for health insurance. It appeared that none of the employees had private medical aid cover, though management reported that a few (non-cleaning) employees were paying for medical aid coverage. Sick leave benefits included two days salary for temporary absence due to illness provided the employee produced a doctor's certificate. Employees and managers alike contributed 1% of their salary to the unemployment insurance fund. Income insurance was available through the company's insurance company to employees at a cost of R6 a month.

3. Method of study

3.1 Conceptualising the cost of HIV/AIDS

Figure 1 shows Rosen et al.'s breakdown of the impact of HIV/AIDS on a population of employees (2000). This figure depicts schematically the various costs a company can incur, and is particularly useful because it demonstrates how the costs of HIV infection are delayed. The virus can incubate for three to five years before an infected person begins to suffer HIV-related illnesses. If unable to get adequate treatment, the person will suffer increasing bouts of more severe and varied opportunistic illnesses and ultimately death seven to 12 years after infection. The cost to any organisation of an employee becoming infected is thus also delayed.

Figure 1. Progression of cases, costs and liability

Source: Rosen et al. (2000:301).

Figure 2 shows Rosen et al.'s more specific breakdown of the cost of HIV/AIDS to companies. It presents a quadrilateral taxonomy of likely scenarios, comparing costs for an individual incident of HIV/AIDS on the one hand and multiple incidents on the other, in terms of both direct (‘out of pocket’) costs and indirect (‘productivity’) costs. The left-hand side of the figure shows the direct costs that the company may incur as a result of HIV/AIDS illness and death of one of its employees and of many of its employees. The four direct individual costs are: retirement benefits (provident fund), medical care, recruitment costs and training costs. The first two apply to an existing employee and the second two to a replacement employee. The three direct collective costs are: benefits premiums, increased accidents and legal costs. These collective costs are more difficult to measure but can be significant in companies where large-scale HIV infection is present. The right-hand side of the figure shows the indirect costs that may result from HIV/AIDS illness and death of one of its employees and of all of its employees. The indirect costs of individual illness are sick leave, supervisor's time, on-the-job productivity costs, recruitment, and productivity losses while positions are vacant and during training. The first three apply to an existing employee, and the second three to a replacement employee. These indirect costs, with the exception of sick leave, are difficult to measure but can be significant.

Figure 2. Costs of HIV/AIDS in the workforce

Source: Rosen et al. (2000).

It is important to note that the company incurs not only the delayed costs that will result from new infections, but also the ongoing costs of infections dating back several years and the potential costs of future infections (Rosen & Simon, 2003; BER 2004, 2005; Rosen et al., 2004; SABCOHA, 2004). Further, the costs to an organisation do not inevitably increase over time. Any assessment of cost impacts has to take into account both the dynamics of epidemics generally (infection rates eventually taper off) and the development of responses (the prevention and anti-retroviral and other treatment programmes in South Africa may improve over the next few years).

3.2 The modelling approach

The cost-impact analysis developed by the authors focused on incident HIV infections, rather than prevalent ones. Incidence includes only new infections that begin during a specified time period, such as a year, while prevalence includes all infections that exist at a given point in time, including ones that began in earlier years. The long time lag between being infected with HIV and death from AIDS makes this distinction extremely important for any analysis of the costs of HIV/AIDS. Most of the costs associated with HIV/AIDS are not incurred until long after infection with HIV. However, a company incurs liability for these costs from the moment of infection. Under an incidence-based approach, the future costs of HIV/AIDS are discounted to account for the time value of money in order to generate a ‘present value’ estimate of the total costs of each new infection. The present value of a new infection can be thought of as the amount of money the company should invest in an interest-bearing account now to cover the costs it will incur for the infection in the future. The incidence-based approach is the more useful approach for determining the impact of HIV/AIDS on the value of a company and for assessing the potential benefits of interventions to prevent the disease.

4. Results

4.1 Results of the HIV sero-prevalence study

The HIV sero-prevalence study of company employees was conducted by Dr Mark Colvin from the Centre for AIDS Development, Research and Evaluation and approved by the University of KwaZulu-Natal Medical Ethics Committee. The data for the surveyed population are presented in Tables 1 to 3. The total number of useably completed questionnaires was 147, or 30% of the total workforce. Participation was voluntary and anonymous. Of these 147 participants, 45 (31%) were found to be infected with HIV – two of the 12 men and 43 of the 135 women. Table 1 shows the age distribution of HIV infection. The highest levels are in the 31 to 40 year age group, but infections occur in all the groups. Table 2 shows the distribution of HIV by job band. Cleaning staff are the only employees infected with HIV. Table 3 shows HIV status by racial group as defined by policy in the Republic of South Africa. It demonstrates that 33% of the African workers are infected with HIV. The overall picture of the workforce population with HIV/AIDS is that they are cleaners, therefore predominantly women, mostly 31 to 40 years old and mostly black African.

Table 1: Age distribution of HIV infection

Age group	N	HIV-positive
20–30 years	31	11	35%
31–40 years	54	26	48%
41–50 years	40	5	12%
51–60 years	22	3	13%
Total	147	45

Table 2: HIV status by job band

Job band	N	HIV-positive
Management	8	0	N/A
Admin	4	0	N/A
Cleaner	135	45	33%
Total	147

Table 3: HIV status by race group

Race group	N	HIV positive
African	135	45	33%
Asian	1	0	N/A
Coloured	7	0	N/A
White	4	0	N/A
Total	147

4.2 Future financial sustainability

To add to the sero-prevalence data, the company's Human Resources Director provided an anonymous list of 37 former employees who had terminated their employment because of self-reported HIV positivity or HIV/AIDS-related illnesses such as tuberculosis. These employees were either unable to work or deceased. Thus, a total of 37 former and 45 current employees were identified as being HIV-positive or previously employed with HIV/AIDS. The records of this cohort of 82 were examined closely and provided the basis for the projections of the cost of the disease. The temporal distribution of the deaths of former employees (see Table 4) also formed part of the data for the deployment of the AIDS Projection Model (Matthews, 2007).

Table 4: Deaths of company employees

Year	All deaths	HIV/AIDS-related deaths
2000	3	2
2001	8	4
2002	9	7
2003	4	2
2004	2	2
2005	10	9
2006	13	11
Total	49	37 (72%)
Source: Case study company human resource database.

The estimates shown in Tables 4, 5 and 6 are based on results extracted from the ASSA2003 AIDS and Demographic Model (Provincial Version) of the Actuarial Society of South Africa (ASSA, 2005). These results were used to predict HIV incidence (new infections) and AIDS deaths on the basis of the sero-prevalence data obtained from company employees. The estimates incorporate the demographic composition of the company workforce and HIV sero-prevalence data from the workforce HIV testing programme. These estimates, like any others, are no more than ‘best guesses’ about the future, and they may overstate or understate the actual evolution of the epidemic. They are based on two key assumptions about how HIV progresses, in society and in individuals: firstly, that the rate of new HIV infections (incidence) in South Africa began falling in the late 1990s and is continuing to decline, and secondly (and in contrast), that AIDS-related mortality is climbing and has not yet levelled off, with the average survival time from infection with HIV to death from an AIDS-related illness at the time of the study being nine years, in the absence of life extending treatment. The projected yearly new HIV infections in the company, shown in Table 5, demonstrate that the infection rate is expected to level off in the current decade. The projected decline in HIV incidence in South Africa since the late 1990s is largely a result of the typical dynamics of an infectious disease. Incidence represents the percentage of the susceptible (uninfected) population that becomes infected each year (i.e. the ratio of new infections to the number of uninfected people). Once infected, an individual is removed from the susceptible population. As the epidemic spreads, the population remaining are those at lowest risk of infection. Thus the incidence rate declines as the epidemic matures, though incidence may remain high among new entrants to the susceptible population, who include adolescents and new migrant employees.

Table 5: Projected new HIV infections in company employees

Year	New HIV infections (n)
2006	6.9
2007	6.2
2008	5.5
2009	4.8
2010	4.2
2011	3.7
2012	3.3
2013	2.9
2014	2.6
2015	2.3
Source: AIDS Projection Model estimates.

Table 6 shows the projected number of AIDS-related deaths among company employees. For the next seven years the company should anticipate losing an increasing number of employees per year to HIV/AIDS, in the absence of effective life-extending treatment programmes. For the study, it was assumed that the company will take a ‘business as usual’ approach to dealing with the problem, and that its operations and cost structures and the benefits it provides to employees will be constant. Salaries and all other costs were held constant at real 2006 levels.

Table 6: Projected HIV/AIDS deaths in company employees

Year	Deaths
2007	11.8
2008	14.2
2009	16.5
2010	18.5
2011	20.2
2012	21.3
2013	22.0
2014	22.3
2015	22.1
Source: AIDS Projection Model estimates.

Following the broad breakdown proposed by Rosen et al. (2000), Table 7 summarises the direct and indirect costs to the contract cleaning company for each employee identified as infected. Three points may be noted: the company will incur an average cost of R9000 for each employee who becomes infected with HIV, this cost constitutes 63.2% of an employee's average annual salary, and the bulk of the cost is accounted for by the provident fund payment of R8500. Additional data may be gleaned if we apply the AIDS Projection Model to the same baseline figures. The results of this application are presented in Table 8.

Table 7: Present value of a new HIV infection by cost component, 2006

Type of cost	Amount (R)
Direct
Provident fund	8 515
Recruitment	59
Training	148
Indirect
Sick leave	285
Total	9 007

Table 8: Aggregate present value of all new HIV infections in 2006

Year	New HIV infections	Total cost (present value) (R)
2006	6.9	58 630
2007	6.2	52 682
2008	5.5	46 734
2009	4.8	40 786
2010	4.2	35 688
2011	3.7	31 439
2012	3.3	28 040
2013	2.9	24 641
2014	2.6	22 092
2015	2.3	19 543
Note: ‘Present value’ is the amount the company should invest in an interest-bearing account now to cover the costs it will incur for infection in the future.

Table 8 shows the costs that will be incurred in the second half of this decade and beyond as a result of employees becoming infected in 2006. The cost of the infections that employees acquired before 2006 will be coming due in the first half of this decade. The pattern of costs closely follows the pattern of HIV incidence in the workforce, as was shown in Table 5. The new infection rate (incidence) is believed to be falling as the epidemic matures in South Africa. HIV incidence in the country is thought to have peaked in the late 1990s. It appears to be declining rapidly now, even though the overall infection rate (prevalence) is either stabilising or still rising, and the mortality rate is climbing.

The amounts in Table 8 can be interpreted as the amounts that the company should invest each year in a hypothetical reserve fund at a real interest rate to cover all future costs of new HIV infections acquired in that year. Alternatively, they are the amounts by which HIV/AIDS reduces the asset value of the company each year.

5. Discussion

The key finding of the cost impact assessment is that the company has not incurred and in all likelihood will not incur a large financial burden as a result of HIV infections among its workforce. Although it seems a large expense, the estimated cost of R9000 per individual is spread out over time. Moreover, the total costs of all infections actually decrease in real terms. This conclusion supports evidence from managers and supervisors at other companies who reported that they had not perceived any significant financial effects of HIV/AIDS (see Fraser et al., 2002; Chao et al., 2007; Rosen et al., 2007).

However, it must be noted that the conclusion ignores hidden costs that the study could not determine as envisaged in Rosen's conceptual breakdown (Figure 2), for example lower productivity of sick workers still present at work (‘presenteeism’). These costs could be significant. There is also the possibility that there will be more instances of workers taking time off to obtain treatment at public health facilities, which could reduce work productivity generally because, for example, supervisors might rearrange teams, resulting in greater workloads for those staff present. In addition, there could be increases in direct costs such as sick leave payments.

Yet even if this scenario is likely, one must acknowledge the evidence from the company that their labour practices minimise the threat that HIV/AIDS will increase their indirect costs. For example, the company followed general practice in the industry by having a pool of casual workers to draw on, which offset occasional absences of full-time staff. Also, the demographic profile of its workforce was a de facto means of reducing the financial costs of HIV/AIDS in that it consisted mainly of women over the age of 35, meaning that a large proportion of the employees were not in the younger age groups that are more at risk of infection (NDoH, 2008).

Note that the results of this study should be interpreted and extrapolated with caution because of the non-random (voluntary) basis of participant selection.

6. Implications

6.1 A threat to the provident fund?

The data presented here do not indicate that a significant financial threat to the company exists, given the observed level of HIV/AIDS in their workforce. This situation may or may not be applicable across the industry. However, it does indicate that there is a possible significant threat to BCCCI, the particular provident fund that was contributed to by the company and its employees. The bulk of the costs that arise from HIV infection of an individual worker are met by the provident fund contributions furnishing death, disability and retirement benefit payments to an employee and/or their estate. This cost is insignificant, however, compared to the accumulated costs borne by the fund as a financial entity. There is the threat of an increasing rate and number of payments from the fund in the period before the effects of the pandemic on the workforce taper off, which is projected to be by 2015 (see Table 8). An additional threat is that the fund might not be able to absorb this increase in expenditure if there is a marked increase in claims by members who have been employed for relatively short periods of time by companies across the industry, since the general sustainability of the fund depends on long-term contributions by members and employers.

However, the limited evidence from this study suggests that there is as yet no discernible threat to the fund. On the one hand, evidence of labour turnover at the company indicates that HIV/AIDS is not a principal cause of attrition. In the case study company, the average annual turnover of staff was 20%, with only 6% being attributed to illness or death due to HIV/AIDS. This low rate of attrition due to HIV/AIDS coincides with broader industry studies. For example, a survey of 80 small and medium-sized businesses in Gauteng and KwaZulu-Natal indicated an average rate of 13% staff attrition annually, but only 1.4% of this was HIV/AIDS-related (Connelly & Rosen, 2004). The projections undertaken in this study show that the proportion of HIV/AIDS-related illnesses is expected to increase, yet they also suggest that the rate will eventually level off. Accordingly, the question is whether, despite the decrease in the proportion, there may continue to be a small but notable burden on the fund, and whether this continuing burden could also endanger the fund's overall financial footing. This seems unlikely given the South African insurance industry's standard practice of raising contribution levels marginally in order to cover that burden (and also to reduce benefits). Furthermore, the employment practices of companies indirectly mitigate the threat. If it is, or becomes, common practice for companies to employ primarily middle-aged and older women, as in the case study presented, then the threat of HIV/AIDS to the fund should not increase. Benefit payments for other causes of death could, of course, increase.

Many companies purchase death, disability and medical benefits from outside insurance providers. The employer and the employee share the cost of the premium for these benefits. The insurance provider assumes the risk that the employee will make a claim. As more employees succumb to AIDS in the near term, claims for medical care and death and disability benefits will increase, and the cost of purchasing these benefits will inevitably increase as a result. To avoid losing money, the providers must charge enough in premiums to cover their claims and administrative costs and make a profit. If a particular employer's claims increase, it is likely that the premiums charged by the insurer will increase accordingly.

Various strategies have been implemented by employers to manage the impact of rising benefit costs in provident and pension funds as a result of HIV/AIDS (see for example Dickinson & Stevens, 2005; Stevens et al., 2005; Rosen et al., 2007). To be appropriate a strategy must be in line with the fund's objectives for benefit provision. The following are possible strategies that could be put in place:

•	Capping the contribution rate for all categories of members. This change would cause a reduction in benefits over time should the contribution level remain the same.
•	Increasing employer contributions to maintain current benefit levels. This would only be a short-term solution and might well not be affordable in the longer term.
•	Setting up a risk reserve to mitigate future cost increases and volatility. This would tie up some of the company's capital and it would have equity implications, since one generation would be subsidising another.
•	Restricting access to benefits. Such a strategy would be designed to limit or exclude access to benefits in the case of people who are HIV-positive by introducing pre-benefit testing, exclusions and waiting periods. Before any such strategy could be implemented, various legal and moral issues would have to be considered.
•	Implementing an HIV/AIDS management programme.

6.2 Managing the risk of HIV/AIDS

Yet all the strategies outlined above would manage the symptoms and effects of the epidemic and not the epidemic itself. To manage the epidemic, a holistic HIV/AIDS management programme would be needed, including full treatment and medication. In the long run this could lead not only to savings in the cost of risk benefits but also to other direct and indirect employment-related savings. The company in this study has minimised the threat of HIV/AIDS to their operations, but the costs have been defrayed onto the company's total workforce – a workforce that has very limited financial and social security. This situation is broadly similar throughout the industry and is the starting point for any consideration of where and how changes to the present situation ought to be made (SABCOHA, 2002).

The findings of this study do not provide substantive reasons for companies themselves to support the development of HIV/AIDS workplace programmes. Anecdotal reports by managers during the study supported the general view that most companies are motivated to provide HIV/AIDS services only when they see a negative financial effect on the business as a result of AIDS or became aware of employees who are sick. Surveys such as SABCOHA (2002), George (2003), Connelly & Rosen (2004) and GBC (2006) reveal little appreciation in the private sector of the benefits of proactive prevention, treatment, care and support interventions. The projected financial costs of future HIV infections in the case study company workforce demonstrate how much, hypothetically, the company should invest annually in a reserve fund to cover future costs. Equally, these are amounts that the company could invest in HIV prevention activities to reduce future costs, but the effectiveness of such an investment cannot be predicted.

Irrespective of whether the assessment of the case study company is representative of the industry, there is certainly no incentive for companies to support workplace anti-retroviral therapy programmes or even worker membership of medical aid schemes. The 2008 annual costs of a currently standard treatment regime for an AIDS patient, based on the protocol recommended by the World Health Organisation, were R3000 (first line treatment) and R6400 (second line treatment) (NDoH, 2008). These sums amount to 20 to 40% of the average cleaner's annual wage.

To demand support from companies to cover these costs is unrealistic. Direct support would inevitably pose questions about why HIV/AIDS specifically deserves investment and not other life-threatening diseases. The cost of providing health insurance would be prohibitive for workers and would substantially increase the overall labour costs for a company working in a context that is highly competitive and geared to reducing costs. Contract cleaning companies exist because of efforts by the businesses that hire cleaners from these companies to reduce not only wages but also the associated benefit costs of labour. By not employing the cleaners directly and getting contract cleaning companies to bid for the supply of these cleaners, the company that needs cleaners can avoid paying the potential associated costs of ill employees, such as time off work, health insurance and so on. It becomes the contract cleaning companies' responsibility to supply the required number of cleaners, therefore the company avoids the ‘inconvenience’ and costs associated with ill health. Ultimately it is the workers involved who pay the price because if they become ill and cannot work they do not receive any pay from the contract cleaning companies and further they need to cover their own health care expenses.

The study findings show that there is a need for intervention as the workforce is particularly vulnerable. HIV/AIDS is having an impact on the industry but the costs are largely hidden and are mainly borne by workers who are poorly paid and receive minimal welfare assistance and health insurance. The results and analysis presented here do not provide a strong business case for firms to undertake amelioration activities to deal with HIV/AIDS. However, there is a strong case for industry-level initiatives, especially with regards to managing HIV/AIDS in the workforces of small and medium-sized businesses. To enable this, both regulatory and industry agencies (for example, BCCCI, SABCOHA and the Services SETA – Sector Education and Training Authority) should facilitate an industry-wide plan to assist companies. This would involve at a minimum the development of policies and the design of programmes for prevention (condom distribution, peer education) linking with the government's national anti-retroviral treatment programme to enable their AIDS-ill workers to access treatment.