Restless worlds of work, health and migration: domestic workers in Johannesburg

Abstract

The lives of migrant women have generally received far less attention than those of their male counterparts. Similarly, male migrants have been the focus of research on the relationship between migration and HIV/AIDS. Little attention has been paid to the vulnerability of female migrants themselves to HIV infection and their access to health care and treatment. Domestic work is the second largest sector of employment for black women in South Africa, and the largest for black women in Johannesburg and, as this article shows, most of these workers are migrants. Based on a survey of 1100 domestic workers in Johannesburg, the article explores the lives of domestic workers, focusing on their experience as migrants, their working conditions, use of health-care services and knowledge of and possible vulnerability to HIV/AIDS.

1. INTRODUCTION

The lives of migrant women have generally received far less attention than their male counterparts. So, despite the long history of women's internal and cross-border migration, their stories and lives have remained largely undocumented (Dodson, 2000). Male migrants, and particularly mineworkers and truck drivers, have also been the primary focus of research on the relationship between HIV and migration (Crush et al., 2002; Lurie et al., 2003; Zuma et al., 2003; Lurie, 2004). At times their partners – usually called ‘women at risk’ – have been included in research. Other researchers have started to look at the sexual activities of women ‘left behind’ by their migrant male partners (Lurie et al., 2003), yet little attention has been paid to the vulnerability of female migrants themselves to HIV infection and their access to health care and treatment.

Johannesburg is the largest city in South Africa. In 2001 more than 3 225 000 people were counted in the Census. The city is also home to the largest number of migrants of any city in South Africa. Census 2001 found that 35.2 per cent of the population of Johannesburg were internal migrants born outside Gauteng Province, and 6.7 per cent were cross-border migrants or had been born outside South Africa. When people think of migrant workers, they usually think of male migrants, yet women have a long history of migration to Johannesburg. If place of birth is used as a marker of migrancy, Census 2001 shows that in Johannesburg women constitute a significant proportion of migrant workers in the city. For some provinces (Eastern Cape, Free State, Northwest and Western Cape) women migrants to Gauteng exceed the number of men (Figure 1).

Figure 1: Population of Johannesburg by sex and place of birth (%), 2001.

Domestic work, although often characterised as ‘atypical work’ in the service sector, provides significant opportunities for employment for black women in South Africa. In 2004, it was the second largest employment sector for South Africa's black female workforce, employing some 755 000 women (StatsSA, 2005). Census 2001 found that work in private households is the largest source of employment for black South African women in Johannesburg, with 88 000 women so employed (or 31 per cent of employed black women).

Available evidence suggests that domestic work has traditionally been, and remains, a significant area of employment for internal and cross-border female migrant workers (Miles, 1991, 1996; Cockerton, 1997). Census 2001 data obtained from Statistics South Africa shows that 42 per cent of employed black women from the Southern African Development Community (SADC) who lived in Johannesburg worked in private households, although they comprise only 4.9 per cent of women working in private households in the city (StatsSA, 2004). Census 2001 also shows that some 35.6 per cent of employed black South African women born outside Gauteng Province who lived in Johannesburg worked in private households compared to 9 per cent of employed black women who were born in Gauteng and lived in Johannesburg (StatsSA, 2004).

Many domestic workers are migrant workers. They endure poor working conditions and low incomes despite attempts by the Department of Labour to set minimum standards. Many live in isolation on their employers' properties and lack opportunities for collective action to improve their working conditions. Low incomes and arduous working conditions mean that access to health services may be limited, as time away from work may mean lost income. Domestic workers could be at increased risk of HIV infection as a result of their gender, migrancy, social isolation, poverty, low levels of education, lack of access to health-care services and lack of power at work and possibly at home (Peberdy & Dinat, 2005).

This article explores the vulnerability of migrant domestic workers to HIV/AIDS. It is based on a survey of 1100 female domestic workers working in the Johannesburg Magisterial District (JMD) undertaken in 2004 by the Southern African Migration Project (SAMP) and Baragwanath Perinatal HIV/AIDS Unit. The sample was identified using a cluster sampling technique from 94 randomly selected census enumerated areas in the JMD. The random selection of enumeration areas included only those areas with average household monthly incomes of more than R2500 and which had more than 40 households living in detached, semi-detached or town houses. Houses where interviews would take place were then randomly selected from the selected enumerator areas. On average 11.6 domestic workers were interviewed per enumerator area. Fieldworkers were instructed not to undertake the interview if the employer was present. Following the interview, all participants were provided with a list of clinics in the area, as well as HIV/AIDS specific services available in the city. An average of 16.8 per cent of domestic workers in the selected households in each area could not be contacted despite repeat visits, and field workers recorded a refusal rate of 18.3 per cent.

The JMD was chosen after examining the Labour Force Survey, which indicated that most areas lying outside the JMD but inside the boundaries of the City of Johannesburg report extremely low rates of employment of domestic workers. In effect, this meant that Soweto was excluded. However, the sample included a range of suburbs from Linksfield Ridge in the northwest of the city to Eldorado Park in the southeast. Owing to financial constraints it was not possible to employ rigorous sampling techniques in the flatlands (areas of apartment blocks) of Johannesburg, so these areas were also excluded. Therefore, all enumeration areas lying in Soweto, areas with average incomes of less than R2500 per month, areas with less than 40 households living in houses, and domestic workers employed by apartment dwellers were excluded from the survey. The survey thus provides a profile of domestic workers employed in houses in Johannesburg in areas with average incomes of more than R2500. It is possible, therefore, that there may be an over-representation of women who live in and work full-time for a single employer.

2. JOBS FOR MIGRANT WOMEN

Domestic work provides significant employment opportunities for black female migrant and non-migrant women in Johannesburg and, as noted above, Census 2001 shows that work in private households provided employment for more than 88 000 black women in the city. The census also shows that domestic service was the largest sector of employment for black South African women who had moved to Johannesburg from other provinces and other countries.

One of the defining features of participants in this study was that the overwhelming majority were migrant workers. So, although more than 50 per cent of participants called the Johannesburg area home, some 86 per cent said they had another home somewhere else (Table 1). Participants with homes elsewhere showed strong ties to their other homes. Of those with other homes, 72 per cent said they would rather be living there if the same job and working conditions were available (Table 1). Although less than 10 per cent visited their other home more than once a month, nearly 90 per cent visited at least once a year (Table 2). Opportunities to visit their other home may be constrained by cost, distance and time.

Table 1: Domestic workers as migrants (%)

	Total non-South Africans	South African migrants	Total sample
Call the Johannesburg area home (N = 1100)	51.5	44.3	52.2
Have a home somewhere else as well (N = 1100	98.5	100.0	86.1
Would move home if could have same job there (N = 945)	58.2	72.8	71.9

Table 2: Frequency of visits home (%)

	Total non-South Africans	South African migrants	Total sample
More than once a month	1.5	7.2	6.8
Once a month	0.0	17.7	16.5
Once every few months	7.5	26.1	24.7
Once or twice a year	67.2	39.4	41.4
Less than once or twice a year	6.0	6.4	6.3
I have been just once	4.5	0.9	1.2
I have not been home yet but I would like to	7.5	1.6	2.0
Never – I cannot return home	4.5	0.2	0.5
Never – I have no desire to return home	0.0	0.3	0.3
N = 945.

It is sometimes assumed that cross-border migrants from SADC countries constitute a significant proportion of domestic workers in the city. However, this study, together with Census 2001, suggests that the overwhelming majority of migrant domestic workers are internal migrants (Table 3). Only 68 (6 per cent) of the women interviewed were not South African citizens. When asked about the location of their other home, 6 per cent said it was in another country. Of those who had homes in other countries all were from SADC countries. Almost 50 per cent of foreign domestic workers in the study came from Lesotho, almost a third from Zimbabwe and the rest were from Mozambique (6 per cent), Botswana (4 per cent), Swaziland (4 per cent), Malawi (4 per cent) and Zambia (2 per cent).

Table 3: Place of birth and place of other home (%)

Province	Province of other home (survey)	Place of birth of black women in Johannesburg (Census 2001)
Gauteng	16.3	22.5
Eastern Cape	12.2	10.5
Free State	12.1	9.4
KwaZulu-Natal	10.9	10.9
Limpopo	10.5	15.9
Mpumalanga	6.2	5.2
Northern Cape	3.2	1.5
North West	20.4	17.9
Western Cape	1.1	1.3
Other country	6.0	5.0

The majority of migrant domestic workers had grown up in rural areas, particularly those from other countries. However, many had lived in an urban area for some time. More than three-quarters had lived in Johannesburg for five years or more (Table 4). Unemployment seems to have been a significant motivator for moving to Johannesburg. Almost 70 per cent had been unemployed before they left home for Johannesburg (Table 5).

Table 4: Length of time in Johannesburg (%)

	Total non-South Africans	South African migrants
Less than one year	5.9	4.9
1 year	8.8	2.7
2 years	4.4	5.5
3 years	8.8	4.8
4 years	5.9	4.2
5 years	7.4	8.0
More than 5 years	57.4	69.9
N = 943.

Table 5: Employment status prior to coming to Johannesburg (%)

	Total non-South Africans	South African migrants
Employed	19.4	23.6
Unemployed	70.2	68.7
Self-employed	3.0	0.8
Student/scholar	4.5	6.3
Other	1.5	0.3
N = 943.

The majority of non-South African migrant workers seem to retain relatively strong ties with their home countries, as 88 per cent regularly send money and goods to their home country. Not surprisingly, the main reason they came to South Africa was to find a job. The choice of South Africa may have been influenced by family experience and contacts, as more than half said their parents had worked in South Africa and a third said their grandparents had.

3. DOMESTIC WORKERS IN PROFILE

Who are the women who clean the homes of the city and ensure that families leave their homes pressed and tidy? The majority are of an age at which people have already established or are establishing long-term relationships and are developing families. Respondents were mostly aged between 21 and 50 years (Table 6). The overwhelming majority (almost 60 per cent) were aged between 31 and 50. Five of the workers interviewed were under 20, while 35 were still working when they were aged more than 60 years. Non-migrants tended to be older than migrants. Some 64 per cent of the former were more than 40, compared with 38 per cent of foreigners and 49 per cent of internal migrants.

Table 6: Age (%)

	Total non-South Africans	South African migrants	South African non-migrants	Total South Africans	Total sample
15–20 years	0.0	0.3	1.3	0.5	0.5
21–30	25.0	20.5	13.3	19.4	19.8
31–40	36.8	30.4	21.2	26.1	29.6
41–50	22.0	29.3	39.1	30.7	30.2
51–60	10.3	16.4	21.5	17.2	16.6
Over 60	5.9	3.0	3.3	3.0	3.2
N = 1100.

The women live solitary lives marked by separation from family members. Although 58 per cent had a husband or partner, more than half were temporarily living apart from them. So, only 27 per cent of the women interviewed were living with a spouse or partner. Some 26 per cent were single and 16 per cent divorced, separated or widowed (Table 7). A third of the married women and 60 per cent of those with partners were living on their own.

Table 7: Marital status (%)

	Non-South Africans	South African migrants	South African non-migrants	Total South Africans	Total sample
Single	27.9	24.8	29.1	25.5	25.8
Divorced/separated	7.4	7.0	7.3	7.0	7.0
Widowed	8.8	8.3	13.9	9.1	9.1
Married but temporarily living apart	5.9	6.2	3.3	5.7	5.8
Married and living together	11.8	9.6	15.9	10.5	10.5
Living with partner	14.7	17.1	16.6	17.0	16.8
Partner but temporarily living apart	23.5	27.0	13.9	25.1	24.9
N = 1100.

Many women are forced to live apart from their partners and spouses in their working and living arrangements. Almost 45 per cent were not allowed to have a partner stay with them where they were living, mainly because of restrictions imposed by employers. Some 17 per cent of non-migrants were temporarily living apart from their partners, which suggests that these partners are migrant workers. The majority of the 378 women who were temporarily living apart from their partners and spouses saw them fairly frequently. Some 64 per cent saw them more than once a month, and 14 per cent once a month. Some 9 per cent saw them only once every few months, and 8 per cent once or twice a year. And 4 per cent said they saw their partner less than twice a year, while four women said they never wanted to see their partners again.

Almost 70 per cent of the women had children, but were likely to live apart from them. Non-migrants were more likely to be childless (42 per cent compared to 28 per cent of migrants), but were more likely to live with children (52 per cent compared to 13 per cent of migrants). Overall, less than 20 per cent of the women interviewed lived with their children. Separated children were most likely to be in the care of their grandmothers. Almost a third said their children were not allowed to stay with them where they lived, because of restrictions imposed by employers. However, although many domestic workers live alone, separated from their partners and children, this does not mean that they do not have family responsibilities. Almost 95 per cent were financially supporting other people in full or in part. The majority of dependents were their own children. Other financial dependents included parents, siblings, grandchildren and nieces and nephews and their partners.

Domestic work is often considered to be low-skilled and new entrants to the sector do not usually have to meet educational entry requirements. Certainly, this survey confirms that domestic work provides employment for women with low levels of education. Almost one-third had had no schooling or only some primary education. A further 16 per cent had completed primary school, while more than 40 per cent had been to secondary school. Domestic workers are generally less educated than other women in the city. Literacy levels are important, as they not only affect employment opportunities but also may affect access to health information and ability to deal with employers and bureaucracies.

In sum, this demographic profile of domestic workers in Johannesburg indicates that many live on their own, and even those with partners and children are likely to live apart from them. Their lives are shaped, at least in part, by their profession as well as their migrant status, since where they live prevents them from living with their partners and children. Their choice of profession may be shaped by their low level of educational attainment which may restrict their job opportunities and their access to information.

4. WORLDS OF WORK

The South African Government has recognised that the working conditions and incomes of many domestic workers are poor. To that end it has made efforts to improve and formalise these (Hardy & Kleinsmidt, 2004). At the time of the study, the official minimum wage for domestic workers employed in urban areas and working more than 27 hours per week was R4.87 per hour, or R861,90 per month. For those working 27 hours or less per week the minimum wage was R4.87 per hour or R569.79 per month. Maximum working hours have been set by the Department of Labour at 45 hours per week (nine hours per day for those working one to five days per week and eight hours a day for those working six to seven days per week) plus ten hours of overtime. As of October 2003, employers of domestic workers have had to make contributions to the Unemployment Insurance Fund for their employees.

4.1 Working conditions

However, these minimum standards set by the Department of Labour do not guarantee that the working conditions of domestic workers are easy. Certainly, the women interviewed for this study lived hard lives, and worked long hours for low pay, and their responses indicate that many employers are not meeting the statutory minimum standards of employment. The majority worked for one employer only (88 per cent) and lived at the place where they worked (64 per cent). These responses may reflect the parameters of the sample, which excluded women working in apartments.

Domestic workers' employment seems to be stable, or at least long term. More than 40 per cent of the women interviewed had been employed by their main employer for more than five years (Table 8). Non-South African migrants seem to have less stable employment or were newer entrants to this sector of the labour market, as more than 30 per cent had been employed by their main employer for less than a year.

Table 8: Length of time working for main employer (%)

	Total non-South Africans	South African migrants	South African non-migrants	Total South Africans	Total sample
Less than 6 months	17.7	13.4	10.6	13.0	13.4
More than 6 months and less than a year	14.7	8.4	15.2	9.4	9.7
1–3 years	30.9	26.9	23.8	26.4	26.7
4–5 years	4.4	9.7	12.6	10.1	9.8
Over 5 years	30.9	41.6	37.8	41.0	40.4
N = 1100.

The domestic worker's week tends to be long. On average, respondents worked 5.4 days per week (Table 9). More than 30 per cent worked five days per week, more than 20 per cent worked a six-day week, while almost 20 per cent worked seven days per week. The data suggests that migrant workers worked the longest weeks and that non-migrant workers were most likely to work a five-day week.

Table 9: Number of days worked (%)

	Total non-South Africans	South African migrants	South African non-migrants	Total South Africans	Total sample
1 day per week	10.4	7.9	9.9	8.2	2.8
2 days per week	5.2	9.4	10.5	9.6	5.9
3 days per week	7.8	7.8	10.5	8.2	7.9
4 days per week	5.2	2.8	6.4	3.3	3.7
5 days per week	32.3	29.5	34.3	30.2	34.7
6 days per week	16.9	23.7	17.4	22.8	25.9
7 days per week	18.2	17.2	7.6	15.7	18.5
Half a day per week	1.3	0.2	2.3	0.5	0.5
Other	0.0	0.1	0.6	0.2	0.1
N = 1100.

Domestic workers also work long days. Some 46 per cent worked nine hours or more per day and 31 per cent worked 10 hours or more per day. Some, it seems, never go off duty. Migrant workers, whether South African or foreign, are likely to work the longest days. A significant proportion of the employers were thus exceeding the maximum working hours set by the Department of Labour. People who work long weeks and days for little pay can find it hard to access health care, particularly if taking time off to attend a clinic results in a loss of pay.

4.2 Incomes and expenditure

A significant proportion of domestic workers appear to earn less than the minimum wage for urban areas as set out by the Department of Labour. As the majority of women surveyed worked more than 27 hours a week they should have been earning more than R860 per month. More than 20 per cent of respondents earned less than R500 per month, and just over 55 per cent made between R501 and R1000 per month (Table 10) (StatsSA, 2005). The September 2004 Labour Force Survey found that nationally (including rural areas), 41.2 per cent of domestic workers earned between R1 and R500, 40.9 per cent between R501 and R1000 and 13.2 per cent between R1001 and R2500 (StatsSA, 2005). One elderly woman who lived at her employer's property said she earned nothing, but was provided with food and accommodation. Almost 25 per cent received some food as part of their pay, and 61 per cent had free accommodation on their employer's property. Only 5 per cent of the women had another source of income, which on average brought them a further R240 per month. Few who said they had children of eligible age said they received the child income grant.

Table 10: Monthly income (%)

	Non-South Africans	South African migrants	South African non-migrants	Total South Africans	Total sample
Less than R100	0.00	0.3	1.3	0.5	0.5
R100–500	30.9	19.9	22.5	20.2	21.2
R501–1000	47.1	56.6	51.7	55.9	55.7
R1001–1500	14.7	18.0	19.9	18.3	18.1
R1501–2000	7.4	3.4	4.6	3.6	3.8
R2001–2500	0.0	0.6	0.0	0.5	0.5
R2501–3000	0.0	0.1	0.0	0.1	0.1
Not answered/don't know	0.0	0.0	0.9	0.0	0.7
N = 1100.

4.3 Accommodation

Almost two-thirds (64 per cent) of the workers received accommodation with their jobs but few paid rent to their employers. The remaining 36 per cent who had to pay for their own accommodation in Johannesburg, paid an average of R152 per month. Those who lived away from their place of employment paid an average of R175 per month for transport to work.

If the working week is long and hard for the majority of these women, what do they go home to at the end of a long day of cleaning up after other people? The majority go nowhere but stay at their place of employment in accommodation provided by their employer (64 per cent). Not surprisingly, migrant workers, but particularly South Africans (69 per cent), were most likely to live on their employer's property. The slightly lower proportion of non-South African migrants living in (62 per cent) is probably because they were more likely to have been employed for a shorter time and to work part-time. Only 36 per cent of non-migrants lived-in.

While the majority lived on their employer's property, the remainder were most likely to live in shacks or rent a room. Therefore, most of the domestic workers who left their place of employment at the end of the day went home to a shack (16 per cent). Another 8 per cent went home to a room. Overall, more than two-thirds of respondents lived in just one room, and another 11.5 per cent lived in two rooms. Some of those who live in places with more rooms were living inside their employers' houses. However, even if a domestic worker lives on her employer's property this does not necessarily guarantee good and healthy living conditions. Almost 37 per cent of the women interviewed had no access to a bathroom with running water, and only just over half had access to an inside tap where they lived.

These women clearly work hard for a living, toiling through long days. Many never leave their place of employment at the end of the day, and wherever they live conditions are not good. Their living and working conditions may not increase their vulnerability to HIV infection but, if they are infected, have the potential to compromise their health.

5. WORLDS OF LEISURE: A LONELY LIFE?

Many domestic workers live relatively solitary lives and are often isolated, even though they live in South Africa's largest city. This isolation, in part, reflects the location of their homes and workplaces in the suburbs of Johannesburg and the length of their working weeks and days, but it also reflects their migrant status.

Of the women interviewed, almost 30 per cent of migrants and almost a quarter of non-migrants did not have any friends near where they worked. Despite their separation from family and friends, more than half said they never felt lonely and only 16 per cent felt lonely often or most of the time. The loneliest times for domestic workers were the evenings and weekends. Again, this may be because their living arrangements precluded them having friends (and, of course, partners and children) to visit. Almost half were not allowed to have visitors where they lived. Friends were primarily fellow domestic workers, other neighbours, and relatives and friends from home. Sadly, and perhaps expressing the isolation of some of these workers, eight said their employer was their best friend. Finding friends in their neighbourhood meant that almost two-thirds said they saw their most important friend three or more times a week.

The social lives of domestic workers, like their working lives, are relatively constricted, at least for those who live in. Most of their social activity (whether migrant or non-migrant) took place in homes or at church, and very few regularly visited a bar or a shebeen. Despite the clusters of domestic workers often seen sitting on the grass verges of Johannesburg's northern suburbs, the street played only a small role in their social lives.

The main leisure activities for domestic workers when not working were watching television and listening to the radio (49 per cent), their two main sources of information. A further 14 per cent spent time alone. Some 12 per cent spent time with friends, and only 3 per cent (or 33 women) said they spent time with a male friend. Domestic workers do not seem to be big gamblers in their leisure time, with only 8 per cent reporting that they had played fafee (a local gambling game based on predicting dream imagery). However, dreams of big wins are not far away, as almost half had played the Lotto (the South African national lottery) in the previous three months.

6. WORLDS OF HEALTH

In South Africa, health care is available to South Africans through state or private health-care services. Citizens attending state facilities are required to pay fees and pay for medicines unless they are able to prove that they cannot. Non-citizens, if identified as such, may be required to pay higher fees and deposits for services provided by the state. The policy of charging foreigners different rates varies from facility to facility. Medical aid or health insurance will pay for health-care treatment (if required, and according to the plan paid for) at private health facilities. Domestic workers, because of the long days they work and their working conditions, isolated lives and low pay, may have particular health problems, and also problems accessing health care.

6.1 Health status

Just over a fifth of respondents had taken a day or more off work in the previous three months because of ill-health. Non-migrants were more likely than migrants to have taken a day off work. Their illnesses may have prompted visits to the doctor, as overall 20 per cent had visited a clinic or doctor in the previous three months. Almost a hundred women (9 per cent) had been admitted to hospital in the previous year.

Participants reported that they had been told by a doctor or nurse that they currently had specific medical conditions. Some 12 per cent (133 women) reported that they had been diagnosed with a sexually transmitted disease at some point in their lives. However, overall, their most significant health problems were related to work (joint, back and limb problems). More than 20 per cent reported high blood pressure problems and more than 5 per cent that they had been diagnosed with heart problems. South African non-migrant women were most likely to have reported heart and blood pressure problems.

Domestic workers generally do not compromise their health through smoking or find solace in alcohol or drugs. Snuff seems to be the tobacco of choice, and less than 8 per cent said they currently smoked cigarettes. The 130 women who had drunk alcohol in the four weeks prior to the interview were asked how often they had had a drink. For most, having a drink was a weekend leisure activity. Some 60 per cent of those who had used alcohol drank every weekend. Another third were occasional drinkers, having drunk alcohol on less than three days in the previous month. Only one respondent said she drank alcohol every day. Three others had a drink nearly every day. These four women were all long-term Johannesburg residents.

6.2 Use of health-care services

Usage of health-care services suggests that, overall, these relatively healthy women can find health services if they need them. The majority choose to use allopathic (i.e. standard Western) health services for their health problems. Almost a third had been to see a doctor in the year prior to their interview. Almost 50 per cent had been to a clinic and 15 per cent had been to a hospital outpatients' department. Family planning services and clinics were the venues of choice for getting help with sexual or reproductive health issues. Almost 30 per cent of the sample had attended a family planning service in the six months prior to the interview, although less than 4 per cent had had a baby in the previous year. Only 15 per cent had visited a traditional healer, and only 48 per cent of these women had done this for health reasons. Almost half of those who had visited traditional healers for health reasons had also used allopathic services for the same problem.

What is the cost of using health services and who usually pays? Only 13 of the women interviewed were members of medical aid schemes, which would enable payment for private medical services, so most would have to pay for private health services themselves. Very few employers (15) helped pay for medical treatment. Payments by domestic workers for health services suggest that state services are the most affordable, and private doctors the least. Those who saw private doctors either had medical aid or were likely to get assistance from their employers. The low usage of traditional healers may have something to do with cost. Respondents reported the cost often exceeded R100 and three women had paid traditional healers between R1000 and R3000.

6.3 Contraceptives and condoms

As noted above, more than 70 per cent of respondents had children. More than 90 per cent of these women had received antenatal care from a doctor or clinic for their last pregnancy. A significant proportion had also attended family planning services in the six months prior to interview and/or had been to a clinic or doctor for sexual or reproductive health issues; yet the majority said they are not using contraceptives, and do not seem to be receiving or following any advice about using condoms. While only 11 per cent of respondents wanted a child at the time of interview, less than 40 per cent were using anything to delay or avoid pregnancy. Only 12 per cent of the women used condoms for contraception (Parker et al., 1998).

It could be that many of these women may not need to use contraceptives or to protect themselves from sexually transmitted infections (STIs) by using condoms because they do not have active sex lives. Many live apart from their partners, seeing them only irregularly, and more than 40 per cent were single (although some of these women did say they had boyfriends). Respondents were asked how many sexual partners they had had in the past five years and on average they had slept with just 1.7 men. Thirty per cent had had two or three sexual partners. In addition, 24 had had between six and ten sexual partners and two women had had 17 or more partners. Overall, just over half (52 per cent) were in an ongoing sexual relationship with their main partner.

Although the majority of the domestic workers appear to have only one partner, only a quarter with long-term partners or husbands actually live with them. Women who live with partners may still be at risk for HIV infection or infection with other sexually transmitted infections (STIs) if they or their partners are unfaithful and do not have safe sex with other partners. Those who live separately from their partners may be at risk if either they or their partners are unfaithful and have unsafe sex with their other partners (Lurie, 2004).

South Africa has very high rates of sexual violence and domestic abuse. Studies have indicated an association between sexual violence and HIV (UN, 2004). The relationships that women have may be violent, and more than 80 per cent said that they had been pushed, shoved, slapped and/or had things thrown at them. Some 18 per cent of all respondents had been assaulted in the previous year. Also, although just over half of those said it had happened only once, almost 30 per cent said it had happened a few times, and 18 per cent said it had happened many times. Three of the women had been assaulted by their employer (all South Africans). Fewer women reported that anyone had ever threatened to use a gun, knife or other weapon against them (10 per cent). It is unclear whether these threats came from partners or other people.

Not altogether surprisingly in the South African context, almost 6 per cent of participants, or 64 women, said they had been raped. Six per cent, or 66 women, had been forced to have sex with their current boyfriend, husband or other partner because they were afraid of what he might do. One woman said she had been sexually assaulted by her employer. These figures are consistent with national data, indicating that domestic workers are not at greater risk of violence than other women in the general population but nor are they more protected (Jewkes et al., 2001).

Perhaps reflecting their less restricted lives, South African non-migrant women were more likely than migrant women to have been assaulted in the previous year (26 per cent). They were also more likely to have been raped (13 per cent) or forced to have sex by a partner (9 per cent). Levels of violence reported by women in long-term relationships suggest that it may be difficult for women to negotiate condom use in their relationships (Jewkes et al., 2001). Given that the majority of domestic workers live apart from their partners, condom use may be particularly important to protect them from infection. Disturbingly, their use of condoms in sexual relationships was low, and lower than among young women in South Africa (Parker et al., 1998).

More than 60 per cent of the women had never used a condom in their lives. Of those who had used condoms, almost 30 per cent had never used a condom with a new partner. Of those who had used a condom in the past, only 71 per cent had used one the last time they had sex. Only 65 per cent had used one the time before that; therefore, it seems that condom use is somewhat haphazard. Only 20 per cent of these domestic workers who used condoms used them all the time. Others said that they used condoms with some men and not others, while some said that they start with condoms and then stop using them.

Some of those who had used condoms but who were not using them regularly had only ever had one partner (27 per cent of condom users). While monogamy can be seen as protection from HIV infection, it may be less effective if the partner is not faithful. Living separately from their partners may also encourage sex with multiple partners (Lurie et al., 2003). Less than a quarter of respondents did anything to protect themselves against contracting an STI. Of the 250 women who did something, the majority said they used a condom (56 per cent), and some had only one sexual partner or did not ‘sleep around’ (20 per cent). Others said they abstained from sex (21 per cent). A few used traditional medicine (Table 11). Only 12 per cent of the women interviewed did not know where to get condoms.

Table 11: Protection from STIs (% of those using protection)

	Total non-South Africans	South African migrants	South African non-migrants	Total South Africans	Total sample
Use a condom	70.0	55.1	37.0	53.2	59.8
One sexual partner/boyfriend only	15.0	16.5	14.8	16.4	13.5
Do not sleep around	0.0	1.7	11.1	2.7	2.0
Abstain from sex/no sexual partner	15.0	19.5	29.6	20.5	20.3
Use traditional medicine (drink muti)	0.0	1.3	3.7	1.5	1.6
Healthy lifestyle/take care of myself	0.0	0.4	0.0	0.4	0.4
Avoid high-risk areas	0.0	0.4	0.0	0.4	0.4
I had an AIDS test done	0.0	0.4	0.0	0.4	0.4
N = 251.

6.4 Knowledge of HIV/AIDS

Low use of condoms for protection can only partially be explained by abstinence, monogamy, lack of access to condoms, or abusive or unbalanced relationships. Although the majority of the women interviewed spent most of their leisure time watching television and listening to the radio, they did not seem to be taking on board HIV education and prevention messages and information available through the media. Furthermore, a significant proportion had given birth to children in the past and used family planning services in the previous six months, while others had attended other health services. In theory, they should have had relatively good knowledge of at least some HIV/AIDS issues from health-care providers.

Participants were asked if they had heard of a variety of issues relating to HIV/AIDS and whether they could explain what they were. More than 30 per cent could not explain or describe how to have sex safely. If people do not understand the importance of safe sex, or how to have sex safely, they are unlikely to be able to protect themselves or ask their partners to wear a condom. Although more than a quarter of South Africa's adult population is infected by HIV, and women (because of childbirth, antenatal care and use of family planning facilities) are likely to have come into contact with health-care workers and may even have been tested for HIV, only 45 per cent of respondents knew about mother-to-child transmission.

The domestic workers also showed relatively low knowledge about HIV treatment issues. Only 16 per cent could explain what antiretroviral treatment is and only 20 per cent could explain treatment for opportunistic infections. Their lack of knowledge about issues related to treatment is surprising, as a significant proportion of participants said HIV/AIDS had touched their lives in intimate ways. Furthermore, the survey took place at a time when there was considerable debate in the media about treatment for AIDS and the roll-out of antiretroviral therapy.

While there have been a number of campaigns around HIV/AIDS education, prevention and treatment and these issues have been integrated into popular locally produced soap operas and dramas, it seems that this group, who watch television and listen to the radio, are not being reached by these campaigns. As some commentators have noted, approaches which focus on behavioural change are often inadequate as they ignore the context within which people live and have to negotiate behaviour change. They suggest that it is necessary to take into account issues of power, poverty and gender relations in people's social and working lives, so as to better understand how to enable people to act on information received, or even how to enable them to receive the information in the first place (Parker et al., 1998).

7. CONCLUSION

How do migrancy, work, health and HIV/AIDS intersect in the lives of domestic workers living in Johannesburg? And what are the implications of this study for domestic workers and health service providers?

First, migrancy is a defining feature of the lives of the majority of the domestic workers surveyed in this study. It shapes their lives and relationships in many ways. Many, although they have lived in Johannesburg for a long time and see it as a home, are constantly looking to a home elsewhere where their children, and sometimes their partners, live. More than 40 per cent of the sample described themselves as single, widowed, divorced or separated (although this did not prevent them from having children and boyfriends). Further research is needed to understand whether being single provides an incentive for women to migrate for work.

As migrant workers, many live in accommodation provided by their employers. Restrictions imposed by employers prevent them being joined by their children, partners and boyfriends. Only a quarter of the women interviewed live with their partner. While living with a partner does not prevent people from having multiple relationships, it does reduce the likelihood.

Secondly, working conditions are hard. The women work long days and long working weeks for low pay. Despite attempts by the government to improve working conditions for domestic workers, many employers pay under the minimum wage and require their employees to work more than the maximum working hours. Most domestic workers only have access to one room, and the majority of those who live off their employer's property are living in shacks. Only a third have access to a bathroom with running water and almost half have to go outside to fetch water. While none of this necessarily increases their vulnerability to HIV infection, low incomes and poor living conditions do have implications for people who are living with the virus as these may compromise their health status.

Working and living conditions, particularly for those who live in, appear to provide some protection to domestic workers as they limit their opportunities to socialise with friends and meet new partners. It seems that many live restricted lives, as most of their friends are likely to be the neighbouring domestic workers they meet at their homes or in the street. Church provides another significant opportunity for social interaction. Otherwise, the majority of these women workers spend most of their few leisure hours alone watching television and listening to the radio.

Thirdly, the majority of these women, whatever their migrant and national status, do not have problems accessing health services. The majority use allopathic health services provided by the state. Traditional healers are used by only by a few, and almost half of those who use a traditional healer for a health problem also go to allopathic services for the same problem. They seem to have only minor health problems particular to their work, and perhaps their isolation in the workplace reduces their chance of contracting infectious illnesses such as colds and flu.

Fourthly, despite their use of health services, the majority of these women do not appear to be protecting themselves from HIV infection. A defining characteristic of this group is the lack of condom use. More than 60 per cent of the sample had never used a condom in their lives. Also disturbing is that the majority of those who used condoms used them irregularly, with only a fifth of condom users saying they used condoms all the time when they had sex. Low levels of condom use may also reflect experiences of violence in relationships.

Fifthly, health promotion campaigns do not seem to be reaching these women workers, or it may be that they are just not listening. Low levels of condom use may be related to low levels of knowledge about HIV/AIDS issues, including safe sex and inaccurate perceptions of their own personal risk of HIV infection. Almost a third were unable to describe how to have safe sex and only 11 per cent thought they might have been infected. Levels of knowledge about treatment issues were also low and only 16 per cent knew about antiretroviral therapy. Low levels of knowledge about this therapy, and the fact that it had yet to be introduced in public health services in the Johannesburg area, may have affected attitudes to testing. Less than a third had been tested for HIV and only 26 women in the sample had tested positive. Further research is necessary to find out whether the effectiveness of campaigns which encourage voluntary testing and counselling will be increased with the roll-out of antiretroviral treatment.

Sixthly, low levels of knowledge and condom use are inconsistent with the experiences of these women with the virus. Many have been touched by HIV/AIDS in their lives. More than a third knew someone who had died of AIDS; a similar proportion said a member of their family was HIV-positive or had died of AIDS; and almost a fifth had physically cared for or supported someone with AIDS. Probably reflecting their closer connections to their communities and wider social networks, non-migrant women were most likely to know someone with HIV/AIDS and to have cared for someone with AIDS.

Overall, it seems that migrancy and work shape these women's lives and affect their vulnerability to HIV. It seems that for many, particularly those who live in on their employer's property, their social lives are restricted by their working and living conditions. This social isolation may actually protect these domestic workers as it reduces opportunities for starting new relationships. Unlike many other migrant workers, it seems that the live-in status of many migrant domestic workers, and their working conditions, may mean that their chance of becoming infected by the virus could be lower than that of their non-migrant counterparts. However, this does not mean that they are not vulnerable. Conversely, their migrant status, separation from partners and, for many, restrictions on when and where they can see their partners and boyfriends may make them more vulnerable.

The low levels of condom use, given the circumstances of their relationships and low levels of knowledge about HIV/AIDS, are of concern. The majority of these women rely on television and radio for information, and the majority attend health services at some point during the year. Therefore, it seems that these women workers in Johannesburg are not being reached by health promotion campaigns relating to HIV/AIDS education, prevention and treatment. It may be that their isolation, socio-economic circumstances and lack of power in their working lives affect their ability and willingness to act on the information they have.

Notes

¹Respectively, Head of Palliative Care, Palliative Medicine, Department of Medicine, University of the Witwatersrand; and Project Manager, Southern African Migration Project, and Visiting Research Fellow, School of Public Development & Management, University of the Witwatersrand. The authors would like to thank the domestic workers who took time out of their busy working days to participate in this research, and to acknowledge the financial support from DFID (British Department for International Development) and CIDA (Canadian International Development Agency) through the Southern African Migration Project (SAMP) that made the study possible.