drugs, sex work, and HIV in three South African cities
Ted Leggett, J.D., M.Soc.Sci.
School of Development Studies
University of Natal, Durban
South Africa

Introduction
This study was designed to test the hypothesis that a positive association exists between HIV infection and drug use among commercial sex workers in South Africa. Since heterosexually transmitted HIV has become one of the most serious threats to development in the country, it is crucial to define any impact that the growing illicit markets for sex and drugs may have on HIV prevalence. Prior to this study, the connection between these phenomena in South Africa had never been assessed.

Due to the lack of available data in this area and the urgency of the situation, a small, low-cost assessment was conducted. Using a short questionnaire, a total of 349 female street sex workers were interviewed in the three largest cities in South Africa-Johannesburg, Cape Town, and Durban. The saliva of 249 of these women was submitted to GAC Elisa testing.

Contrary to expectations, an inverse relationship was found between hard drug use and HIV serostatus. While drug-using sex workers did have higher client volumes, they had lower levels of HIV than nonusers did. This may be due to the fact that most of the drug users were white, while HIV is most severe in the black community in South Africa.

A Brief History of Drugs in South Africa
Prior to 1994, many of the street drugs most common internationally were not readily available in South Africa. In particular, cocaine and heroin use was not widespread. In their place, certain drugs were abused in South Africa that saw very little popularity overseas. Mandrax and Wellconal were the two hard drugs commonly used or abused by sex workers in the late-apartheid era.

Mandrax is sold in tablet form but is crushed and smoked in a combination known as "white pipe." It provides a short-term "rush" during which users lose control of their bodies (including muscular collapse and drooling), followed by a prolonged feeling of intoxication. Wellconal (known as "pinks" on the street, for the pink color of the tablets) is crushed, dissolved in water, and injected. It also provides a rush, in addition to opiate effects. Both drugs are considered addictive, but have a natural "saturation point" in any given period of time.

Neither of these drugs is particularly well-suited to use during sex work. While Mandrax has a disinhibitory effect, it impedes coordination and judgment in much the same way as alcohol does. Wellconal, like other opiates, makes human contact uncomfortable. Despite the fact that it comes in pharmaceutically regulated doses, "pinks" tend to kill off users for no apparent reason, perhaps due to the fact that they are injected with unsterile water. As a result, a strong stigma exists against injection drug use in South Africa.

After the borders opened following the first democratic elections in 1994, crack cocaine was introduced to the sex work industry. The first police seizure of crack occurred in Johannesburg in 1995. The drug was not seen in Durban until 1996. Crack immediately began to displace Mandrax as the preferred hard drug of choice among inner-city sex workers in all three of the largest cities. Mandrax is still used by many users to "come down" following a crack binge, or sometimes in combination with crack.

Heroin has been slower to gain popularity, perhaps partly due to the stigma against injection. Indeed, in the areas where it has emerged, use has largely been limited to smoking in combination with cannabis or the "chasing the dragon" technique of vaporizing the drug on aluminum foil. Since the same syndicates control the markets for both crack and heroin, it may be that dealers are not pushing heroin as strongly as crack due to its lower profitability. Heroin has become most popular among white youth and among white sex workers in Johannesburg. It is marketed to sex workers as a come-down drug for crack cocaine.

It should also be noted that South Africa is a major producer of high-quality cannabis, which is in common use among most ethnic groups and social classes. Total alcohol consumption is moderate by international standards, but this is due to a large number of teetotalers among the black majority. Alcohol abuse is common and consistently heads the list of problem substances seen in rehabilitation centers.

Drugs and HIV
From the above, it should be clear that South Africa does not have a major injection drug use problem at present, although the potential exists for the rapid spread of such an epidemic. However, it does have a large and growing problem with crack cocaine, especially among sex workers. Research from another country with both HIV and crack problems-the United States-indicates that crack use may be as closely associated with HIV as injection drug use due to its association with high-risk sexual behaviors (Ross et al. 1998). There is also conflicting evidence about the effects of intoxication from drugs and alcohol during sex work on the ability of sex workers to negotiate safer sex practices, with some studies finding an impact and some not.

It was hypothesized that the association between crack and HIV would hold true in South Africa as well. Previous research (Leggett 1999) suggested that crack users are compelled to maintain higher client volumes to ensure adequate income, increasing the chances of exposure to HIV. Additionally, the compulsion of addiction might lead to a willingness to engage in unsafe sexual practices in exchange for more money, and anecdotal evidence from street sex workers interviewed in the past indicates that this is indeed the case. It was also suggested that addicted sex workers are less likely to bother to clean themselves between clients and generally suffered from diminished hygiene, which could have an impact on HIV transmission. Crack users were also viewed as more reckless in their client choice and more likely to be raped, thus having a greater chance of suffering violent intercourse that could lead to blood-to-blood contact.

Since many of the men who patronize crack-using sex workers also use the drug, the effects on their sexuality are also relevant. Sex workers interviewed stated that crack made men more violent. It led to impotence, with the resulting frustration sometimes ending in violence, including rape, once an erection was achieved. It was also suggested that men on crack found it difficult to achieve orgasm, which meant sexual sessions could be prolonged, leading to increased wear and tear on the sex worker.

Some of these arguments would also hold true for Mandrax, heroin, and pinks, given that these are all strongly addictive drugs. It was also hypothesized that other intoxicants, if used during sex work, could cloud judgement and reduce self-control. Additionally, although all data prior to the study suggested that intravenous drug use was not widespread among sex workers, if the contrary were to be found then it could clearly have an impact on HIV levels.

HIV in South Africa
It is generally agreed that HIV is primarily heterosexually transmitted in South Africa, and that South Africa has one of the fastest growing HIV epidemics in the world. As Figure 1 demonstrates, antenatal testing shows that HIV prevalence is highly regionalized. On the low end, the Western Cape (where Cape Town is located) displays antenatal rates of about 8 percent. KwaZulu-Natal, where Durban is located, has rates of nearly 33 percent. The third site, Johannesburg, is in Gauteng, where the rate is 24 percent.

Figure 1: Antenatal HIV rates by province

Although antenatal prevalence rates by ethnic group are no longer released due to low numbers of nonblack subjects, early figures showed the disease disproportionately impacting the black community. Apartheid-era legislation forbade sexual contact between certain delineated ethnic groups, and while such contact is no longer prohibited, prejudice and ingrained social and cultural divisions remain.

Given this regionalization and ethnic specificity, any practice that serves to promote sexual contact between people of different provinces or different ethnic groups, such as sex work, is of great interest to researchers. Despite the pressing need for such information, very little research into sex work and HIV has been conducted and it is presently impossible to gauge whether sex workers contribute significantly to the spread of the disease. No study has attempted to determine the extent of drug use among sex workers nationally, or the effect drug use might have on their business practices.

Sample
As a "hidden" population, it is extremely difficult to estimate the total number of female street sex workers operating in the major urban areas of South Africa. Previous studies (Leggett 1999) have used the "tag and release" technique by tallying the number of re-arrests recorded in police records, but there are many flaws to this approach.

Given limited resources, it was decided to make use of the expertise of authorities proficient in working with sex workers in each of the three cities, and these experienced individuals administered the questionnaires. As the nature of street sex work requires that the sex workers congregate at specific areas where customers know to find them, a sampling of all known street sites was seen as the best way of obtaining a fair cross section of the population.

Using the ethnic categorizations imposed during apartheid, the overall study sample was just over 50-percent black, just over 38-percent colored, over 8-percent white, and less than 3-percent Indian. This small Indian sample, drawn exclusively from Durban, was too small for separate analysis. The ethnic mix varied sharply by region, with the Cape Town sample being largely colored and black, Johannesburg being mainly black, and Durban being more varied. These samples were considered by the interviewers to be representative of the overall sex work population in each area, although the Johannesburg sample was drawn exclusively from the inner city and does not represent the sex work population of the city as a whole.

Reported ages ranged from 10 to 53 years old. Half the population was under 24, and over 10 percent were underage (less than 18 years). All but one of these children was black or colored. On the other end of the scale, almost a quarter of the interviewees were over the age of 30. Most of the older women were white or colored.

The Johannesburg sample was notably younger than the other two, which is most likely due to its ethnic composition. In Cape Town, the mean age was 27, the median 25, and the mode 19, with a standard deviation of 7.8. In Johannesburg, the mean age was 24, the median 23, and the mode 18, with a standard deviation of 6.5. The Durban sample had a mean age of 27, a median of 26, a mode of 23, and a standard deviation of 6.6.

The Cape Town sample, while the largest and most geographically diverse, may suffer from a certain amount of redundancy in that the several individuals were re-interviewed in the second (saliva-linked) set of interviews. It appears that nearly all the sex workers at several sites were interviewed.

It is impossible, given the imbalances discussed, to generalize about the sex-working population as a whole on the basis of this sampling. But it is possible to draw conclusions about the differences between drug-using sex workers and their abstinent colleagues in each of the urban centers.

Sex work practice
As HIV in South Africa is so highly regionalized, the home origins of these women are extremely relevant to a discussion of their HIV status. Analysis of home origins shows significant amounts of migration, including migration from higher HIV areas to lower ones (e.g., from Eastern Cape to Western Cape or from KwaZulu-Natal to Johannesburg). One of the unexpected findings of this survey was the high number of sex workers in each city originating from the Eastern Cape. Over 40 percent of Johannesburg sex workers were from either Port Elizabeth, East London, the former Transkei, or other parts of the Eastern Cape. Durban was the second largest contributor of sex workers to the Johannesburg Central Business District (CBD), with nearly a 20 percent share. Only a small minority (6 percent) claimed Johannesburg as their hometown. The same is true to a lesser degree in the Durban inner city, where 14 percent said they were from Johannesburg, and 14 percent from Cape Town. Similarly, in Cape Town, 47 percent of the women polled were from outside the metro area. A surprising number of women working in Cape Town were from small towns throughout the country, but especially from the Western and Eastern Cape.

Cape Town showed the greatest range of housing options, with women often living together with a large number of housemates in a free-standing home. The Johannesburg sample was almost entirely housed in daily accommodation hotels in the immediate area of the interview. Seventy-five percent of the sex workers in the Durban CBD resided in residential hotels, while a similar percentage of black women interviewed outside the inner city resided in informal settlements.

Asking questions about illegal activity is always problematic, particularly when the subject is as intimate as sex work. Nonetheless, the women interviewed were quite forthcoming about their practices, and this was facilitated by the experience of the interviewers in working with this population group. Certain response areas should be treated with some skepticism, however, and the motivations of the sex workers must be kept in mind at all times. Some women clearly exaggerated their popularity and income as a way of emphasising their desirability, while others minimized their participation in the market due to modesty.

Overall, the modal trends reflected a credible pattern of business activity for the average street sex worker, or rather, two patterns. A distinction needs to be made between inner city sex workers, who tend to be lighter in complexion, and more remotely based black women.

The black women report being in the business for shorter periods of time than the other groups, with 58 percent having been in business two years or less. Only 26 percent of the white women were so recent to the business.

Overall, 20 percent of the women reported having only one to five acts of vaginal sex with clients per week, but this figure is heavily skewed by statistics from poorer black women in more remote areas. Thirty-one percent of black women had five or fewer clients per week, compared to 16 percent of white women and 7 percent of colored women. On the other extreme, 16 percent of white women reported more than 30 clients a week, compared to 15 percent of coloreds and 13 percent of blacks.

As might be expected, client volumes were tied to drug use. Only 16 percent of those women who had less than 10 clients a week used hard drugs, while 43 percent of those who had more than 20 clients a week had a drug habit.

Rates charged were also racially skewed-over 75 percent of white women reported charging over R90 (U.S. $13) for vaginal sex, while 83 percent of black women charged less than that. As a result, over half the white women surveyed reported making over R1,500 (U.S. $218) per week, while 80 percent of black women reported earning less than R500 (U.S. $23) each week.

The women were understandably reluctant to admit to unsafe sexual activity, although many eventually did so with some coaxing. Twenty-eight women, or about 8 percent, admitted to occasional condom-free sex with clients, at least for oral sex. Eighty-two percent of this group were black women, and 78 percent of this group were HIV positive. Over 70 percent of these cases were found in Durban, but this likely has more to do with interview technique than actual prevalence.

This data would seem to call for greater HIV education among the black community, but other studies have assessed the general knowledge levels of the disease to be high among this grouping (Varga 1997), and attribute inconsistent condom usage to sociocultural factors and gender-related coercion.

Ten women, or under 3 percent, admitted to performing anal sex with clients, but there did not appear to be any special correlation with ethnicity or HIV status in this category.

All of this must be weighed against the fact that over 70 percent of those sex workers with boyfriends said they did not use condoms with these partners, and that over a third of these men were using drugs other than alcohol.

Also disturbing were the number of women who reported not using any form of contraception other than condoms (50 percent), as the women interviewed indicated a high rate of condom breakage. Although this information was not directly solicited, there were many complaints about government-issued condoms, including the fact that they were generally too small and thus subject to breakage.

A number of the sex workers contacted were pregnant at the time of the interview or had recently given birth. One sex worker indicated that crack use caused amenorrhea, and that users would only become pregnant following a period of incarceration.

Drug-use habits
The sex workers were questioned about which drugs they used, how much they used, how often they used them, when they used them, where they used them, and how much they spent on a single session of use. A question aimed at understanding whether certain substances were used prior to entering into the field of sex work was invalidated due to early misapplication.

There are considerable difficulties in acquiring accurate information on drug use via self-reporting. Nonetheless, the responses from the sex workers with regard to drug use formed a consistent pattern. White, high-income, high-client-volume, inner city women reported high levels of drug use, while poorer, black women in more remote areas did not. Higher income colored and Indian women in inner city areas tended to follow the white pattern, while those in more remote areas followed the black pattern.

As might be expected, there was a strong link between drug abuse and ethnicity. Eighty-one percent of the white women interviewed reported at least weekly use of crack or Mandrax, as compared to 25 percent of colored women and 6 percent of black women.

There was considerable regional variation between reported drug abuse prevalence. Twenty-six percent of Cape Town interviewees, 14 percent of Johannesburg interviewees, and 41 percent of Durban interviewees reported using "hard drugs" (drugs other than alcohol and cannabis, generally Mandrax or crack) at least occasionally. With regard to frequency of use, 73 percent of those who said they had ever smoked cannabis said they currently did so "daily" or "frequently," compared with 72 percent of lifetime Mandrax users and 81 percent of crack users who were current daily or frequent users. A further 20 percent of lifetime Mandrax users and 7 percent of lifetime crack users said they had quit using the drug after a period of at least weekly use. Thus, the women were sharply divided between those who had never tried drugs, those who used them chronically, and reformed chronic users. Very little casual use of hard drugs was seen. For the purposes of this analysis, a "hard drug user" is considered anyone who uses or previously used either Mandrax or crack cocaine regularly.

Three women reported heroin use and three women reported "pinks" use, and only one of these women reported daily use of each drug. All six of these women were white, and only one (a "monthly" user of pinks) was HIV positive.

The regional variation in numbers of interviewees reporting drug use may have been due in part to interview technique, but is largely attributable to sampling. For example, the Johannesburg sample, although drawn from a notorious drug area, was 86-percent black. Of the 12 white women who were interviewed in Johannesburg, eight admitted hard drug use.

Within each of the three metro regions there were pockets of women in which no interviewee reported using drugs. This was in line with expectations-some of these areas are remote from any known drug outlets. There were also areas where reported drug use was very high. In the Durban CBD, 75 percent of the women reported the use of drugs other than alcohol and cannabis.

There is a very strong relationship between the sex worker's place of origin and her likelihood of using hard drugs, regardless of interview site. Nearly 65 percent of women originating from Johannesburg reported hard drug use, as did 60 percent of women from outside South Africa. Women originating from Durban (18 percent), rural areas (19 percent) and the Eastern Cape (8 percent) were far less likely to report using drugs.

Just under two-thirds of the women reported using alcohol, with 38 percent reporting drinking more than three drinks at a sitting. Dagga is the most commonly consumed illegal drug, with 26 percent of the women admitting some use, and 18 percent reporting daily use. Twenty-three percent of the women reported using alcohol before working, and 15 percent used dagga before or during sex work.

There is a significant relationship between use of hard drugs and high-client volumes, with about half of the nonusers having 10 clients a week or less, and 44 percent of the users having more than 20 clients per week. Clearly, a drug habit requires greater client numbers in order to fund the addiction.

The modal value for crack use was 10 rocks, or R500 worth of crack, per night. This total cost may not come from the women's income, though. Many women have effectively shifted out of full-time sex work into the business of procuring drugs and providing company during drug use. About one quarter of the women reported having over 20 percent of their clients requesting drugs. Sometimes these requests also included sex, but many others were more requests for "company." This often involved disrobing and the occasional attempt at oral sex, but the inability of most men to attain an erection while high on crack or Mandrax made sex-to-orgasm very rare in these encounters.

There is a significant relationship between the number of years spent in sex work and hard drug use, with women in the business for more than 4 years being more than twice as likely to use hard drugs as women who have entered the field in the last year. This suggests that drug use is adopted as a coping mechanism to deal with the work, and eventually becomes an important reason for staying in the business.

HIV/AIDS and drug use
This research was commissioned principally to test whether a positive association exists between use of hard drugs and HIV seroprevalence among commercial sex workers in South Africa. This was not found to be the case. Hard drug users were actually less likely to be HIV positive (27 percent) than nonusers (56 percent) were.

A similarly negative relationship was found between self-reported drug use and HIV status among arrestees in South Africa by Parry et al. (1999). Parry noted the high tendency to deny drug use as one possible explanation for this finding. In the present study, however, levels of drug reporting are much higher in areas where this would be expected and denials of drug use generally correspond with low-income levels. For example, in the Durban CBD, 75 percent of the women admitted to hard drug use, but 70 percent of the hard drug users were HIV negative, while half of the nonusers tested positive. This can be compared to those interviewed in the Durban suburbs of Chatsworth and Wentworth, where the mean income was R357 per week; none of the women in this area reported using drugs, and 67 percent tested positive for HIV.

A far more accurate predictor of HIV seroprevalence within this population is ethnicity. Black sex workers show substantially higher levels of HIV than other ethnic groups. Over 66 percent of black sex workers in the sample were HIV positive, compared to 18 percent of whites and 17 percent of coloreds. Areas of sex work that were exclusively or almost exclusively black had by far the highest rates of HIV. One drug-free location truck stop contributed disproportionately to the HIV rate for Cape Town. All the workers tested at this site were black, and none of them reported using drugs. Of the 11 women tested at this site, 10 were HIV positive. Most reported weekly incomes of between R150 and R200. This group comprised nearly half of the 21 women who tested positive in Cape Town.

This strong association between ethnicity and HIV status seems to be behind the inverse relationship between HIV and drugs, as whites (71 percent) and coloreds (42 percent) are more likely to use hard drugs than blacks (10 percent). Even within an ethnic group, however, use of drugs makes no difference in HIV status. Isolating just the black sample showed equal portions of drug-user types tested positive (66 percent). Drug use within this population apparently has no bearing on whether the individual is HIV positive or not. Similar tests could not be run for other ethnic groups due to lack of sufficient numbers. However, combining all nonblack sex workers into one category likewise showed no relationship between drug use and HIV.

Conducting an analysis of variance (ANOVA) on the results, with HIV as the dependent variable and race (excepting the small Indian sample) and drug use as dependent variables, showed that race is highly associated with HIV, drug use is not, and the interaction between race and drug use is not significant (p = 0.057).

Thus, it is precisely the sex workers least likely to be using drugs-the poorest black women-who are most susceptible to HIV/AIDS. These are also, ironically, the women with the lowest client volumes, as there appears to be an association between high-client volumes and drug use.

References

1. Abdool Karim, Q.; Abdool Karim, S.; Soldan, K.; and Zondi, M. Reducing the risk of HIV infection among South African sex workers: Socio-economic and gender barriers. American Journal of Public Health 85(1), 1995.
2. Leggett, T. Poverty and sex work in Durban, South Africa. Society in Transition 30(2), 1999.
3. Parry, C.; Louw, A.; Vardas, E.; and Pluddemann, A. "Phase 1 Report: Medical Research Council/Institute for Security Studies 3-Metros Arrestee Study." Unpublished Report, 1999.
4. Ross, M.; Hwang, L.; Leonard, L.; Teng, M.; and Duncan, L. Crack cocaine as a major risk for HIV transmission in a crack house population. Texas Commission on Alcohol and Drug Abuse Research Briefs, University of Texas, Houston, 1998.
5. Varga, C. The condom conundrum: Barriers to condom use among commercial sex workers in Durban, South Africa. African Journal of Reproductive Health 1(1), 1997.