At present in the Democratic Republic of Congo (DRC), commercial sex workers (CSW) constitute the group documented as being the most exposed to sexually transmitted infections (STI) and to HIV/AIDS. High rates of risky behaviour, STI and HIV have consistently been reported among them,1 2 pointing to the fact that CSW are the group who are most exposed to HIV/AIDS and that their clients constitute a group that might be playing an important role in the spread of HIV to the general population.3

Targeted interventions have been designed and implemented for more than a decade. Their impact on risk-taking behaviours has not, however, been formally evaluated.

HIV prevalence in the DRC is thought to be stable,4 5 but no clear indication can be provided to explain the observed stability.

Consistent condom use has been shown to reduce the incidence of HIV among CSW6^–8 and it has been reported that condom programmes oriented to CSW might be highly cost-effective.9

Consistent condom use depends on a variety of factors, including the knowledge that condoms are effective for preventing HIV,10 the partner’s acceptance, condom beliefs and self-efficacy,11 12 sociodemographic characteristics of the CSW, such as age, duration of commercial sex work, education, age at first sexual experience, intensity of sexual activity,13^–15 marital status,16 the capacity to negotiate safe sex with clients who offer extra money in order not to use a condom,17 18 the type of sexual partners (whether paying or not paying),10 19 20 socioeconomic disadvantages and the threat of violence from clients.21

Knowing the pattern of condom use and its determinants in the DRC might shed light on the course of the epidemic and on the role played by CSW and their sexual partners in the spread of the disease.

This paper presents the prevalence of consistent condom use by CSW during a 30-day period and its correlates, information that is of importance in the design of intervention programmes targeting CSW.

PARTICIPANTS AND METHODS

Settings

The present cross-sectional study was conducted in all the provincial capital cities in the DRC between January 2005 and March 2006.

Sample size and sampling strategies

Based on the assumption that one CSW out of two is involved in risky sexual behaviour, a minimum sample size of 267 CSW per site was expected. The CSW were selected using a time location sampling technique.22 In each city, all bars, nightclubs, walkways, and brothels were mapped. Considering CSW as “floating” populations that visit the sites at different times during the evening and that might have different behaviours, each day of the week was subdivided into two discrete time intervals, namely 18:00 to 20:00 hours and 21:00 hours to midnight. The primary sampling unit (PSU) was the site during each different time interval specified, and each PSU was entered separately into the listing of PSU for the sampling frame. From the list, 30 PSU were selected with equal probability using a systematic random selection procedure.

All CSW present during the day and the time interval selected were solicited to give an interview.

MEASURES

CSW were interviewed face-to-face by experienced interviewers who received additional training focused on the present research. The interviewers used a standard questionnaire designed by Family Health International22 that was adapted and pre-tested on a few CSW a year before the actual data collection.

Sociodemographic measures

Information on age, education, past and present marital status and dependents was collected.

Intoxication

Respondents were asked whether they had consumed intoxicants such as alcohol during the four weeks preceding the interview, and if they had ever used drugs.

Knowledge of condoms

Respondents were asked if they had ever heard of HIV/AIDS, and if they could cite any of the three traditional individual-level prevention measures, namely abstinence, being faithful and consistent condom use. Those who were considered as knowing about condoms were the CSW who spontaneously cited condoms as one of the effective means to avoid getting STI and HIV/AIDS.

Exposure to intervention programmes

Exposure to intervention programmes was measured as having been exposed to voluntary counselling and testing (VCT) and having seen an HIV/AIDS explicit message during the preceding six months.

Sexual behaviour

Information on age at first sexual experience, duration of commercial sex work, number of sexual partners during the preceding seven days, the categories of sexual partners (specifically paying partners and non-paying partners), and the pattern of condom use during the preceding four weeks was collected.

Sexual intensity was computed according to the number of sexual partners a respondent had had during the seven days preceding the interview. CSW who had had none or only one partner were considered as having a low sexual intensity, those with two to seven partners were regarded as having a moderate sexual intensity, and those with more than seven partners were classified as having a higher sexual intensity.

Sexually transmitted diseases

A history of STI was considered as a history of genital ulceration or a report of a foul-smelling vaginal discharge.

Risky sexual behaviour

CSW who admitted to not having used a condom at each sexual intercourse during the four weeks preceding the interview were considered as being involved in risky sexual behaviour.

Data analyses

CSW were subdivided into three age groups to take into account the difference that might exist between teenagers (age less than 20 years), adults (aged 20–44 years) and those likely to be at the menopausal stage (45 years or more).

Frequencies and descriptive statistics were computed for qualitative and continuous variables. Comparisons were made using Student’s t-test, Z-score, chi-square and one-way analysis of variance when appropriate. Stepwise logistic regression was used to identify correlates of consistent condom use. The following variables were entered in the model: age, education, number of sexual partners, duration of commercial sex work, knowledge of condoms as a protective means against HIV, geographical accessibility to condoms, having dependents, consumption of alcohol, having had an HIV test, and exposure to an HIV/AIDS explicit message. Variables with a p value of 0.05 or less were entered in the model and the removal criterion was a p value of 0.10 or more.

Ethical issues

The study protocol was approved by the Kinshasa School of Public Health Internal Review Board. CSW anonymity was ensured by not taking personal identifiers. A written consent was obtained from each CSW.

RESULTS

Sociodemographic characteristics, knowledge of HIV/AIDS and exposure to explicit messages and services regarding HIV/AIDS

Out of 2937 CSW solicited to give an interview, 2638 (89.8%) gave an informed consent and participated in the study. Sociodemographic characteristics differed across age groups. Overall, one in five CSW (18.7%) in the DRC were younger than 20 years, whereas the majority (76.4%) was aged between 20 and 44 years (table 1). Approximately three CSW out of four (71.8%) had been involved in commercial sex for more than one year, and close to two-thirds (63.2%) had regular sexual partners.

Three CSW out of five (60%) had attained a secondary school education, whereas four out of 10 (39%) never attended school or had dropped out at the elementary stage. Compared with younger CSW (40.9%), an important proportion of older CSW (92.2%) had dependents. Fewer than one in two (44.2%) had another source of income besides being involved in commercial sex work. A substantial proportion of CSW (70.9%, 95% confidence interval (CI) 69.2 to 72.6) had used alcohol and one in four (22.2%, 95% CI 20.6 to 23.7) had ever tried marijuana.

At the time of the survey, almost all of the CSW (98.3%) had heard of HIV/AIDS, but only a few (8.8%, 95% CI 7.7 to 9.9) could spontaneously cite all three means of prevention (consistent condom use, being faithful to one HIV-negative partner, and abstinence).

A sizeable proportion of the CSW (74.5%, 95% CI 72.8 to 76.1) had recently been exposed to a message on HIV/AIDS. Approximately four CSW in 10 (38.7%, 95% CI 36.8 to 40.5) had been counselled and had undergone an HIV test. CSW aged 45 years or older were less likely to have easy access to condoms (35.2%) compared with other age categories (63.7%; Z-score 6.5, p<0.001).

Sexual history, sexual behaviour and history of symptoms suggestive of STI

The CSW began their sexual activity at an early age (median age 15 years; table 2). The mean number of sexual partners differed among age groups (mean 1.20, SD 1.44 among those aged 45 years or older; mean 3.79, SD 5.55 among those aged 20–44 years; and mean 4.13, SD 5.31 among those aged less than 20 years; F  =  18.01, p<0.001).

In general, four CSW out of 10 (40.0%, 95% CI 38.1 to 41.8) were consistent condom users. The pattern of condom use, however, differed among the age categories (chi-square 37.1, p<0.001). Moreover, consistent condom use depended on the category of sexual partners. Condoms were used less with non-paying partners (38.2%, 95% CI 36.3 to 40) than with paying partners (61.4%, 95% CI 59.5 to 63.2). Condom breakage/slippage was common as 36.2% (95% CI 34.4 to 38.0) of CSW had ever experienced it. It was rare, however, among older CSW. The periodic prevalence of symptoms suggestive of STI was high among CSW, at 16.3% (95% CI 14.9 to 17.7) for a foul-smelling genital discharge and 9.8% (95% CI 8.7 to 10.9) for genital ulceration. A history of genital ulceration was associated with non-consistent condom use (chi-square 14.28, p<0.001).

Correlates of consistent condom use

Consistent condom use correlated with the duration of commercial sex work, the intensity of sexual activity, having cited condoms as a means of prevention of HIV, age, having had an HIV test, and condom accessibility (table 3). The duration of sex work was inversely correlated with condom use. Compared with CSW who had taken up prostitution recently, all those who had been prostitutes for a longer time were less likely to have consistently used condoms during the preceding four weeks (OR 0.68, 95% CI 0.49 to 0.94; OR 0.62, 95% CI 0.48 to 0.81; OR 0.53, 95% CI 0.39 to 0.72, respectively). Compared with those with low sexual intensity, CSW with moderate sexual intensity and those with high sexual intensity were more likely to have used condoms consistently over the preceding four weeks (OR 1.86, 95% CI 1.52 to 2.27; OR 3.83, 95% CI 2.95 to 4.96, respectively). Knowledge that condom use could prevent HIV was correlated with consistent use (OR 2.88, 95% CI 2.09 to 3.96). CSW aged 20–44 years were more likely to have consistently used condoms (OR 1.34, 95% CI 1.06 to 1.69). Having undergone an HIV test was associated with consistent condom use during the four weeks preceding the interview. Those who had been tested for HIV were twice as likely to be consistent condom users compared with those who had not (OR 2.02, 95% CI 1.70 to 2.42). Easy access to condoms correlated with their consistent use. Those who were close to distribution sites were more likely to have used condoms consistently over the preceding four weeks (OR 1.51, 95% CI 1.25 to 1.82).

DISCUSSION

The present study shows that in the DRC, the knowledge of traditional prevention means is still low among CSW and that consistent condom use is not common. Consistent condom use correlates with age, the duration of prostitution, the intensity of sexual activity, knowledge of condoms, easy access to condoms and experience with HIV counselling and testing.

These findings are consistent with previous reports.13^–15 Nonetheless, some of the associations found might be artifactual because of the power of the test and because of multiple comparisons. In addition, as in all studies based on voluntary participation and on self-report, it is not possible to rule out selection and misclassification biases. CSW who refused to participate might have different behaviours and those who participated might have given socially acceptable answers. The precautions taken, however, namely working with experienced interviewers, interviewers having been introduced to this group by members of local organisations working with CSW, having achieved a high response rate, and having minimised the memory bias by not asking information on events that had happened more than 6 months before the interview, we assume that the selection bias is likely to have a minimal effect and that misclassification, if any, might be non-differential, suggesting that the associations found might be even bigger than observed.

The perception of own risk has been found to be associated with condom use.23 This was not formally measured in our study. We think that CSW who have been counselled and tested might, however, have a higher perception of their risk of getting HIV. Nevertheless, further work to identify psychosocial correlates of condom use in DRC is needed.

The knowledge of traditional means of prevention of HIV being low points to the need to find genuine ways of educating this group.24 Low access to prevention services, ie a low proportion of CSW who have ever tested for HIV and low condom accessibility suggest that VCT should be promoted aggressively and that condom promotion should target more younger and older CSW, and that distribution sites should be brought very close to where CSW live and work.

The fact that condoms are used less consistently with regular partners implies that steady partners are more at risk and that they should be identified in order to be targeted with specific messages.25 26

That condom breakage/slippage is common among CSW indicates that condom promotion and provision programmes should include activities aimed at improving their correct usage.

As previously reported, condom use decreases with increasing age,19 and with the duration of commercial sex work. This suggests that ageing CSW and those with more time as CSW are probably biased about their risk assessment. It might also be that older CSW, who are less educated, at the menopause, or less successful at attracting clients, might be more prone to compromise or might perceive condoms as only being about preventing pregnancy and therefore do not use them consistently.18 27

The increase in consistent condom use with the number of sexual partners suggests that those with more clients are likely to perceive that they are at a high risk of getting HIV28 and that those with fewer clients might not correctly perceive their risk of getting HIV.

Having mentioned condom use as being effective in preventing HIV/AIDS, having been counselled and tested for HIV, and having had easy access to condoms correlating with consistent condom use, stress one more time that knowledge and access to prevention services can positively impact on behaviour.16 29 The government should feel compelled to scale up prevention services, namely VCT, condom promotion and provision.