SEXUALLY TRANSMITTED DISEASES (STDs) and AIDS continue to rise rapidly, particularly in Asia. The World Health Organization projected that Asia will have the largest number of HIV-infected persons in the world by the year 2000.1 The main mode of transmission in Asia is heterosexual contact, with sex work being the most important source of heterosexual transmission.2-4 Given the low probability of an AIDS vaccine in the near future, AIDS-prevention efforts for sex workers have concentrated on promoting safer sexual practices and condom use. A recent metaanalysis showed that condoms reduced risk of HIV transmission by approximately 69%.5
The effectiveness of condoms in preventing sexually transmitted infections is dependent on consistent as well as correct use to prevent breakage and slippage. Studies on condom breakage and slippage among sex workers in developed countries found that condom-breakage rates for vaginal sex ranged from 0% in Nevada6 to 0.8% in Sydney7 and the Netherlands8 and 5% in Denmark.9 The study in Nevada on brothel-based sex workers found that only 0.6% of condoms completely fall off during withdrawal, 3.4% slip down during intercourse, and 4.3% slip down during withdrawal.6
Some studies have been carried out on specific behaviors associated with condom slippage and breakage among male condom users. A collaborative study on male condom users from three international sites in Mexico, Philippines, and the Dominican Republic10 found condom breakage to be associated with opening condom packages with sharp objects and unrolling condoms before use. Condom slippage was found to be associated with lengthy or intense intercourse and unrolling condoms before use. Another study on condom behaviors among male university students in the United States11 found condom problems among both consistent and inconsistent condom users. Condom breakage during intercourse or withdrawal occurred in 4.1% of episodes of condom use, whereas slippage occurred in 1.5% of episodes.
Data on condom breakage and slippage among commercial sex workers in Asia are limited. A prospective study in Thailand of 326 sex workers found a breakage rate of 5.9% per act of intercourse among 5,559 condoms used.12 The slippage rate was 0.4% per act of intercourse among 1,296 condoms used with clients.
In Singapore, all brothel-based sex workers are required to maintain contact with the only public STD clinic and with designated general practice clinics for regular monthly screening for gonorrhea and chlamydia and three monthly screenings for HIV and syphilis. Sex workers found to have STDs are treated immediately to prevent transmission. In 1992, the Department of STD Control (Singapore) implemented a behavioral intervention project to promote consistent condom use among all sex workers.13 The project has succeeded in increasing consistent condom use from 44.4% in 1994 to 90.3% at 2 year follow-up period with a corresponding decline in gonorrhea.14 Presently, all brothel-based sex workers are required to attend health-education sessions on AIDS and STDS, condom negotiation skills, and condom-use techniques at the Department of STD Control. They are also required to use condoms with all clients, and brothel keepers have been instructed to support this behavior. Although gonorrhea rates have decreased markedly, cases of cervical gonorrhea, albeit few, are still reported among sex workers who report consistent condom use. The occurrence of infection could have resulted from condom breakage or slippage. Presently, there are no data on condom breakage and slippage rates among these sex workers.
In recent years, some critics of condom-promotion programs have also questioned the effectiveness of condoms in protecting users against HIV and other STDs, and have therefore argued against condom promotion because of anecdotal reports of breakage or slippage. Anecdotal reports do not reflect the true incidence of breakage and slippage, and studies to quantify the incidence of breakage and slippage are needed. The determinants of condom breakage and slippage among sex workers are complex and incompletely understood. We conducted this study to determine the condom slippage and breakage rates among brothel-based sex workers in Singapore. We also investigated the relationship between condom breakage and slippage and sociodemographic characteristics of the sex workers. This relationship could help to identify high-risk groups among the sex workers and target programs to reduce condom failure, such as breakage and slippage. In addition, local data on condom failure rates would be useful to policy makers and STD program managers in Singapore to assess the risk of STD transmission with consistent condom use.
Methods
The study was conducted among 219 brothel-based sex workers attending health-education workshops on condom use and condom negotiation skills at the Department of STD control from May to December 1997. Following the workshop session, sex workers were given a monitoring form to record the total number of condoms used and the total number of condoms that broke or slipped off during vaginal intercourse with paying clients during 1 working day. They were also asked to record whether the client for each condom use was a spontaneous condom user.
A condom use was defined as a sexual episode during which a condom was used; an episode in which two condoms were used simultaneously was counted as one condom use. We used the definitions of clinical breakage and slippage from the standardized protocols for condom breakage and slippage trials14 to define condom breakage and slippage. A clinical condom breakage was thus defined as a condom reported to have broken during intercourse or withdrawal. Likewise, a clinical condom slippage was defined as a condom reported to have slipped off during intercourse or withdrawal. A spontaneous condom user was defined as a client who initiated condom use or allowed the sex worker to put a condom on his penis without hesitation, resistance, or the need for persuasion. The sex workers were instructed to put the condom on their clients using the techniques taught during the health-education sessions. A nonspontaneous condom user was defined as one who had to be persuaded to use condoms; this definition was used to assess the client's readiness to use condoms according to the stages of change model.16 Thus, a spontaneous condom user had the intention to use condoms and was in the ready-for-action stage or action stage of behavior change. A nonspontaneous condom user did not have the intention to use condoms and was in the precontemplation or contemplation stage of using condoms. The question on spontaneous condom use was pretested, used in our previous survey, and was well understood by sex workers.17
Almost all sex workers were literate and able to record the data on condom breakage and slippage on the self-monitoring form. Only 1.4% of the sex workers were illiterate. The questions were explained to sex workers, who were asked to record a condom breakage or slippage event with a check. To ensure accuracy of the self-reported data, sex workers were instructed to check the condom before use so that broken condoms would not be counted or used. They were also told to give truthful answers and record the information after each act of sexual intercourse to reduce recall bias. We explained that the data would be used to help find solutions to problems on condom use encountered by sex workers in Singapore.
A self-administered questionnaire was written in simple colloquial Chinese, Malay, Thai, and English, and was distributed to all sex workers in the sample. Information was collected on age, ethnic group, educational level, duration of sex work, socioeconomic class, and perceived reasons for condom breakage and slippage. Respondents with limited understanding of the written language were interviewed by trained interviewers. Qualitative investigations, using in-depth interviews, were conducted with 25 sex workers to explore possible reasons for breakage and slippage.
Statistical Analysis
Condom breakage and slippage were analyzed using the condom as a unit of analysis. Condom breakage and slippage rates were calculated as the percentage of the total condoms used with clients that had broken or slipped off during vaginal intercourse or withdrawal. As each sex worker (cluster) had a daily mean of nine sexual encounters with clients in which condoms were used, the intracluster correlation of measurement observations taken on each condom such as breakage, slippage, or age within each sex worker has to be taken into account. The generalized estimated equation method18,19 was thus used to adjust for the cluster effect of condom measurements within subjects in the computation of the 95% CIs for the rate (percentage) of condom breakage and slippage per condom use with clients during the 1-day monitoring period.
The chi-square test was used to compare differences in proportion of condom breakage and slippage among groups. The generalized estimated equation for longitudinal data, using the logit link function, was used to compute the crude and adjusted odds ratio of condom breakage and slippage (dichotomous outcome, yes or no) by selected sociodemographic characteristics.
All P values calculated were two tailed. Analysis was carried out using the STATA 6 software program (Stata Corporation, College Station, TX). The α level of significance was set at 0.05.
Results
Sociodemographic Characteristics
A total of 235 sex workers were given forms to monitor condom breakage and slippage. Of these 235 sex workers, 219 (93.2%) attended the follow-up session and returned the completed forms. Approximately half were Chinese, and the others were Malay or Thai. Ages ranged from 21 to 49 years with a mean of 25.2 years (Table 1). The median duration of sex work was 2 months (range, 1-42 months). Sex workers saw a median of 10 clients per day with the maximum being 30 clients per day. Approximately two thirds were middle-class sex workers who charged less than US$30.00 per client. One third were upper-class sex workers who charged US$30.00 or more per client. Rate of anal intercourse was negligible (0.9%).
Source of Condoms
Over the 1-day period, 1,885 condoms were used for vaginal intercourse. The number of condoms used per sex worker during the 1-day period ranged from one to 31 (mean, 8.6 condoms; median, 9.0 condoms). The majority of sex workers (96.8%) consistently used condoms with their clients. All sex workers used prelubricated, good-quality latex condoms. Approximately two thirds of sex workers (68.9%) bought the condoms themselves, whereas the remaining one third got condoms from brothel keepers. A negligible proportion (1.4%) got condoms from their clients. Among those who bought their condoms, the main sources were the pharmacy or the STD clinic (49.0%), peddlers who come to the brothels to sell condoms (25.2%), and the provision shops (23.2%). The majority of sex workers (98.2%) used water-based lubricants; none used oil-based lubricants. Less than 1% of sex workers used two condoms for a single sexual episode.
Condom Breakage and Slippage Rates
Table 2 shows the rates of condom breakage and slippage per condom use with a client. During the 1-day monitoring period, 1,885 condoms were used. The breakage rate per condom use was 1.2% and the slippage rate was higher at 2.1%. Of the 219 sex workers, 19 (8.7%) reported at least one incident of condom breakage, and 23 (10.5%) experienced at least one incident of condom slippage. Four sex workers (1.8%) self-reported more than one condom breakage (maximum, two condom breakages). Ten sex workers (4.6%) experienced more than one slippage (range, 2-4) (data not shown).
Rates and Odds Ratios of Condom Breakage
Table 3 shows the rates and crude and adjusted odds ratios of condom breakage by sociodemographic characteristics of the sex workers. Malay sex workers had a significantly higher breakage rate than Chinese or Thai sex workers, who shared similar condom-breakage rates of 1%. Breakage rates increased significantly with age and duration of sex work. Higher breakage rates were also found among middle-class sex workers and nonspontaneous condom users, but these differences were not statistically significant. After adjusting for the confounding variables shown in Table 3, ethnic group and age were not significantly associated with condom breakage. However, duration of prostitution and nonspontaneous condom-using status of the client showed a significant association. The odds of condom breakage was 4.9 times higher among those who worked more than 25 months (95% CI, 1.03-24.30) compared with sex workers who worked less than 6 months after adjusting for ethnic group, age, class, educational level, and spontaneous condom-using status of client. The odds of condom breakage was 2.4 times higher among nonspontaneous condom users (95% CI: 1.02-6.02) compared with spontaneous condom users after adjusting for confounders.
Rates and Odds Ratios of Condom Slippage
Table 4 shows the rates and odds ratios of condom slippage by sociodemographic characteristics of the sex workers. Unlike condom breakage, condom slippage rates were higher among Chinese sex workers compared with Malay and Thai sex workers. Condom-slippage rates were significantly higher among new sex workers (i.e., those working less than 6 months) and those who had worked longer than 25 months, with no slippage reported by those working between these two periods. Like condom breakage, slippage rates were also significantly higher among nonspontaneous condom using clients. After controlling for confounding variables, duration of prostitution and spontaneity of client to use condoms were significantly associated with increased condom slippage. Sex workers who had worked more than 25 months and nonspontaneous condom-using clients were more likely to experience condom slippage. The odds of increased condom slippage among Chinese sex workers compared with sex workers belonging to other ethnic groups was also close to statistical significance.
Reasons for Condom Breakage and Slippage
Sex workers who reported having experienced a condom breakage or slippage during the 1-day monitoring period were asked to provide perceived reasons for the slippage and breakage. The reasons given for condom breakage included rough intercourse (81.5%), lengthy intercourse (55.6%), a dry vagina (37.0%), and condoms that were too small (11.1%). Reasons for slippage included lengthy intercourse (60%), condoms that were too big (59.1%), delayed withdrawal after loss of penile erection (40%), and incorrect condom use (2.5%).
In-depth Interviews
In-depth interviews were carried out with 25 sex workers of different ethnic groups, class, and work duration to explore reasons for breakage and slippage. Sex workers reported that they could convince almost all of their clients to use condoms because of the 100% condom policy and support by brothel keepers.14 Clients, especially immigrant workers, would agree to condom use after persuasion because they did not want to "get in trouble with the law." However, because they were not spontaneous condom users and because they were initially reluctant to use condoms, some of these client deliberately pricked the condoms with their fingernails.
Many of the sex workers also described some clients as being "greedy." These clients wanted longer sessions with the sex workers. After intercourse, they intentionally delayed withdrawal, thereby leading to loss of penile erection and increased slippage.
Some sex workers reported using the following approaches to decrease slippage: (1) holding the rim of the condom against the base of the penis during withdrawal; or (2) contracting their pelvic muscles to hasten ejaculation and thereby reduce lengthy intercourse, which could lead to slippage. To reduce breakage, sex workers (1) applied more lubricant inside their vagina; or (2) pinched the tip of the condom to expel any air after putting the condom on the client's penis.
Discussion
The breakage and slippage rates of 1.2% and 2.1%, respectively, were low. The breakage rate was slightly higher than that found among brothel-based sex workers in the United States,6 Sydney,7 and the Netherlands8 where the rates were less than 1%. However, the rate observed was much lower than the breakage rate of 5.7% reported by brothel-based workers from Thailand.12 This difference could be explained by the different assessment techniques used. Unlike the study in Thailand in which condoms were examined for breakage by the researcher, condom breakage in our study was reported by sex workers. Our study participants may have neglected to find broken condoms or may have intentionally underreported breakage. Another possible reason for the lower breakage rate in our study was that the rate was assessed after sex workers received health-education sessions on proper condom-use techniques.
In our study, we found that sex workers experienced higher breakage and slippage rates with nonspontaneous condom-using clients. A possible explanation is as follows: Because these clients were not ready to use condoms and used condoms reluctantly because of pressure from the sex workers, they might have exercised less caution in using condoms to prevent infection. For example, sex workers have reported clients who deliberately pricked the condoms with their fingers or delayed withdrawal after ejaculation. In contrast, spontaneous condom users, being in the ready-for-action or action stage, intended to use condoms. They would therefore be more likely than nonspontaneous condom users to take precautions to prevent breakage and slippage. This hypothesis needs to be tested further to compare the behaviors of nonspontaneous and spontaneous condom-using clients.
Other than spontaneous condom-using status, risk factors for condom breakage and slippage appear to be different. For example, higher slippage rates were found among new sex workers and those working longer than 25 months compared with those working between the two periods. The higher slippage rates among newer sex workers, probably due to inexperience, are consistent with other studies.6 However, unlike the study in Nevada that showed higher slippage rates among sex workers working less than 2 years, we found sex workers working longer than 25 months to have a significantly higher risk of condom slippage and breakage.6 It is unclear from our study why sex workers who had worked longer experienced higher slippage rates. Interpretation could have been hampered by the small numbers (there were only five sex workers in this category who used a total of 43 condoms). If the finding is true, it could be because they have become more complacent with longer duration of work, and have not taken precautions to prevent slippage by holding the rim of the condom at the base of the penis during withdrawal. The main self-reported reason for increased slippage among this group was the loss of penile erection resulting from delayed withdrawal. Client characteristics of the sex workers could also differ by duration of sex work; however, we did not collect sufficient data on client characteristics.
The higher slippage rates among Chinese compared with non-Chinese sex workers could be due to differences in sociodemographic and other characteristics of their clients. Our recent study carried found that most of the clients of Chinese sex workers (92%) were Asian (e.g., Chinese, Japanese, or Taiwanese) compared with non-Chinese sex workers, among whom only 67% of clients were Asian (other clients were Bangladeshi, Indian, or Indonesian. We did not look into penis size of the clients; however, preliminary results from ongoing condom trials conducted by the Los Angeles Regional Family Planning Council suggest a strong relationship between penis circumference and condom slippage.15
Our study has a few limitations. Apart from looking at the readiness of the client to use condoms, we did not examine other client characteristics such as age, ethnic group, penis size, and specific behaviors that could have caused slippage and breakage. Future studies should investigate these factors in-depth and record the specific behavior with each event of condom use so that the breakage and slippage can be linked to specific client behaviors and characteristics. Another limitation is bias arising from self-reporting. We tried to reduce self-reporting bias by assuring the sex workers about the confidentiality of the data collected and explaining the purpose of the study, which was to help them reduce their risk of infection and self-monitor condom use. Finally, our study findings cannot be generalized to street-based sex workers, whose working conditions may be very different.
Conclusion
Condom breakage and slippage rates were low, with the breakage rate (1.2%) being lower than the slippage rate (2.1%). Nonspontaneous condom-using clients and sex workers who worked longer than 25 months had higher risks for condom breakage and slippage. The risk of condom slippage, but not breakage, was also higher among new and Chinese sex workers. Sex workers should be told to take extra precaution with nonspontaneous condom users. Presently, only new sex workers were required to attend classes on STD and HIV prevention and condom use. Our findings suggest that sex workers who have worked longer than 25 months are at increased risk of condom failure, and should be given refresher talks on HIV and STD prevention. Future research should be carried out on clients to identify characteristics and specific behaviors of clients that are associated with condom slippage and breakage. This research would provide information that could be used to develop more effective interventions to reduce condom failure among sex workers. Finally, we should continue to promote condom use because of low breakage and slippage rates.
References
1. The Global AIDS Policy Coalition. Status and Trends of the HIV/AIDS Pandemic as of January 1996. Cambridge: Harvard School of Public Health, 1996.
2. Fontanet A, Piot P. State of our knowledge: the epidemiology of HIV/AIDS. Health Trans Rev 1994; 4(suppl):11-23.
3. Hanenbery RS, Rojanapithayakorn W, Kunasol P, Sokal DC. Impact of Thailand's HIV control programme as indicated by the decline of sexually transmitted diseases. Lancet 1994; 3344:243-245.
4. Chew SK, Snodgrass I, Monteiro EHA. HIV infection in Singapore: the second wave. Epidemiol. Singapore News Bulletin 1992; XVIII: 25-27.
5. Weller SC. A meta-analysis of condom effectiveness in reducing sexually transmitted HIV. Soc Sci Med 1993; 36:1635-44.
6. Albert AE, Warner DL, Hatcher RA, et al. Condom use among female commercial sex workers in nevada's legal brothels. Am J Public Health 1995; 85:1514-1520.
7. Richters J, Donovan B, Gerofi J, Watson L. Low condom breakage rate in commercial sex (letter). Lancet 1998; 2:1487-1488.
8. De Graaf R, Vanwesenbeeck I, Van Zessen G, Straver CJ, Visser JH. The effectiveness of condom use in heterosexual prostitution in the Netherlands. AIDS 1993; 7:265-269.
9. Gotzsche PC, Harding M. Condoms to prevent HIV transmission do not imply safe sex. Scand J Infect Dis 1988; 20:233-44.
10. Spruyt A, Steiner MJ, Joanis C, Glover LH, Piedrahita C, Alvarado G, Ramos R, Maglaya C, Cordero M. Identifying condom users at risk for breakage and slippage: findings from three international sites. Am J Public Health 1998; 88:239-244.
11. Warner L, Clay-Warner J, Boles J, Williamson J. Assessing condom use practices: implications for evaluating method and user effectiveness. Sex Transm Dis 1998; 25:273-77.
12. Rugpao S, Pruithithada N, Yutabootr Y, Prasertwitayakij W, Tovanabutra S. Condom breakage during commercial sex in Chiang Mai, Thailand. Contraception 1993; 48:537-547.
13. Wong ML, Chan R, Koh D, Wong CM. Theory and action for effective condom promotion. Illustrations from a behaviour intervention project for sex workers in Singapore. Int Q Commun Health Educ 1995; 15:405-421.
14. Wong ML, Chan Roy KW, Koh D. A sustainable behavioural intervention to increase condom use and reduce gonorrhoea among sex workers in Singapore: 2-Year follow-up. Prev Med 1998; 27:891-900.
15. Steiner M, Trussell J, Glover L, et al. Standardized protocols for condom breakage, slippage trials: a proposal. Am J Public Health 1994; 84:1897-1900.
16. Schnell DJ, Galavotti C, Fishbein M, Chan DKS. Measuring the adoption of consistent use of condoms using the stages of change model. Public Health Rep 1996; 111(suppl 1):59-66.
17. Wong ML, Wong TC, Tan ML, Ho JY, Lim S Wan, R. Chan, Factors associated with sexually transmitted diseases in Singapore, Int J STD AIDS; 1992; 3:323-328.
18. Liang K-Y, Zeger SL. Longitudinal analysis using generalized linear models. Biometrika 1986; 773:13-22.
19. Zeger SL, Liang K-Y. Longitudinal data analysis for discrete and continuous outcomes. Biometrics 1986; 42:121-130. AQ1: Au: Short title OK (see p. 3)?