Given the role of male condoms in preventing the sexual transmission of HIV, condom distribution has become central to HIV prevention activities in both developed and developing countries [1–5]. In South Africa, where antenatal HIV prevalence increased dramatically during the 1990s [6,7], male condom distribution has been central to the public sector's response to the growing HIV epidemic.
During 1999, approximately 198 million male condoms were distributed free of charge to the public by the South African Department of Health compared with only six million in 1994 . Condoms purchased by the government are distributed primarily through public health services such as primary care clinics and AIDS Training Information and Counseling Centres. The government is the largest distributor of condoms in South Africa. Social marketing programmes and commercial retailers together distributed 10–20 million condoms in 1997 .
From cross-sectional studies, the prevalence of reported condom use ever among South Africans ranges from 14% in a community-based survey of a rural area  to 22% among women of reproductive age nationwide . These data provide little insight into what percentage of condoms distributed to the public are ever used in sex and whether increased condom distribution is contributing to the fight against the growing HIV/AIDS epidemic in South Africa. The aim of this study was to assess the proportion of public sector condoms used in sex and to investigate the determinants of condom use among those procuring condoms from the public health service.
Materials and methods
To select a sample of sites which could represent public sector condom distribution points across the country, South Africa's 52 demarcated health regions were divided into four strata according to geographic location (east versus west) and population density (urban versus rural). One region from each of the resulting strata was then selected randomly for research. Within each of the four selected regions a list of the busiest public sector condom distribution sites was generated through consultation with relevant Department of Health officials. A total of 12 condom distribution points were selected randomly from these lists, ranging from small rural clinics which distribute an average of 200–500 condoms each month to large urban and peri-urban clinics which distribute as many as 6000–8000 condoms each month.
Fieldwork was conducted from June 1998 to March 1999. Participants in each region were recruited over a 2-week period, or up to approximately 160 participants, whichever came first (the mean time period for participant recruitment in each region was about 9 days). Upon leaving the selected site with condoms, consecutive individuals were approached by fieldworkers who explained the purpose and process of the study. For those who consented to participate, a semi-structured baseline questionnaire was administered to collect demographic information as well as data on recent sexual behaviour and condom-related knowledge and attitudes.
Immediately after completing this interview, participants were recruited for follow-up interviews over the next 5 weeks. Fieldworkers numbered each of the condoms taken by participants using permanent ink on the condom packaging to facilitate identification and recall during follow-up interviews. Participants were also given a simple condom-use journal for self-recording the fate of individual condoms. The procedure for using these was explained to participants in detail, and the text of the journal was translated into the appropriate local languages. It was carefully explained to participants that the condoms in the study should be treated normally, and that the ability to participate in the study was not related to their condom use or sexual behaviour.
At follow-up, participants used calendars and their condom-use journals to report on the fate (e.g. used, wasted, given away, etc.) of each condom taken at the initial interview and to count the number of days between the condom's procurement and its eventual fate. All interviews took place in a private room at the distribution sites and were conducted in the participants’ home language by trained fieldworkers.
Data were analysed using Stata (Stata Corporation, College Station, USA). Chi square and Fisher Exact tests were used for comparing proportions and Kruskal–Wallis tests were used to compare medians. Kaplan–Meier analyses were used to analyse the rate at which condoms were used over time. Poisson regression analysis was used to determine the effect of the potential risk factors on the incidence rate of condom use. Study region appeared as an effect modifier for several risk factors of interest and in these instances the regression model was adjusted accordingly.
A total of 594 individuals procuring condoms from the 12 clinics were recruited. Of these 40 (6.7%) declined to participate either because they would not be available for the follow-up interviews or because they did not feel comfortable with discussing condom use. The 554 consenting participants had procured a total of 8164 condoms at the beginning of the follow-up period. The median and modal number of condoms taken was 10 (range, 1–450 condoms). Similar numbers of condom procurers were recruited and followed-up in regions B, C and D (164, 149 and 162 participants recruited, and 122, 109 and 107 followed-up, respectively) and the rates of follow-up here were comparable. However in region A (from the rural, western stratum), small clinic sizes limited the number of condom procurers recruited over the 2-week period (79). In addition, follow-up rates were lower in region A compared with the other three regions, with 46 subjects (58%) followed-up.
Over all regions, 384 subjects and their 5528 condoms were successfully followed for up to 5 weeks, representing follow-up rates of 69.3% of people and 67.7% of condoms. The median time for which participants were followed was 35 days (range, 25–40 days). Comparing subjects followed-up with those lost to follow-up, participants successfully followed were more likely to be men, were more likely to have procured their condoms actively, reported more sexual partners, and were more likely to have used condoms previously, than those procurers lost to follow-up (Table 1).
The 5528 condoms traced during the study met one of five fates: (i) 2418 condoms were used in sex [43.7%; 95% confidence interval (CI), 42.4–45.1]. This includes condoms which were worn during sexual intercourse, as well as 144 condoms which broke during sex or while the condom packaging was being opened immediately before sex; (ii) 1202 condoms were given away (21.7%; 95% CI, 20.7–22.9). This involves condoms which were given to another individual; no condoms in this study were reported as sold; (iii) 473 condoms were wasted or lost before use (8.6%; 95% CI, 7.8–9.3). This includes condoms which were thrown away without use, or which were used for purposes other than sexual intercourse, as well as (vi) 12 condoms whose fate was unknown at the end of the study period; (v) 1435 condoms remained unused (26.0%; 95% CI, 24.8–27.1). This includes all condoms which were still available for use by a participant at the end of the follow-up period.
In Kaplan–Meier survival analysis, 34% of condoms were used 15 days after the start of the study and 53% after 30 days (condoms which were lost, discarded or given away were excluded from this analysis as they were not considered available for use by a participant). Subjects who asked for or took condoms (active condom procurers) used condoms at a rate of 5.5 per 30 days compared to a rate of 4.4 per 30 days among subjects who received their condoms without requesting them from health facility staff (or passive condom procurers) (crude incidence rate ratio, 1.22). In a multivariate model which included region as an effect modifier (Table 2), this trend appeared statistically significant in three of four study regions – although the direction of effect was opposite in one region. Other risk factors which were associated with the rate of condom use included gender, number of partners reported during the study, concern over pregnancy prevention, and prior education about condom use.
To the best of our knowledge, this is the only investigation of the fate of public sector condoms based on tracking individually identified condoms after they are procured. Our results on condom waste appear encouraging in the light of concerns about widespread condom waste associated with free male condom distribution [12,13]. The passage of condoms between peers after they have been distributed from health facilities has received little attention in the published literature on condom distribution or use. Such informal distribution networks may be used to help address well-documented problems in facility-based public sector condom accessibility .
The results of multivariate analysis indicate that the rate of public sector condom use is higher among subjects with more than one sexual partner during the study period compared with those reporting only one partner (in two regions), and higher among participants concerned with pregnancy prevention compared with those who were not (in one region). Meanwhile the rate of condom use was lower among women than in men across all regions, and among subjects who had received formal education about condom use compared with those who had received no such education (in one region). These findings suggest the diverse range of possible factors associated with condom use or non-use which require further investigation. However it is important to emphasise that these measures of association reflect the rate of public sector condom use through time, but do not reflect the proportion of protected sexual events, and thus the risk of contracting HIV or another sexually transmitted disease, among subjects. In addition, in interpreting these and other aspects of the multivariate analysis, the lower number of participants from small clinics in region A means that we do not have the same power here to detect associations as in other regions.
In addition, the data suggest that the way in which public sector condoms are procured may be associated with the rate at which they are eventually used. The unadjusted incidence rates shown here indicate that condoms procured actively are likely to be used more quickly than those procured passively, although the direction and magnitude of this effect varied between study regions when included in the regression model. While this interaction suggests that the precise nature of the association may vary, this is important preliminary evidence for the possible relationship between the kind of condom procurement and the quantity of public sector condoms which go on to be used in sex. Further research is required to explore the role of different forms of condom procurement and condom distribution on the proportion of condoms which are used in sex.
Based on these data we estimate that of the 198 million public sector condoms distributed in South Africa during 1999, at least 87 million condoms were used in sex (including those broken in sex). In a country of approximately 40 million people, this translates into slightly more than two condoms used in sex for each South African, or almost four condoms used in sex in 1999 for every individual of reproductive age (15–49 years) using 1996 census estimates . During the same period, within 5 weeks of distribution at least 17 million condoms were lost or discarded without being used.
These results must be viewed in light of several methodological limitations. The study's reliance on questionnaire data is likely to introduce an unknown degree of reporting bias, common to all social research on sexual behaviour . Despite a clear explanation of the study's purpose and explicit directions to the contrary, it is possible that participants in follow-up interviews may have tended to overestimate their reported use of condoms or that the study may have increased condom use. Additional bias may be introduced by the loss of participants to follow-up. Participants successfully followed were more likely to be male, were more likely to have procured their condoms actively, had more sexual partners, and were more likely to have used condoms previously, than those lost to follow-up. Both of these factors would probably lead to the results of this study overestimating condom use and underestimating condom non-use and waste.
The methodology developed for this research represents an innovative approach to the study of condom use and waste. For South Africa and other developing countries, the data may be used to estimate the potential impact of existing public sector condom programmes and to plan improvements to condom distribution strategies. In the future these methods can be used to evaluate different approaches to condom distribution and measure the impact of different condom-promotion interventions.
The authors thank C. Lombard, C. Morroni and A. Harrison for their comments on previous drafts of this manuscript.
1. De Vincenzi I. A longitudinal study of human immunodeficiency virus transmission by heterosexual partners. N Engl J Med 1994, 331: 341 –346.
2. Mann J, Quinn TC, Piot P. et al
. Condom use and HIV infection among prostitutes in Zaire. N Engl J Med 1986, 316: 345. 345.
3. Carey RF, Herman WA, Retta SM, Rinaldi JE, Herman BA, Athey TW. Effectiveness of latex condoms as a barrier to human immunodeficiency virus-sized particles under conditions of simulated use. Sex Transm Dis 1992, 19: 230 –234.
4. Nelson KE, Celentano DD, Eiumtrakol S. et al
. Changes in sexual behaviour and a decline in HIV infection among young men in Thailand. N Engl J Med 1996, 355: 297 –303.
5. Davis KR, Weller SC. The effectiveness of condoms in reducing heterosexual transmission of HIV. Fam Plann Perspect 1999, 31: 272 –279.
6. Swanevelder JP, Kustner H, Middelkoop AV. The South African HIV epidemic reflected by nine provincial epidemics, 1990–1996. S Afr Med J 1998, 88: 1320 –1325.
7. Abdool Karim Q, Abdool Karim SS. South Africa: Host to a new and emerging HIV epidemic. Sex Transm Infect 1999, 75: 139 –140.
8. Warren M. Condom use in South Africa: facts and fantasies. AIDS Bull 1997, 1: 4 –6.
9. South African Department of Health, HIV/AIDS & STD Directorate. Condom use in South Africa. Red Hot News 1998, 3: 1 –2.
10. Colvin M, Abdool Karim SS, Connolly C, Hoosen AA, Ntuli N. HIV infection and asymptomatic sexually transmitted infections in a rural South African community. Int J STD AIDS 1998, 9: 548 –550.
11. South African Department of Health. South African Demographic and Health Survey: Preliminary Results.
Tygerberg, South Africa: Medical Research Council, Department of Health and Macro International, 1999.
12. Lamptey PR, Price JE. Social marketing sexually transmitted disease and HIV prevention: a consumer-centred approach to achieving behaviour change. AIDS 1998, 12 (suppl 2): S1 –S9.
13. Meekers D. The implications of free and commercial distribution for condom use: evidence from Cameroon.
PSI Research Division Working Paper No. 9. Washington, DC: Population Services International, Research Division, 1997.
14. Gilmour E, Abdool Karim SS, Fourie H. Availability of condoms in urban and rural areas of KwaZulu-Natal, South Africa. Sex Transm Dis 2000, 27: 353 –357.
15. Statistics South Africa. 1 996 Census in Brief, 3rdEdn.
Pretoria: Statistics South Africa, 1999.
16. Ferry B, Deheneffe J-C, Mamdani M, Ingham R. Characteristics of surveys and data quality. In: Sexual behaviour and AIDS in the Developing World.
Edited by Cleland J, Ferry B. London: Taylor and Francis; 1995: 10 –42.