THERE IS AN URGENT NEED for methods that women can use to protect themselves against sexually transmitted disease (STD). The need is particularly critical in the case of human immunodeficiency virus (HIV) because this disease continues to spread among women in the United States and is devastating communities in Africa and Asia. 1 Given the disappointing results from a large clinical trial evaluating the effectiveness of a vaginal preparation containing nonoxynol-9 against HIV, the need to exploit the potential of every available prevention method is even more critical. 2
A host of acceptability studies across numerous populations and settings have shown that the female condom is a realistic alternative to the male condom for STD/HIV protection. 3–10 Other studies have shown it to be a proven contraceptive. 11,12 Yet the female condom is still not fully accepted as providing dual protection equal to that of the male condom against both pregnancy and STD. 13,14 Our objective was to test the hypothesis that women counseled on the use of, and given, female condoms were no more likely than those counseled on the use of, and given, male condoms to present with incident cases of STD.
There are compelling reasons to expect that the female condom will provide protection that is at least equal to that of the male condom. The female condom is a prelubricated polyurethane pouch that, once inserted, lines the vagina and extends outside it to protect the outer labia. The device operates on the same principle as the male condom in that it prevents contact between mucosal surfaces and therefore would be expected to provide a similar degree of protection against STDs. In fact, the female condom has three design advantages over the male condom that may bolster the female condom’s method-effectiveness as well as increase the chances that it will be used, thereby enhancing its use-effectiveness.
First, polyurethane can be used with either oil- or water-based lubricants—an important feature, given the continued confusion in the general population about how to properly lubricate latex male condoms. 15–17 Second, the female condom provides more extensive coverage of the female external genitalia than the male condom. Finally, unlike male condom use, the female condom can be used at the initiative of the woman, without the active participation of the male partner. 4
Availability of female condoms remains an issue. Although female condoms cost more than male condoms, there are many public-sector distribution programs that provide the female condom at no or low cost.
The distinction of the latex male condom as the gold standard method for STD prevention was codified early in the AIDS epidemic when, in 1987, the US Food and Drug Administration (FDA) expanded the labeling of male condoms to include protection against STDs, including HIV. 18 Although there has been no randomized trial investigating the effectiveness of the latex male condom in preventing STDs, subsequent observational studies of the male condom have supported expanded labeling for several STDs. 19 A meta-analysis of the male condom in preventing HIV suggests that their effectiveness is 87% (60–96%), with the lower bound emphasizing the importance of consistent and correct use in determining use-effectiveness. 20
In 1993 the FDA classified the female condom as a class III medical device and required labeling that read, “If you are not going to use a latex male condom, you can use Reality to help protect yourself and your partner.” This labeling placed the female condom in a distinctly inferior position relative to the male condom. Results from in vitro studies conducted before approval by the FDA have shown the female condom to be impermeable to both cytomegalovirus and HIV. 21 Subsequently, three studies have investigated in vivo STD effectiveness. The first, a small observational study investigating the prevention of trichomoniasis, found no infections among women who used the devise consistently, but it lacked the power to detect differences between users and nonusers. 22 Nonrandom assignment in this study also limited inference.
Randomization was achieved in a Thai study where 71 brothels employing 504 sex workers were assigned to either a male-condom-only group or a group instructed to use female condoms if a male condom could not be used. 23 This study did not allow for a direct comparison between a group randomized to male condoms or female condoms only. However, this study did show that the use of the female condom as an alternative to the male condom was associated with a reduced but not statistically lower STD rate, compared with use of the male condom alone (2.81 vs. 3.69 per 100 person-months of observation;P = 0.18). A third, community-level intervention trial conducted in Kenya found no differences in STD incidence between groups provided with male condoms or with both male and female condoms. 33
Thus, data on the effectiveness of the female condom against STDs are limited. Yet it remains impossible to conduct the strongest possible study design: a randomized double-blind trial comparing the female and male condoms. 24 In the current study, we accomplished a study design that combined the random allocation of women to a male condom or to a female condom group and prospectively followed them for incident STDs, with the screening for these outcomes performed with blinding to study group. Unlike the Thai study, we encouraged those in the female condom group to use female condoms consistently, not only when they had failed with the male condom. We tested the hypothesis that women assigned to counseling on use of the female condom were no more likely than those assigned to counseling on use of the male condom to present with incident cases of any of the following four treatable STDs: gonorrhea, chlamydia, early syphilis, and trichomoniasis.
The study sample comprised the female patient population clinically evaluated on at least one occasion at the main public STD clinic in Philadelphia between May 1, 1995, and April 12, 1996 (intervention period). The clinic provided free evaluation and treatment for STDs on a walk-in basis and served a predominantly low-income, minority population. Exclusions were limited to those women attending solely for the purpose of anonymous HIV testing and those brought to the clinic as part of contact-tracing since women in these two groups were not registered in the medical database. All other women who attended the clinic during the specified period were included in the study. The study comprised a total of three intervention arms: a male condom arm, female condom arm, and safer-sex hierarchy that presented a range of barrier methods. Results from the hierarchy arm have been reported elsewhere. 25 This analysis is limited to the comparison between the two arms that provided education and supplies of either male condoms or female condoms.
Women in this study were assigned to an intervention arm and were prospectively followed through their medical records for the occurrence of STDs. At their first clinic visit during the intervention period, women were assigned to one of three interventions, depending on the week of their initial clinic visit. Assignment to an intervention arm was determined at the study outset and occurred on the basis of the week of initial visit, rather than by person, to minimize cross-contamination of patients during their initial visit. This also allowed for the integration of the study into existing clinic services.
Weeks within a given month were randomly dedicated to the provision of either the male condom or the female condom intervention. The randomization schedule was filed off-site and was divulged to clinic staff only at the beginning of each intervention week. Months dedicated to the condom interventions alternated with entire months dedicated to the third intervention (reported on elsewhere).
Women received the assigned intervention—efficiently delivered in small groups—during the waiting period between registration and clinical examination. At the first clinic visit during the intervention period, each woman’s medical chart was coded with her intervention assignment, which then became part of her medical record. Consistency in intervention assignment was maintained by the study staff through any subsequent clinic visit through October 31, 1996. This corresponded to a minimum of 6 months of follow-up for women who initially attended the clinic at the end of the intervention period and a maximum of 12 months for those who initially attended the clinic at the beginning of the intervention period. Women were not asked to return to the clinic on a specific screening schedule, but whenever a woman returned to the clinic for medical care during the follow-up period, she was screened for all four STDs under study.
The protocol for this study was approved by the institutional review boards of the city of Philadelphia, Columbia University, and Hahnemann University. Approval was granted for the clinic to vary its counseling procedure, without requiring women to provide their expressed, written consent. This procedure made it possible to implement the intervention without disrupting clinic routine and minimized contamination among participants. The counseling procedure in the clinic before the study was brief, and since it was limited to use of the male condom, clients had been deprived of counseling on the female condom.
In our study, those assigned to the male condom arm were given “enhanced male condom counseling,” which was more intensive than prior clinic practice. Those assigned to the female condom arm were virtually all familiar with the male condom, so counseling on the female condom gave them something extra. Thus, both study groups experienced an improvement over prior clinic procedure. This research, which tested alternative, enhanced service-delivery methods, was deemed to present minimal risk to participants and allowed for the waiver of written consent. In addition, outcome data were analyzed only after being linked with identification numbers and stripped of personal identifiers.
Throughout the trial, the prevention intervention was delivered to all study subjects by one of the same three female health counselors. Both interventions were delivered via interactive, multimedia, small-group sessions that included a video, condom samples, posters, pamphlets, and pelvic and penis models to practice condom application. The objective was to deliver intervention messages that differed in content but not in quality. Delivery was usually to groups of up to five women but occasionally occurred on a one-on-one basis. All intervention sessions took place in a private room and lasted between 10 and 20 minutes. Details on the content and approach of the intervention have been described elsewhere. 25–26
Male Condom Message
The male condom sessions included demonstration of the proper application and use of a male condom on a penis model by the health counselor; practice of the application by participants; review of how to avoid common use and lubrication mistakes; and review of condom negotiation language. Participants were encouraged to use male condoms during all types of sexual intercourse. Women were not encouraged to stop using any other method of protection.
Female Condom Message
The female condom sessions included a review of the features of the device; demonstration by the health counselor of insertion on a pelvic model; practice on the pelvic model by participants; review of proper use; provision of troubleshooting tips; and a brief overview of female reproductive anatomy (i.e., location of the cervix and pubic bone) to aid in teaching proper insertion of the device. Participants in the female condom arm were encouraged to use the female condom during vaginal and anal sex and were instructed to purchase flavored male condoms for oral sex (not supplied by the study). Women were not advised to stop using other methods.
Throughout the study, women were given a free supply of their assigned condoms. At each clinic visit, women were given at least a 2-month supply of condoms and a prepaid postcard, stamped with their clinic identifier, to request that additional free condoms be mailed in a plain envelope (passive distribution). When initial postcard return rates were lower than expected, all women were then offered a choice of taking home a 6-month supply of condoms or taking home a 2-month supply and giving consent for clinic staff to telephone at 2-month intervals to determine if they needed additional supplies (active distribution). In addition, women could come into the clinic at any time to obtain additional condom supplies. Thus, women were supplied with their assigned condom for at least 6 months of follow-up.
At the time the study was conducted, female condoms were not widely available. Hence, women in the male condom arm had little access to female condoms, and data from a substudy confirmed that women in this arm rarely used the female condom. 26 In contrast, women in the female condom arm had continued access to male condoms from sources outside the clinic, and findings from the substudy revealed that male condoms accounted for one third of condom-protected sex acts. Thus, in practice, the two arms are most accurately described as a group of women who were counseled on and supplied with male condoms, versus a group of women who were counseled on and supplied with female condoms, who also had access to male condoms obtained from other sources.
Baseline Sample Information
Baseline data were collected at the first clinic visit during the intervention period and therefore represent preintervention sample characteristics. Demographic and contraceptive information, collected immediately preceding the physical examination, included age; race; current use of male condoms, diaphragm, spermicides, or hormonal methods (defined as use of the oral contraceptive pill Depo-Provera [Pharmacia & Upjohn, Kalamazoo, MI] or Norplant [Wyeth-Ayerst Laboratories, Philadelphia, PA]); hysterectomy or tubal ligation; and ever using a diaphragm, cervical cap, contraceptive sponge, or spermicides.
Demographic data were extracted from the clinic records, and STD data were extracted from medical records and the reportable disease repository (described below). All data were extracted for analysis at the end of the follow-up period, and the person extracting the data was blind to the intervention status.
Definition of Outcome
The outcome was defined as a postintervention incident laboratory-diagnosed infection with any of the following four STDs: gonorrhea, chlamydia, early syphilis, and trichomoniasis. The outcome comprised these four STDs because they can be treated, and therefore a new infection during the follow-up period could be considered a truly incident case. Gonorrhea was diagnosed with standard GC culture, chlamydia with the Gen-Probe Pace 2C (AccuProbe, San Diego, CA), and early syphilis with standard RPR/MHA or FTA analysis. Trichomoniasis was diagnosed via wet mount and was considered present if any parasites were seen on microscopy.
All city laboratories and clinicians followed CDC guidelines for STD diagnosis and treatment. 27 All city clinics made blood and cervical samples blind to intervention assignment at off-site laboratories for the diagnosis of gonorrhea, chlamydia, and syphilis. At the intervention clinic, diagnosis of trichomoniasis was performed on-site by the examining clinician, who was aware of the intervention assignment.
Women were screened for all four STDs at each clinic visit (baseline and follow-up). The only exception was when women were asked to return 2 weeks after a screening visit to learn their test result. Women presenting with STD-like symptoms were presumptively treated at the time of their visit. Women who were asymptomatic at screening but found positive for an STD through laboratory diagnosis were treated 2 weeks later at their results visit. If a woman did not return at 2 weeks, she was actively sought and brought to the clinic for treatment by health department personnel.
The intervention site was one of nine health centers in the city’s public clinic system and was the central repository for all reportable diseases diagnosed in the city. Therefore, if a study subject had gonorrhea, chlamydia, or early syphilis diagnosed at any of the nine city health centers or by a private, reporting physician, the diagnosis was captured in the central disease registry. This database allowed us to evaluate STD outcomes for participants who did not return to the intervention clinic. The exception was trichomoniasis, which was captured only from those diagnoses made at city health clinics and not by private physicians, since it is not a reportable disease.
The outcome was dichotomized as either absent or present, and was defined as at least one occurrence of one of the four STDs during the follow-up period. Women presenting with more than one STD at a single visit were counted only once. To avoid double-counting, an infection was coded as incident if it was not preceded by a prior laboratory diagnosis of the same STD that had not been appropriately treated. Baseline STDs were those diagnosed during a first visit to the intervention clinic between May 1, 1995, and April 12, 1996. Follow-up STDs were those diagnosed during a visit subsequent to baseline that occurred between May 15, 1995, and October 31, 1996. Each postintervention STD was manually abstracted from participants’ medical records (with links to the disease registry). Each STD outcome was coded as incident only after review by two experienced STD clinicians, both of whom were blind to intervention assignment at the time of coding.
A programmer blind to intervention status combined the data into a separate database and performed the analysis. Chi-square and t tests were used to compare baseline characteristics of women assigned to each intervention arm. To determine if the observed follow-up data were biased by differential risk profiles of women returning for screening, we evaluated the characteristics of “returners” in each arm. Among those women who returned, we calculated the proportion of baseline covariates and used chi-square P values to compare these proportions between arms. We compared the two arms on demographic characteristics, baseline method use, and STD prevalence, as these variables were thought to reflect STD risk. To check whether there was heterogeneity in the diagnosis of trichomoniasis, we calculated an odds ratio for the association between the intervention arm and the outcome during the follow-up period, with the outcome defined as all of the STDs except trichomoniasis (masked diagnoses), and for trichomoniasis alone (unblinded diagnosis). The odds ratios were similar, and therefore trichomoniasis was included in the outcome in all further analyses.
Two statistical approaches were used to compare the outcome between arms. First, in an intention-to-treat analysis, we calculated a maximum-likelihood estimate of the odds ratio from an unadjusted logistic regression model and from a model that adjusted for variables associated with STDs. Women who had no subsequent visits to the clinic were included in the analysis and coded as being free of STDs. Because of the change from passive distribution of condoms in the first 2 months of the study to more active distribution throughout the rest of the intervention period, this analysis was also performed with stratification by type of condom distribution (passive vs. active). Since findings were similar for the two distribution methods, a Mantel–Haenszel estimate of the odds ratio was used to combine the data across strata.
In the second approach, only those women who returned to the clinic at least once during the follow-up period were included in the analysis. An STD incidence rate for the first postintervention STD was calculated for each arm, and the two arms were compared with use of an incidence rate ratio. A woman contributed person-months of observation between her baseline visit and the date of her first postintervention STD diagnosis or the last date of her disease-free postintervention visit, with all follow-up censored at 12 months of postintervention.
A total of 1442 women were assigned to either the male condom (n = 587) or female condom (n = 855) intervention. The smaller size of the male condom arm was due to the fact that slightly fewer weeks were allocated to this arm and that some of these weeks occurred near major holidays, when there was low clinic attendance. The sample was 85% black and had a mean age of 28 years (Table 1). Apart from a lower history of contraceptive sponge use in the female condom arm, baseline characteristics of the two arms were similar. The prevalence of STDs at baseline was similar across arms (32.8% and 33.1% for any of the four STDs in the female and male condom arms, respectively;P = 0.91).
There was no difference in the rate of follow-up or in the risk profiles across arms (Table 2). Half of the sample returned during the follow-up period for additional STD screening, and women averaged two follow-up visits. Returners in the male and female condom arms were equally likely to be teenagers (23.0% and 21.2%, respectively;P = 0.57), to be black (85.0% and 85.1%;P = 0.91), and to have been diagnosed with an STD at baseline (34.0% vs. 32.6%;P = 0.70).
A total of 199 women were diagnosed at least once with the outcome. The percentages who had an STD diagnosed were 12.4% (106/855) in the female condom arm and 15.8% (93/587) in the male condom arm (unadjusted OR = 0.75 [95% CI, 0.56–1.01];P = 0.06). The point estimate neither changed with adjustment nor differed by method of condom distribution (Table 3). Incidence rates for STDs, per 100 woman-months of observation, were 6.8 in the female condom arm and 8.5 in the male condom arm (0.79 [0.59–1.06];P = 0.11). Results of the two methods of analyses were similar: the point estimates comparing STDs in the two arms were inverse and favored the female condom arm, but the confidence interval narrowly included the null value.
In this study, women were given instructions on how to use a given condom type, were given abundant supplies of condoms, and were then compared for incident STDs over time. The objective was to compare the rate of STDs over time in a group encouraged to use female condoms to that in a group encouraged to use male condoms. In both analyses, point estimates favored the female condom, and the upper limit of the confidence intervals only narrowly included the null value. The results certainly allow us to reject the hypothesis that those in the female condom arm were less protected from STDs than those in the male condom arm. Thus, at the very least, women in the female condom arm fared no worse than those in the male condom arm. However, the data suggest that when male condoms are already available, a small protective association may exist between a program providing counseling on, and supplies of, female condoms and a similar program focused solely on male condoms, as most of the values within the interval were statistically compatible with a finding that favored the female over the male condom arm. 28
Our finding was consistent across two analytical approaches, despite differing assumptions about those who did not return for subsequent STD screening. We used an intention-to-treat analysis and assumed that those who did not return remained disease-free. We also calculated the incident rate ratio, which employed no such assumption, and instead included only those who returned to the clinic at least once. Under both methods of analyses, the point estimates were similar in strength, but the confidence interval widened for the rate ratio, reflecting decreased precision due to fewer contributing subjects.
The delivery of this intervention was integrated into the existing services of an STD clinic. A traditional randomization approach was impossible. We strove to minimize bias by withholding randomization assignment from the clinic staff until the beginning of the first day of each week, and we believe this approach was successful. 29 Therefore, some aspects of the study design were circumscribed by the study setting. Nevertheless, we retained several critical components of a controlled trial. We determined the assignment of women to intervention condition; retained tight control of the intervention “dose” by delivering the intervention via a small, trained staff; maintained integrity of the initial assignment during each subsequent visit to the clinic through coded computerized medical records; and performed diagnoses, outcome coding, and analysis blind to intervention assignment.
In addition, the sample was assembled from the entire care-seeking population at a single clinic and was not based on a group predisposed to a particular condom type (i.e., male vs. female condoms) or to condoms in general. In fact, prevalence of condom use at the study outset was on par with that observed in population-based samples. 30
Two major limitations of the study relate to follow-up procedures and limited data on the use of condoms among all included in this analysis. We did not screen women for follow-up STDs via a prespecified schedule; instead, we relied on capturing this information from the medical records and centralized STD reporting database. There was no evidence of differential return from the two arms in the number or type of women who returned, since the proportion and risk profiles of “returners” were similar across arms. This lessened the possibility that findings merely reflected a preponderance of higher-risk women returning in the male condom arm or lower-risk women returning in the female condom arm. While there was the possibility that we did not detect all outcomes, three of the four outcomes were reportable diseases and should have been captured by the health department’s disease registry. In addition, follow-up across arms was similar, so any undercounting of the outcome was likely to have been nondifferential by arm, allowing us to draw inferences from the data as presented.
We did not have direct measures of the extent to which women actually used condoms during follow-up. Rather, we used randomization to balance the propensity to use condoms across arms and used an intention-to-treat analysis. Nevertheless, we have some information on condom use from data collected during a substudy involving volunteers from each arm. In this study, women in both arms reported increased use of their assigned condom during 6 months of follow-up. 26 Among women in the female condom arm, the mean proportion of condom-protected coital acts at baseline was 26%, and this increased to 74% at 6 months’ postintervention, with female condom use accounting for two thirds of condoms used. Thus, while women in the female condom arm were supplied with only female condoms, they also used male condoms obtained from other sources.
Our data suggest that those in the female condom arm were better protected than women in the male condom arm. One reason may be the female-controlled nature of the female condom. For example, in our conversations with women about the female condom, we heard comments such as: “If I put this in, there is nothing he can do about it.” Another explanation may be the improved negotiating power of women to use any type of condom when the female condom is available as an option. From our substudy, we know that when women in the female condom arm used condoms, they used male condoms about one third of the time. If this was also the case in the larger study population analyzed here, these results suggest that alternating between condom types led to an overall increase in protection.
One reason for this boost may be that, in the absence of another method, if a man refuses to use a male condom, the “negotiation” is over. However, when women have the female condom as an option, men may be more likely to engage in a discussion about protection, since the nature of the negotiation may change from a question of simple use or refusal to a discussion of which type of condom to use. In fact, when we asked women about their condom use, we had responses such as: “We decided to take turns—sometimes he used his, and sometimes I used mine.” Other work has suggested that providing women with “their own” protection options, coupled with basic body education, can lead to an increased sense of entitlement to protection, to self-confidence, and to a perception of empowerment that enables women to take steps to achieve higher levels of protection. 13 Other contributing factors may be features of the device itself, such as the strength of polyurethane, and the additional coverage of the external female genitalia.
Despite our intervention message, which involved the selective encouragement for women in the female condom arm to use female condoms, women in this group used the female condom in alternation with the more familiar male condom—a phenomenon consistent with reports from other studies. 31,32 Our STD results were also in line with the Thai study in that there was a trend toward reduced STDs in the arm having access to female condoms. 23 Taken together with the Thai study, findings on STD effectiveness of the female condom are encouraging since they support a similar conclusion but are from diverse populations.
The current study suggests that women educated about and supplied with female condoms are at least as protected as (if not at lower risk of STDs than) women educated about and supplied with male condoms alone. Results of this study provide no support for the view that the female condom should be offered only as a “second-best” option after the male condom. Rather, this study suggests that the female condom is an effective condom option in the battle against STDs in women.
1. UNAIDS. Report on the global HIV/AIDS Epidemic. Geneva: World Health Organization, June 2000.
2. Van Damme L, Ramjee G, Alary M, et al. Effectiveness of COL-1492, a nonoxynol-9 vaginal gel, on HIV-1 transmission in female sex workers: a randomized controlled trial. Lancet 2002; 360: 971–977.
3. Shervington D. The acceptability of the female condom among low-income African-American women. J Natl Med Assoc 1993; 85: 341–347.
4. Gollub EL, Stein Z, el-Sadr W. Short-term acceptability of the female condom among staff and patients at a New York City Hospital. Fam Plann Perspect 1995; 4: 155–158.
5. Ruminjo JK, Steiner M, Joanis C, Mwathe EG, Thagana N. Preliminary comparison of the polyurethane female condom with the latex male condom in Kenya. East Afr Med J 1996; 73: 101–106.
6. UNDP/UNFPA/WHO/World Bank Special Programme of Research, Development, Research Training in Human Reproduction. The Female Condom. Geneva: World Health Organization, 1997.
7. el-Bassel N, Krishnan SP, Schilling RF, Witte S, Gilbert L. Acceptability of the female condom among STD clinic patients. AIDS Educ Prev 1998; 10: 465–480.
8. Sinpisut P, Chandeying V, Skov S, Uahgowitchai C. Perceptions and acceptability of the female condom [Femidom] amongst commercial sex workers in the Songkla province, Thailand. Int J STD AIDS 1998; 9: 168–172.
9. Witte SS, el-Bassel N, Wada T, Gray O, Wallace J. Acceptability of female condom use among women exchanging street sex in New York City. Int J STD AIDS 1999; 3: 162–168.
10. Artz L, Macaluso M, Brill I, et al. Effectiveness of an intervention promoting the female condom to patients at sexually transmitted disease clinics. Am J Public Health 2000; 90: 237–244.
11. Farr G, Gabelnick H, Sturgen K, Dorflinger L. Contraceptive efficacy and acceptability of the female condom. Am J Public Health 1994; 84: 1960–1964.
12. Trussell J, Sturgen K, Strickler J, Dominik R. Comparative contraceptive efficacy of the female condom and other barrier methods. Fam Plann Perspect 1994; 26: 66–72.
13. Gollub EL. The female condom: tool for women’s empowerment. Am J Public Health 2000; 90: 1377–1381.
14. Mantell JE, Hoffman S, Weiss E, et al. The acceptability of the female condom: perspectives of family planning providers in New York City, South Africa, and Nigeria. J Urban Health 2001; 78: 658–668.
15. Martin DJ. Inappropriate lubricant use with condoms by homosexual men. Public Health Reps 1992; 4: 468–473.
16. Steiner M, Piedrahita C, Glover L, et al. The impact of lubricants on latex condoms during vaginal intercourse. Int J STD AIDS 1994; 5: 29–36.
17. Sikkema KJ, Heckman TG, Kelly JA. HIV risk behaviors among inner-city African American women. Women’s Health 1997; 3: 349–366.
18. Gollub EL, Stein ZA. The new female condom: item 1 on a women’s AIDS prevention agenda. Am J Public Health 1993; 83: 498–500.
19. Cates W Jr. The NIH Condom Report: the glass is 90% full. Fam Plann Perspect 2001; 33: 231–233.
20. Davis KR, Weller SC. The effectiveness of condoms in reducing heterosexual transmission of HIV. Fam Plann Perspect 1999; 31: 272–279.
21. Drew WL, Blair M, Miner RC, Conant M. Evaluation of the virus permeability of a new condom for women. Sex Transm Dis 1990; 17: 110–112.
22. Soper DE, Shoupe E, Shangold GA, Shangold MM, Gutmann J, Mercer L. Prevention of vaginal trichomoniasis by compliant use of the female condom. Sex Transm Dis 1993; 20: 137–139.
23. Fontanet AL, Saba J, Chandelying V, et al. Protection against sexually transmitted diseases by granting sex workers in Thailand the choice of using the male or female condom: Results from a randomized controlled trial. AIDS 1998; 12: 1851–1859.
24. Stein Z, Susser M. Annotation. Prevention of HIV, other sexually transmitted disease, and unwanted pregnancy—Testing physical barriers available to women. Am J Public Health 1998; 88: 872–874.
25. Gollub EL, French P, Loundou A, Latka M, Rogers C, Stein Z. A randomized trial of hierarchical counseling in a short, clinic-based intervention to reduce the risk of sexually transmitted diseases in women. AIDS 2000; 14: 1249–1255.
26. Latka M, Gollub E, French P, Stein Z. Male-condom and female-condom use among women after counseling in a risk-reduction hierarchy for STD prevention. Sex Transm Dis 2000; 27: 431–437.
27. Centers for Disease Control and Prevention. Guideline for treatment of sexually transmitted diseases. MMWR Morb Mortal Wkly Rep 1998; 47(RR-1): 1–116.
28. Goodman SN, Berlin JA. The use of predicted confidence intervals when planning experiments and the misuse of power when interpreting results. Ann Intern Med 1994; 121: 200–206.
29. Schulz KF, Grimes DA. Unequal group sizes in randomized trials: guarding against guessing. Lancet 2002; 359: 966–970.
30. Catania JA, Coates TJ, Stall R, et al. Prevalence of AIDS-related risk factors and condom use in the United States. Science 1992; 258: 1101–1106.
31. Musaba E, Morrison CD, Sunkutu MR, Wong EL. Long-term use of the female condom among couples at high risk of human immunodeficiency virus infection in Zambia. Sex Transm Dis 1998; 25: 260–264.
32. Macaluso M, Demand M, Artz L, et al. Female condom use among women at high risk of sexually transmitted disease. Fam Plann Perspect 2000; 32: 138–144.
33. Felbum PJ, Kuyon MA, Bwayo JJ, et al. Female condom introduction and sexually transmitted infection prevalence: results of a community intervention trial in Kenya. AIDS 2001; 15: 1037–1044.