- Email: [email protected]
Three new female condoms: which do South-African women prefer?
Contraception 83 (2011) 248 – 254
Original research article
Three new female condoms: which do South-African women prefer? Carol Joanisa , Mags Beksinskab , Catherine Harta,⁎, Katie Tweedya , Jabu Lindab , Jenni Smitb b
a Family Health International (FHI), P.O. Box 13950, Research Triangle Park, NC 27709, USA Maternal, Adolescent, and Child Health (MatCH), Department of Obstetrics and Gynaecology, University of the Witwatersrand, Durban 4091, South Africa Received 2 June 2010; revised 29 July 2010; accepted 2 August 2010
Abstract Background: The widespread distribution of female condoms (FCs) in developing countries has been hindered by high unit cost, making new less expensive devices a priority for donor agencies. Study Design: Randomized, crossover study assessing product preference, safety, acceptability and function of three new FCs (PATH Woman's Condom, FC2 and V-Amour) among 170 women in Durban, South Africa. A subsequent “simulated market” study provided participants with free choice of FCs and assessed condom uptake over 3 months. Results: Of the 160 women who used at least one FC of each type, 47.5% preferred the PATH Woman's Condom (WC), 35.6% preferred FC2 and 16.3% preferred V-Amour (pb.001). Women rated the WC better than FC2 and V-Amour for appearance, ease of use and overall fit and better than V-Amour for feel. WC was rated worse than FC2 and V-Amour for lubrication volume. The simulated market demonstrated similar preferences. Total clinical failure rates (i.e., the types of failures that could result in pregnancy or STI) were low (b4%), regardless of condom type. Conclusions: Three new FC types functioned similarly and were generally acceptable. Most participants preferred WC and FC2 over VAmour, and WC was preferred over FC2 in several acceptability measures. © 2011 Published by Elsevier Inc. Keywords: Female condom; Function; Acceptability; Preference; Randomized trial; Barrier methods
1. Introduction Providing women in developing countries with female condoms (FCs) for pregnancy and HIV/STI prevention has been a high priority for donor agencies. The Reality® female condom (now known as FC1) was approved by the United States Food and Drug Association (US FDA) in 1993. Although FC1 was marketed worldwide for almost 20 years, distribution of this polyurethane condom has been hampered by a unit price much higher than that of male latex condoms. Manufacturers have recently developed new FC types made of less costly materials and design features that may make them more acceptable. The Female Health Company, Chicago, IL, USA, has developed the FC2®; a synthetic rubber (nitrile) condom that costs 30% less than FC1 but is similar in specification, function and appearance. A comparative functionality and acceptability study of the FC1 and FC2 devices among 276 women in Durban, South ⁎ Corresponding author. Tel.: +1 919 544 7040x11367. E-mail address: [email protected] (C. Hart). 0010-7824/$ – see front matter © 2011 Published by Elsevier Inc. doi:10.1016/j.contraception.2010.08.002
Africa, demonstrated that they have similar failure rates and are equally acceptable [1]. The FC2 was approved by the US FDA in March 2009, holds the CE Mark1 of the European Union (EU) and is distributed in 105 countries. Manufacture of FC1 ceased in October 2009. MedTech Products (Chennai, India) began developing a latex female condom in the early 1990s, culminating in the V-Amour®. A comparative acceptability study of V-Amour with FC1, conducted among 135 couples in New York City, found the two equally acceptable [2]. A subsequent study in India among 60 couples found that the V-Amour provided comparable or better acceptability results when compared to FC1 [3]. V-Amour bears the CE mark and is available in selected EU countries. PATH (Seattle, WA, USA) developed the WC female condom with several design innovations differentiating it from the aforementioned FCs. The WC has a polyurethane sheath with hydrophilic foam dots to anchor it within the 1 Certification given to products that fulfill the EU safety, health and environmental requirements.
C. Joanis et al. / Contraception 83 (2011) 248–254
vagina. A dissolvable polyvinyl alcohol insertion capsule encases the sheath, making the device easier to insert. The tampon-shaped capsule also aims to improve the aesthetics of the device. Unlike other FCs, the WC is not prelubricated and, for proper function, requires use of additional vaginal lubricant. A comparative crossover study of the WC and the FC1 was conducted in 2005 among 75 couples, with the WC exhibiting less breakage and slippage than the FC1. The WC was favored over FC1 2.6 and 2.0 times by females and males, respectively [4]. A clinical trial to assess the contraceptive effectiveness of the WC in comparison to FC2 is expected to begin in 2010. As new FCs enter the market, donors must decide which to purchase for wide-scale distribution in public health programs. Cost, which varies by volume of FCs purchased, will affect this decision, but consumer preference must be considered as well. In this study, the WC, FC2 and V-Amour were compared to determine which FC was preferred and to gather additional data on device safety, function and acceptability. Part 1 of the study was a three-period randomized crossover study; Part 2 was a “simulated market” in which women completing Part 1 could return to the clinic to receive free female condoms of their choice; and Part 3 employed in-depth interviews to elicit detailed reasons for participants' preference. This paper reports on the results of Part 1 and Part 2. 2. Materials and methods 2.1. Participants This study was conducted by the Reproductive Health and HIV Research Unit Durban (now MatCH) of the University of the Witwatersrand among female clients of the Commercial City Clinic in Durban, South Africa, between March 2007 and April 2008. Potential participants had to be in good general health with no known allergies to the study products, at least 18 years of age, using a reliable, non-barrier method of contraception and free of sexually transmitted infections (STI), as determined by pelvic examination and use of a syndromic diagnostic tool.2 Pregnant (according to urine pregnancy test) or breastfeeding women were excluded. Participants were required to be sexually active, monogamous and could not be practicing sex workers. Since take-home coital logs were used to gather data, participants were also required to be literate. Participants read and signed an informed consent form (ICF) prior to screening and enrollment. The protocol and ICF were approved by the institutional review boards of Family Health International (FHI) and the University of the Witwatersrand. The study was conducted in compliance 2 Syndromic management is the WHO recommended approach of identifying STIs in resource-limited settings. It is based on the identification of a group of symptoms and easily recognized signs (i.e., syndromes) which are associated with a number of well-defined etiologies.
249
with the protocol and the US Code of Federal Regulations [5], using Good Clinical Practice guidelines (ICH-E6) [6]. Study reporting was done according to CONSORT recommendations for improving randomized trials [7]. 2.2. Intervention and procedures To minimize use order bias, enrolled women were randomized to one of six condom use sequences based on a randomly varied permuted-block design with block sizes of 4, 6 and 8 using the SAS® System (SAS Institute, Cary, NC, USA) [8]. Randomization allocations were sealed in separate, opaque, sequentially numbered envelopes and assigned to participants in chronological order of enrollment. Due to the dissimilar appearance of the condom types, it was not possible to blind participants and clinic staff. Only study statisticians were definitively blinded until after primary analysis programs were written. Women were given five devices of the first FC type in their assigned-use sequence and asked to use them during five acts of intercourse with their partner over a three-week period. (During the WC use period, participants were also provided with five packages of Astroglide™ water-based lubricant, manufactured by BioFilm, Vista, CA, USA). Participants were instructed not to use male condoms while wearing a female condom, to complete a coital log after each FC use and to return to the clinic after use of the fifth device. At the follow-up visit, an interviewer-assisted questionnaire was used to ask women their opinions about — and experiences with — that FC type, including any condom device failures or adverse events (AEs). Data from the questionnaire were compared with the coital log to ensure consistency of response and determine FC use. This process was repeated for the second and third FC types according to the assigned use sequence. After use of the third FC type, women were either discontinued from the study or progressed voluntarily into Part 2. In Part 2, women were provided with up to 10 FCs of their choice (all of one type or a mix of types). If a participant ran out of study FCs, she was instructed to return to the clinic, for up to three visits, to collect more. Male condoms were also available at the clinic if the participants ran out of study FCs. Although the FC1 was also available at the clinic, we asked participants not to use the clinic supplies, but to return to the study nurse to obtain their FCs. At each follow-up visit, the participants were administered a questionnaire about their choices, acceptability of the FCs selected and whether they experienced any problems while using the FCs. Participants' self report was used to determine medical problems and FC use. At their last visit, women were either discontinued from the study or progressed voluntarily into Part 3 (qualitative methodology not described here). Women were paid 50 South African Rand (∼US⁎$8.00) per clinic visit for their time and travel expenses. Potential AEs were auto-coded to the MedDRA dictionary (Version 6.1) by FHI's data management system. For
250
C. Joanis et al. / Contraception 83 (2011) 248–254
AEs that could not be auto-coded, the study investigators assigned codes blind to FC type. The principal investigator (PI) and study medical officer subsequently excluded reported conditions that were not considered AEs in an adjudication process. 2.3. Study objectives The primary objective was to determine which, if any, of the three FC types was most preferred. Secondary objectives included gathering detailed information on safety, function and acceptability of each FC type. 2.4. Statistical methods The study planned to enroll 180 women, with the expectation that at least 150 would complete Part 1. With 150 completing Part 1, the study would have 80% power to conclude that a difference in preference exists, as long as one of the FC types was favored by at least 45% of the women (using a two-sided, .05 level of significance test). It was expected that 130 women would continue to Part 2, with at least 100 completing. With 100 completing Part 2, the study would have approximately 90% power to conclude that a
difference in preference exists, so long as one of the three FC types was favored by at least 50% of the women. A two-degree of freedom chi-squared test was used to compare overall product preference probabilities (“Which condom did you like best?”) across the three FC types in Part 1, excluding women who did not use at least one condom of each type. If the overall analysis was significant, pair-wise comparisons were to be performed using data from the subset of participants who preferred one of the FC types. Response options to each of the eight acceptability items ranged from 1=Disliked very much to 5=Liked very much and were compared across FC types using Friedman's test. p Values were not adjusted for multiple analyses. Condom function was assessed based on the clinical failure modes which could lead to semen exposure (i.e., misdirection, breakage, invagination and slippage) defined by Beksinska et al. [9] and endorsed by the World Health Organization. The relative performance of the three FC types in Part 1 was compared for two main function outcomes, clinical breakage and total clinical failure, using generalized linear equation (GEE) methods with identity link to account for clustering of responses within participants. The primary FC uptake analysis during the simulated
Fig. 1. Participant flow diagram.
C. Joanis et al. / Contraception 83 (2011) 248–254
market phase used GEE with a logit link to compare the odds of each FC type being chosen, with FC2 serving as the reference condition. 3. Results 3.1. Recruitment and follow-up One hundred eighty women were enrolled in Part 1 of the study, with 170 attending at least one follow-up visit and providing some follow-up data. One participant was discontinued early because she was unable to read and could not complete the coital log and nine women were lost to follow-up (Fig. 1). Other than being somewhat older (median age 29.5 vs. 26), the 10 women who did not provide data in Part 1 had no obvious differences from the 170 who did. Of the 160 women who completed Part 1 (i.e., used at least one condom of all three FC types), 148 enrolled in the Part 2 simulated market. Of these, four were lost to followup, and 12 were discontinued early due to reasons not related to the study products. There were no important differences in Table 1 Participant characteristics at baseline Age, years Mean (S.D.) Median Range Race, n (%) Black Colored White Indian Years of school completed Mean (S.D.) Median Range Primary occupation, n (%) None Self-employed Unskilled labor Health/medical Office Public service/government Sales Student Other Number of living children, n (%) Mean (S.D.) Median Range Marital relationship, n (%) Married and living together Married and not living together Not married and living together Not married and not living together Contraception method, n (%) Contraception pills Injectable contraception Tubes tied Total contributing data to Part 1 (crossover study)
(98.8) (0.6) (0.0) (0.6)
11.2 (1.43) 12 1–12 77 5 7 13 2 1 7 34 24
demographic or behavioral characteristics between participants who continued to Part 2 and those who did not continue (data not shown). Study participants ranged in age from 18 to 48 years (mean=28.0), and all but two were African (Table 1). Mean educational level was 11.2 years (range 1–12). Almost half (45.3%) were unemployed, 20% were students and the remaining women held various professional or semi-skilled positions. Ninety-one percent of women were not married and 24% resided with their male partners. Most (94%) had been with their male partners for over one year and 82% had at least one child (range 0–6). Previous use of male condoms was high, with 81% reporting use of a male condom in at least half of their sex acts during the previous 6 months. The most common contraceptive method used was progestin injectables (72%). At the time of this study, the FC1 condom was supplied free-of-charge at the clinic [1], although only 24 women (14.1%) reported previous use of a FC. When these women were asked what they would change about FCs, three themes emerged: FCs should be more readily available (n=5); FCs should be easier to insert (n=2), and there should be more FC options from which to choose (n=5). Seven women said nothing about the FCs should be changed. 3.2. Condom preference
28.0 (7.18) 26 18–48 168 1 0 1
251
(45.3) (2.9) (4.1) (7.6) (1.2) (0.6) (4.1) (20.0) (14.1)
1.4 (1.17) 1 0–6 12 3 28 127
(7.1) (1.8) (6.5) (74.7)
35 123 12 170
(20.6) (72.4) (7.1)
A total of 160 women used at least one condom of each type in Part 1. Of these, 76 (47.5%) preferred the WC, 57 (35.6%) preferred FC2 and 26 (16.3%) preferred V-Amour (one woman stated no preference), differences that were statistically significant overall (pb.001). Subsequent pairwise comparisons found that WC and FC2 were significantly preferred over V-Amour (pb.001 and p=.007, respectively), whereas there was a nonsignificant preference for WC over FC2 (p=.10). When asked which FC type they preferred the least, 58.8% cited V-Amour, followed by FC2 (24.4%) and WC (16.9%). Partner preferences (as reported by the participants) followed similar trends. A total of 4320 condoms were distributed in Part 2. The WC was chosen most frequently, with 1898 (44%) condoms distributed versus 1585 (37%) and 837 (19%) for FC2 and V-Amour, respectively. The mean exposure for each FC type was 12.8, 10.7 and 5.7 uses for WC, FC2 and V-Amour, respectively. The overall difference in the proportions of FC types selected was significant (pb.001). Pair-wise comparisons found no significant difference in odds of selecting more WC than FC2 condoms (p=.24), but a significant increased odds of selecting FC2 over V-Amour (p=.001). 3.3. Acceptability Of eight acceptability parameters compared between FC types in Part 1, five were statistically significant: feel/ sensation, lubrication volume, appearance, ease of use and fit (Table 2). There was no significant difference in preference with regard to scent, color or length. The participants rated
252
C. Joanis et al. / Contraception 83 (2011) 248–254
Table 2 Comparative acceptability outcomes by female condom type Acceptability measure
Overall Preference Pair-wise comparisons Feel/Sensation Disliked very much Disliked somewhat Neither liked nor disliked Liked somewhat Liked very much Pair-wise comparisons Amount of Lubrication Disliked very much Disliked somewhat Neither liked nor disliked Liked somewhat Liked very much Pair-wise comparisons Appearance Disliked very much Disliked somewhat Neither liked nor disliked Liked somewhat Liked very much Pair-wise comparisons Ease of Use Disliked very much Disliked somewhat Neither liked nor disliked Liked somewhat Liked very much Pair-wise comparisons Overall Fit Disliked very much Disliked somewhat Neither liked nor disliked Liked somewhat Liked very much Pair-wise comparisons a
PATH WC (n=165)
FC2 (n=164)
n
n
(%)
V-Amour (n=164) (%)
n
(%)
p valuea
76 (47.5) vs. FC2: p=.10 vs. V-Amour: pb.001
57 (35.6) vs. V-Amour: p=.007
26
(16.3)
b.001
7 (4.2) 10 (6.1) 13 (7.9) 68 (41.2) 67 (40.6) vs. FC2: p=.221 vs. V-Amour: p=.014
4 (2.4) 7 (4.3) 29 (17.7) 70 (42.7) 54 (32.9) vs. V-Amour: p=.033
7 12 33 70 42
(4.3) (7.3) (20.1) (42.7) (25.6)
.014
9 (5.5) 29 (17.6) 25 (15.2) 68 (41.2) 34 (20.6) vs. FC2: p=.004 vs. V-Amour: pb.001
5 (3.0) 14 (8.5) 18 (11.0) 75 (45.7) 52 (31.7) vs. V-Amour: p=.307
4 7 21 71 61
(2.4) (4.3) (12.8) (43.3) (37.2)
b.001
6 (3.6) 14 (8.5) 12 (7.3) 66 (40.0) 67 (40.6) vs. FC2: p=.023 vs. V-Amour: p=.011
3 (1.8) 22 (13.4) 20 (12.2) 71 (43.3) 48 (29.3) vs. V-Amour: p=.536
7 28 20 54 55
(4.3) (17.1) (12.2) (32.9) (33.5)
.014
2 (1.2) 7 (4.2) 4 (2.4) 61 (37.0) 91 (55.2) vs. FC2: p=.002 vs. V-Amour: pb.001
3 (1.8) 4 (2.4) 16 (9.8) 74 (45.1) 67 (40.9) vs. V-Amour: pb.001
6 21 24 65 48
(3.7) (12.8) (14.6) (39.6) (29.3)
b.001
5 (3.0) 4 (2.4) 11 (6.7) 78 (47.3) 67 (40.6) vs. FC2: p=.061 vs. V-Amour: p=.002
3 (1.8) 11 (6.7) 11 (6.7) 86 (52.4) 53 (32.3) vs. V-Amour: p=.254
7 16 23 67 51
(4.3) (9.8) (14.0) (40.9) (31.1)
.008
p Values are from Friedman’s test comparing distribution of responses across all 3 FC types.
the WC better than both FC2 and V-Amour in terms of appearance, ease of use and overall fit and better than VAmour in terms of feel. The WC was rated worse than both FC2 and V-Amour with respect to lubrication, notably the inconvenience of self-lubrication and indecision about volume of lubricant to use. 3.4. Functionality Out of 2423 condom packages opened during Part 1 of the study, 2376 were used during intercourse. Clinical function problems were reported in less than 2% of devices used (Table 3). Total clinical failure (functional problems that could lead to a pregnancy or sexually transmitted disease)
occurred in 3.6% of uses overall and was comparable across FC types (3.6%, 3.9% and 3.3% for WC, FC2 and V-Amour, respectively). The rate of clinical failure was higher in the first use period (6.2% overall) than in the second (2.3%) and third (2.3%) periods, regardless of condom type use sequence.! 3.5. Safety There were few medical problems or AEs associated with the use of FCs in Part 1 of the study, with 56 total AEs identified among 18 women following adjudication by the PI and Medical Officer. Of the 56 events, 28 (50%), 19 (34%) and 9 (16%) occurred during V-Amour, WC and FC2 use periods, respectively. An additional 123 reports deemed not
C. Joanis et al. / Contraception 83 (2011) 248–254 Table 3 Condom function by failure mode and female condom type Failure modea
PATH WC (n=814 condoms opened, 798 used) N
%
95% CIb
Nonclinical breakage Clinical breakaged Slippage Misdirection Invagination Other failuree Total breakagef Clinical failureg
4 6 8 5 14 11 10 29
0.49 0.75 1.00 0.63 1.75 1.38 1.23 3.63
(0.13–1.25) (0.28–1.63) (0.43–1.97) (0.2–1.46) (0.96–2.93) (0.69–2.45) (0.59–2.25) (2.45–5.18)
Failure mode
FC2 (n=808 condoms opened; n=794 condoms used)
c
n
%
95% CI
Nonclinical breakage Clinical breakage Slippage Misdirection Invagination Other failure Total breakage Clinical failure
3 4 13 4 14 6 7 31
0.37 0.50 1.64 0.50 1.76 0.76 0.87 3.90
(0.08–1.08) (0.14–1.28) (0.87–2.78) (0.14–1.28) (0.97–2.94) (0.28–1.64) (0.35–1.78) (2.67–5.5)
Failure mode
V-Amour (n=801 condoms opened; n=784 condoms used)
Nonclinical breakage Clinical breakage Slippage Misdirection Invagination Other failure Total breakage Clinical failure
n
%
95% CI
5 2 10 4 12 10 7 26
0.62 0.26 1.28 0.51 1.53 1.28 0.87 3.32
(0.2–1.45) (0.03–0.92) (0.61–2.33) (0.14–1.3) (0.79–2.66) (0.61–2.33) (0.35–1.79) (2.18–4.82)
a
Denominator for percentage is number of condoms used. CIs do not account for correlated nature of the data. c Break noticed before intercourse. d Breaks during intercourse. e Functioned like a male condom and ejaculate spilled. f Clinical or non-clinical breakage. Denominator for percentage is number of condoms opened. g Clinical breakage, slippage, misdirection or invagination. A given condom use may involve more than one type of failure. b
to be AEs following adjudication were similarly distributed across the condom use periods. The most common AE was genital itching, with 9, 10 and 1 events occurring during VAmour, WC and FC2 use periods, respectively. Other AEs included genital pain (ten events from six women) or genital rash (seven events from five women.) All reported AEs resolved without sequelae. There were no AEs reported during Part 2. 4. Discussion Although FCs are available in many countries, their supply and uptake in countries hardest hit by HIV/AIDS is
253
inadequate. At the time of this study, the South African Department of Health supplied FC1 free-of-charge in over 200 family planning clinics, including the study site [1]. However, advertisement or clinic posting of their availability was limited due to fear of stock-outs and/or the potential substitution of the male condom for a more expensive method. This may partly explain the low prevalence of past FC use (14%) in the study population, although the lack of FC choices in South Africa and reported high male condom use could also have contributed to low FC uptake. This study represents the first head-to-head comparison of three new FCs, which may offer an expanded choice of less expensive FC options in the future. Almost all participants expressed a preference for one condom type following the three-period crossover component of the study, with significantly more women preferring WC and FC2 than VAmour. Nonetheless, all three FC types were deemed generally acceptable for use. One potential disadvantage of evaluating product preference and acceptability in a randomized crossover trial is the artificial nature of the research setting. To help address this, we designed Part 2 of the study to more closely resemble a “real world” setting, allowing women to freely choose which condoms to use over a 3-month period. Results of this simulated market agreed with Part 1, indicating that most women had honed in on their FC preferences by the end of the crossover trial. Previous research demonstrated that the efficacy of the FC1 female condom for pregnancy and STI prevention was comparable to other barrier methods [10,11] and that FC1 offered an important alternative when male condoms could not be used [12]. Given the functional performance of the FCs evaluated in this study, we believe that similar benefits can be achieved with the new devices. All three FCs performed comparably, with total clinical failure occurring in 3.6%, 3.9% and 3.3% of condom uses for WC, FC2 and V-Amour, respectively. These results are consistent with those observed in a previous trial of FC1 and FC2 conducted in a similar population [1]. Failure rates also decreased over time, regardless of FC use sequence. Such reductions are supported by earlier research on male and female condoms [1,13,14] and emphasize the need to thoroughly instruct women on correct use of FCs lest they prematurely abandon the device as a viable method of STI and pregnancy prevention. Acknowledgments Support for this work was provided by FHI with funds from the United States Agency for International Development (USAID) under USAID Cooperative Agreement GPO–A–OO–05–00022–00; however, the views expressed in this article are not necessarily those of USAID. FHI is an international non-profit organization that conducts research and provides technical assistance in health, family planning, STIs, HIV and infectious diseases, such as malaria and tuberculosis. The authors appreciate the feedback on earlier
254
C. Joanis et al. / Contraception 83 (2011) 248–254
drafts provided by FHI colleagues Doug Taylor, Debra Weiner, David Grimes and Laneta Dorflinger. We would also like to thank Christine Mauck and Jill Schwartz of CONRAD for their thoughtful reviews. Finally, we would also like to acknowledge the participants and staff of the Commercial City Clinic in Durban, South Africa, for their support and dedication to the study. References [1] Beksinska M, Smit J, Mabude Z, Vijayakumar G, Joanis C. Performance of the Reality® polyurethane female condom and a synthetic latex prototype: a randomized crossover trial among South African women. Contraception 2006;73:386–93. [2] Latka M, Joanis C, Glover L. Acceptability of the Reality® female condom and a latex prototype. J Urban Health 2001;78:614–26. [3] Smita J, Neelam J, Rochelle DY, et al. Comparative acceptability study of the Reality® female condom and the version 4 of modified Reddy female condom in India. Contraception 2005;72:366–71. [4] Schwartz J, Barnhart K, Creinin M. Comparative crossover study of the PATH Woman's Condom and the FC Female Condom®. Contraception 2008;78:465–73. [5] Code of Federal Regulations Title 21: Selected Regulations & Guidance for Medical Device Studies. Philadelphia, PA: Clinical Research Resources, LLC; 2008.
[6] International Conference on Harmonization. ICH E6 - Guideline on Good Clinical Practice. USFDA Federal Register May 1997; 62:25691-25709E7. [7] Moher D, Schultz KF, Altman D, for the CONSORT Group. Revised recommendations for improving the quality of reports of parallel group randomized trials. JAMA 2001;285:1987–91. [8] SAS Institute Inc. SAS/STAT® 9.2 User's Guide. Cary (NC): SAS Institute Inc.; 2008. [9] Beksinska M, Joanis C, Manning J, et al. Standardized definitions of failure modes for female condoms. Contraception 2007;75:251–5 [Epub 2007 Feb 7]. [10] Trussell J. Comparative contraceptive efficacy of the female condom and other barrier methods. Intl Fam Plann Perspect 1994; 26:66–72. [11] French PP, Latka M, Gollub EL, Rogers C, Hoover DR, Stein ZA. Use-effectiveness of the female versus male condom in preventing sexually transmitted disease in women. Sex Transm Dis 2003;30: 433–9. [12] Hoke TH, Feldblum PJ, Van Damme K, et al. Temporal trends in sexually transmitted infection prevalence and condom use following introduction of the female condom to Madagascar sex workers. Intl J STD AIDS 2007;18:461–6. [13] Hollander D. Failure rates of male and female condoms fall with use. Intl Fam Plann Perspect 2005;31:94–5. [14] Valappil T, Kelaghan J, Macaluso M, et al. Female condom and male condom failure among women at high risk of sexually transmitted diseases. Sex Trans Dis 2005;32:35–43.
Original research article
Three new female condoms: which do South-African women prefer? Carol Joanisa , Mags Beksinskab , Catherine Harta,⁎, Katie Tweedya , Jabu Lindab , Jenni Smitb b
a Family Health International (FHI), P.O. Box 13950, Research Triangle Park, NC 27709, USA Maternal, Adolescent, and Child Health (MatCH), Department of Obstetrics and Gynaecology, University of the Witwatersrand, Durban 4091, South Africa Received 2 June 2010; revised 29 July 2010; accepted 2 August 2010
Abstract Background: The widespread distribution of female condoms (FCs) in developing countries has been hindered by high unit cost, making new less expensive devices a priority for donor agencies. Study Design: Randomized, crossover study assessing product preference, safety, acceptability and function of three new FCs (PATH Woman's Condom, FC2 and V-Amour) among 170 women in Durban, South Africa. A subsequent “simulated market” study provided participants with free choice of FCs and assessed condom uptake over 3 months. Results: Of the 160 women who used at least one FC of each type, 47.5% preferred the PATH Woman's Condom (WC), 35.6% preferred FC2 and 16.3% preferred V-Amour (pb.001). Women rated the WC better than FC2 and V-Amour for appearance, ease of use and overall fit and better than V-Amour for feel. WC was rated worse than FC2 and V-Amour for lubrication volume. The simulated market demonstrated similar preferences. Total clinical failure rates (i.e., the types of failures that could result in pregnancy or STI) were low (b4%), regardless of condom type. Conclusions: Three new FC types functioned similarly and were generally acceptable. Most participants preferred WC and FC2 over VAmour, and WC was preferred over FC2 in several acceptability measures. © 2011 Published by Elsevier Inc. Keywords: Female condom; Function; Acceptability; Preference; Randomized trial; Barrier methods
1. Introduction Providing women in developing countries with female condoms (FCs) for pregnancy and HIV/STI prevention has been a high priority for donor agencies. The Reality® female condom (now known as FC1) was approved by the United States Food and Drug Association (US FDA) in 1993. Although FC1 was marketed worldwide for almost 20 years, distribution of this polyurethane condom has been hampered by a unit price much higher than that of male latex condoms. Manufacturers have recently developed new FC types made of less costly materials and design features that may make them more acceptable. The Female Health Company, Chicago, IL, USA, has developed the FC2®; a synthetic rubber (nitrile) condom that costs 30% less than FC1 but is similar in specification, function and appearance. A comparative functionality and acceptability study of the FC1 and FC2 devices among 276 women in Durban, South ⁎ Corresponding author. Tel.: +1 919 544 7040x11367. E-mail address: [email protected] (C. Hart). 0010-7824/$ – see front matter © 2011 Published by Elsevier Inc. doi:10.1016/j.contraception.2010.08.002
Africa, demonstrated that they have similar failure rates and are equally acceptable [1]. The FC2 was approved by the US FDA in March 2009, holds the CE Mark1 of the European Union (EU) and is distributed in 105 countries. Manufacture of FC1 ceased in October 2009. MedTech Products (Chennai, India) began developing a latex female condom in the early 1990s, culminating in the V-Amour®. A comparative acceptability study of V-Amour with FC1, conducted among 135 couples in New York City, found the two equally acceptable [2]. A subsequent study in India among 60 couples found that the V-Amour provided comparable or better acceptability results when compared to FC1 [3]. V-Amour bears the CE mark and is available in selected EU countries. PATH (Seattle, WA, USA) developed the WC female condom with several design innovations differentiating it from the aforementioned FCs. The WC has a polyurethane sheath with hydrophilic foam dots to anchor it within the 1 Certification given to products that fulfill the EU safety, health and environmental requirements.
C. Joanis et al. / Contraception 83 (2011) 248–254
vagina. A dissolvable polyvinyl alcohol insertion capsule encases the sheath, making the device easier to insert. The tampon-shaped capsule also aims to improve the aesthetics of the device. Unlike other FCs, the WC is not prelubricated and, for proper function, requires use of additional vaginal lubricant. A comparative crossover study of the WC and the FC1 was conducted in 2005 among 75 couples, with the WC exhibiting less breakage and slippage than the FC1. The WC was favored over FC1 2.6 and 2.0 times by females and males, respectively [4]. A clinical trial to assess the contraceptive effectiveness of the WC in comparison to FC2 is expected to begin in 2010. As new FCs enter the market, donors must decide which to purchase for wide-scale distribution in public health programs. Cost, which varies by volume of FCs purchased, will affect this decision, but consumer preference must be considered as well. In this study, the WC, FC2 and V-Amour were compared to determine which FC was preferred and to gather additional data on device safety, function and acceptability. Part 1 of the study was a three-period randomized crossover study; Part 2 was a “simulated market” in which women completing Part 1 could return to the clinic to receive free female condoms of their choice; and Part 3 employed in-depth interviews to elicit detailed reasons for participants' preference. This paper reports on the results of Part 1 and Part 2. 2. Materials and methods 2.1. Participants This study was conducted by the Reproductive Health and HIV Research Unit Durban (now MatCH) of the University of the Witwatersrand among female clients of the Commercial City Clinic in Durban, South Africa, between March 2007 and April 2008. Potential participants had to be in good general health with no known allergies to the study products, at least 18 years of age, using a reliable, non-barrier method of contraception and free of sexually transmitted infections (STI), as determined by pelvic examination and use of a syndromic diagnostic tool.2 Pregnant (according to urine pregnancy test) or breastfeeding women were excluded. Participants were required to be sexually active, monogamous and could not be practicing sex workers. Since take-home coital logs were used to gather data, participants were also required to be literate. Participants read and signed an informed consent form (ICF) prior to screening and enrollment. The protocol and ICF were approved by the institutional review boards of Family Health International (FHI) and the University of the Witwatersrand. The study was conducted in compliance 2 Syndromic management is the WHO recommended approach of identifying STIs in resource-limited settings. It is based on the identification of a group of symptoms and easily recognized signs (i.e., syndromes) which are associated with a number of well-defined etiologies.
249
with the protocol and the US Code of Federal Regulations [5], using Good Clinical Practice guidelines (ICH-E6) [6]. Study reporting was done according to CONSORT recommendations for improving randomized trials [7]. 2.2. Intervention and procedures To minimize use order bias, enrolled women were randomized to one of six condom use sequences based on a randomly varied permuted-block design with block sizes of 4, 6 and 8 using the SAS® System (SAS Institute, Cary, NC, USA) [8]. Randomization allocations were sealed in separate, opaque, sequentially numbered envelopes and assigned to participants in chronological order of enrollment. Due to the dissimilar appearance of the condom types, it was not possible to blind participants and clinic staff. Only study statisticians were definitively blinded until after primary analysis programs were written. Women were given five devices of the first FC type in their assigned-use sequence and asked to use them during five acts of intercourse with their partner over a three-week period. (During the WC use period, participants were also provided with five packages of Astroglide™ water-based lubricant, manufactured by BioFilm, Vista, CA, USA). Participants were instructed not to use male condoms while wearing a female condom, to complete a coital log after each FC use and to return to the clinic after use of the fifth device. At the follow-up visit, an interviewer-assisted questionnaire was used to ask women their opinions about — and experiences with — that FC type, including any condom device failures or adverse events (AEs). Data from the questionnaire were compared with the coital log to ensure consistency of response and determine FC use. This process was repeated for the second and third FC types according to the assigned use sequence. After use of the third FC type, women were either discontinued from the study or progressed voluntarily into Part 2. In Part 2, women were provided with up to 10 FCs of their choice (all of one type or a mix of types). If a participant ran out of study FCs, she was instructed to return to the clinic, for up to three visits, to collect more. Male condoms were also available at the clinic if the participants ran out of study FCs. Although the FC1 was also available at the clinic, we asked participants not to use the clinic supplies, but to return to the study nurse to obtain their FCs. At each follow-up visit, the participants were administered a questionnaire about their choices, acceptability of the FCs selected and whether they experienced any problems while using the FCs. Participants' self report was used to determine medical problems and FC use. At their last visit, women were either discontinued from the study or progressed voluntarily into Part 3 (qualitative methodology not described here). Women were paid 50 South African Rand (∼US⁎$8.00) per clinic visit for their time and travel expenses. Potential AEs were auto-coded to the MedDRA dictionary (Version 6.1) by FHI's data management system. For
250
C. Joanis et al. / Contraception 83 (2011) 248–254
AEs that could not be auto-coded, the study investigators assigned codes blind to FC type. The principal investigator (PI) and study medical officer subsequently excluded reported conditions that were not considered AEs in an adjudication process. 2.3. Study objectives The primary objective was to determine which, if any, of the three FC types was most preferred. Secondary objectives included gathering detailed information on safety, function and acceptability of each FC type. 2.4. Statistical methods The study planned to enroll 180 women, with the expectation that at least 150 would complete Part 1. With 150 completing Part 1, the study would have 80% power to conclude that a difference in preference exists, as long as one of the FC types was favored by at least 45% of the women (using a two-sided, .05 level of significance test). It was expected that 130 women would continue to Part 2, with at least 100 completing. With 100 completing Part 2, the study would have approximately 90% power to conclude that a
difference in preference exists, so long as one of the three FC types was favored by at least 50% of the women. A two-degree of freedom chi-squared test was used to compare overall product preference probabilities (“Which condom did you like best?”) across the three FC types in Part 1, excluding women who did not use at least one condom of each type. If the overall analysis was significant, pair-wise comparisons were to be performed using data from the subset of participants who preferred one of the FC types. Response options to each of the eight acceptability items ranged from 1=Disliked very much to 5=Liked very much and were compared across FC types using Friedman's test. p Values were not adjusted for multiple analyses. Condom function was assessed based on the clinical failure modes which could lead to semen exposure (i.e., misdirection, breakage, invagination and slippage) defined by Beksinska et al. [9] and endorsed by the World Health Organization. The relative performance of the three FC types in Part 1 was compared for two main function outcomes, clinical breakage and total clinical failure, using generalized linear equation (GEE) methods with identity link to account for clustering of responses within participants. The primary FC uptake analysis during the simulated
Fig. 1. Participant flow diagram.
C. Joanis et al. / Contraception 83 (2011) 248–254
market phase used GEE with a logit link to compare the odds of each FC type being chosen, with FC2 serving as the reference condition. 3. Results 3.1. Recruitment and follow-up One hundred eighty women were enrolled in Part 1 of the study, with 170 attending at least one follow-up visit and providing some follow-up data. One participant was discontinued early because she was unable to read and could not complete the coital log and nine women were lost to follow-up (Fig. 1). Other than being somewhat older (median age 29.5 vs. 26), the 10 women who did not provide data in Part 1 had no obvious differences from the 170 who did. Of the 160 women who completed Part 1 (i.e., used at least one condom of all three FC types), 148 enrolled in the Part 2 simulated market. Of these, four were lost to followup, and 12 were discontinued early due to reasons not related to the study products. There were no important differences in Table 1 Participant characteristics at baseline Age, years Mean (S.D.) Median Range Race, n (%) Black Colored White Indian Years of school completed Mean (S.D.) Median Range Primary occupation, n (%) None Self-employed Unskilled labor Health/medical Office Public service/government Sales Student Other Number of living children, n (%) Mean (S.D.) Median Range Marital relationship, n (%) Married and living together Married and not living together Not married and living together Not married and not living together Contraception method, n (%) Contraception pills Injectable contraception Tubes tied Total contributing data to Part 1 (crossover study)
(98.8) (0.6) (0.0) (0.6)
11.2 (1.43) 12 1–12 77 5 7 13 2 1 7 34 24
demographic or behavioral characteristics between participants who continued to Part 2 and those who did not continue (data not shown). Study participants ranged in age from 18 to 48 years (mean=28.0), and all but two were African (Table 1). Mean educational level was 11.2 years (range 1–12). Almost half (45.3%) were unemployed, 20% were students and the remaining women held various professional or semi-skilled positions. Ninety-one percent of women were not married and 24% resided with their male partners. Most (94%) had been with their male partners for over one year and 82% had at least one child (range 0–6). Previous use of male condoms was high, with 81% reporting use of a male condom in at least half of their sex acts during the previous 6 months. The most common contraceptive method used was progestin injectables (72%). At the time of this study, the FC1 condom was supplied free-of-charge at the clinic [1], although only 24 women (14.1%) reported previous use of a FC. When these women were asked what they would change about FCs, three themes emerged: FCs should be more readily available (n=5); FCs should be easier to insert (n=2), and there should be more FC options from which to choose (n=5). Seven women said nothing about the FCs should be changed. 3.2. Condom preference
28.0 (7.18) 26 18–48 168 1 0 1
251
(45.3) (2.9) (4.1) (7.6) (1.2) (0.6) (4.1) (20.0) (14.1)
1.4 (1.17) 1 0–6 12 3 28 127
(7.1) (1.8) (6.5) (74.7)
35 123 12 170
(20.6) (72.4) (7.1)
A total of 160 women used at least one condom of each type in Part 1. Of these, 76 (47.5%) preferred the WC, 57 (35.6%) preferred FC2 and 26 (16.3%) preferred V-Amour (one woman stated no preference), differences that were statistically significant overall (pb.001). Subsequent pairwise comparisons found that WC and FC2 were significantly preferred over V-Amour (pb.001 and p=.007, respectively), whereas there was a nonsignificant preference for WC over FC2 (p=.10). When asked which FC type they preferred the least, 58.8% cited V-Amour, followed by FC2 (24.4%) and WC (16.9%). Partner preferences (as reported by the participants) followed similar trends. A total of 4320 condoms were distributed in Part 2. The WC was chosen most frequently, with 1898 (44%) condoms distributed versus 1585 (37%) and 837 (19%) for FC2 and V-Amour, respectively. The mean exposure for each FC type was 12.8, 10.7 and 5.7 uses for WC, FC2 and V-Amour, respectively. The overall difference in the proportions of FC types selected was significant (pb.001). Pair-wise comparisons found no significant difference in odds of selecting more WC than FC2 condoms (p=.24), but a significant increased odds of selecting FC2 over V-Amour (p=.001). 3.3. Acceptability Of eight acceptability parameters compared between FC types in Part 1, five were statistically significant: feel/ sensation, lubrication volume, appearance, ease of use and fit (Table 2). There was no significant difference in preference with regard to scent, color or length. The participants rated
252
C. Joanis et al. / Contraception 83 (2011) 248–254
Table 2 Comparative acceptability outcomes by female condom type Acceptability measure
Overall Preference Pair-wise comparisons Feel/Sensation Disliked very much Disliked somewhat Neither liked nor disliked Liked somewhat Liked very much Pair-wise comparisons Amount of Lubrication Disliked very much Disliked somewhat Neither liked nor disliked Liked somewhat Liked very much Pair-wise comparisons Appearance Disliked very much Disliked somewhat Neither liked nor disliked Liked somewhat Liked very much Pair-wise comparisons Ease of Use Disliked very much Disliked somewhat Neither liked nor disliked Liked somewhat Liked very much Pair-wise comparisons Overall Fit Disliked very much Disliked somewhat Neither liked nor disliked Liked somewhat Liked very much Pair-wise comparisons a
PATH WC (n=165)
FC2 (n=164)
n
n
(%)
V-Amour (n=164) (%)
n
(%)
p valuea
76 (47.5) vs. FC2: p=.10 vs. V-Amour: pb.001
57 (35.6) vs. V-Amour: p=.007
26
(16.3)
b.001
7 (4.2) 10 (6.1) 13 (7.9) 68 (41.2) 67 (40.6) vs. FC2: p=.221 vs. V-Amour: p=.014
4 (2.4) 7 (4.3) 29 (17.7) 70 (42.7) 54 (32.9) vs. V-Amour: p=.033
7 12 33 70 42
(4.3) (7.3) (20.1) (42.7) (25.6)
.014
9 (5.5) 29 (17.6) 25 (15.2) 68 (41.2) 34 (20.6) vs. FC2: p=.004 vs. V-Amour: pb.001
5 (3.0) 14 (8.5) 18 (11.0) 75 (45.7) 52 (31.7) vs. V-Amour: p=.307
4 7 21 71 61
(2.4) (4.3) (12.8) (43.3) (37.2)
b.001
6 (3.6) 14 (8.5) 12 (7.3) 66 (40.0) 67 (40.6) vs. FC2: p=.023 vs. V-Amour: p=.011
3 (1.8) 22 (13.4) 20 (12.2) 71 (43.3) 48 (29.3) vs. V-Amour: p=.536
7 28 20 54 55
(4.3) (17.1) (12.2) (32.9) (33.5)
.014
2 (1.2) 7 (4.2) 4 (2.4) 61 (37.0) 91 (55.2) vs. FC2: p=.002 vs. V-Amour: pb.001
3 (1.8) 4 (2.4) 16 (9.8) 74 (45.1) 67 (40.9) vs. V-Amour: pb.001
6 21 24 65 48
(3.7) (12.8) (14.6) (39.6) (29.3)
b.001
5 (3.0) 4 (2.4) 11 (6.7) 78 (47.3) 67 (40.6) vs. FC2: p=.061 vs. V-Amour: p=.002
3 (1.8) 11 (6.7) 11 (6.7) 86 (52.4) 53 (32.3) vs. V-Amour: p=.254
7 16 23 67 51
(4.3) (9.8) (14.0) (40.9) (31.1)
.008
p Values are from Friedman’s test comparing distribution of responses across all 3 FC types.
the WC better than both FC2 and V-Amour in terms of appearance, ease of use and overall fit and better than VAmour in terms of feel. The WC was rated worse than both FC2 and V-Amour with respect to lubrication, notably the inconvenience of self-lubrication and indecision about volume of lubricant to use. 3.4. Functionality Out of 2423 condom packages opened during Part 1 of the study, 2376 were used during intercourse. Clinical function problems were reported in less than 2% of devices used (Table 3). Total clinical failure (functional problems that could lead to a pregnancy or sexually transmitted disease)
occurred in 3.6% of uses overall and was comparable across FC types (3.6%, 3.9% and 3.3% for WC, FC2 and V-Amour, respectively). The rate of clinical failure was higher in the first use period (6.2% overall) than in the second (2.3%) and third (2.3%) periods, regardless of condom type use sequence.! 3.5. Safety There were few medical problems or AEs associated with the use of FCs in Part 1 of the study, with 56 total AEs identified among 18 women following adjudication by the PI and Medical Officer. Of the 56 events, 28 (50%), 19 (34%) and 9 (16%) occurred during V-Amour, WC and FC2 use periods, respectively. An additional 123 reports deemed not
C. Joanis et al. / Contraception 83 (2011) 248–254 Table 3 Condom function by failure mode and female condom type Failure modea
PATH WC (n=814 condoms opened, 798 used) N
%
95% CIb
Nonclinical breakage Clinical breakaged Slippage Misdirection Invagination Other failuree Total breakagef Clinical failureg
4 6 8 5 14 11 10 29
0.49 0.75 1.00 0.63 1.75 1.38 1.23 3.63
(0.13–1.25) (0.28–1.63) (0.43–1.97) (0.2–1.46) (0.96–2.93) (0.69–2.45) (0.59–2.25) (2.45–5.18)
Failure mode
FC2 (n=808 condoms opened; n=794 condoms used)
c
n
%
95% CI
Nonclinical breakage Clinical breakage Slippage Misdirection Invagination Other failure Total breakage Clinical failure
3 4 13 4 14 6 7 31
0.37 0.50 1.64 0.50 1.76 0.76 0.87 3.90
(0.08–1.08) (0.14–1.28) (0.87–2.78) (0.14–1.28) (0.97–2.94) (0.28–1.64) (0.35–1.78) (2.67–5.5)
Failure mode
V-Amour (n=801 condoms opened; n=784 condoms used)
Nonclinical breakage Clinical breakage Slippage Misdirection Invagination Other failure Total breakage Clinical failure
n
%
95% CI
5 2 10 4 12 10 7 26
0.62 0.26 1.28 0.51 1.53 1.28 0.87 3.32
(0.2–1.45) (0.03–0.92) (0.61–2.33) (0.14–1.3) (0.79–2.66) (0.61–2.33) (0.35–1.79) (2.18–4.82)
a
Denominator for percentage is number of condoms used. CIs do not account for correlated nature of the data. c Break noticed before intercourse. d Breaks during intercourse. e Functioned like a male condom and ejaculate spilled. f Clinical or non-clinical breakage. Denominator for percentage is number of condoms opened. g Clinical breakage, slippage, misdirection or invagination. A given condom use may involve more than one type of failure. b
to be AEs following adjudication were similarly distributed across the condom use periods. The most common AE was genital itching, with 9, 10 and 1 events occurring during VAmour, WC and FC2 use periods, respectively. Other AEs included genital pain (ten events from six women) or genital rash (seven events from five women.) All reported AEs resolved without sequelae. There were no AEs reported during Part 2. 4. Discussion Although FCs are available in many countries, their supply and uptake in countries hardest hit by HIV/AIDS is
253
inadequate. At the time of this study, the South African Department of Health supplied FC1 free-of-charge in over 200 family planning clinics, including the study site [1]. However, advertisement or clinic posting of their availability was limited due to fear of stock-outs and/or the potential substitution of the male condom for a more expensive method. This may partly explain the low prevalence of past FC use (14%) in the study population, although the lack of FC choices in South Africa and reported high male condom use could also have contributed to low FC uptake. This study represents the first head-to-head comparison of three new FCs, which may offer an expanded choice of less expensive FC options in the future. Almost all participants expressed a preference for one condom type following the three-period crossover component of the study, with significantly more women preferring WC and FC2 than VAmour. Nonetheless, all three FC types were deemed generally acceptable for use. One potential disadvantage of evaluating product preference and acceptability in a randomized crossover trial is the artificial nature of the research setting. To help address this, we designed Part 2 of the study to more closely resemble a “real world” setting, allowing women to freely choose which condoms to use over a 3-month period. Results of this simulated market agreed with Part 1, indicating that most women had honed in on their FC preferences by the end of the crossover trial. Previous research demonstrated that the efficacy of the FC1 female condom for pregnancy and STI prevention was comparable to other barrier methods [10,11] and that FC1 offered an important alternative when male condoms could not be used [12]. Given the functional performance of the FCs evaluated in this study, we believe that similar benefits can be achieved with the new devices. All three FCs performed comparably, with total clinical failure occurring in 3.6%, 3.9% and 3.3% of condom uses for WC, FC2 and V-Amour, respectively. These results are consistent with those observed in a previous trial of FC1 and FC2 conducted in a similar population [1]. Failure rates also decreased over time, regardless of FC use sequence. Such reductions are supported by earlier research on male and female condoms [1,13,14] and emphasize the need to thoroughly instruct women on correct use of FCs lest they prematurely abandon the device as a viable method of STI and pregnancy prevention. Acknowledgments Support for this work was provided by FHI with funds from the United States Agency for International Development (USAID) under USAID Cooperative Agreement GPO–A–OO–05–00022–00; however, the views expressed in this article are not necessarily those of USAID. FHI is an international non-profit organization that conducts research and provides technical assistance in health, family planning, STIs, HIV and infectious diseases, such as malaria and tuberculosis. The authors appreciate the feedback on earlier
254
C. Joanis et al. / Contraception 83 (2011) 248–254
drafts provided by FHI colleagues Doug Taylor, Debra Weiner, David Grimes and Laneta Dorflinger. We would also like to thank Christine Mauck and Jill Schwartz of CONRAD for their thoughtful reviews. Finally, we would also like to acknowledge the participants and staff of the Commercial City Clinic in Durban, South Africa, for their support and dedication to the study. References [1] Beksinska M, Smit J, Mabude Z, Vijayakumar G, Joanis C. Performance of the Reality® polyurethane female condom and a synthetic latex prototype: a randomized crossover trial among South African women. Contraception 2006;73:386–93. [2] Latka M, Joanis C, Glover L. Acceptability of the Reality® female condom and a latex prototype. J Urban Health 2001;78:614–26. [3] Smita J, Neelam J, Rochelle DY, et al. Comparative acceptability study of the Reality® female condom and the version 4 of modified Reddy female condom in India. Contraception 2005;72:366–71. [4] Schwartz J, Barnhart K, Creinin M. Comparative crossover study of the PATH Woman's Condom and the FC Female Condom®. Contraception 2008;78:465–73. [5] Code of Federal Regulations Title 21: Selected Regulations & Guidance for Medical Device Studies. Philadelphia, PA: Clinical Research Resources, LLC; 2008.
[6] International Conference on Harmonization. ICH E6 - Guideline on Good Clinical Practice. USFDA Federal Register May 1997; 62:25691-25709E7. [7] Moher D, Schultz KF, Altman D, for the CONSORT Group. Revised recommendations for improving the quality of reports of parallel group randomized trials. JAMA 2001;285:1987–91. [8] SAS Institute Inc. SAS/STAT® 9.2 User's Guide. Cary (NC): SAS Institute Inc.; 2008. [9] Beksinska M, Joanis C, Manning J, et al. Standardized definitions of failure modes for female condoms. Contraception 2007;75:251–5 [Epub 2007 Feb 7]. [10] Trussell J. Comparative contraceptive efficacy of the female condom and other barrier methods. Intl Fam Plann Perspect 1994; 26:66–72. [11] French PP, Latka M, Gollub EL, Rogers C, Hoover DR, Stein ZA. Use-effectiveness of the female versus male condom in preventing sexually transmitted disease in women. Sex Transm Dis 2003;30: 433–9. [12] Hoke TH, Feldblum PJ, Van Damme K, et al. Temporal trends in sexually transmitted infection prevalence and condom use following introduction of the female condom to Madagascar sex workers. Intl J STD AIDS 2007;18:461–6. [13] Hollander D. Failure rates of male and female condoms fall with use. Intl Fam Plann Perspect 2005;31:94–5. [14] Valappil T, Kelaghan J, Macaluso M, et al. Female condom and male condom failure among women at high risk of sexually transmitted diseases. Sex Trans Dis 2005;32:35–43.