Minnis, Alexandra M. PhD*†; Shiboski, Stephen C. PhD‡; Padian, Nancy S. PhD*§
UNIVERSALLY, WOMEN COMPRISE the fastest growing group of individuals with sexually transmitted human immunodeficiency virus (HIV) infection. 1 In the absence of an effective vaccine or treatment, barrier methods capable of preventing sexual transmission of HIV as well as other sexually transmitted infections (STIs) are crucial for protecting the health of women. At the same time, widespread violence against women, double standards of sexual behavior, and the imbalance of power within many sexual partnerships make methods whose use is controlled by women critically important. 2–5 Thus, the need for safe and effective female-controlled topical microbicides and physical barriers that protect against both bacterial and viral STI pathogens is widely recognized. 6–9
Product effectiveness is inextricably bound to use, and, therefore, the needs and preferences of women and men who will be potential consumers must be considered at every step of development. There is an implicit assumption that acceptability serves as a proxy for use and that achievement of a high level of acceptability translates into maximal use. Consequently, before launching an effective intervention to promote existing barrier methods or a new microbicide, factors related to acceptability and use need to be identified.
The concept of acceptability, which is shaped by user perspectives, method characteristics, availability (service delivery and marketing), partner choice, and relationship dynamics, as well as the social and cultural context in which individuals make choices about contraception, has been evaluated in several distinct ways: hypothetical acceptability (before use), product choice, and satisfaction after use. Because few microbicides and female-controlled barrier methods currently exist, much research has focused on the acceptability of hypothetical products. These studies typically assess hypothetical acceptability in one of two ways: (1) by eliciting acceptability of specific attributes of a hypothetical method 10,11 or (2) by asking individuals to make a hypothetical choice among existing barrier methods that serve as product models for future products (e.g., using a spermicide to approximate a future microbicide). 12–14
Real product choice, in which a range of contraceptive barrier methods are presented and participants are offered a choice among them, has also been measured as an indicator of acceptability and as a proxy for future use. This design attempts to assess preference and choice in a more “natural” setting, such as a drug store, where women choose from among a range of products as opposed to being given a single method to try. 15–17 Method satisfaction after use assessed in longitudinal studies offers a third measure of product acceptability. 18–22 Here, satisfaction measures may include features of method use that were liked and disliked as well as future use intentions. The degree to which these measures are reliable indicators of use, however, remains unexamined.
Given the fairly extensive literature on barrier method acceptability, two conclusions shaped the design of our study and prompted us to conduct the analysis presented here. First, the majority of studies examined only one or very few products. Our study offered six physical or chemical barrier methods, which allowed us to consider a much broader range of method attributes and user preferences. Second, those studies that measured acceptability, choice, and/or satisfaction did not correlate these factors with future method use, the behavior ultimately linked to HIV and STI prevention. We assessed the degree to which these acceptability measures predict actual method use among an ethnically diverse population of young women.
We conducted a prospective study of barrier contraceptive method acceptability and use among sexually active women aged 15 to 30 years. Women were recruited from three reproductive health clinics in the San Francisco Bay Area in waiting rooms, during clinic visits, and through flyers posted in the clinic. Research staff also conducted outreach in the communities served by these clinics. 16 Females were eligible if they could provide informed consent; spoke English or Spanish; were currently sexually active (defined as having had vaginal sex at least three times in the previous 3 months); were not pregnant or trying to get pregnant during the next year; did not identify themselves as having tested positive for HIV; and would reside in the Bay Area for the next 12 months. Eligibility was determined through a screening questionnaire, and eligible participants were enrolled from March 1996 through April 1998 and followed for 1 year. Participants could enroll with a current male sex partner, but each member of a couple completed study visits individually. Study protocol and consent procedures were approved by the institutional review boards at the University of California San Francisco and the University of California Berkeley.
Study Protocol and Measures
At enrollment, all participants provided written informed consent and then completed an interviewer-administered face-to-face questionnaire that focused on sociodemographic factors; medical and reproductive health history; risk behaviors; previous contraceptive method use; and hypothetical acceptability of product attributes. To assess hypothetical acceptability, participants were asked about their willingness to use a product with a variety of physical attributes.
Following the baseline interview, trained health educator counselors conducted a standardized one-on-one educational HIV/STI risk reduction presentation with each respondent. The session was designed to provide participants with information to help them choose over-the-counter barrier contraceptive methods offered through the study. These methods included female and male condoms and four nonoxynol-9 vaginal spermicide formulations: film, foam, suppositories, and gel. (When study enrollment began, investigations of nonoxynol-9 as a chemical barrier for STI transmission had suggested that it either decreased risk for infection or had no effect on infection. 23,24 Throughout the study we emphasized that these products were still being tested for their disease-prevention effectiveness). A hierarchy of method effectiveness was presented, but the importance of choosing a barrier method that could be used consistently, even if it was inherently less efficacious, was emphasized. Furthermore, the presentation incorporated a discussion of basic anatomy with a focus on STI and pregnancy prevention. Participants had an opportunity to practice using each of the barrier methods on plastic anatomical models and to discuss method negotiation and product-use skills.
At the completion of the educational presentation, participants were asked to choose the barrier method(s) they preferred (they could also choose nothing) and the quantity of each one they wanted. An attempt was made to provide sufficient quantities of the products to last until the next visit, but at any time during the study, participants could request an additional supply by telephone. Arrangements were then made for a dispersal visit or the products were sent to the participants through the mail. All products dispersed to participants (both during and between study visits) were recorded. If participants preferred prescription methods, such as hormonal contraceptives or diaphragms, our research nurses arranged for these products to be provided before the participants left the clinic.
The 6- and 12-month study visits included a questionnaire that referred to the previous 6-month period. The interview focused primarily on contraceptive method use as well as satisfaction with the methods (male condoms, female condoms, and each spermicidal formulation) chosen at the previous visit. Participants who had not used methods they had chosen during or since the previous study visit were asked to identify reasons why they had not used them. Participants were again given the opportunity to choose from among the same barrier methods as those offered at baseline, and additional method education was offered as needed.
First we examined descriptive statistics that characterized the sociodemographic background, sexual risk behaviors, and STI history of the study population as well as method choice and use over time. Then, we analyzed the associations between three measures of acceptability and barrier method use: willingness to use a product on the basis of hypothetical product attributes; actual choice of over-the-counter barrier methods before and after use; and method satisfaction among users.
We evaluated the degree to which hypothetical acceptability assessed at baseline predicted over-the-counter barrier method use between baseline and 6 months. We used contingency table analysis to test the association between stated willingness to use a product with a given attribute and actual product use during the subsequent 6 months.
Next, we assessed the relationships between method choice and use over time by calculating the proportion of participants who used each over-the-counter barrier method from among those who chose it. We also examined whether the relationship between choice and use changed over time with prevalence ratios and confidence intervals that compared the proportions for each time interval. To examine the relationship between choice and use unconfounded by other factors known to be associated with these contraceptive behaviors, we used Generalized Estimating equations to estimate the overall marginal probability of use for each method, conditional on the methods being chosen over the course of the study. These models account for the repeated measures of use over follow-up. We derived 95% confidence intervals for these estimates from 1000 bootstrap replications for each model where individuals were sampled with replacement, which preserved the number of visits per individual. We adjusted these estimates for sexual behaviors found statistically significant with method choice at baseline 16 and factors that have been documented to influence use in other studies. 25,26 We Sconsidered demographic characteristics (age and race/ethnicity); partner-related sexual behaviors over follow-up (having multiple partners, having a main partner, and partner communication about contraception); use of hormonal contraception; STI incidence; and whether the participant had enrolled in the study with her male partner.
We examined the relationships between method satisfaction after use of a particular product and subsequent use with contingency table analysis. Given the duration of the study and the fact that satisfaction could not be assessed at baseline, these associations could be assessed only once (satisfaction reported at 6 months and subsequent use as reported at 12 months). Satisfaction was assessed on the basis of four questions: whether the participant liked the method; whether she thought her partner liked the method; whether sex with the method was the same, better, or worse; and whether she would recommend the method to a friend. Using these four questions, we created a composite satisfaction scale with which participants could receive a score from 0 to 4, depending on how many measures to which they responded favorably. Small samples available for this analysis precluded multivariate adjustment for potential confounding factors. We conducted stratified analyses, however, to determine whether the associations between satisfaction and use varied by the sociodemographic, relationship, and sexual behavior factors considered in the choice-use analysis described above.
Study Participant Characteristics
Study eligibility was assessed for 826 women; 654 were eligible, and 510 chose to enroll (response rate = 78%). Individuals were ineligible primarily because they were not currently sexually active (67%) or, for those who were sexually active, because they did not meet the study's age criterion (23%). On the basis of data collected in the eligibility questionnaire, the 22% of eligible females who chose not to enroll were comparable to the study population with regard to sexual risk characteristics and most sociodemographic factors; however, they were more likely than study participants to be Latina (P < 0.001).
Among participants, the median age was 21 years (interquartile range = 18–24). Non-Latina whites comprised approximately half of the participants and Latinas represented 25%, which reflects the ethnic distribution of the target populations within the study communities. The median number of previous sex partners was four (interquartile range = 2–9; range = 1–110). Twenty percent of females reported a history of STI, and 35% reported having been pregnant. Study retention proportions were 0.91 at 6 months and 0.86 at 12 months. Thirty-one percent of females (n = 160) enrolled with a male partner.
Hypothetical Acceptability and Use
Hypothetical acceptability, which was assessed by asking participants about their willingness to use a method with particular attributes, was not predictive of method use between baseline and 6 months (Table 1). In fact, the majority of significant associations between hypothetical acceptability of each attribute and method use were not relevant to actual use of the method. For example, among females, use of female condoms was significantly associated with hypothetical willingness to use a method that caused leakage after intercourse and that a partner could smell (P < 0.05). Neither of these qualities, however, is pertinent to use of the female condom. Similarly, a significantly greater proportion of participants who stated they would use a method that required insertion 15 minutes before intercourse used every spermicide, regardless of its properties or formulation. This attribute, however, is relevant only for film and suppositories. The one factor meaningfully predictive of spermicide use was a willingness to use a method that caused leakage after intercourse (P < 0.05).
Choice Versus Use
Across all female-controlled barrier methods, the proportion of women who used a method was always smaller than the proportion who chose it through the research study (Table 2). We expected that the majority of individuals who chose a method would use that method; however, apart from male condoms, this was not the case. Only 44% of participants who chose female condoms at baseline, for instance, used them (adjusted summary use estimate). As demonstrated by the confidence intervals for the estimated adjusted proportions (all of which either fall below or include 50%), with every female-controlled barrier method, the majority of individuals who chose the product did not use it. In addition, the relationship between choice and use did not improve over time because the proportion of participants who used a method among those who chose it did not change (data not shown). Male condoms, in contrast, were used by 87% of participants who chose them (adjusted summary use estimate).
Product Satisfaction and Use
High satisfaction scores (at least three of four) generally were not associated with use of female condoms and spermicides (Table 3). Fewer than half of participants who reported a high level of satisfaction with each product used it again. Although close to 80% of users reported satisfaction with gel, for example, only 37% of those females used it in the subsequent 6 months. An unexpected finding was that for male condoms, in contrast to the female-controlled barrier methods, more individuals used them between 6 and 12 months’ follow-up than reported satisfaction with them at 6 months. Only 52% of male condom users at 6 months reported high satisfaction ratings, but 87% of these individuals used male condoms over the next 6 months.
In contrast, low satisfaction (one or fewer of four measures) with use of female-controlled barriers was related to not using the method in the future. Females who reported a method to be unfavorable were unlikely to use the product again. Only 15% of participants who did not like female condoms used them again, and no users of film or gel who reported dissatisfaction used these methods again. Despite the fact that nearly one third of respondents reported a low level of satisfaction with male condoms, 65% of these individuals used male condoms again in the subsequent 6 months. These relationships were not influenced by the participant's age or partner-associated characteristics assessed through stratified analyses.
Regardless of the efficacy of an HIV/STI prevention product, its effectiveness will be only as good as its use. Although choice and acceptability are thought to be predictive of use, in our study hypothetical acceptability, choice, and product satisfaction following use generally were not associated with use. Young sexually active females may have the ability to ascribe value to specific method attributes such that those decisions influence choice. However, factors that may ultimately influence use likely extend beyond the product itself to the social and environmental context and dynamics of the intimate relationship.
Although product satisfaction among users was not associated with future female-controlled over-the-counter barrier method use, low product satisfaction corresponded with low levels of use of these methods. Thus, strong negative experiences with a product, such as finding that sexual intercourse was worse with the method, presented powerful motivations not to use the product in the future. In contrast, although satisfaction was associated with subsequent male condom use, participants also used male condoms despite not liking them. Furthermore, likely due to their ready availability, even participants who did not choose them used them. It may be that an established belief in their efficacy and a social climate in which their use is accepted and where they are easy to obtain underlie this finding. Indeed, it is difficult to ask people to use a product with unknown efficacy or that might be less efficacious than other available methods.
In addition to obtaining the quantitative data presented above, we asked open-ended questions to elucidate our findings (results not shown). When we asked participants why they did not use a method they had chosen at the previous study visit, concern that the method was not effective enough against pregnancy—not comfort level with a partner or confidence about its proper use—was cited most often as the reason. The finding that fewer than half of participants who reported a high satisfaction level for any female-controlled barrier method chose or used it again further suggests that, as for male condoms, additional factors beyond method satisfaction assumed a more prominent role in influencing future contraceptive behavior.
Several study limitations restrict the generalizability of our findings. With 1 year of follow-up and study visits at 6-month intervals, we had only one time period in which we could examine the association between product satisfaction after use and future use of those products. These relationships may change over a longer period of time or might be reported differently over shorter follow-up intervals. Also, because participants knew the study focused on barrier method acceptability, individuals might have felt compelled to take methods from us, to report being satisfied with the methods, and to report having used the methods. However, 62% of participants reported a low level of satisfaction with the female condom, which suggests that participants felt comfortable reporting their dissatisfaction. In addition, the fact that for all methods fewer people reported using them than the proportion who chose them during the previous time period suggests that the discordance between choice and use is unlikely to be due to reporting bias and the influence of social desirability. Data regarding the lack of a protective effect of nonoxynol-9 spermicides that began to emerge over the study period did not appear to affect use of these products, because no changes in use by visit were reported over time.
In an attempt to assess sustained product acceptability and ultimately usage so as to provide data useful in the development of a microbicide, this investigation presents more uncertainties than solutions. In evaluating the acceptability of methods that do not yet exist, a particular challenge remains for studies that attempt to connect acceptability of or satisfaction with product attributes and actual choice and use of the product. Indeed, use may be more strongly influenced by perceived efficacy, irrespective of formulation. If so, acceptability assessed in the context of an effectiveness trial may not be generalizable to use once efficacy is established.
The findings underscore the fact that hypothetical acceptability; choice of a product; satisfaction among users; and product use may also represent distinct constructs (e.g., one may find a product acceptable and not use it or vice versa). Furthermore, they suggest a need to re-frame the way we evaluate product acceptability by, for instance, examining the context (e.g., relationship or situational factors) in which products are used and in which choices among new and existing products are made so that such assessments are more closely aligned with future method use.
2. Maldonado M. Latinas and HIV/AIDS: implications for the 1990s. SIECUS Report 1990/91; 19( 2): 11–15.
3. Fort A. Investigation of the social context of fertility and family planning: a qualitative study in Peru. Int Fam Plann Perspect 1989; 15: 88–95.
4. Heise LL, Elias C. Transforming AIDS prevention to meet women's needs: a focus on developing countries. Soc Sci Med 1995; 40: 931–943.
5. Buvé A, Bishikwabo-Nsarhaza K, Mutangadura G. The spread and effect of HIV-1 infection in sub-Saharan Africa. Lancet 2002; 359: 2011–2017.
6. McCormack S, Hayes R, Lacey CJN, Johnson AM. Microbicides in HIV prevention. BMJ 2001; 322: 410–413.
7. Elias C, Heise L. Working Papers. The Development of Microbicides: A New Method of HIV Prevention for Women. New York: The Population Council, 1993.
8. Cates W, Stone KM. Family planning, sexually transmitted diseases and contraceptive choice: a literature update. Part I. Fam Plann Perspect 1992; 24: 75–84.
9. Stein, Z. HIV prevention: an update on the status of methods women can use [editorial]. Am J Public Health 1993; 83: 1379–1382.
10. Coggins C, Blanchard K, Friedland B. Men's attitudes toward a potential vaginal microbicide in Zimbabwe, Mexico and the USA. Reproductive Health Matters 2000; 8: 132–141.
11. Darroch J, Frost J. Women's interest in vaginal microbicides. Fam Plann Perspect 1999; 31: 16–23.
12. Eldridge G, St. Lawrence J, Little C, Shelby M, Brasfield T. Barriers to condom use and barrier method preferences among low-income African-Am women. Women Health 1995; 23: 73–89.
13. Hardy E, Jimenez A, de Padua K, Zaneveld L. Women's preferences for vaginal antimicrobial contraceptives III: choice of a formulation, applicator, and packaging. Contraception 1998; 58: 245–249.
14. Moon MW, Msizvo MT, Heiman JE, et al. Evaluation of the feasibility and acceptability of vaginal microbicides for the prevention of HIV/STI in Zimbabwe. National Academies of Practice Forum 2000; 2: 135–139.
15. Allen S, Serufilira A, Carael M. Acceptability of condoms and spermicides in a population-based sample of urban Rwandan women [abstract 5137]. Presented at the 4th International Conference on AIDS. Stockholm, Sweden, 1988.
16. Minnis A, Padian N. Choice of female-controlled barrier methods among young women and their male sexual partners. Fam Plann Perspect 2001; 33: 28–34.
17. Gollub E, French P, Latka M, Rogers C, Stein Z. Achieving safer sex with choice: studying a women's sexual risk reduction hierarchy in an STD clinic. Journal of Women's Health and Gender-Based Medicine 2001; 10: 771–783.
18. Farr G, Gabelnick H, Sturgen K, Dorflinger L. Contraceptive efficacy and acceptability of the female condom. Am J Public Health 1994; 84: 1960–1964.
19. 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.
20. Ruminjo J, Steiner M, Joanis C, Mwathe E, Thagana N. Preliminary comparison of the polyurethane female condom with the latex male condom in Kenya. East Afr Med J 1996; 73: 101–106.
21. Gollub E, 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; 27: 155–158.
22. Coggins C, Elias C, Atisook R, et al. Women's preferences regarding the formulation of over-the-counter vaginal spermicides. AIDS 1998; 12: 1389–1403.
23. Weir S, Feldblum P, Zekeng L, Roddy R. The use of nonoxynol-9 for protection against cervical gonorrhea. Am J Public Health 1994; 84: 910–914.
24. Cook R, Rosenberg M. Do spermicides containing nonoxynol-9 prevent sexually transmitted infection? A meta-analysis. Sex Transm Dis 1998; 25: 144–150.
25. Macaluso M, Demand M, Artz L, Hook ER. Partner type and condom use. AIDS 2000; 14: 537–546.
26. 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.