HIV and its virtually inevitable consequence, AIDS, continues to spread rapidly around the globe. Increasingly, its impact on women gains significance in terms of both the number of women now being infected and the implications this has on the rapid transmission of HIV to their children and sex partners. Worldwide, an estimated 2 million women were newly infected with HIV in 2002, representing 48% of the total new infections. Most of that transmission was through heterosexual sex. (1; 2) Women now comprise 50% of the 38.6 million adults living with HIV/AIDS. Globally, most of the women at risk or infected with HIV are of reproductive age and may wish to have children, thereby reducing their ability to take preventive measures and increasing possible vertical transmission.

The global AIDS epidemic and ongoing threat of other infectious sexually transmitted diseases has heightened the urgency of calls in recent years for products that protect against STIs both in addition to and exclusive of preventing pregnancy. (3–9) Particularly needed are women-initiated prevention methods, such as topical microbicides. A microbicide is any product that can kill or prevent infection from one or more microbes that can be transmitted during sexual intercourse. Microbicides for HIV prevention have been in development for nearly two decades. Numerous products are being developed and tested in clinical trials as gels, creams, foams or suppositories that can be inserted prior to sex to provide a protective cover or that contain some mechanism to inactivate or block the infectious agent. (10–16) Some of the compounds being tested are contraceptive and some are not. Some can be used without partner awareness, offering the user a concealable means of protection. A few are being tested for both vaginal and rectal use. (17) Early safety trials of microbicide candidates are also testing other factors, such as how well they protect from various transmissible diseases in addition to HIV, their longevity and durability (e.g., how well they cover, how long prior to sex they can be inserted, and how long they last), preferred form (gel, foam, etc.), the most effective mechanisms for application, and their effects on sexual sensation and pleasure or discomfort during sexual activity. (18–21)

Several studies have been conducted recently to assess women’s openness to various types of microbicides, their preferences (both in the abstract in the absence of trial products and in early testing of specific products), their willingness to try new products in association with sexual activity in various circumstances, and their responses to preliminary product trials. (3; 10; 22–29) These studies identified variations in preferences on the basis of age, income level, marital status, education levels, ethnic background, and current use of condoms. One U.S. study that explored these issues with women at high risk of HIV infection identified several likely cultural preferences and differences that could potentially affect use of microbicidal products. These included dislike of or disregard for touching the genitalia, preferring wet vs. drier sex, and being concerned with the smell and taste of a product. (22) These studies indicate the need for multiple options on the market to address the changing needs of different groups of women depending on their individual circumstances, the context of use, and personal preferences. These findings also suggest the need for further exploration of factors affecting the probable acceptability of microbicides among high-risk women. Better understanding of the specific contexts that create the need for products with particular characteristics is essential as development of microbicides proceeds.

Our own work in Hartford, Connecticut focused on microbicide acceptability and readiness among women at high HIV risk in an urban U.S. context and the potential of microbicides to be a viable option for preventing sexually transmitted HIV and other STIs. In our study called Project Protect, we were interested in understanding whether such products fit with the needs of these women, and the factors and conditions that influence whether and when high-risk women will use them. We explored their perspectives on HIV/STI risk and prevention, their experiences with contraceptives that are similar to many microbicides being tested, and the social contexts and relationship factors affecting their willingness and ability to use microbicides during sex with various types of partners. We report here on the findings of this study.

In designing Project Protect, we sought to integrate methods successfully utilized in previous studies assessing women’s prior contraceptive experiences and preferences for microbicide characteristics, and to expand on these with exploratory ethnographic approaches designed to assess conceptual framing of microbicides and other prevention methods. (3; 22–26) We also designed the study to explore more fully the personal and social contextual factors that affect current efforts to prevent HIV, perceived interest in and willingness to use microbicides, and sexual practices with the introduction of a vaginally inserted microbicide-like product. For these reasons we used a combination of qualitative and quantitative methods to assess microbicide readiness and acceptability in the study population. These included ethnographic elicitation interviews, a risk and attitudinal survey, and a simulation trial of microbicides using an over-the-counter product. All research activities, including all components of the microbicide simulation trial, were reviewed and approved by an Institutional Review Board for acceptable protections of human subjects; all participants provided informed consent before engaging in any research activities.

Women recruited or referred into the survey and simulation trial were briefly screened to determine if they were sexually active within the prior 30 days before being given an appointment to enter the study, though this was not necessary for women who participated only in the ethnographic exploratory exercises and focus groups. Men recruited for the men’s focus groups were referred from our other ongoing studies of drug users. Other eligibility criteria for all components of the study included being 18 years of age and assessed to be able to understand and provide informed consent. Women entering the simulation trial were additionally screened for a variety of physical symptoms that might indicate an active sexually transmitted infection (STI) and for pregnancy or intention to get pregnant within the subsequent month, any of which resulted in exclusion from the trial. Anyone indicating possible infection with an STI was referred to medical services immediately after completion of the interview. Interviewers also had discretionary ability to refuse participation of anyone who appeared too drunk or high to understand the requirements and risks of the project or other aspects of informed consent.

The elicitation exercises included a free listing of pregnancy/HIV prevention options, followed by a pile sorting exercise (32) with groups of women from the target population. In the free list exercise, we asked a small number of women (n=10) to list every way they could think of or had heard of to prevent pregnancy or HIV/STI. We also asked them which of these they had ever used and their experiences with these methods. The pile sort exercise required placing each of these free-listed methods on a numbered card. We included the name of the item in English and Spanish, and where appropriate, a picture of the item, not possible with most behavioral methods (e.g., selecting clean partners). We then asked 24 women who did not participate in the free list exercise to sort these prevention methods into piles of items they thought were related. The interviewer documented which items each person grouped together and the reasons she gave for their association. The pile sort data were analyzed with ANTHROPAC (33) using Johnson’s hierarchical clustering and non-metric multi-dimensional scaling (MDS), and graphed on a two-dimensional scale representing a “cognitive map” of prevention options. The cluster analyses and MDS are designed to assess how frequently respondents grouped specific items together. We interpreted these clusters and the cognitive map both deductively on the basis of our theoretical model and inductively using the responses women gave as their reasons for grouping items they considered related.

In addition to these elicitation exercises, we conducted four focus group interviews with 6–10 women each, and three with 8–10 high-risk men each. (The men were not related to the women in the focus group samples.) After explaining microbicides and demonstrating their potential forms using contraceptives as similar models, we asked the women about their experiences with contraception and HIV/STI prevention, issues and barriers to prevention or problems they have encountered, their ideas about the potential benefits of microbicides, characteristics of the ideal microbicide, and methods they believed would successfully spread the word about microbicides when these products should become available on the market. We asked the men their reactions to the idea of microbicides after giving the same explanation and demonstration of potential microbicide forms, and explored the potential benefits or problems they imagined these products might have. We asked both men and women to talk about their interest in microbicide use with different types of sex partners, including primary, casual, and paying or paid partners.

Participants who completed the full survey and passed the eligibility screening were shown the surrogate product, provided in single-dose, pre-filled plastic applicators, opened by breaking off a small plastic tab to apply by inserting directly into the vagina. This moisturizing product is composed of glycerin and water and has no contraceptive or microbicidal attributes. Women were given oral and written directions on the proper way to apply the product for this microbicide simulation study, strongly cautioned both verbally and in the written instructions that the product does not prevent either pregnancy or HIV/STI, and asked to apply one dose of the product before every act of vaginal intercourse over the next two-week period and to use a male or female condom during every sexual encounter. We also asked them to bring back the removable plastic tabs from opened product applications and to document each sexual encounter either on an encounter form or to use our anonymous and toll-free 800 telephone number to call in their responses to the same set of questions.

Encounter forms documented the time and place of the sexual encounter, the type of sex partner (primary, casual, paying), whether male or female condoms were used, whether the product was used, and any problems, complications, interruptions, or partner responses to the use of the product during that sexual encounter. Women were provided sufficient doses of the product and more than sufficient condoms for the two-week period, given a booklet of encounter forms, and scheduled for a return appointment two weeks later. The follow-up surveys assessed their sexual activity during that period with all types of partners, their use of the product and condoms, their attitudes and partners’ attitudes toward the product and microbicides in general as a result of the trial, and for those in the first trial, whether they were interested in participating again in a second trial, for which they received a second screening for continued eligibility.

The 24 women who participated in the pile sort exercise included fourteen African Americans, eight Puerto Ricans, and two non-Hispanic Whites. Statistical analysis of their card sorts indicated two large groupings and, within them, a number of smaller sub-groupings. We identified the two large clusters of related items as: 1) medical or mechanical prevention options, and 2) behavioral prevention methods. Under the domain of medical/mechanical were clustered condoms, pills, other hormonal contraceptives, sterilization operations, and all vaginal products. Within this broad category, long-acting contraceptives (Depo-provera, Norplant, IUD, patch) and the diaphragm formed a tight sub-group. In women’s verbal explanations provided during the pile sort exercise, those who grouped these items together described them as working well for contraception and methods they had used. Contraceptive gel, foam, sponge and cream formed another tight cluster within the larger grouping; however, women often described this sub-group as methods that do not work or that they did not like, or frequently, as methods they had never heard of. Also notable was that among the mechanical methods, the female condom was least often grouped with any other prevention methods, and commonly mentioned as unfamiliar. Under the behavioral domain were practices such as avoiding, refusing, or modifying sexual behavior, partner selection, and hygiene. Women’s descriptions of this cluster included “not really using anything” and “still putting yourself at risk for diseases.” This suggests that they recognize the need to take steps beyond partner selection or trying to avoid sex or sexual intercourse to prevent pregnancy or STIs.

When the pile sort data were analyzed using MDS and graphed two-dimensionally, the resulting “cognitive map” illustrated additional aspects of the relationships among these prevention methods for these women (see Figure 1). The two primary domains participants created in their pile sorts are evident as a continuum from the more medical/mechanical to the more behavioral along the x axis of the map. The pattern along the y axis suggests grouping that reflects a general trend regarding degree of internal locus of control, a key factor in our theoretical model hypothesized to influence women’s prevention practices. The trend moves from methods requiring a greater sense of personal power to those methods more commonly employed when a woman feels less empowered to take preventive action. For example, within the behavioral sector refusal of sex (#25, #12, #17), which happens before risk occurs, may be feasible only for women who have more control over the sexual encounter. By contrast, douching (#9) and other attempts to be clean before or after sex (#5, #2, #11) are a last resort for women who are unable to avoid a risky sexual encounter.

Cognitive Mapa of Pregnancy and HIV/STI Prevention Methods for High-risk Women in Project Protect

When asked to describe their experiences using the methods of HIV/STI and pregnancy prevention available to them, the women who participated in these exploratory exercises gave a variety of responses. For contraception they clearly preferred and most commonly used pills or long-lasting hormones (Depo-provera, Norplant) because of their effectiveness and convenience. They recognized avoiding sex as an effective means of preventing both HIV/STI and pregnancy, but considered it a less preferred method. Regarding condoms (predominantly referring to male condoms), they expressed ambivalence, with comments like: “I felt in control [when using them], safe from getting pregnant and sick, including with casual partners... But physically, it [sex] was not the same feeling” (Puerto Rican, age 26); “I can’t control use of a condom; but it’s good for protection–glad to have something” (African American, age 54); “I used them in the ’80s, but not [now] with my partner [of 14 years]; I don’t like them because they break; you’re always taking a risk” (Puerto Rican, age 30). However, despite this ambivalence, women indicated they believed condoms were the safest and most reliable method of STI prevention and frequently chosen for contraception as well.

By contrast, participants had very limited experiences with vaginal contraceptive products, and those who had (and some who had not) ever used them generally expressed negative reactions to them. For example, women made comments like: “I never used [contraceptive gel]; had it but didn’t use it; it was too messy” (African American, age 46); “I used [gel] two times with my oldest son’s father; I used it with condoms; it helped with lubrication, but I don’t like it because it’s too much chemicals” (Puerto Rican, age 30); “I used [contraceptive foam] once; didn’t like how it felt; didn’t like inserting it; I was embarrassed” (non-Hispanic White, age 30); but “I liked the [diaphragm with gel] because I can do it myself anywhere; at first I didn’t like putting it on” (Puerto Rican, age 33). These comments illustrate the general discomfort and disinclination women felt regarding vaginal contraceptive use.

Following the card sorting exercises, we conducted focused group interviews with 28 women, including ten African Americans, eight Puerto Ricans, seven non-Hispanic Whites, and three West Indians. Women in the focus groups identified several problems they encountered protecting themselves from diseases and unwanted pregnancy. These include trust issues, lack of control in relationships, overcoming shame associated with addiction or sex work, fear of harm to themselves or their children, fear of rejection, and loss of self-esteem. They pointed out that benefits of a microbicide include “put[ting] power back in the hands of women” and providing a choice of protection options in addition to condoms. These women thought the ideal microbicide would be inexpensive, simple to use, effective, discreet, comfortable, fast acting, easily accessible, long-lasting, and affordable. In their general enthusiasm for having a prevention option like microbicides, they identified a number of venues for promoting them to women at risk, including billboards, radio and television, public announcements, bus stop posters, the needle exchange van, social service organizations, training parole/probation officers, providing free samples and educational materials in the mail and places such as college campuses, and making it part of health education in schools.

We also conducted focused group interviews with a total of 30 men, including eighteen African Americans, eight Puerto Ricans, and four non-Hispanic Whites, to assess their reactions to the concept of microbicides. Some indicated concern about their primary partner using a microbicide because it raised issues of trust. Others expressed support for the idea to relieve some of their own burden of responsibility for disease prevention, and indicated they would like having an alternative to condoms. Quite a few men, however, raised concerns about the likely efficacy of microbicides for HIV/STI or pregnancy prevention and indicated they would probably still use a condom if their partner were using a microbicide. Concern about efficacy was noticeably lacking in the focus groups with women.

In general, the exploratory ethnographic exercises and interviews we conducted in the first year of the study indicated that these high-risk women were very positive about the idea of vaginal microbicides and expressed interest in having an alternative to male condoms, especially one they could control. High-risk men also indicated support, though they had greater reservations about efficacy. However, despite general support for the idea of having a microbicide available for prevention, overall findings from our initial elicitation exercises suggested very limited preparedness within this high-risk population for use of microbicides for HIV/STI prevention based on limited parallel experience with other vaginal products.

Table 1 Demographic and Risk Characteristics of Women in the Microbicides Acceptability Survey Sample (N=471)

Like the women who participated in our exploratory ethnographic interviews, women in the survey sample had limited experience with vaginal products, especially for contraception. Reported contraceptive use in the prior thirty days included male condoms (56.5%), female condoms (9.8%), and very limited use of oral contraceptives (2.8%) and Depo-provera (2.5%). About two-thirds reported familiarity with vaginal contraceptive foam, cream and gel, but less than one-fifth had ever used any of these products.

Overall responses to the microbicide acceptability items were more positive than negative, resulting in a mean score on the index of 2.73 (SD=.49, scale of 1–4). Microbicide acceptability varied by several factors, most notably by ethnicity (p<.01). Post-hoc analyses revealed that non-Hispanic Whites had significantly higher scores (mean = 2.99, SD=.43) than Puerto Ricans/other Latinas (2.72, SD=.46), African/Caribbean Americans (mean = 2.65, SD=.52), and other/mixed ethnic groups (2.59, SD=.46). None of the other ethnic groups were significantly different from one another. Microbicide acceptability was also significantly correlated with having graduated from high school (r=.10, p<.05), being an injection drug user (r=.10, p<.05), and being HIV-negative (r=.15; p<.01).

We conducted hierarchical regression analyses to examine the differential percentage of variance accounted for by these groups of variables from the personal, relational, and attitudinal domains to predict microbicide acceptability (see Table 2). We were also interested in estimating the adjusted effect size and direction of association for each individual predictor variable. After controlling for the effects of ethnicity, age, marital status, and education level (block 1 in the regression analyses), microbicide acceptability scores were negatively related to having experienced either physical or sexual violence, but not significantly related to other HIV drug- and sex-risk conditions (block 2). Regarding attitudinal factors (block 3), acceptability was negatively related to AIDS risk perceptions, and positively related to an internal locus of control for STI prevention (i.e., a sense that prevention from STI infection results from one’s own actions). Additionally, although only marginally significant, prior experience with vaginal contraceptive products was positively correlated with microbicide acceptability, and having power within the sexual relationship was negatively correlated with microbicide acceptability scores (i.e., those women with perceived less power were more positive about microbicides). An examination of interaction effects between ethnicity and the independent variables revealed no significant interactions. The change in explained variance (R²) between the variables controlled for in the first block and the personal variables in the second block of the regression analyses was not significant. However, the change between the personal variables (block 2) and relational and attitudinal variables (block 3) was, indicating that attitudinal variables accounted for a significant percentage of the variance above and beyond that accounted for by personal risk characteristics. The percentage of variance accounted for by this model was 15.5%, and the overall relationship was significant (F_21,303=2.64, p<0.01).

Table 2 Summary of Hierarchical Regression Analyses for Variables Predicting Microbicide Acceptability (N = 471)a

These findings suggest that while this high-risk population generally considers microbicides to be an acceptable option for HIV prevention in the abstract, microbicide use might depend on women’s prior experience with other vaginal products and their relative sense of power and control in relationships. Notably, women who perceived themselves to be at greater risk for HIV reported lower microbicide acceptability scores, the reasons for which are unclear from these data. More investigation is needed to understand why these women find this HIV-prevention option less attractive. However, even with these variations among women in this study, many (83.4%) indicated a strong interest in participating in clinical trials of microbicides if these should be conducted locally. Participants’ interest in trying microbicides was also evident in their reports during the microbicide simulation trial.

Table 3 Sexual Activity and Use of Product and Condoms During 2-week Microbicide Simulation Trial 1

All the women with primary partners and all those with casual partners used the product at least once with that type of partner during the trial, while nearly all participants with paying partners also did so. Notably, women in the trial reported using the product on average more frequently than they reportedly used male condoms. (Female condom use was negligible.) However, overall reported condom use during the two weeks of the trial was still higher on average than reported use with all partner types during the thirty days before entering the study (see Table 1 and Table 3). About two-thirds of women with each type of partner said they liked using the product with that partner, and about one-third thought it was too messy or drippy. Over half (53.2%) of the women with primary partners said that their partner reported liking the product, though this was true of less than a third of women commenting on casual (30%) and paying (25%) partners. Male partners also reported that it made sex better (9.1% of primary, 10% casual partners and 16.7% of paying partners), though a few also said it was too messy or drippy (16.9% of primary partners, 10% of casual partners, and 8.3% of paying partners). Thus, the overall response of women and their partners to the microbicide simulation product when used before regular vaginal intercourse was generally positive.

The issue of stealth or clandestine use of an HIV prevention method for women concerned about partner resistance is often noted as a significant advantage of microbicides. However, possibly because they were in a research project and were documenting their sexual encounters, women in the simulation trials reported that they commonly informed their partners that they were using the product. Specifically, 85.8% of primary partners knew the participant was using the product and 63.6% of women with casual partners said some or all of them knew, but 61.5% of women with paying partners did not tell those partners when they were using the product. Altogether, partners were aware of the participants’ use of the product during 65.7% of encounters.

Sexual encounter forms women completed to document their sexual activity during the 2-week trials confirmed extensive use of the product with various types of partners and in various settings. Altogether, women who participated in the first trial completed 519 sexual encounter forms. Five forms did not contain adequate data on product usage and were dropped from further analysis. The remaining 514 encounter forms indicated that the 94 women in the trial used the product 432 times, or during 84.0% of encounters, including 274 (82.5%) times with primary partners, 44 (86.3%) times with casual partners, and 110 (88.0%) times with paying partners. Four encounters with a partner whose type was unspecified involved the use of the simulation product. (Women in the trial recorded 6% more product use on encounter forms than they reported in follow-up surveys; however, the number of returned tabs from opened product applicators was within two percentage points higher than the number of uses recorded on encounter forms.) During the first trial, women generally used the microbicide simulation product in a home (377 encounters), but also used it in hotels or motels (55 encounters), in cars (40 encounters), and in public or other areas (38 encounters), reportedly including in the library, a garage, a rooftop, a tent, bars, and clubs.

Participants indicated that use of the simulated microbicide did not interfere with sex during 93.8% of these encounters. The most commonly reported reason for failing to use the product during sexual encounters as documented on the encounter forms was forgetting (27.5%), followed by not wanting to interrupt sex (18.5%), partner refusal (13.4%), and inconvenience or a bad location for use (10.0%). A small number (6.3%) reported being too high or drunk to apply the product before sex. These encounter forms highlight the considerable willingness and ability of this high-risk group of women to incorporate the practice of applying a vaginal microbicide during sexual activity with most any kind of partner in both ordinary and extraordinary circumstances in which they might be having sex.

Mathematical models estimating the potential value of introducing a microbicide into populations facing high rates of new infection suggest that even a relatively ineffective microbicide could have a significant positive effect on reducing sexually transmitted HIV. One model estimated that introduction of a 40% efficacious microbicide into 73 lower-income countries could avert 2.6 million HIV infections over 3 years in men, women, and infants, even assuming that only 20% of people likely to be reached used it in only 50% of sexual encounters in which condoms were not used (38). The high and increasing rates of HIV and other sexually transmitted diseases among women is a compelling reason to support greater efforts directed toward microbicide research and development. Unfortunately, financial support for development and testing of microbicides has been extremely limited and microbicides are still considered with great skepticism, even among many in the field of HIV/AIDS prevention (16; 19; 20). Thus, while many products have been identified as potentially effective microbicides, exceedingly few have made it to the efficacy stage (Phase III) of clinical trials.

Our Hartford study of microbicide acceptability and readiness among women at risk suggests that there could be a substantial benefit to making microbicides available as an additional option for high-risk women to protect themselves from HIV and other STIs. This study and others (3; 10; 12; 22) confirm the willingness and interest of high-risk women to use microbicides, as indicated by the high ratio of microbicide simulation product use to sexual encounters with all types of partners in the trial period. The positive response to the simulated microbicide was surprising, given their prior experiences with vaginal products, although in light of their general concerns about currently available prevention options and their commonly expressed interest in using microbicides, their readiness to try the simulated product makes sense. These findings also suggest the potential acceptability, and hence value, of other woman-initiated, vaginally inserted prevention options such as the female condom. However, while the higher rate of product use over condom use during the trial may be an artifact of participating in the trial itself and not a clear preference for microbicides vs. (primarily male) condoms, it suggests the need to understand better the issue of “condom migration” (39) with the introduction of an alternative prevention option such as microbicides.

Reported differences in microbicide acceptability with different types of sex partners and among women of different ethnic groups indicates that microbicides may not necessarily be appropriate or desirable for some women or in some contexts. Overall rates of use of the simulation product during the trial suggests feasibility for many high-risk women to incorporate the practice of using microbicides into their regular sexual activity. Much greater understanding of the social dynamics of microbicide use and negotiation, particularly for women at high risk, is needed. Also needed is a fuller exploration of the gender role, identity, and sexuality issues that intersect with personal and environmental risk factors to intervene and impede women from using microbicides or compel them to use them as an alternative to male or female condoms. The notable difference between women and men in the focus groups regarding the importance of product efficacy (with men indicating greater concern about this than women) suggests that women may be more focused on increasing their prevention options, self-reliance for prevention, and having a surreptitious option, even if it is somewhat less efficacious than the current male-initiated options. Men, by contrast, may be more comfortable with a product of known efficacy, despite its inconveniences and discomforts, with which they are familiar and able to control. Greater understanding of men’s and women’s preferences and the dynamics of their negotiations of these around prevention is needed.

This study had several limitations. First, the sample, while guided by a targeted sampling plan, (31) is not necessarily representative of all women at risk in Hartford because of the hidden and stigmatized nature of HIV risks and the difficulty of finding and recruiting women who are consequently at highest risk. Second, because we lacked a true microbicidal product for women to view and try, responses to the microbicide acceptability questions were based on women’s perceived preferences for hypothetical characteristics. Related to this, while women’s use of the product in our simulation trial was promising, the trial itself is only able to suggest high-risk women’s likely interest in and ability to use microbicides in their real-life contexts. Thus, in order to gain a more accurate understanding of women’s attitudes toward microbicides and their willingness and ability to use them, acceptability studies in real life contexts that accompany clinical trials remain essential.

The urgency for women locally and globally to have alternative, women-initiated HIV/STI prevention approaches provides a compelling rationale for the promotion of topical microbicides. Such urgency calls us to press for significantly greater financial and political support for rapid microbicide development (both vaginal and rectal), clinical trials, and acceptability testing. This would be enhanced through greater dissemination of information on the status of microbicide development and the potential contribution of this important prevention method for reducing the spread of HIV.

This study was supported by a grant from the National Institute on Mental Health, #R01 MH63631 to the Institute for Community Research. It is an affiliated project of the Center for Interdisciplinary Research on AIDS (CIRA) (#P30 MH62294). The authors wish to thank the other members of the Project Protect research team, including Michelle Corbett, Helena de Moura Castro, Raquel Martinez, Maria Martinez and Robert Rooks. We also greatly appreciate the assistance of Marilyn Stolar, Ph.D., CIRA, for statistical consultation in the writing of this article.

Margaret R. Weeks, Institute for Community Research, 2 Hartford Square West, Suite 100, Hartford, CT 06106, 860-278-2044 x229, 860-278-2141 (fax), Email: mweeks/at/icrweb.org.

Katie E. Mosack, Institute for Community Research.

Maryann Abbott, Institute for Community Research.

Laurie Novick Sylla, Connecticut AIDS Education and Training Center.

Barbara Valdes, Institute for Community Research.

Mary Prince, Institute for Community Research.