|
|
| |||||||||||||||||||||||||||||||||
Background African-American and Hispanic women are disproportionately affected by sexually transmitted diseases, including the acquired immunodeficiency syndrome (AIDS). In the effort to reduce infection rates, it is important to create and evaluate behavioral interventions that are specific to the target populations.
Methods We enrolled women with nonviral sexually transmitted diseases in a randomized trial of a sex- and culture-specific behavioral intervention. The intervention consisted of three small-group sessions of three to four hours each designed to help women recognize personal susceptibility, commit to changing their behavior, and acquire necessary skills. The control group received standard counseling about sexually transmitted diseases. The design of the intervention was based on the AIDS Risk Reduction Model and ethnographic data on the study populations. Participants in both groups underwent screening, counseling, and an interview before randomization and at the 6- and 12-month follow-up visits. The principal outcome variable was subsequent chlamydial or gonorrheal infection, which was evaluated on an intention-to-treat basis by logistic-regression analysis.
Results A total of 424 Mexican-American and 193 African-American women were enrolled; 313 were assigned to the intervention group and 304 to the control group. The rate of participation in the intervention was 90 percent. The rates of retention in the sample were 82 and 89 percent at the 6- and 12-month visits, respectively. Rates of subsequent infection were significantly lower in the intervention group than in the control group during the first 6 months (11.3 vs. 17.2 percent, P=0.05), during the second 6 months (9.1 vs. 17.7 percent, P=0.008), and over the entire 12-month study period (16.8 vs. 26.9 percent, P=0.004).
Conclusions A risk-reduction intervention consisting of three small-group sessions significantly decreased the rates of chlamydial and gonorrheal infection among Mexican-American and African-American women at high risk for sexually transmitted disease.
The African-American and Hispanic populations are disproportionately affected by sexually transmitted disease. The incidence of gonorrhea is substantially higher among African-American and Hispanic persons than among whites1; for AIDS the respective incidence rates are six and three times as high as among whites.10 The proportion of cases of AIDS that occur among minority groups has been increasing since 1992.11 Differences in the incidence of AIDS according to race and ethnic group are especially dramatic for women. In 1996 these rates were more than 17 and 6 times as high for African-American and Hispanic women, respectively, as for white women (61.7 and 22.7 as compared with 3.5 per 100,000).7,9
Changing sexual behavior is critical to controlling sexually transmitted infection. However, success in reducing high-risk sexual behavior has been limited, particularly among women.8,12,13 For women, change is impeded by inequalities between the sexes and dependence on male partners for emotional and economic well-being; minority women are at a further disadvantage because of discrimination and their disproportionate poverty.4,12,13,14,15,16,17,18,19 Of the few controlled trials that have used systematic screening for sexually transmitted diseases at base line and at follow-up to assess the efficacy of interventions,20,21,22 only one21 found significant differences among study groups, and none focused on minority women. We sought to create culture- and sex-specific small-group interventions to prevent sexually transmitted diseases among African-American and Mexican-American women and used clinical markers to assess their efficacy in a randomized, controlled trial. Behavioral changes that lead to prevention of the most common sexually acquired diseases may be considered a proxy for changes that prevent heterosexual transmission of HIV.21 HIV infection could not be used as an outcome variable in our study because of its low prevalence in our heterosexual community.
Methods
Study Procedures
This study was approved by the institutional review boards at the University of Texas Health Science Center and the San Antonio Metropolitan Health District. The efficacy of the intervention was evaluated in a randomized, controlled trial. The chief outcome variable was subsequent infection with Chlamydia trachomatis or Neisseria gonorrhoeae. In order to ensure the high-risk status of the participants, eligibility was limited to women who currently had a nonviral, sexually transmitted disease (chlamydial infection, gonorrhea, syphilis, or infection with trichomonas); eligibility was also limited to speakers of English (8 percent of otherwise eligible Hispanic women were therefore excluded) in order to maximize homogeneity across the racial and ethnic groups.
All 947 eligible women whom we could contact were recruited from public health clinics in San Antonio and were offered the opportunity to participate in this study at our research clinic. After giving their written informed consent, subjects were examined, screened for infection, treated (if necessary), counseled, and interviewed at base line and at their 6- and 12-month visits. Participants were told to return for repeated screening and additional treatment whenever they noticed gynecologic symptoms or were concerned about possible reinfection ("problem" visits). Information was also obtained about infections that had been diagnosed elsewhere. As an incentive to participate, the women received $25 for the initial and 6-month visits and $50 for the 12-month visit.
At each visit a targeted physical examination was performed and genital tract specimens were collected for microbiologic testing; specimens were tested for the presence of N. gonorrhoeae, C. trachomatis, and other pathogens of the lower genital tract. The primary method of screening for gonorrheal and chlamydial infections (at both scheduled follow-up and on- and off-site problem visits) was testing of endocervical samples with DNA probes (Gen-Probe PACE 2 assay, Gen-Probe, San Diego, Calif.). Limited quality-assurance testing was performed for N. gonorrhoeae (cultured in Thayer–Martin medium and detected with Phadebact monoclonal antibody [Boule Diagnostics, Huddinge, Sweden]) and for chlamydia (by culture with McCoy cells, by enzyme immunoassay [Kodak Sure Cell, Eastman Kodak, Rochester, N.Y.], or with fluorescent antibody [Syva Microtrak, Syva, San Jose, Calif.]). Any test that was positive for gonorrheal organisms or chlamydia was considered diagnostic. Testing for the presence of HIV was offered to all participants at each visit.
Standard individual counseling (in a session lasting 15 minutes) was provided by nurse clinicians according to guidelines issued by the Centers for Disease Control and Prevention; counseling was individualized according to the patient's sexual history and her responses to a test of knowledge. Because these clinicians had more time for patient care and education than staff at typical clinics, all participants may have benefited from involvement in the study.
Design of the Intervention
We adapted the AIDS Risk Reduction Model23,24 to guide the collection of qualitative data, the design of the intervention, the structure of the questionnaire, and the overall research strategy. The model builds on and integrates elements of several social and psychological theories, including the Health Belief Model,25 self-efficacy theory,26 decision-making models,27 and diffusion theory.28 Its three stages are recognition of one's risk, commitment to reducing that risk, and following through with that commitment by seeking solutions. Passing from one stage to the next requires knowledge of disease transmission; recognition of personal susceptibility; perception of the costs and benefits of behavioral change, including ways to increase enjoyment from low-risk activities; self-efficacy; and attainment of skills. Social support is helpful at each stage.
Ethnographic data were collected for approximately 18 months from both men and women on conditions of life and lifestyles, values and beliefs, sexual behavior, knowledge and concern about sexually transmitted diseases (including AIDS), perception of risk, male–female relationships, drug and alcohol use, condom use, strategies to motivate behavioral change, and the logistics of intervention. We conducted 25 focus-group interviews and 102 in-depth interviews and made extensive observations in low-income, inner-city communities.29 Ethnograph (Qualis Research Associates, Amherst, Mass.), a computer software package for qualitative data, was used to analyze content.29 Ethnographic data provided insight into how to encourage recognition of risk and motivate change and also identified barriers to change. For example, our data revealed cultural strengths such as strong racial and ethnic identification and pride; a need, particularly among Mexican Americans, to protect one's family from harm or shame; and among African Americans, emphasis on cleanliness and disease prevention, including a proscription against sharing utensils (not "eating behind"), a prohibition that was already extended to needle sharing (not "fixing behind"). Strategies for risk reduction consequently emphasized the disproportionate prevalence of disease among minority groups; possible effects of HIV infection on family members; and, like the importance of not "eating or drinking behind someone," the importance of not having unprotected sex (having sex "behind someone"). Barriers to change, such as fatalism, poor understanding of the transmission of sexually transmitted diseases, and belief that a potential sex partner's risk status can be assessed accurately, were also uncovered. Findings were integrated with the AIDS Risk Reduction Model to create the intervention, with advice from a multiracial and multiethnic team.
Successive iterations of the three-session intervention were pretested with 13 groups (a total of 85 women). The content and objectives of each session (roughly corresponding to the stages in the AIDS Risk Reduction Model) are shown in Table 1. The content of the interventions for African-American and Mexican-American women was largely the same, although emphases and cultural cues varied.
|
Statistical Analysis
Randomization was accomplished at the end of the initial visit by asking participants to select starting times for the intervention from several dates within three weeks of enrollment, without their knowing whether the starting time they selected was a starting time for the intervention or a "dummy" (control) starting time. Starting times for both the intervention and control groups were preassigned to dates randomized and balanced during the enrollment period across times of day, days of the week, weeks of the month, and months of the year. After selecting starting times, participants were informed of their group assignments. This approach to randomization was used to maximize similarity between the groups in motivation (all the women had to agree to select a starting time) and to minimize resentment of group assignment (since the participants believed they had made the group selection). Women in the control group received only standard counseling but were invited to receive the intervention after completing the study. Although the study was not conducted in a blinded manner, group assignments did not appear on interview documents or clinic records. Participants were asked their group assignment only at the end of follow-up interviews, to ascertain the benefits of the intervention.
All laboratory tests that were positive for N. gonorrhoeae or chlamydia, including those performed at on-site and off-site problem visits, were counted in the interval in which they occurred (entry through the 6-month follow-up, or any time after the 6-month follow-up through the 12-month follow-up). The overall count (entry to 12 months) included all tests positive for either infectious agent. Chi-square analysis was used to assess the extent to which behavioral variables were associated with infection and assignment to the intervention group. Multiple logistic-regression analysis was also used to assess associations of behavior with the study group, as well as to determine the effects of intervention, while controlling for differences between the groups in base-line characteristics related to outcome. Analyses were conducted on the basis of the intention to treat.
Results
A total of 424 Mexican-American and 193 African-American women were randomly assigned (after stratification according to race and ethnic group) to the study group (313 women) or the control group (304 women). Most were enrolled before treatment; approximately 10 percent were treated at a public health facility. Of 947 potentially eligible women, approximately 22 percent refused participation and 13 percent did not appear within the required time (30 days from testing, or 14 days from treatment). Rates of participation in the intervention were 90 percent for at least one session, 82 percent for at least two, and 75 percent for three. Enrollment began in January 1993 and ended in July 1994. The 6- and 12-month retention rates were 82 percent (study group, 84 percent, and control group, 80 percent, for a total of 508 women) and 89 percent (study group, 91 percent, and control group, 87 percent, for a total of 549 women), respectively; 26 women with 6-month visits were lost to follow-up at 12 months, and 67 women who missed their 6-month visit returned at 12 months. Repeated screening for chlamydial and gonorrheal organisms was also performed at a total of 260 problem visits (13 at an off-site facility).
Results for women who were lost to follow-up at 6 or 12 months but who had positive laboratory tests at a problem visit before the 6- or 12-month visit are included (five participants at 6 months, of whom three returned at 12 months). Women were excluded from the analysis if laboratory data were missing (because of relocation or a telephone interview), if results were indeterminate, or if any treatments were missed (this was the case for four participants at 6 months and five at 12 months, two of whom previously tested positive and thus were included in the overall count). The analysis included 509 women at 6 months, 545 at 12 months, and 549 for the total study period. Behavioral analysis included results for 477 women who attended both follow-up visits (5 of the total of 482 had missing data). The rates of loss to follow-up did not differ significantly between groups for any subgroup analysis.
Base-line characteristics of the women are provided in Table 2. Low levels of income and education characterized the population, whose ages ranged from 14 to 45 years; 71 percent were younger than 24. Age (<19 years vs. 
19 years) and the number of partners in the three months preceding enrollment were controlled for in multiple logistic-regression analysis because differences between the groups at base line were potentially important and because these variables predicted infection. There were no significant differences between the groups in the type of disease present at base line.
|
|
|
|
This intervention significantly decreased the rates of chlamydial infection and gonorrhea among high-risk African-American and English-speaking Mexican-American women. Prevention of sexually transmitted disease was our goal, but insight into reducing the risk of gonorrheal or chlamydial infection may help reduce the risk of sexually transmitted HIV infection. Because of the very low prevalence of HIV in our population, we did not include HIV seroconversion as an additional outcome measure.
High rates of attendance at intervention sessions and high rates of retention in the study, particularly for a population that may be difficult to follow, strengthen the generalizability of our results. However, generalization to other women at high risk, outside of the groups we studied, should be done cautiously. We attribute the efficacy of our intervention to the integration of behavior-change theory with extensive qualitative data collected in target communities. This approach allowed us to create culturally meaningful strategies to promote the recognition of risk and to stimulate motivation to effect personal change. Other critical elements included a focus on male–female relationships and health-seeking behavior.
Knowledge alone does not necessarily lead to changes in behavior; preventive strategies that focus on providing information have had minimal success.13,24,30,31,32 More successful interventions, tested in randomized, controlled trials, have had a theoretical basis and have promoted self-efficacy and communication skills, as well as skills in the use of condoms, through modeling and role play. These techniques have produced some changes in self-reported behavior, such as increases in the use of condoms.20,21,22,30,32,33,34,35,36,37,38,39,40,41,42 The number of sex partners appears to be more resistant to change.36,40 However, because most studies have not used clinical disease as an outcome variable, evaluation of preventive strategies has been difficult.
Infection is related to both behavior and exposure. Limited changes in behavior may fail to decrease the risk of infection, and the same behavior may yield different results in populations with low and high prevalence of the disease. Moreover, self-reported information about behavior may be biased. Consequently, it is possible to achieve changes in reported behavior without corresponding decreases in rates of infection.43 A prospective cohort study of clients at a clinic for the treatment of sexually transmitted disease found that participants who reported that they always used condoms were as likely to become infected as those who said they never used them.44 The National Institute of Mental Health Multisite Prevention Trial found significantly greater increases in condom use among men and women assigned to an intervention than among control subjects; however, chart review showed a reduction in the incidence of gonorrhea only for men.42 Boyer and colleagues20 showed that the frequency of condom use increased among men who received the intervention; their infection rates, however, were equivalent to those of control subjects.
In our study, sex with multiple partners, sex with an untreated or incompletely treated partner, and five or more acts of unprotected sex in two three-month periods contributed to infection and were affected by the intervention. These factors may affect the rates of infection more directly than increases in the frequency of condom use or decreases in the number of unprotected sexual acts. Changes in condom use may not be sufficient to affect infection rates because the effect probably depends on the context of exposure and may be offset by other factors. For example, persons choosing mutual monogamy as their primary risk-reduction strategy may not derive additional benefit from decreases in the number of unprotected acts because their exposure is already low; persons at higher risk may not derive benefit unless the absolute number of acts is minimal. Moreover, possible benefit may be offset by the pursuit of other high-risk behavior.43 To gauge the usefulness of behavioral measures and the effect of various strategies for change, observation of clinical outcomes is critical.
The well-designed, randomized, controlled trial by Boyer et al.20 of an intervention that was similar in content to ours and based on the same theoretical model found neither behavioral nor clinical differences between women who received the intervention and those who did not. An important difference in their design was the format of the intervention sessions, which were individual, not conducted in small groups. In addition, although their facilitators were sensitive to cultural differences and engaged participants in sex-specific skill-building scenarios, their intervention lacked culturally based preventive strategies derived from a knowledge of the target population, and it did not emphasize relationship issues.
Several studies have demonstrated effects of an intervention for men but not for women.20,45,46 Although both men and women face common barriers to behavioral change, such as misperceptions about risk and concern about poverty,29 women must also confront power imbalances, unrealistic romantic ideals, and a need to minimize disruptions in their relationships.12,13,14,15,16,18,19,47 Interventions are more likely to be successful if they are culturally relevant and specifically address culturally defined aspects of the relationships between the sexes.12,14,29,47,48,49,50,51 Moreover, it has been suggested that women gain strength from interactions with others52; discussing problematic relationships with other women may be particularly beneficial in increasing personal awareness and reinforcing resolve.12,14,52 Incarcerated women who did not receive training in communication skills but who participated in open discussions of issues related to sex roles and power spontaneously shared experiences and personal approaches to communication, and in so doing they improved their communication skills as much as those who had explicit training.53 We believe that the open discussions of relationship issues in our intervention contributed to bonding among the women, helped them confront their own situations, and encouraged empowerment and action.
Timeliness in treating sexually transmitted diseases and full compliance with treatment are also important. A population-based study in rural Tanzania indicated that improved management of cases of sexually transmitted diseases significantly reduced the incidence of HIV infection during a two-year follow-up.54,55 Although we focused on changing sexual behavior, our intervention also addressed health-seeking behavior, early recognition of symptoms, and prompt, complete treatment of both the patient and her partner (or partners). One type of behavior that contributed to reduced rates of infection was avoidance of sex with a partner who was untreated or treated incompletely.
Most studies of interventions have follow-up intervals of one to six months, and the effects of the intervention typically diminish substantially after several months.35 However, in our study, reductions in the incidence of gonorrheal or chlamydial infection were larger at 12 months (49 percent) than at 6 months (34 percent). For some of the women, the effects of intervention may have been delayed. The interviews and examinations at six months could have "boosted" the effects of intervention, and it may have taken some women more time to disengage from high-risk partners and to discontinue high-risk behavior.
Interventions that reduce the risk of contracting a sexually transmitted disease such as gonorrhea or chlamydia may also prevent HIV infection. Heterosexual contact is the leading cause of AIDS in women,8 and persons with other sexually transmitted diseases have a risk of contracting HIV if exposed that is 3 to 50 times that of an uninfected person.2,55,56 This study shows that a culturally relevant behavioral intervention grounded in theory can reduce the frequency of gonorrheal and chlamydial infections for at least 12 months in minority women at high risk.
Supported by a grant (U01 AI31498) from the National Institute of Allergy and Infectious Diseases.
We are indebted to Alan E.C. Holden for assistance with the analysis of infection rates and behavioral outcomes, data editing, and overall statistical analysis; to Barbara Elizondo for assistance with data editing and statistical analysis; to Brooke Goodfriend for coordination of the behavioral project; to Carroll Brooks for past and continuing assistance with clinical issues related to this project; to the members of the Mexican-American and African-American communities who assisted with collection of ethnographic data and the design of the intervention; and to all the staff members at our research clinic and the clinic for sexually transmitted diseases of the San Antonio Metropolitan Health District who made this project possible.
Source Information
From the Departments of Obstetrics and Gynecology (R.N.S., J.M.P.) and Microbiology (S.T.P.) and the School of Nursing (S.T.P., J.D.C.), University of Texas Health Science Center at San Antonio, San Antonio; the Department of Obstetrics and Gynecology, East Carolina University, Greenville, N.C. (E.R.N.); Research and Evaluation, Mujeres Project, San Antonio, Tex. (R.R.); and the San Antonio Metropolitan Health District, San Antonio, Tex. (F.A.G.).
Address reprint requests to Dr. Shain at the Department of Obstetrics and Gynecology, University of Texas Health Science Center at San Antonio, 7703 Floyd Curl Dr., San Antonio, TX 78284-7836.
References
| |||||||||||||||||||||||||||||||||
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved. |
Previous
