HIV COUNSELING AND TESTING have been the cornerstone of HIV prevention efforts in the United States since the development of an HIV antibody test in 1984.1 Although HIV counseling and testing are widely supported, their efficacy for changing behaviors or preventing sexually transmitted diseases (STDs) has only recently been demonstrated.2–8 Until recently, brief HIV counseling models, such as the Centers for Disease Control and Prevention’s (CDC’s) client-centered HIV prevention counseling model,1 were believed to be too brief to have an effect on preventing HIV. On the other hand, longer multisession interventions, believed more likely to be effective, were often impractical in public health settings.
In 1998, results from Project RESPECT, a multicenter, randomized, controlled trial showed that CDC’s brief HIV prevention counseling model and a longer enhanced HIV prevention counseling model based on behavior change theory were both effective at changing high-risk sexual behaviors and reducing new STDs.2 Participants in the trial were sexually active heterosexual men and women who enrolled from 1 of 5 participating STD clinic sites.
Although overall efficacy was similar for the 2 counseling approaches (30% STD reduction at 6 months and 20% reduction at 12 months), the efficacy of counseling in subsets of vulnerable populations is not clear. Some studies suggest that the effect of counseling could be limited in women compared with men4,7,9–13 and that adolescents are resistant to counseling efforts.14 It is also sometimes believed that persons who have been previously HIV tested and counseled are resistant to additional counseling.13,15 Our objective, therefore, was to use data collected from a large prevention counseling and testing trial to estimate counseling efficacy among certain subpopulations.
Materials and Methods
Study Design and Procedures
For this subgroup analysis, we used data collected from July 1993 through September 1996 for a randomized, controlled trial of counseling efficacy (Project RESPECT). The trial methods have been previously described.2 In brief, the multisite study was conducted among clients attending 5 public STD clinics (Baltimore, MD; Denver, CO; Long Beach, CA; Newark, NJ; and San Francisco, CA). English-speaking clients aged 14 years or older who came for STD examinations during the study period and reported vaginal sex during the preceding 3 months were asked to participate in the study. Men who identified themselves as homosexual or who reported having had a male sex partner during the preceding 12 months were excluded.
Participants were randomized to 4 arms with varying intensities of 1-on-1 safer-sex counseling with a trained HIV counselor or clinician.
Arm 1: Enhanced Counseling.
Participants (n = 1438) were assigned to receive 4 counseling sessions with a counselor (1 20-minute session followed by 3 1-hour sessions) aimed at changing attitudes, self-efficacy, social norms regarding condom use and other safe sex behaviors, as well as assisting clients to develop a realistic risk-reduction plan). The enhanced counseling model was based on behavior change theories.16
Arm 2: Brief Prevention Counseling.
Participants (n = 1447) were assigned to receive 1 20-minute session (which included a personalized risk assessment aimed at developing an incremental, realistic behavior change step) followed by a second 20-minute session conducted when HIV test results were provided 7 to 10 days later. Based on successes or challenges to the behavior change step negotiated in the first session, a risk-reduction plan was developed in the second session. This counseling model was based on CDC’s client-centered HIV counseling, first recommended for use with HIV testing in 1993.1
Arm 3: HIV Prevention Education.
Participants (n = 1443) were assigned to receive 2 didactic sessions with a clinician. The first session (5 minutes) was a brief, personalized educational message about HIV and STD; the second posttest counseling session (5 minutes) was another personalized educational message when HIV test results were provided. This session was a standardized version of the educational messages that were typically provided with HIV testing at the time of the trial.
Arm 4: HIV Prevention Education.
Participants (n = 1430) were assigned to receive the same intervention as those in arm 3, but because arm 4 was developed to measure only the behavioral and STD prevention efficacy of repeated interviews, these participants were not followed.
Participants in arms 1, 2, and 3 were actively followed and asked to respond to questionnaires at 3, 6, 9, and 12 months and be tested for STDs at 6 and 12 months after enrollment. All participants could return at any time for an STD examination if they wanted a checkup, had symptoms, or had a partner with symptoms. Because arm 4 participants had no scheduled follow up and did not consistently undergo screening, they were excluded from this analysis.
At the enrollment visit, participants were interviewed to obtain information on sociodemographic characteristics, sexual behavior, and STD history. Genital examinations and laboratory testing for STDs were performed for all participants. The main outcomes were self-reported condom use, number of sex partners, and incident STDs defined by laboratory tests. For STD outcomes, infection with Neisseria gonorrhea was defined as a positive culture or, for men, Gram-negative intracellular diplococci on a Gram stain of a urethral swab; Chlamydia trachomatis was defined as a positive polymerase chain reaction assay result for endocervical swabs (women) and urine samples (men). Syphilis was diagnosed on the basis of suggestive history and physical examination with supportive treponemal and nontreponemal antibody test results. HIV infection was diagnosed as repeatedly reactive enzyme immune assay results for HIV antibody with a confirmatory test result. The Project RESPECT protocol was reviewed and approved by the Institutional Review Boards at the CDC and each participating site.
For this analysis, we used an intent-to-treat approach; that is, data on all participants were included in the analysis, regardless of whether they completed their assigned intervention. We stratified participants on 9 subgroups. Five a priori subgroups were identified in the study’s initial protocol: 1) sex (male, female); 2) age (<20, 20-25, >25 years); 3i) site (Baltimore, Denver, Newark, Long Beach, San Francisco); 4) (intravenously) STD diagnosis at enrollment (no, yes); and 5) report at enrollment of a prior HIV test (no, yes). These subgroups were selected because there could be differences in counseling efficacy for these subgroups. Four additional subgroups that were not considered a priori were included because recent literature or experience of principal investigators suggested a possible effect on counseling, and these are considered as more exploratory The additional 4 subgroups considered later included: 5) race/ethnicity (black, white, Hispanic, other); 6) education (<12th grade, 12th grade, >12th grade); 7) reported history of injection drug use (never, ever); and 8) reported history of exchanging sex for money or drugs (never, ever).
The outcome of interest was STD incidence, based on laboratory diagnosis during the 12 months after the date of enrollment. We used SAS software version 8.2 (SAS Institute, Inc., Cary, NC) for statistical analysis. For each subgroup, we estimated the STD incidence during the 12 months after intervention and compared interventions using chi-squared tests and odds ratios (ORs) with 95% confidence intervals (CIs). The sample size for the trial was chosen to detect overall differences in interventions but not differences in specific subgroups. Thus, statistically significant differences were not expected because the sample size was too small for most subgroups. To control for potential confounding effects, we used multiple logistic regressions and created separate models for each of the subgroups. We entered potential confounders (the other subgroups) into each model and determined the adjusted ORs and 95% CI. To determine the number of STDs prevented per year per 100 persons counseled for each subgroup, we subtracted STD incidence in the counseling interventions from that in the educational intervention.
From July 1993 through June 1995, 13,471 eligible clients were asked to participate in Project RESPECT, and 5833 (43%) agreed to enroll. Those who participated and those who refused to participate in Project RESPECT were similar in age, racial/ethnic background, and education level, although compared with those who refused, participants were more likely to be women (OR, 1.49; 95% CI, 1.44-1.55), to have had an STD at enrollment (OR, 1.19; 95% CI, 1.14-1.24), and to have been previously tested for HIV (OR, 1.13; 95% CI, 1.08-1.18).2 After excluding clients with positive HIV test results at enrollment (n = 75) and arm 4 participants who did not routinely receive STD screening (n = 1430), 4328 enrollees remained for the subset analysis.2
Of 4328 participants in the subset analysis, 2457 (57%) were male and 1871 (43%) were female. Participants were mostly young (median age, 25 years) and of minority race/ethnicity (59% black, 19% white, 16% Hispanic, 6% other races). Less than one third had an education beyond high school (12th grade). Most participants had had a prior HIV test (71%), did not have an STD at enrollment (73%), had never exchanged sex for money or drugs (79%), had never injected drugs (92%), and completed all intervention sessions for their assigned study arm (80%). This overall distribution was similar for participants in each intervention group (Table 1) except for intervention completion rates, which were higher for participants assigned to brief counseling than for those assigned to enhanced counseling.
During the 12-month follow-up period, 549 (13%) of the participants had new STDs: 7.3% had chlamydia, 6.3% gonorrhea, 0.6% syphilis, and 0.2% HIV. Some participants had multiple diagnoses.
For nearly every subgroup analyzed, there were fewer incident STDs among those in the enhanced and brief counseling groups than those in the educational messages group (Table 1). In subgroups by age, for example, 68 (26.6%) adolescents (<20 years old) assigned to the educational intervention had new STDs compared with 44 (17.2%) assigned to enhanced counseling (OR, 0.57; 95% CI, 0.37-0.90), and 44 (17.5%) assigned to brief counseling (OR, 0.58; 95% CI, 0.37-0.92). Similarly, older participants (>20 years old) assigned to enhanced and brief counseling interventions had fewer STDs than those assigned to the educational intervention (Table 1). In subgroups by sex, 117 (14.3%) men assigned to the educational intervention had new STDs compared with 92 (11.4%) men assigned to enhanced counseling (OR, 0.77; 95% CI, 0.57-1.04) and 104 (12.6%) men assigned to brief counseling (OR, 0.84; 95% CI, 0.63-1.12). The findings were similar for women.
All other subgroups assigned counseling interventions (enhanced or brief) also had fewer STDs than those assigned educational messages (Table 1). In crude analysis, differences between brief counseling and educational messages were statistically significant (P <0.05) for adolescents, persons from Baltimore, and persons who had ever exchanged sex for money or drugs. Similarly, differences between those in the enhanced counseling groups and the educational messages group were significant only for persons who never inject drugs and persons with an STD at enrollment.
When we conducted our multivariate analysis by controlling for potential confounders using separate models for each of the subgroups, the adjusted ORs (Table 2) were similar to the crude ORs for most subgroups, suggesting there was little confounding in the groups we compared. The subgroups that had had statistically significant differences in STD incidence in the crude analyses also had statistically significant differences in the adjusted (multivariate) analyses (Table 2). In addition, significant differences in STD incidence were found between blacks in the brief counseling group compared with blacks in the educational messages group (Table 2). Overall, statistically significant differences between counseling (brief or enhanced) and educational messages were observed for adolescents, persons from Baltimore, persons who exchange sex for money or drugs, persons with STD at enrollment, persons who never inject drugs, and blacks (Table 2).
For most subgroups, the efficacy of brief counseling was similar to enhanced counseling (Table 1), with crude and adjusted ORs at or near 1.0 (Table 2). There were a few possible exceptions, although the differences were not statistically significant. Participants who were white and who were assigned to enhanced counseling had fewer STDs than those who were assigned to brief counseling (OR, 0.57) (Table 2). Conversely, persons who reported having ever exchanged sex for money and/or drugs (OR, 1.45) and persons who had reported having ever injected drugs (OR, 3.9) had more STDs if they were assigned to the enhanced counseling group than to the brief counseling group (Table 2).
The numbers of STDs prevented per 100 persons counseled were appreciable for most subgroups assigned to brief and enhanced counseling (Table 3), particularly adolescents (9.4) and persons with an STD at enrollment (8.4). The subgroups with greater numbers of STDs prevented had the highest STD incidence overall: persons with STD at enrollment (24%), adolescents (20%), and blacks (16.5%).
Our findings in this study were unexpected. HIV/STD prevention counseling (brief or enhanced) resulted in fewer STDs for many subgroups for whom counseling was not found to be effective previously, including adolescents,14 persons who had an STD,17–19 and blacks.13,20 We also found counseling had similar efficacy for women and men, although earlier studies have suggested that counseling could be of limited value for women.4,11,12,17,21 However, some of these earlier studies11,12,16 were based on reported behavior change rather than incident STDs. These differences could be because our study measured STD incidence, whereas the others measured sexual behaviors; these 2 outcomes are not always highly correlated.22 An interesting finding was that the brief interactive counseling was similar to enhanced counseling in reducing new STDs for most (92%) subgroups. This was also observed among all participants in the original trial (Project RESPECT), but we did not expect this to be the case for all subgroups. Studies have suggested that certain groups such as adolescents14,20 and persons with STDs or repeated STDs could require multisession counseling. This could occur because people with repeat STDs are sometimes characterized as recidivists and being resistant to using condoms or reducing risky behaviors.19,21 Adolescents have also been characterized as needing specialized counseling14,20 because of their high-risk behaviors and particular vulnerabilities (eg, peer pressure, lack of knowledge). Some researchers have hypothesized that some women might not insist on condom use with male partners or engage in other safe sex practices as a result of power imbalances and the need to minimize disruption in their relationships.23–26 People who have been previously counseled and tested for HIV are sometimes believed not to benefit from additional counseling (even though many are engaging in high-risk behaviors). Our results suggest otherwise; most participants (regardless of situation) benefited more from the enhanced counseling or the brief risk-reduction counseling than from the educational messages. The brief counseling approach and the enhanced counseling approach both prevented similar numbers of STDs, regardless of subgroup.
However, there were a few exceptions, in which the STD reduction in brief counseling was more than in enhanced counseling for persons who exchanged sex for money or drugs and who used injection drugs. This was particularly surprising because conventional thinking is that multiple interventions are required to change highly evolved, long-practiced behaviors. These differences were not statistically significant and could have been the result of chance. On the other hand, ours is not the first study showing that brief intervention can benefit such high-risk individuals; one recent study found that a brief intervention for changing risky behaviors among injection drug users was as effective as a longer session.27 Another study among patients who abused alcohol found that a brief, motivational intervention was as effective as a longer intervention in achieving alcohol cessation.28 Furthermore, a recent metaanalysis on the effect of HIV prevention interventions on the sex behaviors of drug users in the United States found that additional sessions did not appear to increase the effect of the interventions.29
To determine which subgroups would benefit most from prevention counseling, we calculated the number of STDs prevented per 100 persons counseled; an appreciable number of STDs were prevented for all subgroups. However, more STDs were prevented for adolescents and persons with an STD at enrollment than for participants in other subgroups. These 2 subgroups had the highest STD incidence in this trial and also benefited the most from the counseling interventions; thus, those at highest risk benefited the most from counseling.
Our study has some limitations. First, the study was powered to detect differences between interventions and not differences among specific subgroups; thus, we had insufficient sample size in most subgroups to ensure that suggestive findings were statistically significant, so some findings could have been the result of chance. Second, because the eligibility criteria for the study required that all participants get HIV counseling and testing, we do not know if those who refused to enroll were less receptive to counseling than those who agreed to be in the study; this could limit generalizability. Third, because HIV seroconversions were unusual, it is not certain the STD reduction found with counseling would hold true for HIV. Fourth, because we studied subgroups in STD clinics, we do not know if the findings are generalizable to other settings such as private clinics or adolescent clinics. Finally, not all subgroups assessed in the subset analysis were considered a priori; findings from post-hoc subset analysis should be interpreted with caution. This is true especially if the intervention might influence whether a participant becomes a member of a particular subgroup.30 In this analysis, the subgroups that were not considered a priori (education, race/ethnicity, history of exchanging sex for money or drugs, and history of injection drug use) were clearly defined groups that were determined from baseline data and unrelated to the intervention assignment.
In summary, the results of our subset analysis indicate that HIV/STD prevention counseling prevents new STDs for almost all types of STD clinic clients. More importantly, the counseling interventions prevented new STDs for many subgroups commonly felt to be resistant to counseling such as adolescents, persons who already have an STD, blacks, persons who have been previously tested for HIV, injection drug users, and persons who exchange sex for money or drugs. Both counseling interventions had similar efficacy for women and men. These results have important implications for STD programs. Counseling clients in STD clinics can contribute significantly to reducing new STDs if counselors use interactive, risk-reduction counseling that engages clients in an exploration of personal risk and encourages them to develop and commit to a realistic, risk-reduction plan. Clinics with limited resources could focus on members of the 2 groups that benefited the most from counseling: adolescents and persons with an STD. Ideally, interactive risk-reduction counseling approaches would be provided to all persons because counseling is beneficial to all subgroups of STD clinic clients.
1. Centers for Disease Control and Prevention. HIV Counseling, Testing and Referral Standards and Guidelines. Atlanta, GA: US Dept of Health and Human Services, Centers for Disease Control and Prevention, 1994.
2. Kamb ML, Fishbein M, Douglas JM, et al. Efficacy of risk reduction counseling to prevent human immunodeficiency virus and sexually transmitted diseases: A randomized controlled trial. JAMA 1998; 280:1161-1167.
3. Shain RN, Piper JM, Newton ER, et al. A randomized, controlled trial of a behavioral intervention to prevent sexually transmitted diseases among minority women. N Engl J Med 1999; 340:93-100.
4. The National Institute of Mental Health (NIMH) multi-site HIV prevention trial: Reducing HIV sexual risk behavior. Science 1998; 280:1889-1894.
5. O’Donnell CR, O’Donnell L, Sandoval A, et al. Reductions in STD infections subsequent to an STD clinic visit. Using video-based patient education to supplement provider interactions. Sex Transm Dis 1998; 25:161-168.
6. Cohen D, MacKinnon D, Dent C, et al. Group counseling at STD clinics to promote use of condoms. Public Health Rep 1992; 107:727-731.
7. Cohen D, Dent C, MacKinnon D. Condom skills education and sexually transmitted disease re-infection. J Sex Res 1991; 28:139-144.
8. Cohen DA, Dent C, MacKinnon D, et al. Condoms for men, not women: Results of brief promotion programs. Sex Transm Dis 1992; 19:245-251.
9. Wortley PM, Fleming PL. AIDS in women in the United States: Recent trends. JAMA 1997; 278:911-916.
10. Guinan ME, Leviton L. Prevention of HIV infection in women: Overcoming barriers. J Am Med Womens Assoc 1995; 50:74-77.
11. Ickovics JR, Morrill AC, Beren SE, et al. Limited effects of HIV counseling and testing for women: A prospective study of behavioral and psychological consequences. JAMA 1994; 272:443-448.
12. Ickovics JR, Druley JA, Grigorenko EL, et al. Long-term effects of HIV counseling and testing for women: Behavioral and psychological consequences are limited at 18 months posttest. Health Psychol 1998; 17:395-402.
13. St. Lawrence JS, Eldridge GD, Brasfield TL. Behavioral and psychological consequences of HIV antibody counseling and testing with African-American women. Women’s Health: Research on Gender, Behavior, and Policy 1998; 4:135-153.
14. Clark LR, Brasseux C, Richmond D, et al. Effect of HIV counseling and testing on sexually transmitted diseases and condom use in an urban adolescent population. Arch Pediatr Adolesc Med 1998; 152:269-273.
15. Schwarcz SK, Spitters C, Ginsberg MM, et al. Predictors of human immunodeficiency virus counseling and testing among sexually transmitted disease clinic patients. Sex Transm Dis 1997; 24:347-352.
16. Ajzen I, Fishbein M. Understanding Attitudes and Predicting Behavior. Englewood Cliffs, NJ: Prentice Hall, 1980.
17. Osewe PL, Peterman TA, Ransom RL, et al. Trends in the acquisition of sexually transmitted diseases among HIV-positive patients at STD clinics, Miami 1988-1992. Sex Transm Dis 1996; 23:230-233.
18. O’Campo P, DeBoer M, Faden RR, et al. Prior episode of sexually transmitted disease and subsequent sexual risk-reduction practices: A need for improved risk-reduction interventions. Sex Transm Dis 1992; 19:326-330.
19. Zenilman JM, Erickson B, Fox R, et al. Effect of HIV posttest counseling on STD incidence. JAMA 1992; 267:843-845.
20. Stanton BF, Li X, Ricardo I, et al. A randomized, controlled effectiveness trial of an AIDS prevention program for low-income African-American youths. Arch Pediatr Adolesc Med 1996; 150:363-372.
21. Boyer CB, Barrett DC, Peterman TA, et al. Sexually transmitted disease and HIV risk in heterosexual adults attending a public STD clinic: Evaluation of a randomized controlled behavioral risk-reduction intervention trial. AIDS 1997; 11:359-367.
22. Peterman TA, Lin LS, Newman DR, et al. Does measured behavior reflect STD risk? An analysis of data from a randomized controlled behavioral intervention study. Sex Transm Dis 2000; 27:446-451.
23. Aral SO, Wasserheit JN. Interactions among HIV, other sexually transmitted diseases, socioeconomic status, and poverty in women. In: O’Leary A, Jemmott LS, eds. Women at Risk: Issues in the Primary Prevention of AIDS. New York: Plenum Press, 1995:13-41.
24. O’Leary A, Jemmott LS. General issues in the prevention of AIDS in women. In: O’Leary A, Jemmott LS, eds. Women at Risk: Issues in the Primary Prevention of AIDS. New York: Plenum Press, 1995:1-12.
25. Sobo EJ. Finance, romance, social support, and condom use among impoverished inner-city women. Hum Organ 1995; 54:115-128.
26. Miller JB, Stiver IP. The Healing Connection: How Women Form Relationships in Therapy and in Life. Boston: Beacon Press, 1997.
27. Rhodes F, Malotte CK. HIV risk interventions for active drug users. In: Oskamp S, Thompson SC, eds. Understanding and Preventing HIV Risk Behavior: Safer Sex and Drug Use. Thousand Oaks, CA: Sage Publications Inc, 1996:211-214.
28. Project MATCH Research Group. Matching alcoholism treatments to client heterogeneity. J Stud Alcohol 1997; 58:7-29.
29. Semaan S, Des Jarlais DC, Sogolow E, et al. A meta-analysis of the effect of HIV prevention interventions on the sex behaviors of drug users in the United States. J Acquir Immune Defic Syndr 2002; Suppl 1:S73-S93.
30. Oxman AD, Guyatt GH. A consumer’s guide to subgroup analysis. Ann Intern Med 1992; 116:78-84.