Enhancing Condom Use Among Black Male Youths: A Randomized Controlled Trial

Abstract

Objectives. We tested the efficacy of a brief intervention to promote correct and consistent use of condoms among Black male youths attending sexually transmitted infection (STI) clinics in 3 southern US cities.

Methods. In 2010 to 2012, we screened (n = 1102) and enrolled (n = 702) youths aged 15 to 23 years who identified as Black and reported recent (past 2 months) sexual activity and randomized them to a private, brief, interactive intervention (n = 349) or an attention-equivalent control condition (n = 353). Assessments occurred at baseline and 2 and 6 months after the intervention.

Results. At 6 months, with adjustment for age and pretest nonequivalence of the outcome variable, an estimated odds ratio (EOR) of 1.63 (95% confidence interval [CI] = 1.07, 2.49; P = .02) indicated efficacy for correct condom use. An adjusted generalized estimating equations model with both 2- and 6-month condom use variables produced an EOR of 1.49 (95% CI = 1.06, 2.08; P = .02). We did not observe significant effects on chlamydia and gonorrhea incidence.

Conclusions. This brief intervention, delivered as part of STI clinical care, could help alleviate the disproportionate STI–HIV burden among young Black men.

In the United States, young Black men continue to be disproportionally affected by sexually transmitted infections (STIs), including HIV.^1–3 The problem is exacerbated in the southern United States.⁴ Unfortunately, a primary method of preventing disease acquisition and transmission among young Black men is vastly underused: the consistent and correct use of male latex condoms.^5–7 The correct use of condoms is especially important in light of recent findings that only correct use is protective against STIs.⁸

Unfortunately, most behavioral interventions have focused solely on young women^9–13 or on men who have sex with men.^14–20 Few behavioral interventions have focused on young Black men who have sex with women.^21–23 Focus on the Future (FoF),²³ classified by the Centers for Disease Control and Prevention (CDC) as an evidence-based intervention,²⁴ was shown to reduce STI incidence among Black men aged 18 to 29 years. This 1-hour, single-session, clinic-based program was designed for patients testing positive for STIs. We adapted FoF for use with Black male youths aged 15 to 23 years attending STI clinics, regardless of diagnosis.

We conducted a randomized, controlled trial of the adapted version of FoF in 3 US cities. We hypothesized that Black male youths randomized to receive FoF would report greater correct and consistent condom use and be less likely to test positive for urethral gonorrhea and chlamydia. Because of the prevalence of STIs and HIV in the southern United States, we conducted our efficacy trial in that region.

METHODS

We conducted a pretest–posttest, randomized controlled trial with a 6-month observation period. Inclusion criteria were being male, self-identifying as Black, being aged 15 to 23 years, engaging in penetrative sexual intercourse at least once in the past 2 months, and not knowingly being HIV positive. Recruitment occurred from 2010 through 2012 in a primary site (New Orleans, LA) and 2 secondary sites (Baton Rouge, LA, and Charlotte, NC).

Recruitment procedures were standard across all study sites. After a potentially eligible patient completed his medical exam, clinic staff notified a research staff member. Researchers used a private exam room to screen for eligibility. We asked persons younger than 18 years whether they had a parent or guardian who would be able to provide in-person parental consent for enrollment. After an eligible person assented, we asked him for permission to contact 1 parent or guardian to obtain consent. All others provided written informed consent. After enrollment, we administered an audio computer-assisted self-interview in a private area.

After the survey, we randomized participants with a computer-generated algorithm; we concealed the allocation technique. We transposed the randomized sequence to separate index cards, which we placed in sealed envelopes. We placed the envelopes in prepackaged and ID-numbered enrollment folders. For each new enrolled participant, project staff used the next available ID number (and folder) and only opened the envelope after baseline survey completion.

Next, participants provided a urine sample for gonorrhea and chlamydia testing. They received results within 2 weeks. Research staff contacted those testing positive to return to the clinic for prompt treatment. We collected contact information and compensated participants with a $50 gift card. We frequently recontacted participants to return for follow-up assessments (occurring ∼2 months and ∼6 months after baseline). We again provided $50 gift cards at these follow-up assessments. The 6-month follow-up assessment included retesting for gonorrhea and chlamydia.

Interventions

The intervention condition used an adapted version of FoF, an evidence-based intervention.^23,24 We adapted the program for use with a younger population of Black males (adolescent boys aged 15–17 years were not included in the original trial). Primary adaptions involved greater reliance on graphic and visual presentations, as well as a greater emphasis on negotiating condom use. We developed adaptations through formative research to tailor the intervention to Black male youths. We conducted a series of iterative one-to-one testing sessions followed by qualitative interviews. We chose questions that would help us discern which aspects of the program participants felt were most and least applicable to their lives, which behaviors they could see themselves changing as a result of the intervention session, and what suggestions they had for making each program objective relevant to sexually active Black male youths. We excluded participants in the qualitative research phase from participating in the randomized, controlled trial.

FoF is a tailored intervention, delivered in a face-to-face setting. The adapted version had 7 objectives (Table 1). We addressed objectives in various sequences according to individual needs. FoF is a sex-positive intervention program that lasts less than 1 hour and is designed for use in clinics. The average time to deliver our version was 54.1 minutes.

TABLE 1—

Condom Use Intervention Objectives for Black Male Youths: Focus on the Future, Louisiana and North Carolina, 2010–2012

Objective	Implementation
Build rapport	Discuss participant’s goals for the future.
Make participant aware that Black males experience disproportionate share of AIDS cases in the United States	Show posters and graphs illustrating proportions.
	Communicate that Black males also experience a disproportionate number of other STI’s (chlamydia, gonorrhea, etc.).
Capitalize on participant’s positive condom use experience or, if no such experiences, capitalize on existing positive behavior by talking about how often he uses condoms and some issues he has had with condoms.	Use Teen Condom Issues sheet if participant is not forthcoming about issues he has with condoms.
Provide guided practice in correct application and use of condoms using anatomical model.	Have participant apply condom to model.
	Review 8-step process and give reminder card listing correct application steps.
Achieve satisfactory fit and feel.	Explain lubricant, including what it is, how and when to use it.
	Show lubricant pictorial slide show of correct and incorrect types of lubricant.
	Conduct baby oil demonstration to show that oil-based products will cause latex condoms to tear.
Explain that planning and negotiating condom use is critical to HIV prevention.	Discuss appropriate time to put on a condom.
	Discuss how condom use fits into participant’s life (e.g., how he can begin using condoms with current sexual partners), condom negotiation skills and tips, partner communication.
	Use role playing to demonstrate how to communicate about condom use.
Summarize session.	Review intervention and ask participant what 1 thing he will remember.
	Ask for and then answer any questions participant may have.

Note. STI = sexually transmitted infection.

Much of the program was designed to help Black male youths overcome barriers to condom use. Established barriers to using condoms consistently and correctly among Black male youths include perceived reduction in sexual pleasure and ability to perform,^25–29 problems with condom fit and feel,²⁶ and reduced self-efficacy to use condoms.^30,31 To reliably deliver the intervention, we conducted a 2-day training session for 5 health educators. Throughout the study period, we tape-recorded every 10th intervention session for quality assurance monitoring by R. A. C.

The attention-equivalent control condition consisted of 40 PowerPoint slides focused on male and female reproductive anatomy, basic knowledge of various STIs, and prevention (1 slide). Health educators specifically refrained from elaborating on the content of these slides unless participants asked direct questions. In that case, the educators answered in 1 or 2 sentences. The average length of time to deliver this condition was 48.5 minutes.

In both conditions, we gave participants free access to a wide variety of condoms and lubricants and told them to take as many as they could fit into a bag we provided. Control participants received a paper bag; intervention participants received a canvas or velvet bag marked with the study logo (FoF). On average, control and intervention participants took about 19.7 and 25.2 condoms, respectively. They selected about 5.6 (control condition) or 11.2 (intervention condition) single-use vials of lubricant. We provided condoms and lubricants in an identical manner at the 2- and 6-month follow-up assessments. We also informed all participants that condoms could be obtained at the clinic at any time. A complete trial protocol is available by request to R. A. C.

Measures

We shipped urine specimens to Quest Diagnostics (Madison, NJ), which performed the Aptima Combo 2 Assay (Gen-Probe, Bedford, MA), a target amplification nucleic acid probe test that uses target capture for in vitro qualitative detection and differentiation of ribosomal RNA from Chlamydia trachomatis (CT) and Neisseria gonorrheae (GC) to diagnose urogenital disease.

We collected self-reported data at baseline, 2 months, and 6 months. We assessed the primary dependent behavioral variable (correct and consistent use) over a 2-month recall period and constructed it with an established 3-step process.²³ First, we subtracted the frequency of condom use for penile–vaginal intercourse from the frequency of penile–vaginal intercourse. Second, we dichotomized this difference to compare those reporting 100% condom use to those reporting less than 100%. Third, we defined correct and consistent use as 100% use with no condom errors; participants who reported 100% use, but who also reported any of 5 condom use errors (slippage during intercourse, slippage during withdrawal, breakage, late application, or early removal) were classified as not achieving correct and consistent use.

The independent variable was group assignment (intervention vs control). The dependent variables were correct and consistent use at the 2-month and 6-month follow-up assessments and laboratory-confirmed incidence of CT and GC at the 6-month follow-up assessment.

Data Analysis

We adhered to an intent-to-treat design: there were no cases of crossover between groups. First, to identify statistically relevant covariates, we tested bivariate associations between group assignment and various demographic measures and potential confounders for significance. Except where otherwise indicated, we compared groups on nominal and ordinal variables with the χ² or Fisher exact test and on interval and ratio variables with the t test.

After we performed unadjusted group comparisons on the primary dependent behavioral variable, we fit logistic regression models to examine group differences in correct and consistent use at 2 months and at 6 months, with adjustment for correct and consistent use at baseline, history of CT or GC, age, and treatment group. We also fit a logistic regression model with generalized estimating equations to estimate the adjusted association of correct and consistent use with each covariate over both follow-up visits. After we made unadjusted group comparisons of chlamydia and gonorrhea, we fit a logistic regression model to relate a positive test for chlamydia or gonorrhea at 6 months to history of CT or GC, age, correct and consistent use at baseline, and treatment group.

Finally, we used a nonparametric rank sum test to determine whether change scores (baseline to 2 months and baseline to 6 months) significantly differed by group assignment for correct and consistent use as well as whether change scores (history to 6 months) significantly differed by group assignment for CT or GC. We performed data analyses with SPSS version 20.0 (IBM, Armonk, NY) and SAS version 9.4 (SAS Institute Inc, Cary, NC). A P value less than .05 defined statistical significance.

RESULTS

We screened 1139 Black male youths. Figure A (available as a supplement to the online version of this article at http://www.ajph.org) displays the flow diagram for the study. The final sample for all 3 cities totaled 702 participants.

Retention at the 2-month assessment was comparable between the intervention and control conditions: respectively, 70.8% and 73.9% (P = .35). Similarly, we observed no significant differences in retention rates between intervention (78.5%) and control (73.9%) groups at the 6-month assessment (P = .16).

Characteristics of the Sample

Participants' mean age was 19.6 years (SD = 1.9). The majority (64.4%) reported finishing high school or college or earning a general equivalency diploma. About one half (50.3%) reported current school enrollment. The mean number of lifetime sexual partners was 18.4 (SD = 19.1; range = 1–95), and the mean number of sexual partners in the past 2 months was 3.0 (SD = 6.2; range = 1–90). Nearly all participants (92.8%) had been tested for HIV at least once. We also examined whether site differences occurred relative to correct and consistent use and to CT–GC incidence. We detected no significant differences by site on the 2-month behavioral outcome (P = .31), 6-month behavioral outcome (P = .17), or 6-month clinical outcome (P = .56; data not shown). Table 2 displays a baseline comparison between intervention and control participants for demographics, sexual behaviors, and STIs. We found significant differences in age, correct and consistent use, and self-reported history of CT–GC. We therefore controlled for these in multivariable models.

TABLE 2—

Baseline Comparison of Demographic, Behavioral, and Clinical Characteristics Among Black Male Youths: Focus on the Future, Louisiana and North Carolina, 2010–2012

Characteristic	Control Group (n = 353), No. (%) or Mean ±SD	Intervention Group (n = 349), No. (%) or Mean ±SD	P
Age, y	19.5 ±2.0	19.8 ±1.8	.02
Enrolled in school	178 (51.4)	171 (49.1)	.54
Highest grade completed			.48
≤ 9	28 (8.1)	25 (7.2)
10	48 (14.0)	34 (9.8)
11	66 (19.2)	45 (12.9)
12 (high school diploma or GED)	134 (39.0)	165 (47.4)
≥ college	68 (19.8)	79 (22.7)
Living with a parent/guardian	214 (60.6)	211 (60.5)	.96
Qualifies for free lunch	211 (61.7)	228 (66.5)	.19
Ever incarcerated	121 (42.5)	119 (42.7)	.96
Sexual identification			.61
Heterosexual	298 (84.4)	291 (83.4)
Bisexual	17 (4.8)	22 (6.3)
Homosexual or gay	17 (4.8)	26 (7.4)
Sexual behaviors
Correct and consistent condom use	95 (26.9)	59 (16.9)	.04
Sexual partners, lifetime	17.0 (17.9)	19.8 (20.2)	.07
Sexual partners, past 2 mo	2.8 (5.1)	3.3 (7.2)	.34
Sexual intercourse while drunk, past 2 mo	2.8 (1.7)	2.8 (1.7)	.94
Sexual intercourse while high, past 2 mo	2.6 (2.1)	2.6 (2.1)	.92
Sexually transmitted infections
History of chlamydia	83 (25.3)	109 (32.4)	.04
History of gonorrhea	53 (16.2)	92 (27.4)	< .001
History of chlamydia or gonorrhea	121 (34.3)	157 (45.0)	.01
Positive for chlamydia	47 (13.7)	53 (15.5)	.43
Positive for gonorrhea	17 (5.0)	20 (5.8)	.15
Positive for chlamydia or gonorrhea	53 (16.3)	65 (19.1)	.37

Note. GED = general equivalency diploma.

We also analyzed differences between participants who returned for at least 1 follow-up assessment (n = 617) and those who completed only the baseline survey (n = 85). We found no significant baseline differences between those who did and did not return by history of STIs, baseline diagnosis of gonorrhea or chlamydia, level of education, correct and consistent use, lifetime number of sexual partners, or age.

Logistic Regression Findings

Table 3 displays the estimated odds ratios (EORs), their 95% confidence intervals (CIs), and P values for 3 models: 2-month correct and consistent use, 6-month correct and consistent use, and the generalized estimating equations results for the combined 2- and 6-month correct and consistent use assessment. Table 3 provides findings on group assignment, baseline classification of correct and consistent use, age, and history of chlamydia–gonorrhea. At the 2-month follow-up, 53.5% of participants randomized to the intervention condition and 49.6% of control participants reported correct and consistent use (P = .41). The 2-month multivariable model for the dependent behavioral variable analyzed data from 380 participants. After adjustment for age, history of chlamydia or gonorrhea, and baseline levels of the outcome variable, the model yielded an EOR of 1.39 for intervention versus control condition (95% CI = 0.90, 2.13; P = .14).

TABLE 3—

Logistic Regression Results for Condom Use Among Black Male Youths: Focus on the Future Intervention, Louisiana and North Carolina, 2010–2012

CCU Predictor Variable	EOR (95% CI)	P
Model 1, 2-mo follow-up
Assigned to intervention group	1.39 (0.90, 2.13)	.14
Baseline CCU	4.15 (2.36, 7.31)	< .001
Age	0.84 (0.75, 0.95)	.006
Ever diagnosed with chlamydia/gonorrhea	1.14 (0.73, 1.79)	.56
Model 2, 6-mo follow-up
Assigned to intervention group	1.63 (1.07, 2.49)	.02
Baseline CCU	5.07 (2.96, 8.71)	< .001
Age	0.87 (0.78, 0.98)	.017
Ever diagnosed with chlamydia/gonorrhea	0.86 (0.56, 1.32)	.48
Model 3, combined 2- and 6-mo follow-ups
Assigned to intervention group	1.49 (1.06, 2.08)	.02
Baseline CCU	4.55 (2.91, 7.09)	< .001
Age	0.86 (0.79, 0.94)	.001
Ever diagnosed with chlamydia/gonorrhea	1.00 (0.71, 1.41)	.99

Note. CCU = correct and consistent use; CI = confidence interval; EOR = estimated odds ratio.

At the 6-month follow-up, 51.5% of intervention and 46.9% of control participants reported correct and consistent use (P = .29). The 6-month multivariable model for the dependent behavioral variable had data from 420 participants. After adjustment for age, history of chlamydia or gonorrhea, and baseline levels of the outcome variable, the model yielded an EOR of 1.63 (95% CI = 1.07, 2.49; P = .02), which supported intervention efficacy.

The generalized estimating equations model, with the 2-month and the 6-month dependent behavioral variables and adjustment for age, history of chlamydia or gonorrhea, and baseline measures of the outcome variable, produced a significant EOR of 1.49 (95% CI = 1.06, 2.08; P = .02), which supported intervention efficacy.

At the 6-month follow-up, 9.4% of intervention participants and 7.4% of control participants tested positive for either CT or GC (P = .51). The 6-month multivariable model for incident infections with CT or GC analyzed data from 384 participants. After adjustment for age, self-reported history of CT–GC, and the baseline assessment of the primary dependent behavioral variable, the model yielded an EOR of 1.04 for intervention versus control group (95% CI = 0.49, 2.23; P = .91; data not shown).

Figure 1 displays findings from 2 comparisons. First, the figure displays correct and consistent use rates for both groups at baseline, 2 months, and 6 months. Correct and consistent use rates changed significantly more for the intervention than the control group, both from baseline to 2 months (P = .01) and from baseline to 6 months (P = .001). The intervention group began the study with far lower rates of correct and consistent use (16.9%) than the control group (26.9%); however, correct and consistent use in the intervention group exceeded that in the control group by the end of 2 months, and the effect remained at 6 months.

FIGURE 1—

Changes in consistent and correct condom use and chlamydia–gonorrhea incidence among Black male youths: Focus on the Future Intervention, Louisiana and North Carolina, 2010–2012.

Note. Asterisks indicate that the change from initial status differed significantly by group (P = .05). Numbers shown at bottom are effective sample sizes.

Figure 1 also displays rates of self-reported history of CT–GC acquisition and the observed incidence of CT–GC at the 6-month follow-up assessment. The intervention group began at a disadvantage (45.0%) relative to the control group (34.3%). Six months after baseline, the 2 groups had similar incidences; however, the changes from self-reported history to 6 months were not significantly different by group (P = .06).

DISCUSSION

FoF, a brief, single-session, clinic-based intervention, initially targeted Black men and became classified as a DEBI (Diffusion of Effective Behavioral Interventions) by the CDC.²⁴ In accordance with the US National HIV/AIDS Strategy,²⁵ adapting this program for Black male youths became a priority. We found that the adapted program reduced sexual risk behaviors for an adolescent population over a 6-month period of observation. Odds of reporting correct and consistent use were about 60% greater among those receiving the intervention than among those receiving an information-based, attention-equivalent control condition.

The significant intervention effects we observed are compelling because the control group had multiple advantages: they had access to high-quality condoms, and they attended an educational presentation about STIs. Control participants selected from the same broad array of condoms and pocket-size lubricant vials as did the treatment group. Issues of condom affordability and access were therefore evenly matched between groups and negated in our trial. The adapted version of FoF showed that a tailored and brief, interactive, sex-positive, skill-building counseling session increased condom use above and beyond the effect achieved by merely providing a variety of high-end condoms and lubricants. Distributing high-quality condoms, as expected, produced a small effect on control participants; however, the effect for the treatment condition was substantially greater. Thus, clinic distribution of condoms alone fails to take full advantage of an opportunity to enhance condom use through a sex-positive counseling session delivered by health educators.

The priority in the CDC’s strategy for high-impact HIV prevention is interventions that optimally combine evidenced-based cost effectiveness and scalability.³² The brief program we tested satisfies these CDC criteria. As an efficacious single-session intervention, FoF could be efficiently disseminated and implemented with relatively minimal resources. Further, the US Prevention Services Task Force has recommended development of single-session interventions for implementation in clinics.³³ Thus, another advantage of the tested intervention is its potential to be universally applicable and desirable for US clinics serving populations of Black male youths.

Although we observed significant effects at 6 months, the lack of significant differences at 2 months suggests 1 or both of 2 possibilities. It may be that the intervention effect required more than 2 months to fully manifest, because of changes in relationships among participants. Previous research has found that, among young Black men, condom use is more prevalent among casual partners and less so with main partners.^34,35 Thus, it may be that for some intervention participants, correct and consistent use was not readily achieved with a resistant main partner, whereas at 6 months, participants may have experienced turnover in partners, thereby creating new opportunities to negotiate correct and consistent use from the onset of a relationship.

It may also be that the sample size for the 2-month model was inadequate to provide sufficient statistical power for a fair test of a modest effect size at that time point. Fewer participants than expected reported that they had engaged in penile–vaginal intercourse at the 2-month assessment, which compromised statistical power. With the sample size at 2 months (n = 380), a 5% significance level, and approximately 80% power, an underlying odds ratio of 1.85 would have been needed. Thus, the observed adjusted EOR of 1.39 should not be taken to mean that the intervention had no effect at 2 months, but rather that the study was underpowered.

We found no significant effect for the primary outcome of incident CT–GC. However, the analysis of change scores yielded a P value of .06, suggesting that this brief, clinic-based intervention may indeed be efficacious in stemming the incidence of chlamydia and gonorrhea among Black male youths in the southern United States. Nonetheless, the P value of .06 did not warrant rejection of the null hypothesis. Previous research has shown that among populations experiencing health disparities in HIV and STIs, social determinants and social–sexual networks are significant antecedents to disease outcomes, perhaps more so than condom use.^36,37 Although the intervention demonstrated significant behavioral change in the hypothesized direction, it was not ample enough to decrease incidence in this high-seroprevalence population. The mean percentage of intervention participants who used condoms consistently and correctly never exceeded 54%. Although that was a significant improvement from baseline, it was apparently too low to confer added protection against STI acquisition. This is consistent with evidence suggesting that falling short of 100% consistent and correct use will allow for ample STI acquisitions in some populations.^38,39

Limitations

As is true for all sexuality research, our findings were limited by the validity of self-report. The use of a convenience sample limited the generalizability of the findings to other populations of Black male youths. Also, randomization did not produce pretest equivalence between study arms. Although we have no reason to believe that these differences were not attributable to chance, we controlled for baseline differences in the analyses and still found significant effects.

The lack of record availability in one of the secondary recruitment sites (Charlotte, NC) precluded the calculation of an overall study participation rate. The estimated participation rate of 60% was also less than desirable; however, this probably resulted from the extensive time commitment involved rather than disinterest. Working with this hard-to-reach population necessitates a willingness to be flexible in expectations relative to the interest of Black male youths in study participation.⁴⁰

Conclusions

Ours is the first published study reporting significant findings from an efficacy trial designed to promote the correct and consistent use of condoms specifically among Black male youths attending clinics that diagnose and treat STIs. Our findings suggest that a brief, clinic-based intervention delivered face to face with Black youths is efficacious in increasing the use of male condoms.

The CDC has termed AIDS a health crisis for Blacks,² so the availability of this low-intensity and low-cost approach to prevention is an important part of a larger public health response to the crisis. By hiring a qualified health educator and offering highly desirable condoms and lubricants to Black male youths, clinics may achieve an appreciable increase in safer sex practices. Along with supporting forms of intervention (e.g., prompt treatment of Black youths and their partners), this program could be readily implemented in clinics serving this vulnerable US population. Adaptation of this prevention approach for vulnerable populations of young men in other nations may also be feasible.