Parent-Centered Prevention of Risky Behaviors Among Hispanic Youths in Florida

Abstract

Objectives. To evaluate the effectiveness of an evidence-based, parent-centered intervention, Familias Unidas, delivered by nonresearch personnel, in preventing substance use (alcohol, illicit drugs) and sex without a condom among Hispanic adolescents.

Methods. A randomized controlled trial (n = 746) evaluated the effectiveness of Familias Unidas among Hispanic eighth graders (age range = 12–16 years), relative to prevention as usual, within a public school system. School personnel, including social workers and mental health counselors, were trained to deliver the evidence-based intervention. Participant recruitment, intervention delivery, and follow-up ran from September 2010 through June 2014 in Miami–Dade County, Florida.

Results. Familias Unidas was effective in preventing drug use from increasing and prevented greater increases in sex without a condom 30 months after baseline, relative to prevention as usual. Familias Unidas also had a positive impact on family functioning and parental monitoring of peers at 6 months after baseline.

Conclusions. This study demonstrated the effectiveness of a parent-centered preventive intervention program in preventing risky behaviors among Hispanic youths. Findings highlight the feasibility of training nonresearch personnel on effectively delivering a manualized intervention in a real-world setting.

Interventions that target family processes, regardless of racial or ethnic group, are often efficacious in preventing and reducing a broad range of risk behaviors, including drug use,¹ yet are rarely evaluated in effectiveness trials. Even with the recent emphasis on implementation science in prevention, and more generally in public health, interventions are still not being implemented broadly. This translation gap is often attributable to the fact that preventive behavioral interventions found to be efficacious in highly controlled research laboratories are often not evaluated for effectiveness in real-world settings.² This shortage of effectiveness research is a barrier to wide-scale dissemination³ to populations in most need of these interventions. This is especially concerning among groups that experience health disparities, such as Hispanic adolescents.⁴ Moving efficacious interventions beyond highly controlled laboratory settings to testing in real-world milieus is important for the prevention of behavioral health disorders and reduction of health disparities.

Hispanic adolescents disproportionately engage in risky behaviors such as alcohol use, drug use, and unprotected sexual behavior. Among sexually active youths, national epidemiological surveys indicate that Hispanics were the least likely to report using a condom at their last sexual encounter.⁵ Furthermore, in school-based surveys, Hispanic eighth graders reported the highest lifetime and current prevalence of alcohol and illicit drug use, compared with their African American and non-Hispanic White counterparts.⁶ Given these statistics, eighth grade is a good intervention point to target prevention efforts among Hispanic youths. Furthermore, from a developmental perspective, eighth grade is an opportune timeframe for preventive intervention (i.e., before the transition from middle to high school, where adolescents will be exposed to a wider panorama of risk).

Familias Unidas is a parent-centered, family-based preventive intervention found to be efficacious in reducing alcohol use, drug use, and unsafe sexual behavior in 3 completed and published randomized clinical trials, all conducted in highly controlled research environments.^7–9 Improvements in family functioning (e.g., parental monitoring of peers, positive parenting, and family communication) partially mediated the relationship between intervention effects and the outcomes. Familias Unidas was specifically designed for Hispanic families and consists of 8 parent-only group sessions and 4 family sessions with the adolescent. Through a participatory learning approach, parenting skills that are discussed and role-played in the parent group sessions are later enacted with the adolescent, with facilitator guidance, during the family sessions. The Institute of Medicine cited Familias Unidas as an efficacious preventive intervention ready for evaluation in real-world settings and subsequently ready for wide-scale dissemination.¹⁰ However, like most parent-centered preventive interventions, Familias Unidas has not previously been evaluated for effectiveness. Therefore, we evaluated the relative effectiveness of Familias Unidas, compared with prevention as usual, in a community sample of Hispanic adolescents. We hypothesized that Familias Unidas would decrease alcohol use, drug use, and sex without a condom, relative to prevention as usual, over time, and that intervention effects would be partially mediated by family functioning.

METHODS

Participants were recruited from 18 middle schools with letters that were sent home with students. Participant recruitment, intervention delivery, and follow-up ran from September 2010 through June 2014 and all took place in Miami–Dade County, Florida. Middle schools had to have a majority of Hispanics in their student body and be within the Miami–Dade County Public School (MDCPS) district to meet the inclusion criteria at the school level. To be eligible for the study, adolescents were required to

1. be of Hispanic origin,

2. attend eighth grade at the time of the baseline assessment,

3. live with an adult primary caregiver who was willing to participate,

4. live within the catchment areas of the participating middle schools, and

5. plan to live in South Florida for the duration of the study.

Table 1 provides demographic information about the sample. Parents were, on average, aged 41 years (SD = 6.3). Most parents (88%) were born in a Spanish-speaking country from the Americas. Fifty-two percent of youths were male, and 83% of parents were female. Adolescent age ranged from 12 to 16 years.

TABLE 1—

Baseline Comparisons by Condition on Demographic Characteristics and Outcome Variables Among Hispanic Eighth Graders: Miami–Dade County, FL, September 2010–June 2014

Variables	Familias Unidas (n = 376), No. (%) or Mean ±SD	Prevention as Usual (n = 370), No. (%) or Mean ±SD	P
Gender			.88
Male	195 (51.9)	194 (52.4)
Female	181 (48.1)	176 (47.6)
Youth’s age	13.88 ±0.66	13.83 ±0.68	.32
US-born (youths)	199 (53.2)	210 (56.8)	.47
Family income, $			.26
0–9999	100 (26.6)	82 (22.2)
10 000–19 999	94 (25.0)	108 (29.2)
20 000–29 999	54 (14.4)	65 (17.6)
≥ 30 000	127 (35.8)	115 (31.1)
Adolescent-specific factors
Alcohol use in lifetime	57 (15.2)	58 (15.7)	.85
Alcohol use in the past 90 d	26 (6.9)	22 (5.9)	.59
Illicit drug use in lifetime	2.01 ±21.60	2.11 ±21.81	.48
Illicit drug use in the past 90 d	12 (3.2)	18 (4.9)	.25
Illicit drug use in the past 90 d	1.26 ±20.48	0.86 ±10.51	.29
Sex in lifetime	26 (6.9)	31 (8.4)	.45
Sex without a condom in the past 90 d (sexually active cases)	0.89 ±1.16	1.59 ±1.32	< .001
Parent-specific factors
Family communication	7.14 ±1.79	7.41 ±1.64	.029
Parental monitoring of peers	11.75 ±4.76	11.66 ±4.67	.79
Positive parenting	24.10 ±4.54	23.56 ±4.55	.10

Note. One participant in the Familias Unidas group did not report income.

Study Design

This study consisted of a randomized controlled trial with 2 intervention arms (Familias Unidas and prevention as usual) and 4 time points (baseline and 6, 18, and 30 months after baseline). Research study staff completed consent and enrollment. Participants were randomized by using stratified randomization within schools to ensure balance of the treatment groups with respect to gender and risky behaviors (i.e., lifetime illicit drug use and lifetime sexual activity). The study’s biostatistician conducted all randomization procedures, including intervention assignment. The allocation ratio was 1 to 1. Using Mplus version 7.2 (Muthén & Muthén, Los Angeles, CA), we conducted Monte Carlo simulation to calculate power with a latent growth curve model framework.¹¹ With 4 time points (baseline and 6, 18, and 30 months after baseline), and assuming that the expected mean trajectories were 0.2 for all targeted outcomes with a 10% attrition rate at each assessment, and a sample size of 746 at baseline, there is 96% power to detect a regression coefficient equal to 0.11 in the regression of the slope growth factor on intervention condition, therefore producing a moderate effect size (d = 0.34).¹²

Neither participants nor research personnel were blinded to the condition assignment. Parent participants were given an incentive of $40, $45, $50, and $55 for completing assessments at each time point, respectively. Adolescent incentives included a movie ticket at the baseline assessment and 2, 3, and 4 movie tickets, respectively, at each subsequent time point. Assessments were completed at the adolescents’ school and were administered by research staff via audio computer-assisted self-interview software, an automated system that improved the rate of honest responses to sensitive questions in previous studies involving youths.¹³

Study Conditions

Familias Unidas.

Familias Unidas was delivered through 8 multiparent group sessions (10–15 parents) and 4 family sessions that took place in the evening at the school from which the adolescent was recruited. The intervention was delivered on a weekly basis, therefore, running 3 months in duration (a session outline can be found in Table A, available as a supplement to the online version of this article at http://www.ajph.org).

Facilitator training.

Facilitators were master’s level social workers and mental health counselors employed by MDCPS. Spanish fluency was required for facilitators to deliver the intervention. Facilitators were recruited through a series of presentations given by research staff in MDCPS mental health district meetings. Of 58 facilitators who expressed interest in delivering the intervention, 31 were trained and 27 delivered the intervention. Twenty-seven facilitators were not trained because of competing demands (n = 10), lack of Spanish fluency (n = 3), distance from school (n = 7), or for an unspecified reason (n = 7; M. Tapia, written communication, February 2, 2016). The research team’s clinical supervisor, a master’s-level experienced Familias Unidas facilitator, delivered training across 4 days that included didactic instruction, role plays, and group discussion of recorded sessions. Facilitators received 2-hour, weekly face-to-face group supervision for 12 weeks. Facilitators were compensated for their time in delivering the intervention, training, and supervision.

Adherence.

To assess adherence to the Familias Unidas intervention, raters evaluated all videotaped group sessions and 25% of the family sessions on a 7-point scale. In addition, a second independent rater rated 20% of all rated sessions to assess interrater reliability. The raters were master’s-level graduate students who were trained by the clinical supervisor. The overall average adherence rating was 3.61 (SD = 0.56); parent groups ranged from 2.04 to 4.30¹⁴ and family visits ranged from 2.60 to 5.20. Interrater reliability was good across all domains (κ > 0.80). Ratings were discussed with facilitators during weekly supervision to improve intervention delivery.

Prevention as usual.

The prevention-as-usual group consisted of the HIV risk reduction intervention provided by the MDCPS system to students. Science teachers delivered the MDCPS intervention in a classroom setting and it consisted of 6 lessons designed to decrease HIV/AIDS and other sexually transmittable diseases via a science-based education. Parental permission was required.

Measures

Demographics.

All participants completed a demographics questionnaire (Table 1). Parents also responded to items concerning household income.

Main outcomes.

We assessed adolescent alcohol and drug use by using items from a population-based epidemiological study.⁶ Adolescent participants were asked whether and how many times they had used alcohol and illicit drugs (e.g., marijuana, LSD, cocaine) during the previous 90 days.

For sex without a condom, adolescents were asked whether they had engaged in sex (vaginal, anal, or oral) in the previous 90 days and how often they had sex without a condom (vaginal or anal), rated on a 5-point scale: 0 = never; 1 = less than half of the time; 2 = about half of the time; 3 = not always, but more than half the time; and 4 = always.

Secondary outcomes.

We assessed family functioning by using parent reports of 3 indicators: (1) parental monitoring of peers (Parent Relationship With Peer Group Scale,¹⁵ 5 items; α = 0.86), (2) positive parenting (Parenting Practices Scale,¹⁶ 9 items; α = 0.71), and (3) family communication (Family Relations Scale,¹⁷ 3 items; α = 0.68). These measures assessed parental attempts to actively monitor adolescents and know their adolescents’ friends; positive parenting, rewards, and acknowledgments given in response to positive behaviors; and communication, respectively. We used a single family-functioning variable by creating a latent construct consisting of parental monitoring of peers, positive parenting, and family communication.

Data Analytic Strategy

The analytic plan consisted of several steps. First, we used a χ² test to determine differences in attrition by condition. Second, we conducted a χ² test (for categorical variables), independent t test (continuous and normally distributed variables), or Mann–Whitney U Test (for count variables) to examine baseline differences on demographic characteristics and outcome variables by condition. Third, we used linear growth curve analyses to examine the study hypotheses. To test binary outcomes for past-90-day alcohol use, we estimated a growth curve model by using the probit-LINK function in Mplus. Also, because of frequency of drug use in the past 90 days being a count variable with a large number of zeros, we estimated a zero-inflated Poisson growth model that examined the effect of the 2 study conditions on frequency of drug use as well as the effect of study condition on the likelihood of using drugs.¹⁸ The proportion with zero inflation was fixed across condition but varied by time. For past-90-day sex without a condom, the analyses only included participants who reported being sexually active. Throughout all study time points, there were 130 adolescents who reported having sex in the 90 days before any of the assessments. We used Cohen's d as a standardized effect by dividing the slope differences with the standard deviation of the residual slope. We based all other analyses on intent-to-treat.

Next, to examine the potential effects of study condition on the family functioning latent variable and its indicators from baseline to 6 months after baseline (i.e., postintervention), we conducted analysis of covariance with family functioning and each of the family-functioning indicators. We used the 6-months-postbaseline data as the dependent variable and study condition as the predictor. We controlled the baseline level of each family-functioning indicator in each respective model. We tested the homogeneity assumption for regression slopes by examining whether there was a significant interaction between the baseline level of the family-functioning latent construct and study condition. Finally, we used the product of coefficients method¹⁹ to test whether the family functioning indicators mediated the effects of study condition on the outcomes.

We addressed missing data for the repeated measures by using full information maximum likelihood.²⁰ To obtain accurate standard error (SE) estimates, we used a complex survey design to adjust for the clustering effect of students within schools.

RESULTS

Figure 1 shows the CONSORT flow diagram for this trial. Analyses indicated no significant differences at baseline by condition on demographic characteristics nor on some risky behaviors, such as percentage of past-90-day alcohol or illicit drug use and prevalence of lifetime use. However, youths randomized to prevention as usual had significantly higher sex without a condom in the past 90 days (U = 3.197; P < .001; d = 0.280) and higher parent-reported family communication (t = −2.19; P < .05; d = 0.163). There were no differences in attrition rates across the 2 study conditions (13.3% and 12.2% for Familias Unidas and prevention as usual, respectively; χ² = 0.216; P = .642).

FIGURE 1—

CONSORT Flow Diagram of Trial Comparing Evidence-Based, Parent-Centered Intervention, Familias Unidas, Relative to Prevention as Usual in Preventing Substance Use and Sex Without a Condom Among Hispanic Eighth Graders, Miami–Dade County, Florida, September 2010 to June 2014

Note. The reasons for lost to follow-up or discontinued study were (1) work issues, (2) time issues, (3) moved out of Miami, (4) declined to continue in the study, (5) health issues, (6) adolescent dropped out of school, (7) did not want to do assessment, (8) unable to reach the family, (9) out of country, and (10) did not like the study.

The average overall attendance, including parent groups and family visits, was 6.4 sessions (SD = 4.2) out of a possible 12 sessions; 12.9% of participants did not attend any sessions. We did not collect attendance information for prevention as usual.

Main Outcomes

Intervention effects on alcohol and drug use.

We tested intervention effects on alcohol and drug use separately. The trajectories for past-90-day alcohol use between Familias Unidas and prevention as usual were not statistically different (b = 0.075; 95% confidence interval [CI] = −0.142, 0.291; P = .499; d = 0.24). However, the trajectories for past-90-day drug-use frequency between Familias Unidas and prevention as usual were statistically different (b = −0.20; 95% CI = −0.298, −0.105; P < .001; d = 0.27). Table 2 details Familias Unidas intervention effects on drug-use trajectories. Specifically, multigroup analysis (i.e., Familias Unidas and prevention as usual) showed that the count trajectory of drug use in the past 90 days among Familias Unidas youths was stable across time points (mean trajectory = 0.176; 95% CI = −0.001, 0.354; P = . 201617498 051) whereas it increased over time (mean trajectory = 0.184; 95% CI = 0.064, 0.304; P < .01) among prevention-as-usual youths. It should be noted that we dropped 7 participant cases from the past-90-day drug use analyses because they were outliers and demonstrated an invalid response pattern.

TABLE 2—

Condition Effects on Drug Use and Sex Without a Condom Trajectories for Familias Unidas vs Prevention as Usual, Hispanic Eighth Graders: Miami–Dade County, FL, September 2010–June 2014

	Drug Use vs Non–Drug Use, Slope		Drug-Use Frequency, Slope		Model Fit Indices
Variable	Est. (95% CI)	SE	Est. (95% CI)	SE	Log-Likelihood (df)	AIC/BIC
Drug use					–0.836.25 (18)	1708.50/1991.39
Intercept	0.06 (–0.36, 0.49)	0.22	0.31 (0.16, 0.45)	0.13
Condition (FU vs PAU)	–0.12 (–0.28, 0.51)	0.20	–0.20 (–0.29, –0.10)	0.04
Mean trajectory
Sex without a condom					–726.718 (11)	1475.44/1506.97
Intercept			0.05 (–0.27, 0.84)	0.05
Condition (FU vs PAU)			0.09 (0.02, 0.16)	0.03

Note. AIC = Akaike information criterion; BIC = Bayesian information criterion; CI = confidence interval; Est. = estimation; FU = Familias Unidas; PAU = prevention as usual. Unstandardized coefficients are presented in the table. The zero-inflated growth model was used for drug use whereas the conventional growth curve model was used for sex without a condom. Seven cases were not included in the drug-use analyses because of an invalid response pattern.

Intervention effect on sex without a condom.

The trajectories for past-90-day sex without a condom among participants who reported being sexually active (n = 130) between Familias Unidas and prevention as usual were statistically different (b = 0.093; 95% CI = 0.024, 0.162; P < .01; d = 0.98). Specifically, multigroup analysis (i.e., Familias Unidas and prevention as usual) showed that the trajectory of sex without a condom among prevention-as-usual youths (mean trajectory = 0.24; 95% CI = 0.154, 0.281; P < .001) increased more than those of Familias Unidas youths (mean trajectory = 0.14; 95% CI = 0.078, 0.207; P < .001). The trajectories of sex without a condom for the sexually active sample can be seen in Table 3.

TABLE 3—

Sex Without a Condom (Past 90 Days) Among Sexually Active Youths Over Time and by Intervention Arm: Miami–Dade County, FL, September 2010–June 2014

Time and Intervention	Never, No. (%)	Less Than Half of the Time, No. (%)	About Half of the Time, No. (%)	Not Always, but More Than Half of the Time, No. (%)	Always, No. (%)	Total, No. (%)
Total	86 (43.9)	32 (16.3)	34 (17.3)	21 (10.7)	23 (11.7)	196 (100.0)
Baseline
Familias Unidas	4 (44.4)	4 (44.4)	0 (0.0)	1 (11.1)	0 (0.0)	9 (100.0)
Prevention as usual	6 (27.3)	2 (9.1)	5 (22.7)	4 (18.2)	5 (22.7)	22 (100.0)
6 mo after baseline
Familias Unidas	5 (45.5)	0 (0.0)	2 (18.2)	0 (0.0)	4 (36.4)	11 (100.0)
Prevention as usual	7 (35.0)	2 (10.0)	2 (10.0)	6 (30.0)	3 (15.0)	20 (100.0)
18 mo after baseline
Familias Unidas	15 (48.4)	6 (19.4)	7 (22.6)	2 (6.5)	1 (3.2)	31 (100.0)
Prevention as usual	12 (44.4)	7 (25.9)	5 (18.5)	1 (3.7)	2 (7.4)	27 (100.0)
30 mo after baseline
Familias Unidas	23 (65.7)	3 (8.6)	4 (11.4)	3 (8.6)	2 (5.7)	35 (100.0)
Prevention as usual	14 (34.1)	8 (19.5)	9 (22.0)	4 (9.8)	6 (14.6)	41 (100.0)

Secondary Outcomes

Intervention effects on family-functioning indicators.

At 6 months after baseline, parents of youths randomized to Familias Unidas reported significantly higher levels of family functioning relative to parents of youths randomized to prevention as usual (b = 0.148; 95% CI = 0.030, 0.266; P = .014; d = 0.47), after we adjusted for baseline levels of family functioning. We decomposed the family functioning latent construct to examine intervention effects on each family-functioning indicator. After we adjusted for baseline levels of parental monitoring of peers, at 6 months after baseline, parents of youths randomized to Familias Unidas reported a significantly higher level of parental monitoring of peers relative to parents of youths randomized to prevention as usual (b = 0.502; 95% CI = 0.043, 0.961; P = .032; d = 0.14). There were no significant intervention effects for the other 2 family-functioning indicators, positive parenting or family communication.

Mediation effects of family functioning and parental monitoring of peers.

Results showed that family functioning did not significantly mediate intervention effects for frequency of drug use (a × b = 0.018; SE = 0.01; 95% CI = −0.002, 0.038; P = .083; d = 0.05), nor did it significantly mediate intervention effects for sex without a condom (a × b = 0.01; SE = 0.014; 95% CI = −0.017, 0.037; P = .476; d = 0.12). However, parental monitoring of peers significantly mediated intervention effects for frequency of drug use (a × b = 0.006; SE = 0.003; 95% CI = 0.001, 0.012; P = .041; d = 0.03).

DISCUSSION

The Familias Unidas intervention maintained drug-use levels from baseline to 30 months after baseline compared with prevention as usual, which evidenced increases in drug use during this same time frame. Although there were increases in sex without a condom from baseline levels to 30 months after baseline for both groups, these differences were statistically greater for prevention as usual. There were no intervention effects for alcohol use. Results also showed that family functioning and parental monitoring of peers significantly increased for families randomized to the Familias Unidas intervention compared with prevention as usual. In addition, drug-use intervention effects were mediated by parental monitoring of peers. This is in line with a robust literature that shows that parental monitoring has an inverse relationship with substance use.²¹ Lower levels of parental monitoring have been associated with a host of adolescent substance use behaviors across a wide range of ages including alcohol, marijuana, and cocaine use.²² Like in previous Familias Unidas trials,⁸ family-functioning indicators demonstrated changes immediately after baseline, whereas changes in the main outcomes occurred at a later time point. Therefore, there was a delayed effect in the main intervention outcomes.

Delayed effects have been seen in previous family-based preventive interventions (e.g., Sandler et al.²³), and are consistent with the suggestion that family-based preventive intervention effects, at least in trials with a universal population, are evident several years after the intervention.⁸ In this trial, the mechanism of change for Familias Unidas was parental monitoring of peers; therefore, it may be that there has to be sustained changes in this variable to see an impact on the main outcomes, which requires the passage of time.

To our knowledge, this is the first drug-use and sexual-risk-behavior preventive (behavioral) intervention targeting Hispanic adolescents with proven efficacy and effectiveness. In general, preventive interventions have remained at the efficacy level without being tested for effectiveness in real-world settings, particularly among minority populations such as Hispanics.²⁴ Beyond findings that Familias Unidas prevented substance use increases and prevented greater increases in sex without a condom, this study is significant in that the research team successfully trained non–research-oriented community members to deliver a manualized intervention with fidelity—an important factor in the delivery of evidence-based behavioral interventions in community settings.²⁵ Indeed, several behavioral interventions have not been successful in effectiveness trials because of low fidelity, which leads to poor intervention outcomes.^14,26 A highlight of the current study is the success in delivering, with fidelity, an evidence-based intervention within a community system. That being said, facilitators were compensated for their services, which is a factor to consider in the sustainability of Familias Unidas within school settings. Future research should conduct a full cost analysis to determine the costs–benefits of this intervention.

Another factor in translating preventive interventions such as Familias Unidas into community settings is participant session completion. Similar to other parent-centered interventions, participation in Familias Unidas across parent group and family sessions was 50% when we included participants that did not attend any sessions. Previous research points to participation barriers including transportation, time, scheduling limitations, and participant characteristics at baseline.^27,28 eHealth interventions offer a viable solution to circumvent participation barriers. Indeed, research with Familias Unidas has begun to look at this via an Internet adaptation of this intervention. Preliminary findings show higher parent group session completion compared with previous Familias Unidas trials and identify predictors of participation including family stress, acculturation, and effective parenting.²⁹

Participation rates notwithstanding, the impact of preventive interventions should not be underestimated. For example, a systematic review of parent-centered interventions for reducing or preventing substance use found that, across 42 randomized studies, parenting interventions were effective in reducing or preventing substance use and that these effects persisted across multiple years.³⁰

Limitations

The current study has several limitations. First, all measures were self-reported. However, to minimize potential social desirability bias, measures were administered through the Audio Computer-Assisted Self-Interview software. Second, Miami is composed of 70% Hispanics, of which approximately half are Cuban; therefore, our sample is not representative of the entire United States nor of all Hispanics.³¹ Study participants, however, did represent a diverse range of Latin American and Caribbean countries.

In addition, we did not collect information for the prevention-as-usual group such as the length of the control condition or whether youths participated. These are important data to collect in future trials. Finally, it is possible that there were differences between families who participated in the study and those who did not enroll. Unfortunately, because families who were ineligible or refused to participate were not consented or assented, we did not collect data that would allow for comparisons with participants enrolled in the study. As in all prevention trials, there exists the possibility of a self-selection bias from participants who decide to engage in a research study.

Public Health Implications

Our findings add to the limited body of research on the effectiveness of family-based interventions for Hispanic adolescents and extend the Familias Unidas program of research 1 more step toward dissemination and implementation. Findings in this study highlight the feasibility of training nonresearch personnel to effectively deliver a manualized intervention in a real-world setting and the opportunity offered by parent-based interventions to target multiple risk behaviors among Hispanic adolescents (i.e., drug use, sex without a condom, low family functioning, and low parental monitoring of peers). Given that substance use and unsafe sexual behavior have similar underlying determinants of risk,¹ interventions that have an impact on multiple risk behaviors demonstrate great potential for translation into community settings.