Culturally Informed and Flexible Family-Based Treatment for Adolescents (CIFFTA)

Study 1

The initial version of the program relied on office visits for the treatment sessions.

Summary

Santisteban et al. (2011) conducted a randomized controlled trial that examined 28 Hispanic youths with a diagnosed substance use disorder. The participants were randomized to the intervention group or a control group that received traditional family therapy. Assessments at baseline and four months after the program end measured youth substance use and behavior problems.

Santisteban et al. (2011) found that, relative to the control group youths, the intervention group youths reported significantly

• Lower marijuana use and marijuana plus cocaine use
• Improved parenting (a risk and protective factor).

Evaluation Methodology

Design:

Recruitment: The study recruited 28 Hispanic adolescents (ages 14-17) living with a parent or guardian who came to the United States from a Spanish-speaking country and meeting DSM-IV criteria for a substance abuse disorder. At least one parent joined each adolescent in the study. Participants were referred by a local Juvenile Addictions Receiving Facility and by the Department of Juvenile Justice Diversion programs.

Assignment: In this randomized controlled trial, the 28 participants completed the baseline assessment and were randomly assigned within strata defined by gender and severity of drug use to an intervention group (n = 14) or a control group (n = 14). The control group met once per week for traditional family treatment, while the intervention group met for treatment twice per week.

Assessments/Attrition: The assessments occurred at baseline, four months (posttest), and eight months (four-month follow-up), but only the baseline and the eight-month follow-up assessments were examined. The completion rate for the follow-up was 89%.

Sample:

The study did not report demographic characteristics other than the sample was 100% Hispanic.

Measures:

The study examined five behavioral outcomes relating to substance use and behavior problems. Four of the five were reported by the adolescent, and one was reported by the parent (the parent report may not be independent given the parenting component of the program). The two risk and protective factors measured the same parenting practices as reported by both adolescents and parents. Another measure was based on all parent-reported parenting subscales. The reported alpha values for scales were high.

Analysis:

The analysis used repeated measures ANOVA with baseline externalizing behaviors as a covariate. Time-by-treatment effects in the repeated measures models tested for the intervention effects.

Missing Data Method: The complete-case analysis did not employ multiple imputation or FIML.

Intent-to-Treat: All participants with complete data were included in the analysis following their group assignment.

Outcomes

Implementation Fidelity:

All but one of the 14 intervention participants received the intervention, and the mean number of sessions was 16.25 of 32. The authors reported that ratings of a subset of the intervention sessions "showed that therapists delivered the intended individual therapy and didactic interventions."

Baseline Equivalence:

The authors reported that "There were no significant differences between conditions on adolescent gender and age, annual family income, and adolescent and parent Spanish-language preference." For the outcomes, one showed a significant difference, with control participants reporting significantly greater externalizing problems than the intervention participants. The analyses included baseline externalizing problems as a covariate for all models except for youth self-reports of externalizing behavior.

Differential Attrition:

The study did not test for differences between dropouts and completers. The rates of attrition at posttest were 14% for the intervention group and 7% for the control group. The difference does not meet the WWC cautious standard but meets the WWC optimistic standard.

Posttest:

Two of the five behavioral outcomes showed significant effects. The intervention group reported significantly lower marijuana use (eta squared = .33 ) and lower marijuana plus cocaine use (eta-squared = .20) than the control group. The two risk and protective factors, adolescent-reported parenting practices and parent-reported parenting practices, were also significant.

Along with three non-significant effects on cocaine use, parent-reported behavior problems, and adolescent-reported externalizing, there were several non-significant effects mentioned in the text. Neither urine tests for marijuana use nor a composite measure of all parent-reported parenting subscales reached statistical significance.

Long-Term:

Not examined.

Study 2

This study evaluated a hybrid computer-based and face-to-face version of the program. The face-to-face psychoeducational sessions were replaced by recorded modules delivered online.

Summary

Santisteban et al. (2017) conducted a randomized controlled trial that examined 80 Hispanic and Black youths with behavioral or mental health problems. After assigning the youths and their families to an intervention group or a waitlist control group, the study measured conduct and mental health problems at posttest.

Santisteban et al. (2017) found that the intervention group relative to the control group had significantly

• lower parent-reported child conduct disorder,
• lower parent-reported child aggression,
• lower youth-reported externalizing,
• higher parent-reported family cohesion (protective factor),
• higher youth-reported family cohesion (protective factor).

Evaluation Methodology

Design:

Recruitment: The study recruited 80 Hispanic and Black non-Hispanic adolescents and their families who were referred by school counselors and community treatment agencies. The adolescents were 12 to 15 years old, self-identified as Black, non-Hispanic, or Hispanic, and had at least two areas of behavioral or mental health problems (e.g., depression, conduct problems,  family conflict). At baseline, 77% of adolescents received a T-score of 60 or higher or met the clinical cutoff on at least one outcome measure.

Assignment: After completing the baseline assessment, participants were randomly assigned within strata defined by race and ethnicity, gender, and presence of a diagnostic condition to the intervention group (n = 40) or a delayed or waitlist control group (n = 40). The control group received only information on crisis interventions available through a local community agency. After 12 weeks, the control group was offered the intervention.

Assessments/Attrition: All participants completed assessments at baseline and posttest (12 weeks post baseline). Additional assessments came after the control group had joined the intervention. The intervention group was assessed six weeks after the posttest, and the control was assessed 12 weeks after posttest and starting the intervention. A total of 85% of participants completed the posttest, and 84% completed the follow-up.

Sample:

The sample was 56% male, 80% Hispanic, and 20% Black non-Hispanic, and the mean age was 13.6 years. Fully 72% of the adolescents reported that their preferred language was English, 4% preferred Spanish, and 24% were comfortable with both languages.

Measures:

The 11 outcome measures included five behavioral outcomes related to externalizing and internalizing youth behavior. Three of the measures came from parents and two came from the youths. The six risk and protective factors measured the family environment and parenting. Most reliabilities were high, but a few fell below .70.

Analysis:

The analyses used Generalized Estimating Equations with Full Information Maximum Likelihood estimation. The models included all time points, including those coming after the control group received the intervention. However, the models allowed for estimation of between-condition effects at posttest separately from estimation of longer-term effects without condition comparisons.

Missing Data Method: The FIML estimation adjusted for missing data.

Intent-to-Treat: The analyses included all participants according to their assigned condition.

Outcomes

Implementation Fidelity:

The intervention group received an average of 6.8 face-to-face sessions, 6.05 family sessions, .75 individual sessions, and 5.0 online psychoeducational modules. Using a scale ranging from 0 (never) to 7 (extensively), mean adherence ratings for a subsample of sessions were 1.33 for the family items, .16 for the individual items, and .56 for the psychoeducational items.

Baseline Equivalence:

Table 1 shows no significant differences for eight sociodemographic measures. For the outcomes, the authors stated, "There were no significant differences in the pretreatment assessment (i.e., T1 for the Immediate CA CIFFTA and T2 for the Delay CIFFTA) for the two conditions." That comparison does not compare the conditions at the same T1 time point but is suggestive of baseline equivalence.

Differential Attrition:

The study did not test for differences between dropouts and completers but used FIML to adjust for potential bias from attrition. The rates of attrition at posttest were 20% for the intervention group and 10% for the control group. The difference does not meet the WWC cautious boundary and barely meets the WWC optimistic boundary of 10.7%.

Posttest:

Table 2 presents results for the 11 outcomes. Three of five behavioral outcomes were significant: The intervention group relative to the control group had significantly lower parent-reported conduct disorder (d = .54), parent-reported aggression (d = .45), and youth-reported externalizing (d = .59). Of six risk and protective factors, parent-reported family cohesion (d = .54) and adolescent-reported family cohesion (d = .58) were statistically significant. Additional tests for longer-term effects using only the intervention group found the initial gains were sustained. Tests for moderation found no significant differences in program effects by race/ethnicity.

Long-Term:

Not examined.

Study 3

The study evaluated a face-to-face version of the program that included 24 sessions.

Summary

Santisteban et al. (2022) conducted a randomized controlled trial that examined 200 Latino youths with symptoms of a diagnosed disorder or high family conflict. After randomly assigning the youth and their parents, the study measured youth externalizing and internalizing and family conflict at posttest.

Santisteban et al. (2022) found that, relative to control group parents, intervention group parents reported significantly

• lower family conflict (a risk & protective factor).

Evaluation Methodology

Design:

Recruitment: The study recruited 200 Latino adolescents (11-14 years of age) with symptoms of two or more DSM-IV disorders (e.g., conduct disorder, ADHD, and major depressive disorder) and/or high family conflict. Participants were referred by school counselors, community treatment agencies, or civil citation programs.

Assignment: Adolescents and their families were randomly assigned after baseline to the intervention group (n = 101) or the control group (n = 99). Randomization was stratified by number of disorders reported and the level of family conflict. The control group received typical individually oriented services offered in the community that, like the intervention group, offered  24 sessions over the 16-week period.

Assessments/Attrition: Assessments were administered at baseline and posttest, four months after baseline. About 82% of participants completed the posttest.

Sample:

For primary caregivers, 91% were mothers or mother figures, 81% preferred Spanish or Spanish and English equally, and 14% were born in the United States. For the target youths, 61% were male, 31% preferred Spanish or Spanish and English equally, and 65% were born in the United States. The average age was 12.5 years.

Measures:

The outcomes included seven behavioral measures: two from youths, two from parents, and two from both youths and parents. The outcomes measured externalizing, internalizing, depression, ADHD, and conduct disorder. Four risk and protective factors obtained from both youths and parents measured family cohesion and conflict. The scales had good alpha reliability values, though one fell below .70.

Analysis:

The analysis used generalized estimating equations with full-information maximum likelihood estimation. The repeated measures model assumed an AR-1 covariance structure and tested for intervention effects with time-by-condition terms.

Missing Data Method: The analysis used FIML to adjust for missing data.

Intent-to-Treat: The analysis included all participants.

Outcomes

Implementation Fidelity:

The intervention group had significantly higher retention (83%) than the control group (71%), with retention defined as attending more than seven sessions. Ratings of adherence reported means of 3.08 for highlighting family interactions and communication patterns in family therapy sessions, 1.47 for exploring risky and problem behaviors in individual therapy sessions, and 3.27 for providing substantial information in a didactic manner in psycho-educational sessions.

Baseline Equivalence:

Table 1 shows no significant differences between conditions on the seven sociodemographic measures and four outcome measures.

Differential Attrition:

The study did not test for baseline differences between completers and dropouts but used FIML to adjust for potential bias from attrition. The combination of the overall attrition rate (18%) with the difference in attrition rates between conditions (4%) meets both the WWC's cautious and optimistic standards for minimal attrition bias.

Posttest:

Table 2 shows no significant time-by-condition effects for the seven behavioral outcomes. One of the risk and protective factors reached statistical significance, with the intervention parents reporting lower family conflict than the control parents. Tests for moderation found stronger intervention effects for parents with low acculturation than high acculturation.

Long-Term:

Not examined.