Health Education Research Advance Access originally published online on June 28, 2006
Health Education Research 2007 22(1):95-107; doi:10.1093/her/cyl052
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Fidelity of implementation in a treatment effectiveness trial of Reconnecting Youth
1 Pacific Institute for Research and Evaluation, Chapel Hill Center
2 Department of Health Behavior and Health Education, School of Public Health, The University of North Carolina at Chapel Hill, Chapel Hill, NC 27599, USA
3 Pacific Institute for Research and Evaluation, Chapel Hill Center, Chapel Hill, NC 27514, USA
* Correspondence to: V. Sánchez. E-mail: vsanchez46{at}aol.com
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionConclusionConflict of interest statementAcknowledgementsReferences |
|---|
In ‘a treatment effectiveness trial’, a program is evaluated in a real-world setting, with an emphasis on achieving high implementation fidelity. Through fidelity assessment, the link between program implementation and outcomes is systematically evaluated and ultimately leads to a greater understanding of program success or failure. This paper reports the results of an implementation fidelity study of the ‘Reconnecting Youth’ (RY) prevention program. The research questions were (i) was the program implemented with fidelity? and (ii) did better fidelity predict better outcomes? RY is an indicated drug abuse prevention program for high school students that seeks to ‘reconnect’ high-risk youth to school before they drop out. The results reported here were part of a randomized controlled effectiveness trial of the RY prevention program conducted in two urban school districts in which 15 teachers taught 41 RY classes. Overall, implementation fidelity was high with an average 90% of core lessons being taught. Unexpectedly, increased quality of implementation predicted increased alcohol use and anger. Adherence (teaching more of the curriculum) predicted increased marijuana use, while exposure (student attendance) significantly increased bonding to high-risk peers, alcohol use and anger.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionConflict of interest statementAcknowledgementsReferences |
|---|
As school-based prevention research moves from small efficacy trials into larger effectiveness trials, prevention researchers are challenged to evaluate prevention programs in real-world settings. In ‘a treatment effectiveness trial’, a program is evaluated in a real-world setting, with an emphasis on achieving high program implementation fidelity. This type of study represents the middle ground between implementation in a tightly controlled efficacy trial and an ‘implementation’ effectiveness trial, where implementation is allowed to vary naturally or by planned comparison [1].
Fidelity research may be especially important as programs move from efficacy to effectiveness trials, where the research emphasis is not limited to testing an intervention model, but based on whether the trial was ‘sufficient to permit a good test’ [2] under real-world conditions. Through fidelity assessment, the link between program implementation and outcomes is systematically explored and evaluated and ultimately leads to a greater understanding about program success or failure.
Fidelity studies can document program implementation failure (when programs have not been implemented as designed—referred to as a Type III error) and test the possibility of theory failure (when the program is delivered as planned yet hypothesized outcomes are not achieved). They can also assure that the contributions of those implementing a program (e.g. teacher) and the program (e.g. curriculum) are adequately measured [3]. Assessing fidelity also reveals information about the feasibility of a program (a low-feasibility program being one in which it is hard to achieve fidelity in practice) [4]. Finally, demonstrating implementation fidelity is central to the internal validity of the study, and closely related to statistical power [5].
In many drug abuse prevention studies, fidelity of implementation has been associated with improved outcomes [6–10]. In school settings, this research indicates that the more completely a program is implemented, the greater the effects; inversely, programs not implemented well are less effective [4]. Fidelity of implementation has also been associated with changes in mediating variables believed to be responsible for outcomes [11].
Fidelity of implementation has been described as the degree to which a program is implemented as intended by the program developers [4]. Although intervention researchers agree that fidelity of implementation is related to the success of empirically tested interventions, there are no universal standards to assess fidelity. At a minimum, fidelity assessment includes evidence that program components were delivered consistently across participants (e.g. individuals or classrooms) and that the implementation was true to the program model and theory [5]. Dane and Schneider [12] identify five dimensions of fidelity: (i) adherence, (ii) exposure (i.e. participant attendance), (iii) quality of delivery, (iv) participant responsiveness and (v) program differentiation (i.e. component analysis to determine which elements of prevention programs are essential). (Note that this last dimension was adapted by Dusenbury et al. [4] and we have chosen their adapted definition.) Together, these measures provide a comprehensive understanding of a program which can be especially useful in effectiveness trials, as assessed within real-world implementation.
The purpose of this paper is to report on the results of a fidelity of implementation study within a treatment effectiveness trial of the Reconnecting Youth (RY) prevention program. Our two main research questions were (i) was the program implemented with fidelity and (ii) did better fidelity predict better outcomes? The fidelity evaluation framework is organized using the dimensions of fidelity of Dane and Schneider [12], through which we examined the implementation of RY in two urban schools districts. As yet, no single published study of drug abuse prevention has included all five measures of fidelity [4].
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionConflict of interest statementAcknowledgementsReferences |
|---|
Program description
RY is an indicated (i.e. targeting persons who are already experimenting with drugs or other risk-related behaviors [13]) drug abuse prevention program for high school students that seeks to ‘reconnect’ high-risk youth to school before they drop out. Students are selected based on school record reports of high truancy (top 25% for their grade level) and low grade point average (GPA; bottom 50% of grade level) or by teacher referral. The program consists of a one-semester class, taken for academic credit, with the objectives of improving academic achievement (GPA and attendance), reducing or preventing drug use and improving mood management (depression, anger and anxiety).
The RY program model, guided by strain, social learning and control theories [14], posits that individual behavior is developed and maintained within networks of social relationships [15], where schools are natural environments for increasing social attachment. The intervention consists of two main components: (i) skills training, including skills application, and (ii) group development, including social support from the teacher and from peers. Social support is critical for students to be willing and able to acquire the skills that will allow them to reconnect and prevent drug use and school failure, i.e. to initiate and institute behavioral change [14]. These changes are measured as intermediate changes in both internal resources (e.g. personal control) and social resources (e.g. peer group interaction) as well as through proximal changes (e.g. academic measures and drug involvement).
A specially trained RY teacher leads RY students in skills training for pro-social behaviors and in giving and receiving social support. RY teachers, ideally recruited from the pool of classroom teachers, are selected because of their facility in working with high-risk students; ability to create and sustain a positive, trusting and supportive environment; appropriate use of discipline and ability to effectively teach life skills using participatory, experiential teaching methods.
RY is implemented in a small group of 10–12 students. Students learn to self-monitor their attendance, moods and drug use on a daily basis, and to develop achievable goals based on these data. RY is a five-unit written curriculum, with 55 core and 24 booster lessons. The first unit is comprised of 10 lessons introducing students to the RY model. Students learn about self-monitoring and begin the process of establishing a positive group environment involving trust, support, confidentiality and individual and group limit setting. Four units follow: self-esteem, decision making, personal control and interpersonal decision making. Each individual unit balances skills training and positive individual and group development. In turn, each lesson follows a basic structure where each part of the lesson is timed, the daily concept is outlined and all activities are detailed. In all, 50–60% of each class session is allocated to skills building; 20% to monitoring and reporting back on homework and 20% to a structured discussion/feedback process for student issues, problems and celebrations.
Sample and setting
The data reported here were collected as part of a randomized controlled effectiveness trial (RCT) of the RY prevention program conducted in two urban school districts beginning in fall 2002 (for a detailed description of study methodology, see [16, 17]). Findings from intent-to-treat analyses across all schools indicated that there were no main program effects at immediate post-test, and only one variable (progression of drug use) showed a school-by-group interaction effect [17]. Small positive effects were found in four schools, while small negative effects were found in five schools. At the 6-month follow-up, three negative main program effects were found (conventional peer bonding, high-risk peer bonding and pro-social weekend activities).
School District A was situated in a large city in the southwest; District B was located in a large metropolis on the Pacific coast. Site B completed the RCT in spring 2003, while Site A completed the RCT in fall 2004 due to later entry of two schools into the study. Data were collected over three semesters from 41 RY classes in nine high schools across the two sites. Outcome data are from 369 students who enrolled in the semester-long class and who attended >1 day.
The study population included 15 teachers who taught one or more of the 41 classes. Teachers implemented the RY curriculum during at least one semester. Although school principals and designated RY teachers had committed to teach for three semesters, only six teachers (40%) met this commitment for several reasons. Two schools dropped out of the program (one in each site) and other RY teachers left the school district or were reassigned to teach another course.
Mean teacher age was 40.3 years (SD = 13.3, median = 37 years). Ten of the 15 teachers (66.7%) were female; eight (53.3%) were Caucasian, three (20.0%) were African–American, one each (6.7%) was Asian and Hispanic and two (13.3%) were classified as other race/ethnicity. Teachers had an average of 13.4 (SD = 11.6) years teaching experience and an average of 7.4 (SD = 8.1) years working at their particular school. In Site A, all RY teachers were certified high school teachers; RY teachers in Site B included certified teachers, school counselors, a social worker and a school nurse.
To ensure close replication of the RY program model, research staff worked closely with school representatives at each stage of the study. Because the RY teacher is considered to be key to a successful RY program, the research site coordinators worked closely with school administrators to identify and select school staff that met the developers' suggested criteria for effective RY teachers. All RY teachers completed an intensive 4-day training workshop, conducted by the RY program development staff; most teachers also completed at least 1 day of follow-up training. The training consisted of both didactic and experiential (e.g. role play) teaching methods that provided an overview of program theory and applied training in key components related to group process and skills development. The two research site coordinators completed two of these intensive teacher trainings in addition to receiving training in observation protocols by program development staff. RY coordinators held monthly supervision meetings with teachers and provided individual feedback following classroom observations. All teachers and students had appropriate manuals and workbooks provided by the research study.
Measures
Because we sought to replicate conditions of the efficacy trial, we used the program developers' process and outcome evaluation instruments with some limited adaptations (e.g. we combined individual group development and skills building items into fewer assessment categories). Table I provides an overview of tools used to measure fidelity of implementation.
|
Adherence
Teachers completed a Daily Lesson History Log, a form designed to minimize the teacher's burden and to motivate daily recording. All core and booster lessons were listed on the log; the log was then used to calculate the percentage of lessons completed. Teachers recorded lesson date and circled a number that corresponded to a range of minutes spent on skill training. To meet protocol, established by the developers, teachers had to spend half the class period on skills building.
Exposure
Student attendance was calculated based on RY teacher attendance records. The attendance results reported here are the average student attendance across all RY class sessions.
Quality
Research site coordinators planned four classroom observations per RY class each semester, in their respective sites. Observations were scheduled for Weeks 3, 5, 7 and 12 of the RY curriculum. The two research site coordinators conducted a total of 150 classroom observations, out of a possible 164, or 91.5% of the observation target. District school personnel, trained in the RY curriculum and in coding observations, observed RY classes on a less regular basis. In these cases, ratings were completed independently and discussed; final observation scores resulted in a score based on consensus.
Observers attended RY classes unannounced to assess the quality of life skills building and group development and to record the number of minutes spent on the skills segment of the lesson. Observers assessed the quality of life skills delivery in three main areas: daily skill introduction (what and why), teacher modeling/student practice and skill application (classroom and homework). The observers also assessed quality of delivery for group development, including the extent to which teachers facilitated the group's caring/helpful behaviors and applied appropriate discipline, encouraged the group to accept responsibility for tasks/processes and facilitated a format for the group to share and respond to problems (and successes) within the group. The structured observation protocol followed the developers' criteria to assess quality of implementation, based on a 0- to 5-point scale (not observed to truly exceptional), using the life skills and group development observation forms. A score of ‘3’ indicated satisfactory completion of the daily lesson as described in the detailed written curriculum (life skills) and adequate facilitation of a positive group environment (group development). Teachers who received a score of ‘1 or 2’ did not complete the core content of a lesson or did not achieve group development goals. A score of ‘4 or 5’—very rare in this study—meant that the teacher exceeded expectations through unusually skilled leadership and facilitation.
Participant responsiveness
Students completed an anonymous satisfaction survey twice per semester—in the early group formation stage and at the end of the semester. The survey consisted of three open-ended questions about their experience in the class and eight questions (four about the teacher and four about the group) on a 1- to 5-point Likert scale (never to always) assessing the quality of teacher and peer help: (i) the teacher/group supported and encouraged me, (ii) I trusted the teacher/group and shared my problems, (iii) the teacher/group made me feel that I really belonged and (iv) I learned how to solve problems from the teacher/group. Scores on the participant responsiveness instrument therefore ranged from 5 to 20 on each component (teacher help, peer help) or from 10 to 40 overall.
Program differentiation
Program differentiation was based on scores from coded classroom observations, and sought to discriminate the effect of the two main program components, life skills training and group development (social support).
Outcomes
The High School Questionnaire (HSQ) is an instrument developed by Eggert et al. to evaluate RY [14]. The HSQ was administered at four time points: at student invitation (Time 1), at the end of the next semester (Time 2), the end of the following semester (Time 3) and 1 year after program end (Time 4). For the present analysis, we used data collected at Times 1 and 3 (comparing baseline with 6 months post-program) because adequate time should have elapsed to observe program effects on both proximal and distal outcomes. HSQ multi-item scales have demonstrated acceptable reliability and validity in a previous efficacy trial and the current effectiveness trial [14, 16, 18–20]. For the effectiveness trial, the questionnaire was adapted to an audio computer-assisted interview format. A total of 25 outcome variables were included. Demographic variables assessing age, gender and grade (9–11) were also included.
From the 25 outcome variables, for the present analysis, we selected five proximal outcomes representing internal and social resources and six distal outcomes representing substance use, psychological distress and school performance (Table II). These outcomes were selected based on their importance in the RY conceptual model [15].
|
Statistical analyses
Based on the RY conceptual framework, we developed an analytical model to guide data analysis. As depicted in Fig. 1, we measured fidelity of implementation of skills building and group development, using constructs of adherence, exposure, quality, participant responsiveness and program differentiation. Higher scores on these constructs indicated higher fidelity, which should be associated with better outcomes in proximal indicators (increased personal control and personal coping skills, increased school connectedness and conventional peer bonding and decreased bonding to high-risk peers) as well as better distal outcomes (decreased substance use and psychological distress and improved GPA).
|
We used descriptive statistics to summarize each fidelity measure and computed Pearson correlation coefficients for associations between fidelity measures and teacher characteristics (age, gender, race, years of experience in teaching) and outcome variables. Since teacher characteristics were not associated with fidelity measures, we dropped them from further analyses.
Because students were nested within classes, intra-class correlation of participants would violate the assumption of independence for ordinary least squares regression analyses. We used a mixed regression model to account for problems related to hierarchically arranged data [21, 22].
In the mixed regression model, we tested effectiveness of classroom-level fidelity measures, which include adherence, exposure, quality, participant responsiveness and program differentiation. Since both the quality and program differentiation measures were based on observer scores (i.e. quality grouped them as one item, while program differentiation tested them separately), we tested two different models, each including one or the other of these two measures and all other fidelity measures. In these models, intercepts and slopes were allowed to vary across individuals, that is the models included random intercepts and random effects for time. Significant interaction terms (time x fidelity component) in the models meant that the fidelity components were predictive of change from baseline to 6-month follow-up in the (individual level) dependent variables. Each model included covariates of student gender, race/ethnicity, parental education, school and site.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionConflict of interest statementAcknowledgementsReferences |
|---|
We begin with descriptive and bivariate analyses related to the fidelity measures (see Table III).
|
Adherence
The mean percentage of total lessons taught (one measure of adherence) was 74% (SD = 0.11, range = 56 to 92%). Teachers taught, on average, 90% (range = 51–100%) of the core lessons, but only 38% (range = 4–83%) of the booster lessons (data not shown). The mean percentage of time spent on life skills building recorded by observers (a second measure of adherence) was 56% (range = 0–88%), just exceeding the 50% mark denoting adequate adherence on this program component.
Exposure
The average student attendance for the RY class was 79% (SD = 22.43), with a range of 4–100%.
Quality/program differentiation
Mean scores on group development and life skills building reported by observers were 2.43 (SD = 0.66, range = 1.65–3.58) and 2.49 (SD = 0.53, range = 1.42–3.33), respectively, out of the total possible scores of 5 on each measure (see Table III). These scores corresponded to a quality of implementation that was between ‘below expectations or protocol’ (a score of 2) and ‘meets study protocol’ (a score of 3). Significant positive correlations (0.66) were noted between the two components of quality (life skills building and group development).
Among teachers who taught more than one semester, there was little evidence that teacher scores improved with subsequent semesters of teaching. The only difference detected (P < 0.05) was between the first (mean = 2.39) and third (mean = 2.85) semester scores related to life skills building.
Participant responsiveness
Teacher help and peer help scores were 17.07 (SD = 1.03, range = 14.75–18.93) and 15.70 (SD = 1.86, range = 12.00–18.07), respectively, out of a total possible score of 20, suggesting relatively high levels of student satisfaction. Significant positive correlations were noted (0.79) between the two components of participant responsiveness (peer help and teacher help). Cronbach's alpha on the four items comprising the teacher help score was 0.82 and on the four items comprising the peer help score alpha was 0.87. The combined alpha for all teacher help and peer help items was 0.90.
Relationships to outcome variables
Table IV presents parameter estimates for interaction terms for fidelity measures and time, controlling for participant gender, race, parental education, school and study site. In this model, significant temporal effects were detected for several fidelity components. Adherence (teaching more of the curriculum) predicted increased marijuana use, while exposure (student attendance) significantly increased bonding to high-risk peers, alcohol use, anger and GPA. RY attendance was also marginally significant in increasing both personal control and marijuana use. Quality of implementation predicted increased alcohol use and anger from baseline to 6-month follow-up. Participant responsiveness predicted increased personal coping strategies and decreased marijuana use.
|
Although we attempted to assess program differentiation by looking at the separate contributions of group development versus life skills, neither of these separate components predicted outcomes. However, when we combined them as a single quality variable, we found that quality predicted two of our outcomes, as described above.
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionConflict of interest statementAcknowledgementsReferences |
|---|
Was RY implemented with fidelity?
The RY effectiveness trial was a rigorous test of the program, conducted by independent evaluators in two large urban environments. As a treatment effectiveness trial, we emphasized high implementation fidelity through extensive teacher training by expert RY trainers, ongoing monitoring and feedback to teachers and extensive support to the schools for teacher and student selection, as well as teacher and student curriculum and all resource needs.
Overall, teacher adherence and student exposure were high. Teachers taught an average of 73% of the RY lessons, including core and booster lessons. Almost all the teachers covered all the core lessons (90%). On average, students attended the RY classes 79% of the time. A small group of students (n = 11) attended just a few classes. Thus, most students (selected for truancy) attended >80% of the time. In short, RY was implemented with high adherence and student exposure.
In implementing a complex program like RY in the real world, some adaptations were needed. For example, the ideal RY teacher has excellent classroom teaching skills and experience working with high-risk students, and is well respected by both students and teachers. However, we found that when we identified such teachers, they were the least likely to be released from regular teaching assignments to teach a very small elective class.
Ultimately, all the teachers had strengths in some areas but not necessarily in all areas. For example, some were highly skilled in traditional teaching methods which helped them in RY's cognitive mapping for skill development. Others had counseling or coaching background which prepared them for group development. Based on these different strengths among teachers, we differentiated skills training and group development as two separate components of fidelity.
It is unknown whether teacher scores in the effectiveness trial approximate the same standard of quality as in the efficacy trial; there are no parallel published studies on implementation from the efficacy trial. We used trained observers, the developer's criteria and their adapted instruments for assessing the quality of implementation. The average scores for group development and life skills building, 2.43 and 2.49, respectively, approached the developers' standard of ‘3’ for achieving implementation quality. Because teachers often had greater strength in either skills training or group development, we found wide variation in these scores. But we also found that the same teacher could have very different scores in two different classes during the same semester or across semesters. Surprisingly, practice and experience did not necessarily improve quality from one semester to the next. The only improvement over time was in skill development and only in the third semester (only six of 15 teachers continued teaching for this long). We found no improvement in the quality of delivery for life skills in teachers in the second semester. In addition, we found no improvement in the quality of group development skills, no matter how long they taught the class.
Participant responsiveness (student feedback) showed high mean scores for both teacher and peer help (see Table III), and there was a strong, positive correlation between these two constructs (0.79, P < 0.001). In other words, students indicated that they felt they were receiving the support that the RY model intended for them to receive.
In sum, across the five components of fidelity, e.g. adherence, exposure, quality of delivery, participant responsiveness and program differentiation, the data suggest a ‘good enough’ implementation of the program in the real world.
Did greater fidelity of implementation predict better outcomes?
We expected that those teachers who implemented the RY program with greater fidelity would have demonstrated better outcomes. For example, teachers who were rated higher by observers on group development and skills training would be associated with better outcomes. This was not the case; in fact, the reverse was true. Higher quality was associated with higher student alcohol use and higher anger. Additionally, we expected that teachers' strength in group development may have predicted better outcomes in personal or social resources or that strength in skill development may have predicted better outcomes in school performance. Again, this was not the case.
We also found a negative relationship between higher adherence (i.e. number of lessons taught) and outcomes. Higher adherence predicted higher marijuana use. The results for student exposure, however, were somewhat mixed. Higher exposure predicted negative outcomes of greater high-risk peer bonding, higher alcohol use and higher anger. But it also predicted higher GPA. The relationship between exposure and GPA is not surprising: students with higher attendance are likely to do better in school.
The only fidelity measure that predicted better (and not worse) outcomes was participant responsiveness. A higher mean score for peer and teacher help (including support, trust, belonging and problem solving) predicted better outcomes in personal coping strategies and decreased marijuana use.
The most unexpected result was the relationship between exposure (student attendance) and negative outcomes, including increased high-risk behaviors, increased alcohol use and increased anger. That is, intervention exposure was the greatest predictor of negative student outcomes. Thus, while students whose attendance was high experienced better GPA, their grouping with other high-risk peers appears to have been harmful. This finding is consistent with that of Cho et al. [16], who examined longitudinal data for the RY treatment effectiveness trial, which included the entire intent-to-treat study population. Although mixed program effects were observed at immediate post-test, only negative effects were found at the 6-month follow-up. That study, part of a growing body of research on group contagion, suggests that bringing high-risk adolescents together increases the likelihood of high-risk behaviors [23]. In response to this emerging research, the revised National Institute on Drug Abuse Prevention publication, Preventing Drug Use Among Children and Adolescents: A Research Based Guide, cautions that grouping high-risk youth in peer group interventions can produce negative effects [24].
Limitations
This study has several limitations. First, we relied on the program developers' implementation instruments and did not assess the validity of these instruments for our study. The developers' created these instruments based on their program model and what they expected to see in program implementation. We applied their comprehensive and systematic assessment of program implementation over a three-semester treatment effectiveness trial.
We did not account for observers' skill in our analysis. However, research site coordinators were trained by the developers, followed the same curriculum during observations and used the developers' forms to code classroom observations. During the initial implementation at each site, each research site coordinator observed at least two classes with the project director to ensure that observation procedures were uniformly implemented.
Student feedback, an average score for each class, and completed anonymously, was at the class level and was not at the individual student level. Thus, we did not account for missing data by students who did not complete the survey, mainly because they were absent the day of the survey. However, the range of open-ended responses to the surveys did not indicate a social desirability bias; student responses included negative comments as well as positive ones. Additionally, GPA was measured differently by site, with some evidence of grade inflation at Site A. In addition, in Site A, GPA calculations only included core, not elective, courses.
|
Conclusion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionConflict of interest statementAcknowledgementsReferences |
|---|
Fidelity of implementation research is an important, yet not fully developed, branch of prevention research. This study represents a specific type of fidelity research—a treatment effectiveness trial—with a focus on achieving high implementation quality. Such trials, ideally conducted by independent investigators, are an important step in evaluating how a program is implemented in the real world and its contribution to targeted outcomes. This study raises questions about how interventions actually play out in the real world.
In summary, this study demonstrates the importance of systematically evaluating fidelity, including linking fidelity to expected outcomes. In general, we assume that our social and behavioral interventions at the very least will do no harm. The results of this study suggest that this is not necessarily the case. Thus, practitioners and researchers should consider the possibility of unexpected and possible negatives consequences when adopting and implementing new programs.
|
Conflict of interest statement |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionConflict of interest statementAcknowledgementsReferences |
|---|
None declared.
|
Footnotes |
|---|
At the time of this study, Dr Sánchez was at the Pacific Institute for Research and Evaluation, Chapel Hill Center. 
|
Acknowledgements |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionConflict of interest statementAcknowledgementsReferences |
|---|
This research was supported by National Institute on Drug Abuse grant DA1366. We acknowledge the contribution of Shereen Khatapoush, one of our research site coordinators, to the fidelity study. We also acknowledge Juanita Cuffee for her contribution in managing complex databases and John Rose and Samruddhi Thaker who provided data management support and feedback on earlier versions of the manuscript. We also thank the district and school administrators and the RY teachers from the two school districts for granting us entrée and for their support to conduct the study.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionConflict of interest statementAcknowledgementsReferences |
|---|
1. Flay BR. Efficacy and effectiveness trials (and other phases of research) in the development of health promotion programs. Prev Med 1986 15:451–74.[CrossRef][Web of Science][Medline]
2. Orwin RG. Assessing program fidelity in substance abuse health services research. Addiction 2000 95:S309–27.
3. Nelson GD, Poehler D, Johnson L. Implementation of the teenage health teaching modules: a case study. Health Educ 1988 19:14–8.[Medline]
4. Dusenbury L, Brannigan R, Falco M, et al. A review of research on fidelity of implementation: implications for drug abuse prevention in school settings. Health Educ Res 2003 18:237–56.
5. Dumas JE, Lynch AM, Laughlin JE, et al. Promoting intervention fidelity. Conceptual issues, methods, and preliminary results from the early alliance prevention trial. Am J Prev Med 2001 20:38–47.[CrossRef][Web of Science][Medline]
6. Botvin GJ, Baker E, Dusenbury L, et al. Preventing adolescent drug abuse through a multimodal cognitive-behavioral approach: results of a 3-year study. J Consult Clin Psychol 1990 58:437–46.[CrossRef][Web of Science][Medline]
7. Battistich V, Schaps E, Watson M, et al. Prevention effects of the child development project: early findings from an ongoing multisite demonstration trial. J Adolesc Res 1996 11:12–35 [special issue: Preventing adolescent substance use].[Abstract]
8. Abbott RD, O'Donnell J, Hawkins JD, et al. Changing teaching practices to promote achievement and bonding to school. Am J Orthopsychiatry 1998 68:542–52.[Web of Science][Medline]
9. Haynes NM. Lessons learned. J Educ Stud Placed Risk 1998 3:87–99.
10. Whitehurst GJ, Crone DA, Zevenbergen AA, et al. Outcomes of an emergent literacy intervention from head start through second grade. J Educ Psychol 1999 91:261–72.[CrossRef]
11. Hansen WB, Graham JW, Wolkenstein BH, et al. Program integrity as a moderator of prevention program effectiveness: results for fifth-grade students in the adolescent alcohol prevention trial. J Stud Alcohol 1991 52:568–79.[Web of Science][Medline]
12. Dane AV and Schneider BH. Program integrity in primary and early secondary prevention: are implementation effects out of control? Clin Psychol Rev 1998 18:23–45.[CrossRef][Web of Science][Medline]
13. National Institute on Drug Abuse. Preventing Drug Use Among Children and Adolescents: A Research Based Guide.Washington, DC: US Government Printing Office 1997.
14. Eggert LL, Thompson EA, Herting JR, et al. Preventing adolescent drug abuse and high school dropout through an intensive school-based social network development program. Am J Health Promot 1994 8:202–15.[Medline]
15. Eggert LL, Thompson EA, Herting JR, et al. Prevention research program: reconnecting at-risk youth. Issues Ment Health Nurs 1994 15:107–35.[Medline]
16. Cho H, Hallfors D, Sánchez V. Evaluation of a high school peer group intervention for at-risk youth. J Abnorm Child Psychol 2005 33:363–74.[CrossRef][Web of Science][Medline]
17. Hallfors D, Cho H, Sanchez V, Khatapoush S, Kim H, Bauer D. Comparison of efficacy and effectiveness trial results in an indicated "model" program: implications for public health. Am J Public Health 2006 96:8.
18. Eggert LL, Thompson EA, Herting JR. A measure of adolescent potential for suicide (maps): development and preliminary findings. Suicide Life Threat Behav 1994 24:359–81.[Web of Science][Medline]
19. Eggert LL, Herting JR, Thompson EA. Measurement document and questionnaire item identification for high school questionnaire, reconnecting at-risk youth NIDA project. Seattle, WA: Department of Psychosocial Nursing, SC-76, University of Washington 19981–25.
20. Thompson EA and Eggert LL. Using the suicide risk screen to identify suicidal adolescents among potential high school dropouts. J Am Acad Child Adolesc Psychiatry 1999 38:1506–14.[CrossRef][Web of Science][Medline]
21. Byrk AS and Raudenbush SW. Hierarchical Linear Models: Applications and Data Analysis Methods.Newbury Park, CA: Sage 1992.
22. Hox JJ. Applied Multilevel Analysis.Amsterdam: TT-Publikaties 1995.
23. Dishion TJ, McCord J, Poulin F. When interventions harm: peer groups and problem behavior. Am Psychol 1999 54:755–64.[CrossRef][Medline]
24. Robertson EB, David SL, Rao SA. Preventing Drug Use Among Children and Adolescents: A Research-Based Guide for Parents, Educators, and Community Leaders. 2nd edn Washington, DC: DHHS, NIH, National Institute on Drug Abuse 2003.
Received on October 19, 2005; accepted on May 10, 2006
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J. R. Masuda, K. Robinson, S. Elliott, and J. Eyles Disseminating Chronic Disease Prevention "to or With" Canadian Public Health Systems Health Educ Behav, December 1, 2009; 36(6): 1026 - 1050. [Abstract] [PDF]
|
|
|
|
 |
|
 L. Buckley and M. Sheehan A process evaluation of an injury prevention school-based programme for adolescents Health Educ. Res., June 1, 2009; 24(3): 507 - 519. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Thaker, A. Steckler, V. Sanchez, S. Khatapoush, J. Rose, and D. D. Hallfors Program characteristics and organizational factors affecting the implementation of a school-based indicated prevention program Health Educ. Res., April 1, 2008; 23(2): 238 - 248. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||