The Center's mission will continue to be accomplished through the following activities: 1) Design, implement, and test integrated education and prevention programs through rigorous randomized field trials, that yield evidence-based education and prevention programs, and that add to knowledge about the causes and prevention of academic, behavioral, and psychological problems and disorders. 2) Design and test methods of putting into practice on a community and population scale the scientifically supported practices identified through this research. 3) Design and test unified systems for integrating education and preventive services beginning with the universal level and integrating with back-up selective, indicated, and reparative/treatment services needed to support child and student development. 4) Train the next generation of inter-disciplinary investigators, policy makers, and service providers necessary for this broad mandate. 5) Contribute scientific evidence to inform the policies needed to institutionalize the integration of education and public health prevention science and programs at the international, national, state, and local levels.

These risk factors are also strongly related to a host of other risk factors that separately or together are predictive during adolescence and young adulthood of not only drug abuse, but also conduct disorders and violence, depression, school drop out, and high-risk sexual behaviors. Ineffective parenting around discipline and homework; classrooms with high levels of aggressive, disruptive behavior; antisocial classmates and peers; poverty at the family level and at the school and community levels, and individual differences such as sensation seeking (Wills, Sandy, & Yaeger, 2000; Palmgreen, Donohew, Lorch, Hoyle, & Stephenson, 2001), all have been found to increase the risk of drug abuse and related comorbid problems (Ary et al., 1999; Dishion et al., 1999; Kellam, Ling, Meriska, Brown, & Ialongo, 1998; Reid, Patterson, & Snyder, 2002).

Developmental epidemiologically based, randomized field trials have directed interventions at decreasing the early risk factors in the classroom, family, and peer-group settings. Our strategy has been to conduct carefully designed, epidemiologically based, randomized field trials beginning in first grade and to target specific sets of SAS risk factors in stages. In the three generations of preventive trials in Baltimore our first stage has been to test interventions separately; the second stage to test combinations; and now in the third stage to fully integrate the preventive interventions with one another and with the actual quality of a teacher's instructional and classroom management practices. The first stage involved the matching of schools in clusters and random assignment of schools to intervention or standard program (control condition) and randomization of teachers to classrooms within schools, with balancing of children across classrooms, and the random assignment of classrooms to intervention conditions. The second and third generations involved within school designs, with random assignment of teachers, children, and classrooms to intervention or standard program conditions. It has required a full partnership with the Baltimore City Public School System, under their aegis and that of the Baltimore Teachers Union, as well as the strong support and active consent of the parents.

The Baltimore trials and others with rigorous designs indicate that school-based universal interventions (i.e., those addressing all children, not merely those at higher risk) can have short- and long-term beneficial effects on aggressive behavior and achievement (Dolan et al., 1993; Ialongo et al., 1999; Reid, et al., 1999), off-task behavior (Brown, 1993a, 1993b), and depressive symptoms (Kellam, Rebok, Mayer, Ialongo, & Kalodner, 1994). Impact has been reported from first grade interventions leading to reduced aggression in middle school (Kellam, Rebok, Ialongo, & Mayer, 1994; Kellam, Ling et al., 1998), delinquency (Eddy et al., 2000), and substance use (Eddy et al., 2003). Longer-term impact has been reported in meta-analyses on illicit drug use by Nan Tobler and colleagues (1986; 2000; Tobler et al., 2000) plus other meta-analyses (Derzon & Lipsey, 1999; Gorman, 1995). Higher-fidelity implementation of the interventions also led to higher impact (Ialongo et al., 1999, Ialongo, Poduska, Werthamer, & Kellam, 2001). Further, we have found compelling evidence that these universal interventions often have the greatest impact on those at highest risk of substance abuse and aggression (Brown & Liao, 1999; Curran & Muthèn, 1999; Muthèn & Curran, 1997; Muthèn et al., 2002; Stoolmiller, Eddy, & Reid, 2000).

Although little was known about prevention science or effective prevention programs at that time, the mandate from the Board was to direct research at the problem of helping children reach their full potential, particularly in school. Preventing school failure and behavioral and mental disorders was felt by the Board to be strongly related to later schooling and success or failure in the lives of their children. This was in stark contrast to the hopes of the neighborhood health and social service agencies whose agency headquarters were outside of Woodlawn, and whose leaders hoped for a back-up by the new Mental Health Center for their mentally ill clients. Woodlawn Mental Health Center services programs for the ill, based in churches and other community settings as well as in the Center, were to come later, but only after research into prevention and promotion for children was well underway.

The first problem was that little if any rigorous data existed in 1964 in regard to early risk factors for later mental, educational, or behavioral problem outcomes. With the Board's support, the first developmental epidemiological randomized field trial was begun in 1964 in all first grade classrooms of the 9 public and three Catholic parochial schools in the neighborhood. Schools were matched into pairs and a program tested consisting of a psychiatrist or psychologist assigned as classroom meeting leader and as consultant to each first grade teacher in a randomly chosen intervention school from each of the pairs of schools. From the 1964-65 school year through 1968-69 four cohorts of first grade children participated in this design. The third cohort has been followed up in third grade, then age 16-17, and then at ages 31 and 42. These last two follow-up efforts have been led by Margaret Ensminger, PhD., one of the early senior leaders of the Woodlawn Mental Health Center.

In November, 1982 Kellam moved to the then Johns Hopkins School of Hygiene and Public Health and developed an NIMH funded Prevention Research Center in strong partnership with the Baltimore City Public School System (BCPSS), with the same agreement developed with the community of Woodlawn. Research to be done would be rigorous, but in the interest of the mission of the BCPSS, as well as contributing to public health prevention science. In March, 2000 Kellam came full-time to AIR, where the tradition of education research allowed the integration of public health prevention science and education research concepts and methods in the form of the Center for Integrating Education and Prevention Research in Schools. He had just retired as Professor Emeritus from the Johns Hopkins Bloomberg School of Public Health, where he had carried out with colleagues two prior generations of preventive trials under the aegis of the BCPSS starting in 1984.

Kellam was followed soon by Jeanne Poduska, ScD, deputy director, who received her doctorate from the School of Public Health and had a long history with the BCPSS/Hopkins based prevention research in schools. Carla Ford, PhD, joined as chief of school-based interventions, with over 32 years experience in child development and curriculum and instruction in BCPSS. Natalie Keegan joined the Ed/Prev team as coordinator of community and institutional base-building and co-ordinator of the family/classroom partnership component of the current trial, after long experience in community base-building and earlier work on the BCPSS/Hopkins based preventive trials. Amy Windham, ScD, received her doctoral degree also from the School of Public Health and came to the Ed/Prev Center as chief of field assessments and as a scientific collaborator in the design and testing of new programs. Hendricks Brown, PhD, working from the University of South Florida and director of the prevention science methodology group (PSMG) joined in the development of the Center as a senior collaborator, after many years of collaborating in Chicago and then in BCPSS in randomized field trials. John Reid also joined the team developing the Center from his base in Eugene, Oregon, at the Oregon Social Learning Center. As the phases of the Center development proceeded, others have joined in various roles.

Three generations of preventive trials have been carried out by the Ed/Prev team in partnership withBCPSS. The first two generations were administered from the Johns Hopkins Bloomberg School of Public Health, Prevention Research Center. The third and current trial is based at AIR in the Ed/Prev Center.

In the 1st generation, we used a separate theory driven intervention aimed at each of two correlated proximal target risk factors-- poor academic achievement and aggressive, disruptive behavior in first grade school children. These two risk factors were substantiated over the last three decades as antecedents of later depression, drug abuse, school failure, and antisocial outcomes. Mastery Learning (a curriculum, instruction method) was directed at poor achievement, and the Good Behavior Game (a classroom management method) was directed at aggressive, disruptive behavior. The research was designed to determine if improving achievement led to improved behavior and/or the reverse, and if by improving either or both, the risk of the long term outcomes could be improved Using RFTs in this way allows such field trials to inform etiology by testing the causal role of the proximal risk factor and to provide evidence-based prevention programs.

In the 2nd generation, these two interventions were combined to answer whether the combination yielded results that were synergistic (more than the sum of the predicted outcomes), additive (merely the sum of results of each), or redundant (no additional benefit from the two interventions beyond one of them). In addition, a Family-Classroom Partnership intervention was tested separately.

In the 3rd generation, the Ed/Prev team and the BCPSS partners are combining three components previously tested separately (the Whole Day (WD) Program for First Grade Classrooms ) to test their effectiveness in preventing school failure, drug abuse, and the other problem outcomes. In the same design, we are following teachers over consecutive cohorts of first grade children to test whether a multi-level model of mentoring and monitoring teachers can lead to institutionalizing the WD.

1. A strong partnership was formed that enabled research applications for funding education/prevention trials, including rigorous population based randomized designs among the Ed/Prev Center and:
   
   
   1. Baltimore City Public School System
   2. Baltimore Community and Institutional Board
   3. Baltimore Teachers Union
   4. Public School Administrators and Supervisors Association
   5. Morgan State University School of Education and Urban Studies
   6. Prevention Science Methodology Group
   7. Oregon Social Learning Center
      
      
2. A three-year planning grant awarded by NIH, administered by NICHD in the fall of 2000 to design the next generation of preventive trials in schools.
   
   
3. A three-year grant from NIMH to analyze data gathered in following up the 1st generation of Baltimore education and prevention field trial from age 6 through 21.
   
   
4. A contract with Johns Hopkins Bloomberg School of Public Health to continue collaborating in the analyses of 2nd generation Baltimore Prevention Trial results.
   
   
5. A five-year grant from NIDA, with contributions from NICHD, to design and test the effectiveness of an integrated education/prevention program, the Whole Day Program for First Grade Classrooms (WD) based on bringing together separate components previously tested separately, with a focus on mastering learning to read. The components integrated in WD include: 1) enhanced classroom instruction of reading throughout the day integrated with all 1st grade academic subjects; 2) classroom behavior management; and 3) classroom/parent partnerships around reading and behavior. The design is a multi-level randomized block design involving first grade classrooms of 20 elementary schools, with 8 schools serving as partners in developing the interventions and measures, while 12 have been the sites for randomization and testing the interventions.
   
   
6. The 20 schools were all of those in two areas of the city that were not confounded by other large programs. Schools were randomly assigned to either the sites for the random assignment of children and teachers or to the role of helping with the development of measures and the training of teachers. In the spring of 2003 teachers who taught the first grade classrooms in the following fall of 2003 were assigned randomly within each school to either the Whole Day classroom or to a standard program classroom. Children were assigned to classrooms randomly in the 12 schools in late summer just before the start of school.
   
   
7. Informed written consent was obtained from over 98 percent of parents for the additional non-educational measures obtained from the children, those involving psychological well-being. The baseline measures were obtained after consent, then the WD administered for the remainder of the last school year. Data are now being analyzed for main effects and for variation in impact, answering the question: what works, for which children, and under what circumstances?
   
   
8. This same grant, in addition to following the first cohort of children through 3rd grade, includes following the first grade teachers over consecutive cohorts of first graders to determine whether the social, political, and mentoring/monitoring structure in place can sustain teacher practices.
   
   
9. In its 4th year, this grant will allow testing the training modules required for expanding beyond the initial 1st grade teachers to include the standard program classroom teachers, as the data warrant. Such social and political structure and the training mentoring/monitoring modules will, we predict, provide a foundation for models for going-to-scale, not only useful in Baltimore but other school districts as well.
   
   
10. The development and application of new research methodology has been a continuous theme, seamlessly woven into all aspects of the above work. The partnership with PSMG led by Hendricks Brown at USF and the support from Bengt Muthèn at UCLA has been a leading and continuing part of the Center development cited here.
    
    
11. A full day workshop at the 2004 AERA annual meeting was sponsored by AIR and involved the growing Ed/Prev faculty from the above partner institutions.
    
    
12. A multi-site Ed/Prev Center training grant for pre and post-docs in education/prevention population based randomized field trials has now been submitted to NIDA, led by Hendricks Brown and PSMG to be based at University of South Florida, and involving the above partner organizations. The scientific parameters that training will include are:
    
    
    • Strategies and phases for population based randomized ed/prev field trials.
    • Theory and practice for community and institutional partnerships required for this kind of research on effectiveness and for institutionalizing effective programs.
    • Role of causal theory, previous research findings, and intervention theory.
    • Role of epidemiology, particularly developmental epidemiology.
    • Defining and specifying the intervention and control groups, so that the core elements are observable and measurable in both.
    • Measurement of baseline; implementation; mediating and moderating variables; periodic outcomes.
    • Design, Sampling, estimating power in RFTs.
    • Analyzing main effects, but also variation in impact over time.
    • Economic analysis: cost, cost effectiveness, cost benefit.
      
      
    The training will focus on integrating concepts, methods, and results from education and prevention in public mental health; but also it will focus on integrating methodology research with substantive research, just as it has since the beginning of the collaboration of these researchers and institutions.
    
    
13. An application was submitted July 1, 2004, for a competitive supplement to the parent NIDA grant that funds the BCPSS randomized Field Trial. These funds would broaden support for assessing the children in the next two cohorts of first grade children.
    
    
14. The next grant to be submitted will be directed at the follow-up of the WD three cohorts of children to assess periodically the WD impact along with assessing the impact of the teacher practices under the differing training, mentoring, and monitoring conditions. This grant will be followed quickly with one directed at the organization of back-up services for children, such that valid criteria can be used for automatically attending to the special needs of children who need more than the WD universal intervention program.

• At the Level of Targeting Early Antecedents Along Developmental Paths Leading to Behavioral and Mental Disorders. For many prevention researchers, a strategy built upon integration of concepts and methods from community epidemiology, life course development, and experimental intervention designs and methods holds great promise. The substantive base for this developmental epidemiological strategy has been the growing identification of early antecedents along developmental paths leading to health or disorders. In many cases preventive interventions have already targeted these antecedents in rigorous experimental, randomized population based designs. This strategy requires long-term periodic follow-up and growth modeling. It is particularly relevant for studying effectiveness of early instructional strategies for teaching children to read, or to behave as students in early grades. The 3 generations of Baltimore trials are examples, where schools, classrooms, teachers, and students have been randomly assigned.
  
  
• At the Level of the Individual in High Risk Situations. A second strategy is directed at reducing risky behaviors as individuals enter periods or situations of increased risk. Examples include increasing adolescents' resistance to peer pressure for drug abuse and increasing use of condoms to prevent HIV infection. Random assignment is made to intervention and control groups. The resulting data are more focused on the short time period between intervention on the proximal risk target and the distal outcome, such as HIV infection.
  
  
• At the Community Level. Community-wide strategies assessing the impact of multiple interventions through media, schools, and other social institutions been developed and tested. Programs have been tested such as efforts to prevent drug and tobacco abuse through enforcing age limits to purchasing. Pairs of communities are matched and frequent baseline measures are made, followed by intervention in one randomly chosen of the pair. Replication is done across such pairs of communities. This strategy could lend itself to studies across school districts as well as communities.
  
  
• At the Policy Level. Concepts and methods have been developed and applied to the study of broad policies that have prevention goals, and also could be applied to education policies.. The impact of policies tested can include laws regulating the age of legal alcohol purchases and their enforcement, drunk-while-driving laws, gun regulation, and school policies regarding class size. Computer simulation modeling with multiple imputations is used as an approach to rigor.

These prevention science strategies provide a scientific foundation for the integration of prevention science with education research. The Ed/Prev Center team makes most use of the developmental epidemiological strategy, our work also will involve other strategies as these become the best approaches for research questions of the future. While prevention scientists have begun to address risk factors using each of the strategies discussed above, many more risk factors require attention. A growing number of randomized preventive trials have already demonstrated malleability of specific targeted early risk factors (such as low achievement or aggressive behavior), with consequent reduced risk of the potential problem outcome such as drug abuse, school failure, or delinquency. Education/preventive trials can test theory-driven causal models by aiming at hypothesized causal early antecedents, determining if they can be improved, and if so whether the risk of the problem outcome is reduced. Such trials can build theory as well as provide useful preventive and education outcomes (Kellam and Rebok, 1992; Kellam, Koretz, and Moscicki, 1999).

Like prevention science more broadly, the Ed/Prev Center must be interdisciplinary in order to discover and experimentally manipulate the potential target risk factors at levels from the individual to the cultural. We must integrate the disciplines concerned with human development, risky behavior, and educational, mental, behavioral, and social health. Educational, biological, psychological, sociological, economic, bio-statistical, and anthropological perspectives are all needed to understand and prevent educational, mental, and behavioral disorders and to promote human development. Integrating quantitative with qualitative methodologies is essential in the pursuit of prevention science. Indeed, the focus of the Ed/Prev Center must be as much about research methods as substance.

The developmental epidemiological approach can encompass the nested influences that culture, broader social structure, community, and specific small social fields such as classroom and peer group have on the social task demands and the variation in adaptive responses of individuals. The prevention strategies outlined in the preceding section can all be applied within this framework, with the primary focus on the small social fields and individual variation in meeting the social task demands.

The theoretical framework through which the developmental epidemiological approach is applied is that of life course/social field theory (Kellam, Branch, Agrawal, & Ensminger, 1975; Kellam and Ensminger, 1980; Kellam and Rebok, 1992; see Life Course/Social Field graphic attached). The central concept is that at each stage of life individuals are involved in a few main social fields, such as the parental family, then the classroom and peer group in the younger years. Later, the intimate/marital and work social fields become more salient. In each social field there are social task demands specific to that social field. In the classroom, sit still, pay attention, and learn are important demands, while in the play ground playing with appropriate control of physical aggression and playing according to the rules of the game are important. The adequacy of our responses to the various demands we face is a judgment external to the individual, rated by others empowered by their role to make such ratings. These natural raters, as we have named them include parents in the family, teachers in the classroom, peers in the peer group, supervisors at work, and partners or spouses in the intimate, or marital social field. This process of demand/response is called social adaptation, and ratings of the adequacy of one's performance in these social tasks as judged by one's natural raters is termed social adaptational status(SAS).

How one responds to the social task demands and how the natural rater views the adequacy of responses is conceptually distinct from how one feels psychologically. This latter concept is called psychological well-being (PWB). It includes internal states such as anxiety, depression, and self-esteem. If we intend to study the relationships between SAS and PWB, it is crucial to measure SAS and PWB separately in order to avoid confounding one with the other. Although these two domains are hypothetically highly interrelated, problems in each may, and often do, have very different long-term outcomes (SXDRUG; FSDELR; GBGAGG; ACHDEP REFS).

Within the framework of life course/social field theory, there are several basic predictions about change:

• Early successful social adaptation predicts later success. Early successful social adaptation tends to establish competencies through the sense of mastery that follows reinforcement by natural raters of successful performance of social task demands. These feelings of mastery and the competencies that accompany them remain accessible in the face of later social task demands, and particularly during times of stress, crisis, novelty, and demands for adaptive innovation (REFS).
• Success in one earlier social field predicts success in other later social fields. Social competencies are available for later adaptation as the individual traverses the life course in new social fields.
• Social adaptational status and psychological well-being are intimately and often reciprocally related. The way we feel affects how well we perform social tasks, just as our performance affects how we feel.

The WD intervention currently being tested in Baltimore is directed at the classroom social field. Hypothetically, improving teacher instructional practices, classroom management, and parent/classroom partnering around homework and behavior should result in improved social adaptational performance by the children, and this in turn should improve their psychological well-being as well as their later SAS in school as well as in other later social fields.
--------------------

* There are two major branches of epidemiology: analytic epidemiology, which studies the causes of nonrandom variation in populations; and demographic (or descriptive) epidemiology, which describes rates and trends in populations. Developmental epidemiology is a subset of community epidemiology, concerned with following defined populations over time and studying their variation.