Science-Based Substance Abuse Prevention: A Guide - PRACTICES
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
1 GUIDE TO SCIENCE-BASED PRACTICE S Science-Based Substance Abuse Prevention: A Guide Substance Abuse and Mental Health Services Administration Center for Substance Abuse Prevention Division of Knowledge Development and Evaluation
Acknowledgments This document was produced under the guidance of Stephen E. Gardner, D.S.W., project officer, Paul J. Brounstein, Ph.D., and Deborah B. Stone, Ph.D., Division of Knowledge Development and Evaluation, Substance Abuse and Mental Health Services Administration, Center for Substance Abuse Prevention (SAMHSA/CSAP), through contract #282-98-0023 to ROW Sciences, Inc., Carol Winner, project director. Science-Based Substance Abuse Prevention: A Guide is based on A Guide to Science-Based Practices in Substance Abuse Prevention, a seminal monograph developed and written by Paul J. Brounstein, Janine M. Zweig, and Stephen E. Gardner, with substantial contributions from Maria Carmona, Paul Florin, Roy Gabriel, and Kathy Stewart. Special thanks go to a dedicated review panel composed of Eric Einspruch, Roy Gabriel, and Katherine Laws, RMC Research Corporation; Carol Hays, Illinois State Incentive Grant Program; Patricia Post, Central Regional Center for the Application of Prevention Technologies; and Mary Joyce Prudden, CSAP, who offered invaluable suggestions for improving the document. The Department of Health and Human Services (DHHS) has reviewed and approved policy-related information in this document but has not verified the accuracy of data or analyses presented in the document. The opinions expressed herein are the views of the authors and do not necessarily reflect the official position of SAMHSA or DHHS. DHHS Publication No. (SMA)01-3505 Printed 2001 For single copies of this document, contact SAMHSA’s National Clearinghouse for Alcohol and Drug Information (NCADI), P.O. Box 2345, Rockville, MD 20847-2345; 1-800-729-6686, 301-468-2600, or TDD 1-800-487-4889; or visit the Web site at www.samhsa.gov. ii Science-Based Substance Abuse Prevention
Foreword Although recent reports show a leveling or decrease in substance use among our nation’s youth, drug abuse remains a problem in our country. There were 14.8 million current users of illicit drugs in 1999. This figure represents 6.7 per- cent of the population 12 years and older. The 1999 National Household Survey also found increases in illicit drug use among adults ages 18–25. Although the rates for those 26-34 years old and 35 years and older have not changed significantly since 1994, overall statistics indicate that there is still work to be done in preventing substance abuse. The Substance Abuse and Mental Services Health Administration’s (SAMHSA) Center for Substance Abuse Prevention (CSAP) developed this series of products in response to the ongoing substance abuse problems. The three components in this series support CSAP’s mission to provide resources that are based on science, with measurable outcomes, and designed to help community and state leaders formulate targeted programs. CSAP is committed to sponsoring, accumulating, and integrating knowledge regarding scientifically defensible and effective prevention practices. The primary foci of each document in this series is CSAP grantees, constituent organi- zations, and the communities these groups serve. We are pleased to release this guide to CSAP’s conceptual framework and methodology for designing and assess- ing scientifically defensible programs for substance abuse prevention. The results reflect the findings of 10 years of CSAP-funded demonstration programs, as well as other advances in the design and evaluation of successful preven- tion strategies. These findings together lay the foundation for a new, empirically based approach to prevention pro- gramming. The booklet highlights the risk and protective factors that help determine an individual’s vulnerability to substance abuse. It also examines CSAP’s qualitative and quantitative strategies for evaluating existing substance abuse preven- tion programs and developing scientifically defensible best practices. This booklet is one in a series of products developed to help key stakeholders structure and assess scientifically defensible programs. It is designed to serve practitioners and others involved in the development, implementation, and evaluation of substance abuse prevention programs as we work together on innovative and effective solutions that respond to the unique needs of our individual communities. Joseph H. Autry III, M.D. Ruth Sanchez-Way, Ph.D. (Acting) Administrator Director Substance Abuse and Mental Health Services Center for Substance Abuse Prevention Administration Substance Abuse and Mental Health Services Administration A Guide iii
Table of Contents Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .ii Foreword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .iii Science-Based Substance Abuse Prevention : A Guide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1 The CSAP Conceptual Framework: Risk and Protective Factors . . . . . . . . . . . . . . . . . . . . . . . . . .2 Issues in Defining Scientifically Defensible Knowledge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4 Application of Specific Principles and Criteria to Research Findings . . . . . . . . . . . . . . . . . . . . . .6 The CSAP Approach to Identifying Scientifically Defensible Prevention Interventions . . . . . . . . . .8 Resources for Identifying Scientifically Defensible Prevention Principles and Programs . . . . . . .10 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .12 Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15 A Guide v
Science-Based Substance Abuse Prevention: A Guide The Substance Abuse and Mental Health Services This booklet, Science-Based Substance Abuse Administration’s Center for Substance Abuse Prevention: A Guide, is intended to assist preven- Prevention (CSAP) is responsible for identifying tion practitioners and others involved in the and disseminating scientifically defensible knowl- design, implementation, and evaluation of sub- edge about proven prevention models and princi- stance abuse prevention programs. This document ples to the substance abuse prevention field. provides the following: Because there is a wide range of opinion about ■ Describes the evolution of the conceptual model appropriate methodological criteria, standards, that CSAP uses to provide a unifying framework and expectations—ranging in rigor from requiring for substance abuse prevention. experimental studies to observing systematically ■ Discusses the principles that provide the founda- to accepting clinical judgments—determining what tion for understanding and defining research constitutes that knowledge can be challenging. findings as scientifically defensible. Over the past 10 years, CSAP demonstration ■ Illustrates how CSAP applies specific principles programs have mirrored the field’s growth in and criteria to research studies, program evalua- understanding what works in substance use pre- tions, and scholarly efforts to identify scientifical- vention. Initially, programs implemented and test- ly defensible findings, prevention principles, and ed prevention approaches based on practitioner prevention models. observation. The realization, for example, that The Guide also provides specifics about the most people had little knowledge about the harm- CSAP methodology for selecting model programs ful effects of alcohol and illicit drugs led to the and identifying scientifically defensible principles. development of public information campaigns It is one in a series of three products that together and school-based curricula designed to replace assist prevention practitioners, evaluators, state myths with facts. With increasing sophistication, and local program administrators, policymakers, this knowledge-centered focus enlarged to include and funders in designing and assessing scientifical- innovative interventions based on theories derived ly defensible programs. The other two products from research in related fields (e.g., Bandura’s are: social learning theory). More recently, data have ■ Promising and Proven Substance Abuse become available from substance abuse preven- tion-specific program evaluations that provide Prevention Programs, a comprehensive compila- empirical evidence about those strategies that suc- tion of both proven and promising interventions ceeded in changing knowledge, attitudes, and in an easy-to-scan grid organized by risk factor behavior. Those findings, many of which stem and domain; and from CSAP-funded demonstration programs, have ■ Principles of Substance Abuse Prevention, a nar- laid the foundation for a new empirically based rative and selected listing of proven and cited approach to prevention programming—one that interventions organized by domain. CSAP is challenging the field to use (Hansen and State and federal agencies, local governments, McNeal, 1996; Pandina, 1998). and private foundations have become increasingly interested in funding substance use prevention A Guide 1
Guide to Science-Based Practices programs with measurable outcomes. The new young adulthood (Bry & Krinsley, 1990; Newcomb emphasis on performance means that prospective & Felix-Ortiz, 1992). Risk factors include biologi- funding applicants must demonstrate that the pro- cal, psychological/behavioral, and social/environ- grams they propose are both outcome-oriented mental characteristics such as a family history of and likely to achieve the outcomes predicted. substance use, depression or antisocial personality Using scientifically defensible interventions can disorder, or residence in neighborhoods where help prevention practitioners respond to demands substance use is tolerated. Researchers have also for accountability and improve their capacity to found that the more the risks in a child’s life can provide effective services. be reduced—for example, by effectively treating CSAP’s demonstration programs have been the mental health disorders, improving parents’ family- testing ground for many of the most promising management skills, and stepping up enforcement ideas for intervening with those at high risk for of laws related to the sales of illicit drugs to substance use. As a part of its knowledge develop- minors or to drinking and driving—the less vulner- ment and application role, CSAP has accepted the able that child will be to subsequent health and responsibility for assessing those interventions and social problems (Hawkins, Catalono, & Miller, disseminating results to the field for replication 1992). and adaptation. In 1999, CSAP issued a primer on effective programs titled Understanding Substance Protective Factors and Resilience Abuse Prevention, Toward the 21st Century: A Protective factors, such as solid family bonds and Primer on Effective Programs that described eight the capacity to succeed in school, help safeguard model programs and briefly summarized the youth from substance use. Research has also methodology guiding their selection. Given the demonstrated that exposure to even a substantial increased emphasis on incorporating scientifically number of risk factors in a child’s life does not defensible principles and interventions in both necessarily mean that substance use or other prob- new and existing programs, it is important that lem behaviors will inevitably follow. Many chil- prevention practitioners understand the conceptu- dren and youth growing up in presumably al framework and criteria used in rating preven- high-risk families and environments emerge rela- tion interventions and attesting to their tively problem-free. The reason for this, according effectiveness. to many researchers, is the presence of protective factors that reduce the likelihood that a substance The CSAP Conceptual Framework: Risk use disorder will develop (Hawkins et al., 1992; and Protective Factors Mrazek & Haggerty, 1994). The research on pro- Theory and theoretical frameworks in the sub- tective factors explores the positive characteristics stance use prevention field have been evolving and circumstances in a person’s life and seeks over time, often through induction based on opportunities to strengthen and sustain them as a applied empirical research. Among the most preventive device. Among these resilient children, important developments in substance abuse pre- protective factors appear to balance and buffer the vention theory and programming in recent years negative impact of existing risk factors (Anthony & has been a focus on risk/protective factors as a Cohler, 1987; Hawkins et al., 1992; Mrazek & unifying descriptive and predictive framework. Haggerty, 1994; Wolin & Wolin, 1995). From a substance abuse prevention perspective, protective Risk Factors factors function as mediating variables that can be targeted to prevent, postpone, or reduce the Put simply, one often tested and supported impact of use. hypothesis derived from this framework is that the Taken together, the concepts of risk and more risk factors a child or youth experiences, the resilience enhance understanding of how and why more likely it is that she or he will experience sub- youth initiate or refrain from substance use. stance use and related problems in adolescence or Although not all risk and protective factors are 2 Science-Based Substance Abuse Prevention
amenable to change—genetic susceptibility to sub- Domain Subcategory of Risk stance use, for example—research demonstrates that their influence can often be assuaged or ■ Individual biological and psychological dispositions, attitudes, values, knowledge, skills, problem enhanced. behaviors ■ Peer norms, activities, bonding Domains ■ Family function, management, bonding Risk and protective factors exist at every level at ■ School/work bonding, climate, policy, performance which an individual interacts with others and the ■ Community bonding, norms, resources, society around him or her. Clearly, the individual awareness/mobilization brings a set of qualities or characteristics to each ■ Society/ norms, policy/sanctions interaction, and these factors act as a filter, color- environmental ing the nature and tone of these interactions— Research has also revealed that domains are not whether positive or negative. One useful way to static in their impact, but interact with each other look at this interplay is to organize interactions by and change over time. As an individual develops, the six life or activity domains in which they his or her perceptions and interactions with family, chiefly occur. Based on more than 30 years of peers, schools, work, and community alter (Botvin study, researchers have delineated specific subcat- et al., 1995; Donaldson, Graham, & Hansen, egories of risk within each domain. They include 1994; Hawkins et al., 1992; Kumpfer, Molgaard, & the following: Spoth, 1996). CSAP graphically depicts this more intricate set of relationships through its Web of Figure 1 Web of Influence Teen School Substance Teenage Violent Pregnancy Dropout Use Suicide Crime Individual Risk and Protective Factors Society/ Family Community** Risk Peer Association Environment* School Risk and Risk and Protective and Protective Risk and Protective Related Risk and Protective Factors Factors Factors Factors Protective Factors *Society/Environment: Refers to the total complex of external social, cultural, and economic conditions affecting a community or an individual. **Community: Refers to the specific geographic location where an individual resides and to the conditions within that particular area. A Guide 3
Guide to Science-Based Practices vide a conceptual and organizational scheme for Institute of Medicine Prevention Categories identifying risk groups and targeting outcomes. Universal programs (e.g., mass media, school- based health curricula): Target the general popula- Issues in Defining Scientifically tion. Defensible Knowledge Selective programs (e.g., mentoring programs Scientific inquiry stems from the need to under- aimed at children with school performance or stand the world at large. The strength of science behavioral problems): Target those at higher-than- and the scientific method is that it makes use of average risk for substance abuse. strictly defined, standardized procedures to deter- Indicated programs (e.g., parenting programs for mine how events are causally related. As science parents with substance abuse problems): Target those improves its methods, levels of certainty about the already using or engaged in other high-risk behaviors nature and extent of cause-and-effect relationships (such as delinquency) to prevent chronic use. increase and more is understood about the resources and effort required to achieve specific changes in existing relationships. Using the scien- Influence model (figure 1). The Web model illus- tific method more systematically to identify knowl- trates the complex series of interactions that edge also fosters recognition of the diversity of occurs between the individual and the six external approaches involved in implementing prevention domains that can result in substance use and other programs and extracting data. problem behaviors. Different Ways of Knowing Institute of Medicine Prevention Classifications A key CSAP objective is to ensure that the inter- Just as the Web of Influence can be used to illus- ventions it recommends to the field are out- trate relationships among risk and protective fac- growths of rigorous evaluation and are described tors across domains, the Institute of Medicine’s in credible terms, comprehensible to the multiple (IOM) prevention program classification system is target audiences involved in substance abuse pre- useful in understanding the differing objectives of vention. Emerging out of the cultural, geographic, various interventions and matching them to the and strategic diversity of high-risk population pro- needs of a targeted population (Kumpfer et al., gramming is a recognition of the different ways in 1997). which people know or understand that a program The IOM system classifies prevention interven- is having an impact on youth, families, and com- tions according to the population they affect munities. (Gordon, 1987). Universal interventions target Like good medicine, the practice of prevention general population groups without reference to is both an art and a science. In assessing preven- those at particular risk. All members of a commu- tion programs as a whole and attempting to under- nity benefit from a universal prevention effort, not stand whether the strategies and interventions just specific individuals or groups within a com- used are exerting an impact and how they affect munity. Selective interventions target those who human behavior, it is critical to consider both are at greater-than-average risk for substance use. quantitative and qualitative evidence. Quantitative Targeted individuals are identified on the basis of data supply the raw material for the extensive sta- the nature and number of risk factors for substance tistical analyses that lend scientific credence to use to which they may be exposed. Indicated program results. Qualitative data provide the rich, interventions are aimed at individuals who may descriptive information needed to explain the already display signs of substance use or abuse effects of program interventions. and are designed to prevent the onset of regular or heavy substance use. Together, the Web of Influence and the IOM classification system pro- 4 Science-Based Substance Abuse Prevention
Data Types and Research Strategies ine qualitative data from prevention practice cases Although much discussion of knowledge focuses in their efforts to assess interventions such as on the results of quantitative outcome evaluations, tobacco prevention/cessation and family-oriented qualitative information can also be extremely use- prevention. ful even if it is not always amenable to strict out- come evaluation. Qualitative data may describe Data Collection Techniques program process or identify contextual mediating Figure 2 identifies the numerous data collection variables that affect outcome results. Such process techniques that are used to gain knowledge in the information embellishes findings from programs, substance abuse field. The techniques are mapped providing an enhanced understanding of program onto a pyramid to provide some idea of the pro- results. When researchers and the field, in general, portion of total information available through par- ignore qualitative data, valuable information can ticular techniques that have been sufficiently be lost. implemented to yield credible findings. Reviews of qualitative information can produce Paradoxically, as indicated in the figure, the credible findings and recommendations. For more traditionally accepted and more quantitative example, expert consensus panels convened by scientific approaches represent a small proportion many government agencies (e.g., Center for of data collection efforts, yet the information Substance Abuse Treatment, Food and Drug derived from such studies constitutes a significant Administration, National Institute on Alcohol portion of the formal knowledge base. Although Abuse and Alcoholism) and private organizations the tendency is to ascribe more validity to quanti- review and use both qualitative and quantitative tative efforts, CSAP recognizes that important con- data to reach conclusions and formulate recom- tributions to the prevention knowledge base are mendations affecting the health and well-being of also made by qualitative studies. Independent of the nation as a whole. CSAP has developed how quantitative or qualitative they are, when Prevention Enhancement Protocol Systems (PEPS), findings are supported by sound scientific princi- which convene expert consensus panels to identi- ples, they can be used legitimately and effectively fy what is known about various dimensions of to construct and implement prevention interven- substance use prevention. PEPS panels also exam- tions. Figure 2 Application of Specific Principles and Criteria to Research Findings Replications and Multisite Studies Single Site Experimental and Quasi-experimental Studies Process Evaluations Focus Groups Expert Panels Key Informant Activities Participant and Program Staff Observations A Guide 5
Guide to Science-Based Practices Application of Specific Principles and Criteria 7 and 8 took the form of summary eval- uations. Criterion 7 captured how well the total Criteria to Research Findings design and execution ruled out alternative Research studies and the findings they produce hypotheses, while criterion 8 reflected overall vary in credibility. The level of observed credibili - confidence in results inspired by methodological ty of research findings regarding the causes of aspects of study design and execution. Criterion 9 intervention program effects hinges on whether addressed the utility of study results. Considered the methods employed provide at least a reason- together, the nine criteria provide one scheme for able means of assessing change over time attribut- evaluating all aspects of a quantitative or qualita- able only to the program. This simple criterion has tive intervention. a number of components that can be applied to assess the credibility of research. The development of the High Risk Populations (HRP) Databank, a previous CSAP endeavor that critically reviewed Program Review Criteria the HRP Demonstration Grants, provides an exam- 1. Theory: the degree to which findings are ple. grounded in sound theory, reflect clearly The HRP Findings Bank was an evaluation- stated hypotheses, and are operationally oriented information system with a comprehen- relevant sive, unifying framework. The Findings Bank 2. Sampling strategy and implementation: the consisted of three primary components: quality of sampling design and implemen- ■ Descriptive information (e.g., location, number tation and strength of evidence concerning and type of sites, target group demographics) sample quality (e.g., data on attrition) about CSAP’s high-risk populations grants; 3. Measures: operational relevance, psycho- ■ Compilation of interventions and outcomes pro- metric quality of measures used in the eval- duced by the grants; and uation, and quality of supporting evidence ■ Compilation of interventions and outcomes pro- 4. Data collection: quality of data collection duced by similar, non-CSAP projects. implementation (e.g., amount of missing data) The Findings Bank allowed users to relate findings from demonstration programs to outcome mea- 5. Fidelity of interventions: evidence of high- sures and to the data that support the findings. fidelity implementation of program, as designed, and sufficiency of dosage (e.g., Nine criteria were used by groups of trained, duration, intensity, frequency) to effect pos- expert evaluators to assess the rigor of grantee itive change program evaluations. Six criteria were used to 6. Analysis: appropriateness and adequacy of rate various aspects of methodological rigor (1 statistical techniques used in analysis through 4, 6, and 7), one was used to rate fidelity (5), one for utility (9), and one for overall integrity 7. Plausible threats to validity: degree to which evaluation design and implementa- and credibility (8). tion address and eliminate reasonable alter- Viewed in an alternative fashion, criteria 1 native hypotheses about program effects through 3 rated the design or planning of the and warrant strong causal attributions study. They encompassed theoretical bases, sam- 8. Integrity: overall level of confidence in ple design, and the outcome measures chosen. project findings based on the research These three criteria pertained to the planning of design and implementation both quantitative and qualitative research. 9. Utility: strength of findings and strength of Criteria 4 through 6 related to execution of the evaluation to determine if findings were study; in other words, how it was implemented. consistent with respect to expectations or Criterion 5 was a hybrid, addressing both fidelity predictions from theory of program/intervention implementation and the design issue of dosage of program/intervention. 6 Science-Based Substance Abuse Prevention
When a few problems were encountered—for Confidence Scale example, attrition was modest, intervention imple- mentation was solid, and analysis was accept- 5 = strong confidence able—an integrity rating of 4 (i.e., confident) was 4 = confident given. 3 = some confidence An integrity rating of 3 reflected some confi- 2 = little confidence dence in resultant data. This rating was often used when program characteristics were strong enough 1 = no confidence to inspire some confidence, but because of imper- fect implementation of the program, moderate attrition rates, data analyses that were not compre- Application of Criteria hensive, uncorrected differences between treat- For the review of CSAP’s HRP demonstrations, ment and comparison samples, or secular events pairs of trained evaluators rated each of the nine that contaminated samples, confidence was not criteria on a 5-point scale. Of particular impor- sufficient to assure reviewers that the results were tance was the rating of integrity, on which review- wholly attributable to the program intervention. ers were required to come to consensus. This If reviewers had little confidence in an experi- rating reflected how much confidence reviewers mental study in which there was high and differ- had that the intervention alone was responsible for ential attrition, the integrity rating received by yielding the findings observed. Confidence was such a study was 2—little confidence. If attrition derived from the quality of the intervention imple- was egregiously pronounced, the study rating may mentation as well as the design of the evaluation have dropped to 1—no confidence. study and how well the evaluation was conduct- Because well-designed studies can be imple- ed. For example, when problems in intervention mented poorly, well-implemented programs can implementation and research design and execu- be evaluated poorly, and findings can be overstat- tion were minimal, reviewers assigned integrity ed, it is important to use criteria capable of expos- ratings reflecting strong confidence in the findings ing these problems. Although overstated findings (i.e., 5). make no positive contribution to the knowledge Figure 3 Sample Application of Scientific Criteria to an Intervention Summary Matrix for Grant ZZZ NA 1 2 3 4 5 1. Theory X 2. Sampling Strategy and X Implementation 3. Measures X 4. Data Collection X 5. Fidelity of Intervention X 6. Analysis X 7. Plausible Threats to Validity X 8. Integrity X 9. Utility X A Guide 7
Guide to Science-Based Practices base, from the perspective of the prevention pro- mixed confidence, this intervention was considered gram designer, policymaker, and funder, negative promising and could be cited as corroborative evi- findings from well-implemented, rigorously evalu- dence of the effectiveness of grant ZZZ in altering ated interventions have enormous value because youth’s knowledge about the harmful effects of they identify interventions that do not appear to substance use. work in naturalistic settings. When objectively In addition to applying the criteria to the HRP reported, they save programs from expending demonstration grants, which include both the High resources in the application of ineffective Risk Youth (HRY) programs and programs for approaches. Pregnant and Postpartum Women and their Infants, The last rating of utility captured data patterns. CSAP is using them in a slightly modified form to Ideally, the field would want to adopt or adapt extract findings from the Community Partnership interventions that yield consistent changes across Program. Likewise, CSAP’s National Registry of domains affected by the intervention. For exam- Effective Prevention Programs (NREPP)—an ongo- ple, program X could demonstrate a dramatic ing repository of scientifically defensible guidance change on a specific targeted behavior such as for the substance abuse field—has incorporated the marijuana use. In response, prevention practition- criteria in the 15 dimensions used in its consensus ers in a community where marijuana use was process when assessing the quality of a particular increasing might rush to adopt program X without prevention program. These 15 criteria are an considering changes across other outcome expansion of the original nine criteria described on domains such as skyrocketing underage use of page 6 of this guide. For a full review of these cri- alcohol (drug substitution) or plummeting self- teria, please refer to the NREPP Web site at efficacy. Reviewers used the utility rating to gauge www.preventionregistry.org. NREPP has evolved the pattern of outcomes and thereby avoid focus- from the HRP Findings Bank and is available to ing on isolated positive or negative outcomes. review programs funded from any source, not only from CSAP. Hypothetical Review Illustrating Application of Criteria The CSAP Approach To Identifying Figure 3 depicts a hypothetical review. In the fig- Scientifically Defensible Prevention ure, program ZZZ has an overall rating of 3. The Interventions program was rated well on many of the criteria; The CSAP HRP Databank review used a qualita- however, reviewers did not think that program tive meta-analytic technique, one of two types of control and treatment groups were comparable meta-analytic techniques currently favored to prior to the intervention, hence the evaluation of organize information and extract defensible princi- 2 for sampling strategy and implementation. ples and practices. In the past, literature reviews Furthermore, the statistical analyses testing differ- were frequently analyzed to determine whether ences between the groups after the intervention substance abuse prevention programs worked. did not attempt to control for what might have Meta-analysis offers a major refinement on that been meaningful pretest differences. For that rea- approach. son, analyses received a 2. In addition, the age In a literature review, researchers scrutinize and group differences observed on outcome measures critique original papers, determine the merit of were large and not predicted by theory. However, specific items, and then integrate the findings in this intervention, along with a number of others, discussions with each other or alone. showed differences in youth’s knowledge of the Occasionally this thinking results in the proposi- harmful effects of substance use as a result of par- tion of a critical experiment that, if performed, ticipating in classroom-based drug education, an would shed light on the true phenomenon under outcome that, given the measurement protocols, scrutiny. More often, the result is a simple summa- could not be attributed to any event or occurrence ry of the field, with the authors’ conclusions sup- other than the intervention. Therefore, despite 8 Science-Based Substance Abuse Prevention
ported by the amount and consistency of the data tive meta-analytic techniques focus on specific assembled. effects, they are extremely useful in identifying Literature reviews make important contributions consistently moderate or large outcomes, and then to understanding the state of the art in the field in relating these observations to specific program and help identify gaps in the knowledge base that characteristics. Quantitative techniques are also should be filled. However, because literature useful in deriving prevention principles and identi- reviews are, for the most part, essentially narrative fying specific prevention interventions that are most and subjective, use few samples, and lack scientif- closely related to consistently favorable outcomes. ic rigor, they cannot respond adequately to ques- Like any other procedure, meta-analysis is not tions about the efficacy of substance abuse without potential hazards (e.g., relying solely on prevention programs (Tobler, 1994, p. 343). journal publications, averaging equally over differ- entially important measures). Recent quantitative Meta-Analysis meta-analytic efforts, including Tobler’s and To address concerns about program effects, CSAP Stratton’s (1997) analysis of the effectiveness of turned to meta-analysis, a conceptual approach school-based substance use prevention programs that reviews the results of the full range of primary and CSAP’s National Center for the Advancement research, including both published and unpub- of Prevention meta-analyses of the Correlates of lished, experimental and quasi-experimental stud- Marijuana Use and the Correlates of Alcohol and ies of programs that succeeded and programs that Tobacco Use, have advanced the knowledge base failed. Meta-analysis uses qualitative as well as considerably. quantitative methods to produce aggregated Qualitative meta-analysis is based on programs results from multiple programs that can be used to as a whole. As an example, in the CSAP HRP study relationships and test hypotheses (Cook et Databank review, trained, expert reviewers used al., 1992; Light & Pillemer, 1984; Tobler, 1994, standard instruments to evaluate individual pro- p. 357). Because meta-analysis allows researchers grams on the basis of source documents in the to use results from small studies, from studies that form of project final reports, as well as articles are quasi-experimental in design, and from studies published in journals, and come to consensus with incomplete information in some areas, it about their credibility. Inclusion of source materi- offers important advantages when evaluating sub- als overcame biases that may have been inherent stance abuse prevention programs (Tobler, 1994, in using journal articles only. p. 350). In applying meta-analysis, CSAP makes a Qualitative techniques are not limited to main distinction between the two types. While both effects models of causation but are robust across forms rely on qualitative judgments as well as interactions and nonlinear models since judg- quantifying information, they differ in the extent to ments of effectiveness are based on the integration which they rely on translating initially observed of implementation, evaluation, and findings within outcomes into hard estimates of effect sizes. the context of the program. While ratings may be Quantitative meta-analysis systematically codes attached to program efforts (e.g., integrity, utility), the results from each study for every variable that they are composites weighted by the judgment of may influence program outcomes and converts expert evaluators and not the product of decon- those results to a standardized score or effect size structed program characteristics averaged across (Tobler, 1994, p. 345). The end product of the multiple programs or interventions. Systematizing analysis consists of a single best quantified esti- procedures and categorizing outcomes across pro- mate of effect for a specific intervention or the grams via qualitative meta-analytic technique not impact of a single contextual factor. Standardi - only offers decided advantages to traditional litera- zation enables meta-analysis to compare programs ture review efforts, but also ensures that both qual- with different sample sizes, and if effect sizes are itative and quantitative information is scrutinized consistent and positive, to aggregate them and so that researchers, policymakers, and program achieve statistical significance. Because quantita- A Guide 9
Guide to Science-Based Practices designers have access to needed outcome and Qualitative Meta-Analyses process data. CSAP culled several well-implemented critical reviews to identify credible scientifically defensi- Distinctions Between Quantitative and ble principles and programs, including the follow- Qualitative Techniques ing CSAP-sponsored efforts: Like quantitative meta-analytic techniques, quali- ■ Environmental Strategies for Substance Abuse tative procedures use rules for inclusion of infor- Prevention: Analysis of the Effectiveness of mation. In general, these rules mirror the nine Policies to Reduce Alcohol, Tobacco, and Illicit criteria described previously in reference to the Drug Problems; HRP Databank. No matter who conducts the ■ The Role of Education in Substance Abuse review, research protocols are screened for quality Prevention (Implementation Guide); of design and implementation. What may differ is ■ The Role of Information Dissemination and Mass the set of procedures, level of data disaggregation, population of studies eligible for inclusion, and Media in Youth-Oriented Prevention rigor of the review itself. (Implementation Guide); While both quantitative and qualitative tech- ■ The Role of Problem Identification and Referral niques have important roles to play, their purposes in Youth-Oriented Prevention (Implementation are different. The more quantitative techniques are Guide); and useful in identifying interventions and contextual ■ A Review of Alternative Activities and Alternative factors that influence effectiveness, whereas quali- Programs in Youth-Oriented Prevention (CSAP tative techniques are most useful in identifying Technical Report 13, 1996). effective programs and models of intervention. As a result, qualitative techniques can be used to identi- Expert Panels fy both successful program models and prevention In addition to information derived from critical lit- principles. Both quantitative and qualitative tech- erature reviews, CSAP also incorporated informa- niques can be supported by field observation and tion derived from expert consensus procedures. As careful review of complete program documenta- part of the consensus process, consensus panel tion (e.g., process analyses). Also, both are valu- members assessed implementation for fidelity and able in developing and deriving principles reviewed research for rigor. Outcomes and review- regarding successful program implementation. er confidence in those outcomes were rated using relatively objective, standard methods. Expert con- Resources for Identifying Scientifically sensus resources included the following: Defensible Prevention Principles and ■ National Structured Evaluation (NSE): The NSE Programs was one of the first expert consensus reviews of CSAP uses both qualitative and quantitative meta- prevention programs. Using standard procedures analyses, as well as findings from critical and measures, the NSE reviewed and rated the overviews of published research and expert con- level of methodological rigor of substance use sensus procedures, to identify scientifically defensi- prevention programs implemented through ble prevention principles and programs. Currently, 1991. The NSE effort included federally funded those principles and programs are described in two programs as well as programs sponsored by uni- CSAP publications, Principles of Substance Abuse versities, foundations, and state or local govern- Prevention and Promising and Proven Substance ments. When the analysis was complete, 10 Abuse Prevention Programs. Following are descrip- percent of the studies reviewed met or exceeded tions of the specific methods CSAP employed to the criteria set for moderately rigorous studies identify the scientifically defensible principles and (i.e., a rating of 3). programs in both documents. ■ William B. Hansen & L. A. Rose (1997): In his work, Dr. Hansen continued to use techniques 10 Science-Based Substance Abuse Prevention
employed by the NSE to identify and review sub- tion of program effectiveness. Like the NSE, stance abuse prevention program evaluations approximately 10 percent of program evalua- and field studies that reflected effective program- tions met the criteria set for moderate rigor. Of ming. Data from individual projects were used to those, an estimated 2 percent of studies reviewed derive more general principles concerning pro- met the more stringent standards that CSAP gram effectiveness. developed for identifying well-implemented, ■ National Institute on Drug Abuse (NIDA, 1997): solidly evaluated, effective model programs. NIDA produced a review document titled Drug Abuse Prevention for At-Risk Individuals that cit- Quantitative Meta-Analyses ed NIDA-sponsored substance abuse prevention The most stringent coding and rating protocols can programs viewed as effective and offered general be followed when there are a wealth of studies with recommendations about key elements of effec- concordant data. As the discipline of substance tive interventions. abuse prevention continues to evolve, quantitative ■ Lawrence W. Sherman et al. (1997): Dr. Sherman efforts like these will become more commonplace. and colleagues conducted a methodologically CSAP used data from the following quantitative rigorous review and rating of U.S. Department of meta-analyses to identify current scientifically Justice programs modeled after the NSE defensible principles and programs, as follows: approach. The authors also evaluated grant and ■ Tobler’s meta-analyses of school-based preven- funding mechanisms for adequacy. Principles tion programs (1986, 1992). pertaining to effective strategies were derived ■ Tobler & Stratton, 1997. and presented as key findings. ■ CSAP meta-analysis: CSAP initiated a series of ■ PEPS: Over the past several years, CSAP has large-scale meta-analyses, each focusing on a sponsored the PEPS, a knowledge development specific topic area relevant to substance use and effort that convenes a panel of experts to review substance use prevention (e.g., Correlates of the literature in a specific area using a strict evi- Marijuana Use Among Youth, Correlates of dentiary procedure to evaluate and either Alcohol and Tobacco Use Among Youth and include or exclude individual research findings. Meta-Analysis of the Effectiveness of School- Findings are then assessed for consistency of Based Programs). valence and magnitude to arrive at a defensible CSAP accepts data from these qualitative and conclusion. quantitative sources as credible, relying on the ■ National Evaluation of the Community review processes implemented and the expertise Partnership Prevention Grant Program (2000): of the review panels convened to extract impor- Extensive cross-site and other evaluations of the tant research findings. With these data as building CSAP Community Partnership Program have blocks, it is possible to generalize across agencies yielded considerable information about commu- and the time periods in which the data were col- nitywide prevention efforts and have identified lected to identify replicable model programs and specific prevention strategies that produce mea- scientifically defensible principles and interven- surable reductions in substance use. tions. ■ The CSAP HRP Databank review: This review of CSAP has developed a series of comprehensive the HRP demonstration grants program was data matrices (Promising and Proven Substance designed to identify credible evidence of pro- Abuse Prevention Programs) that organizes inter- gram effectiveness. An expert consensus process ventions by risk and protective factors, age, strate- evaluated each final report filed through gy, IOM classification, original program name, and December 1995 on the criteria defining program source citation. This inclusive effort includes pro- credibility. In addition, they rated utility of study grams supported by government agencies, includ- information for each program and weighed the ing CSAP, National Institute on Drug Abuse consistency of results in determining the evalua- (NIDA), National Institute on Alcohol Abuse and A Guide 11
Guide to Science-Based Practices Alcoholism, the U.S. Department of Education, References and the U.S. Department of Justice, that have undergone strenuous reviews (such as the CSAP Anthony, E. J., & Cohler, B. J. (1987). The invulner- model programs) or have been cited by experts in able child. New York: Guilford Press. the field as examples of particular intervention strategies. Interventions entered into the matrix Bandura, A. Social Learning Theory. Englewood vary in the levels of rigor assigned by reviewers. Cliffs, NJ: Prentice Hall, 1977. For example, the CSAP model programs incorpo- rated in the matrix represent the highest level of Botvin, G. J., Baker, E., Dusenbury, L. D., Botvin, rigor and credibility. Interventions from other CSAP E. M., & Diaz, T. (1995). Long-term followup programs, such as the Community Partnerships, results of a randomized drug abuse prevention which may have more moderate levels of credibili- trial in a white middle-class population. Journal ty, are included because careful evaluation attests of the American Medical Association, 273(14), to their ability to produce positive outcomes. 1106–1112. Interventions from other federal agency programs are included because their sponsors have identified Bry, B., & Krinsley, K. (1990). Adolescent substance them as noteworthy through their own evaluative abuse. In E. Feindler & G. Kalfus (Eds.), processes. The matrix is a work in progress that Adolescent behavior therapy handbook. New will continue to expand as findings from other York: Springer. promising interventions become available. As the information in the data matrix increases, the inter- Center for Substance Abuse Prevention. (2000). ventions cited will provide a broader range of Prevention works through community partner- interventions and more specific guidance for ships: Findings from SAMHSA/CSAP’s national selecting those most likely to prove effective with evaluation. Rockville, MD: CSAP. particular populations and in particular domains. Cook, T., Cooper, H., Cordray, D., Hartmann, H., Also available is a second, simpler compilation Hedges, L., Light, R., Louis, T., & Mostellar, F., of scientifically defensible principles called (Eds.) (1992). Meta-analysis for explanation: Principles of Substance Abuse Prevention. As its A casebook. New York: Russell Sage Foundation. name implies, this brief booklet organizes princi- ples and interventions by domain and links them to Donaldson, S. I., Graham, J. W., & Hansen, W.B. the prevention strategies identified in the Substance (1994). Testing the generalizability of interven- Abuse Prevention and Treatment block grant. ing mechanism theories: Understanding the Service providers can refer to Principles for ideas effects of a school-based prevention program about what works in substance use prevention, to for potential high school dropouts and drug identify proven interventions in a particular abusers. International Journal of Addictions domain, and to justify the use of one or more inter- 25(7), 773–801. ventions in a program. Because each principle and intervention is cited, it is relatively easy to locate Gordon, R. (1987). An operational classification the full article for more detailed information. of disease prevention. In J. A. Steinberg & M. M. Evaluators, grant reviewers, and policymakers may Silverman (Eds.), Preventing mental disorders. also find it useful as a quick overview of the cur- Rockville, MD: U.S. Department of Health and rent state of the art in prevention programming and Human Services. as a tool useful in ascertaining whether a program is employing scientifically defensible principles Hansen, W.B., and McNeal, R.B. The law of maxi- and interventions. mum expected potential effect: Constraints placed on program effectiveness by mediator relationships. Health Education Research 11(4):501-507, 1996. 12 Science-Based Substance Abuse Prevention
Hansen, W. B., & Rose, L. A. (1997). Issues in Pandina, R.J. Risk and protective factor models in classification in meta-analysis in substance abuse adolescent drug use: Putting them to work for prevention research. In W. J. Bukoski (Ed.), Meta- prevention. In S.L.David, (Ed.), National Analysis of Drug Abuse Prevention Programs. Conference on Drug Abuse Prevention Research Monograph No. 170. Rockville, MD: Research: Presentations, papers, and recommen- National Institute on Drug Abuse, 183–201. dations. National Institute on Drug Abuse, Rockville, MD: 17-26, 1998. Hawkins, J. D., Catalano, R. F., & Miller, J. Y. (1992). Risk and protective factors for alcohol and other Sherman, L. W., Gottfredson, D. C., MacKenzie, drug problems in adolescence and early adult- D. L., Eck, J., Reuter, P., & Bushway, S. W. hood: Implications for substance abuse preven- (1997). Preventing crime: What works, what tion. Psychological Bulletin 112(1), 64–105. doesn’t, what’s promising. Research in Brief, Washington, DC: National Institute of Justice Kumpfer, K. L., & Baxley, G. B. (1997). Drug (NCJ 171676). abuse prevention: What works? Rockville, MD: National Institute on Drug Abuse, National Tobler, N. (1986). Meta-analysis of 143 adolescent Institutes of Health. drug prevention programs: Quantitative out- comes results of program participants compared Kumpfer, K. L., Molgaard, V., & Spoth, R. (1996). to a control or comparison group. Journal of The Strengthening Families Program for the Drug Issues 16(4):537–567. prevention of delinquency and drug use. In R. Peters & R. McMahon, (Eds.), Preventing Tobler, N. (1992). Drug prevention programs can childhood problems, substance abuse, and work: Research findings. Journal of Addictive delinquency. Thousand Oaks, CA: Sage Disorders 11(3):1–28. Publications, 241–267. Tobler, N. (1994). Meta-analytical issues for pre- Light, R., & Pillemer, D. (1984). Summing up: vention intervention research. In L. Seitz & L. The science of reviewing research. Cambridge, Collins (Eds.), Advances in data analysis for pre- MA: Harvard. vention intervention research. Research Monograph No. 142. Rockville, MD: National Mrazek, P. J., & Haggerty, R. J. (Eds.) (1994). Institute on Drug Abuse, 342–403. Reducing the Risk for Mental Disorders: Frontiers for Preventive Intervention Research. Tobler, N. S. (1997). Meta-analysis of adolescent Washington, DC: National Academy. drug prevention programs: Results of the 1993 meta-analysis In W. J. Bukoski (Ed.), Meta- National Institute on Drug Abuse. (1997). Drug Analysis of drug abuse prevention programs. abuse prevention for at-risk individuals (NIH Research Monograph No. 170. Rockville, MD: Publication No. 97-4115). Rockville, MD: National Institute on Drug Abuse, 5–68. The Author. Tobler, N. S., & Stratton, H. H. (1997). Newcomb, M.D., & Felix-Ortiz, M. (1992). Multiple Effectiveness of school-based drug prevention protective and risk factors for drug use and abuse: programs: A meta-analysis of the research. Cross-sectional and prospective findings. Journal Journal of Primary Prevention 18(1), 71–128. of Pers Soc Psychology 63(2), 280–296. Wolin, S., & Wolin, S. (1995). Resilience among youth growing up in substance abusing families. Pediatric Clinics of North America 42(2), 415–429. A Guide 13
Glossary Antisocial and other problem behaviors: Can describe behavior-related problems (e.g., poor conduct and impulsive- ness), behavior-related disorders (e.g., attention deficit hyperactivity disorder), or both. Approach: A set of prevention strategies that typifies a program and can be used in an intervention setting without adopting the entire program. Assignment: The process by which researchers place study subjects in an intervention, control, or comparison group. Experimental design studies randomly assign study subjects to both intervention and control conditions. In quasi- experimental studies, study subjects are nonrandomly assigned to intervention and comparison conditions. Random assignment increases the likelihood that the intervention and control groups are equal or comparable and have simi- lar characteristics. Attrition: An unplanned reduction in the size of a study sample caused by participants dropping out of the evaluation (e.g., they moved away from the study location). Behavior-related disorder: A specific behavioral problem that occurs in persistent patterns and characteristic clusters and causes clinically significant impairment. Behavior-related problem: A behavioral problem that is isolated or intermittent, is not part of a persistent behavior pattern, and varies in severity and seriousness of its consequences. Community: A group of individuals who share cultural and social experiences within a common geographic or politi- cal jurisdiction. Community-based approach: A prevention approach that focuses on the problems or needs of an entire community, be it a large city, small town, school, worksite, or public place. Community readiness: The degree of support for or resistance to identifying substance use and abuse as significant social problems in a community. Stages of community readiness for prevention provide an appropriate framework for understanding prevention readiness at the community and state levels. Community tolerance: Community norms that view problematic behavior as socially acceptable or actively encour- age it. Conduct disorder: A behavior-related disorder that has a repetitive and persistent pattern of violating the basic rights of others or major age-appropriate societal norms or rules. The disorder can include aggression to people and ani- mals, destruction of property, deceitfulness or theft, and serious violation of rules. Construct: An attribute, usually unobservable (e.g. educational attainment or socioeconomic status), that is represent- ed by an observable measure. Control group: In experimental evaluation design, a group of participants that is essentially similar to the intervention group but is not exposed to the intervention. Participants are designated to be part of either a control or an interven- tion group through random assignment. Credibility of findings: Derives from the quality of intervention implementation plus the methodological rigor of the research. When both are high, findings are attributable to the intervention and therefore have high credibility. Data: Information collected according to a methodology using specific research methods and instruments. Data analysis: The process of examining systematically collected information. Design: An outline or plan of the procedures to be followed in scientific experimentation and research studies to reach valid conclusions. A Guide 15
You can also read