Does Prevention Work (and is this even a fair question)?
Michael P. Levine, Ph.D., FAED
If you want to truly understand something, try to change it.
– Kurt Lewin (quoted in APA Policy and Planning Board, 2007)
The prevention of eating disorders (EDs) has been a controversial topic since efforts began approximately 30 years ago. Prevention has an increasing number of staunch supporters throughout the United States, as well as in Australia, Austria, Canada, Great Britain, Iceland, Mexico, Norway, and Spain. After all, it is easy to demonstrate that (a) prevention is the only way to reduce the number of new cases (i.e., the incidence of EDs); and (b) clinical personnel and financial resources are both far too limited to enable the “detect it-treat it” model to ever make anything but a small dent in the prevalence of EDs, that is, the number of people suffering from an ED at any given time (Levine & Smolak, 2006). Moreover, as the opening quotation proclaims, prevention of outcome Y (e.g., bulimia nervosa) by eliminating or modifying precursor X (e.g., body dissatisfaction) is an integral step in research demonstrating that X is a causal risk factor for Y.
Nevertheless, there continue to be a number of influential researchers, clinicians, and activists who are skeptical about the possibility of effective prevention. These stakeholders in the field are concerned that prevention programming and research will siphon precious financial and other resources from basic efforts to understand and treat eating disorders. In addition, over the years several parents have challenged me by insisting that well-intentioned but misguided information aimed at prevention actually triggered their child’s ED.
Thus, the question “Does Prevention Work?” is an important one to address. This essay, the first of two in a series, offers a detailed analysis of that question, a reasonable answer, and the implications of both. The second part will provide specific examples of programs and outcome research that highlight promising future directions for the prevention of EDs.
Key Terms and Concepts
To prevent a disorder or disease is to keep it from occurring, or at least to delay its development or minimize its severity. The first step is using theoretical models and research to clarify, understand, and thus anticipate the conditions that lay the foundation (i.e., the vulnerability or diathesis) for EDs and that trigger their onset (i.e., the stressors or precipitators). The second step involves using theoretical models of change in attitudes and behaviors, along with principles of sound evaluative research, to develop systematic interventions. Thus, one important aspect of prevention is clarifying specific risk factors and then taking specific steps to minimize or eliminate them (Stice, Rohde, & Shaw, 2013). However, prevention also involves the encouraging of more generalized mechanisms for protecting health and for reinforcing resilience and adaptive functioning. This includes the ability to cope effectively with both expected developmental challenges and unpredictable but inevitable hardships (Committee on the Prevention of Mental Disorders, 2009; Levine & Smolak, 2006).
The Prevention Spectrum
There are different forms of prevention, depending upon the size and nature of the group for whom programming is intended. The “mental health intervention spectrum,” developed by the USA’s National Academy of Sciences (Committee on the Prevention of Mental Disorders, 2009), ranges from general health and resilience promotion à universal prevention à selective prevention à indicated prevention. The latter, sometimes referred to as “targeted” prevention, shades into the traditional steps of clinical treatment: case identification à intervention à aftercare.
Universal programs are designed to transform and strengthen public institutions and policies, and as well as normative cultural attitudes and practices. The goal is to prevent EDs from developing in specific but extremely large groups comprising people with varying degrees of risk, such as all youth ages 11 through 14 in California. Selective prevention also seeks to change the developmental contexts created by the ecologies of children, youth, and young adults. However, in contrast to universal prevention programs, the intended audience is a large group of people who are selected for participation based on two criteria: (1) they do not yet have the symptoms of an eating disorder; and (2) biological, psychological, or socio-cultural factors put them at risk. For example, a selective prevention program for EDs could focus on girls ages 10 through 14 who are coping with pubertal development, who live in a society that defines female beauty in terms of slenderness, and who have a parent or older sibling who suffers from one or more of the following: severe depression; substance abuse; and an eating disorder. The prototypical multi-lesson, classroom-based curriculum designed to prevent EDs in middle or high school girls would fall between universal and selective prevention on the spectrum (Levine & Smolak, 2006).
Indicated/targeted prevention programs are designed for people who are at high risk due to the documented presence of warning signs (e.g., mild ED symptoms) and/or clear precursors (e.g., high levels of weight/shape concern). Their “definitely at-risk” status “indicates” the need for an individual or group intervention tailored to the variables that have put them at high risk. Compared to universal and selective prevention, indicated prevention aims to target and change the issues of specific individuals, rather than changing social policies, social systems, and interpersonal behavior.
Ideally, of course, a prevention program will “work”: it will have a strong and meaningful effect in the desired direction—and it will do no harm. But how do researchers demonstrate—and how can policy makers know—that a program has the desired outcomes?
Together, a number of sources (Becker, Stice, Shaw, & Woda, 2009; Levine & Smolak, 2006; Piran, McVey, & Levine, 2014; Society for Prevention Research, 2004) indicate that we can reasonably be confident that a prevention program “really works” when each of six criteria are met:
- The program has a cogent rationale based on a valid model of environmental and attitudinal change.
- The program has been implemented as planned with samples that permit adequate generalization to the population of interest.
- Use of valid measures and of well-controlled research designs demonstrates that participants without an ED at baseline show a low(er) rate of the onset (again, a lower incidence) of EDs over a fairly long time frame than a meaningful sample of the population that is comparable in terms of risk level(s) and that does not receive the program.
- The reduced incidence of EDs in the group receiving the program is mediated by decreases in the risk factors—and/or increases in the protective factors—emphasized by the theoretical model guiding program development. That is, careful assessment demonstrates that the desired prevention outcome is probably attributable to the social and psychological processes activated by the program’s influential components.
- Criteria 2-4 have been met in two or more studies. This suggests that the predicted effects of the prevention program, when faithfully implemented, are replicable and robust.
- The program can be effective in a wide variety of real-world settings, and it can be disseminated broadly because it is not too costly or complicated to purchase the program and to train people to use it. In this regard, note that in thinking about the process of establishing whether a program works, researchers often distinguish between a program’s efficacy under ideal and highly controlled conditions, and its effectiveness under less ideal, less well-controlled, and more varied real world conditions.
Do No Harm
Although not listed as a criterion, it is assumed that a prevention program that works will not be harmful. The principle of “first, do no harm” (primum non nocere), or of at least taking all possible steps to minimize the risk of harm, is a fundamental aspect of many ethical systems in disciplines such as medicine, public health, and psychology. Systematic reviews (e.g., Stice, Shaw, & Marti, 2007) of a large number of ED prevention studies indicate that it is unlikely that carefully designed programs will backfire. Nevertheless, unintended negative consequences of prevention programs have been documented (Levine & Smolak, 2002; O’Dea, 2002). Consequently, preventive interventions should be carefully and cautiously designed with input from multiple constituencies, such as institutional stakeholders, experts, and members of the intended audience. A program that works is an ethical program that can be carefully monitored throughout the process of implementation and evaluation.
Strict Application of the Criteria
If we insist on applying all six criteria, then there are two answers to the question “Does ED prevention work?” The first is “At least one program does.” This program, developed by Eric Stice and colleagues at the Oregon Research Institute and elaborated by Carolyn Becker and colleagues at Trinity University in San Antonio, uses a critical social perspective and the social psychological theory known as Cognitive Dissonance (CD; see, e.g., Draycott & Dabbs, 1998) to reduce internalization of the slender beauty. Extensive research demonstrates that the CD program works for high-risk adolescent girls in high school and for young women in college, across a wide range of risk levels (Becker, MacKenzie, & Stewart, in press; Becker et al., 2007; Stice, Marti, Rohde, & Shaw, 2011; Stice et al., 2013).The CD program will be described in more detail in the second essay in this series.
The second answer is “Not enough research has been done to establish whether the most promising programs do indeed meet all six criteria.” For a typical example, Australian researchers Simon Wilksch and Tracey Wade (2009) have developed, implemented, and carefully evaluated an interactive 8-lesson program called Media Smart. Integrating media literacy, activism, and advocacy, Media Smart is a universal-selective program intended for a mixed-sex audience of students age 13 and 14. A large-scale randomized controlled study revealed that Media Smart produced significant reductions in shape and weight concern and in dieting at 30-month follow up for girls and at 6-months follow up for boys. Although the Media Smart researchers have met criteria 1, 2, 4, and 6, the research has not yet been done to determine if the very important criteria of replication and prevention of eating disorders can be met. This program will also be described in more detail in Part 2 of this series.
Application of Less Strict Criteria
It is reasonable to loosen the positive outcome criteria in order to determine whether a prevention program has enough promise to warrant evaluation of whether it meets all six criteria. One reasonable way to do this while retaining an emphasis on good research design and long-term positive outcomes is to insist on application of Criteria 1 and 2, and a blend of Criteria 3 and 4. Specifically, it is reasonable to argue that a prevention program has the potential to work if controlled outcome research with a reasonable follow-up period (e.g., a minimum of 6 months after completion of the program) shows that the implementation of the program, as it is intended, produces a reduction in risk factors.
There are many detailed analyses of the current state of this more limited (i.e., less well-developed) type of prevention research (see, e.g., Bailey et al., 2014; Levine, McVey, & Piran, 2014; Levine & Smolak, 2006, 2008; Piran, McVey, & Levine, 2014; Stice, Becker, & Yokum, 2013). One important type of review uses the statistical technique called meta-analysis (see, e.g., http://www.lyonsmorris.com/MetaA/abstract.htm). This involves calculating the average comparative statistical effect, across studies, of the prevention program versus the control. The effect contributed by each study is weighted by sample size. Meta-analysis also allows examination of the degree to which variability in effect sizes across a group of studies is correlated with variables of interest, such as age of the participants or whether the program is basically interactive or didactic.
According to meta-analyses by Stice et al. (2007) and by Fingeret, Warren, Cepeda-Benito, and Gleaves (2006), both universal-selective and indicated programs tend to have beneficial effects on measures of risk (e.g., internalization of the thin ideal, body dissatisfaction, and negative mood) and on reductions in current eating pathology. The average effect sizes tend to be statistically significant (i.e., very likely greater than zero) but small for universal-selective prevention, while they tend to be moderate but not large for indicated prevention programs. Overall, indicated prevention programs for female participants ages 15 through 25 who already have body image issues and other weight concerns tend to be the most effective, especially those multi-session, interactive programs that challenge continued internalization of the slender beauty ideal. It should be noted that, although universal-selective programs are less effective than indicated programs overall and at decreasing dieting at follow-up, both types of programs tend to be equally effective in reducing thin-ideal internalization and negative affect (Stice et al., 2007).
Conclusions and Future Directions
No disease or disorder has ever been eliminated or even significantly reduced by detecting and treating individual instances once they have arisen. Prevention is a necessity if we are ever to minimize the number of cases of eating disorders and thereby minimize, if not eliminate, untold amounts of suffering for individuals and their families and friends.
Is it a fair question to ask: “Does prevention work?” The answer is “yes, it certainly is a fair question” – but only if sufficient attention is given to the complexity of the question, while sufficient resources are being devoted to the opportunity for researchers to have a fair shot at meeting the 6 criteria for evaluating whether any particular prevention program works. Assuming that it can be a fair question, the answer to “Does it work?” is “yes, it can work, and more program development and outcome research are needed to clarify the conditions under which prevention works in a meaningful way.” In the second part of this series I will explore a number of exciting developments in the prevention of eating disorders and consider the lessons we can extract for the components of effective programs.
It is more sensible, humane, pragmatic, and cost-effective to build psychological health and prevent maladjustment than to struggle valiantly and compassionately to stay its awesome tide.
– Emory Cowen (1983, p. 14).
We enthusiastically look forward to the second of Dr. Levine’s two-part series which will be posted in the April 2015 Gurze/Salucore enewsletter, and can be found by clicking here.
About the Author:
Michael P. Levine, Ph.D., is Emeritus Professor of Psychology at Kenyon College. In the field of eating disorders his research, writing, and activism focuses on sociocultural risk factors and prevention. He has authored, co-authored, or co-edited five books, three curriculum guides, and 100 articles and book chapters.
APA Policy and Planning Board (2007). Who cares about APA policy and does it have an impact? American Psychologist, 62, 491-503.
Bailey, A. P., Parker, A. G., Colautti, L. A., Hart, L. M., Liu, P., & Hetrick, S. E. (2014). Mapping the evidence for the prevention and treatment of eating disorders in young people. Journal of Eating Disorders, 2(5). Retrieved from http://www.jeatdisord.com/content/2/1/5
Becker, C., MacKenzie, K., & Stewart, T. (in press). Cognitive and behavioral approaches to the prevention of eating disorders. In L. Smolak & M. P. Levine (Eds.), The Wiley-Blackwell handbook of eating disorders (Vol. 2): Assessment,
prevention, treatment, policy, and future directions. Sussex, UK: Wiley-
Becker, C. B., Stice, E., Shaw, H., & Woda, S. (2009). Use of empirically supported interventions for psychopathology: Can the participatory approach move us beyond the research-to-practice gap? Behaviour Research and Therapy 47, 265-274. Committee on the Prevention of Mental Disorders and Substance Abuse Among Children, Youth, and Young Adults [National Research Council & Institute of Medicine of the National Academies]. (2009). Preventing mental, emotional, and behavioral disorders among young people: Progress and possibilities. Washington, DC: The National Academies Press.
Cowen, E. L. (1983). Primary prevention in mental health: Past, present, and future. In R. D. Felner, L. A. Jason, Moritsugu, J. N., & Farber, S. S. (Eds.), Preventive psychology: Theory, research, and practice (pp. 11-25). New York: Pergamon.
Draycott, S., & Dabbs, A. (1998). Cognitive dissonance 1: An overview of the literature and its integration into theory and practice of clinical psychology. British Journal of Clinical Psychology, 37, 341-353.
Fingeret, M. C., Warren, C. S., Cepeda-Benito, A., & Gleaves, D. H. (2006). Eating disorder prevention research: A meta-analysis. Eating Disorders: The Journalof Treatment & Prevention, 14, 191–213.
Levine, M. P., McVey, G. L., & Piran, N. (2014). Eating disorders during adulthood. In T. P. Gullotta & M. Bloom (Eds.), Encyclopedia of primary prevention and health promotion (2nd ed., Vol. 4, pp. 1670-1686). New York: Springer.
Levine, M. P., & Smolak, L. (2006). The prevention of eating problems and eating disorders: Theory, research, and practice. Mahwah, NJ: Lawrence Erlbaum
Levine, M. P., & Smolak, L. (2008). “What exactly are we waiting for?” The case for universal-selective eating disorders prevention programs. International Journal of Child & Adolescent Health, 1, 295-304.
O’Dea, J. (2002). Can body image education programs be harmful to adolescent females? EatingDisorders: The Journal of Treatment & Prevention, 10, 1-13.
Piran, N., McVey, G. L., & Levine M. P. (2014). Eating disorders in adolescence. In T. P. Gullotta & M. Bloom (Eds.), Encyclopedia of primary prevention and health promotion (2nd ed., Vol. 3, pp. 1134-1147). New York: Springer.
Society for Prevention Research. (2004). Standards of evidence: Criteria for efficacy, effectiveness and dissemination. Retrieved April 6, 2009, from http://www.preventionresearch.org/StandardsofEvidencebook.pdf
Stice, E., Becker, C. B., & Yokum, S. (2013). Eating disorder prevention: Current evidence-base and future directions. International Journal of Eating Disorders, 46, 478-485.
Stice, E., Marti, C. N., Rohde, P., & Shaw, H. (2011). Testing mediators hypothesized to account for the effects of a dissonance-based eating disorder prevention program over longer term follow-up. Journal of Consulting and Clinical Psychology, 79, 398-405.
Stice, E., Rohde, P., & Shaw, H. (2013). The Body Project: A dissonance-based eating disorder prevention intervention (Updated ed.). Facilitator guide. NY: Oxford University Press.
Stice, E., Shaw, H., & Marti, C. N. (2007). A meta-analytic review of eating disorder prevention programs: Encouraging findings. Annual Review of Clinical Psychology, 3, 207-231.
Wilksch, S. M., & Wade, T. D. (2009). Reduction of shape and weight concern in young adolescents: A 30-month controlled evaluation of a media literacy program. Journal of the American Academy of Child and Adolescent Psychiatry, 48, 652-661.