As a 2000 report by the Institute of Medicine (IOM) observed, health interventions based on social and behavioral research have great potential to improve health and prevent disease. Unfortunately, to date, such intervention efforts have had a mixed record of success, with some interventions making significant improvements in the health of the target group and others having little or no effect. In this chapter, we discuss some of the major challenges facing such interventions, describe some of the progress that has been made in the 13 years since the IOM report was released and offer a number of suggestions for how researchers might improve their chances of success in future interventions.
THE 2000 IOM REPORT
In 1998, one of us (Syme) was asked to chair a committee at the Institute of Medicine to evaluate the success of interventions aimed at improving health. The IOM committee was asked to focus specifically on interventions that were based on findings from social and behavioral research. The 10-person committee spent two years evaluating intervention accomplishments. Members reviewed the literature and interviewed dozens of experts. The committee also commissioned 12 groups of scholars to prepare detailed papers that summarized intervention work in their area of expertise, and, toward the end of the committee’s work, 33 distinguished researchers and practitioners were invited to assess these 12 papers in the presence of a large audience in Atlanta, Georgia. Finally, the committee published a 508-page
report presenting the papers and discussing the committee’s findings and recommendations (Institute of Medicine, 2000).
The report concluded that there was a great deal of evidence indicating that social and behavioral interventions could lead to significant improvements in health, stating that “Behavioral and social science research has provided many new advancements in the effort to improve population health, and offers promise for the development of new interventions with even greater utility and efficiency in the years to come” (Institute of Medicine, 2000, p. 33). In addition to making a general argument that social and behavioral interventions are promising ways to improve health, the report made a number of specific recommendations concerning promising intervention strategies. It suggested, for example, that interventions should be carried out to address the health of potential mothers before they get pregnant and that efforts should be made to increase the social capital of communities and neighborhoods in order to improve the effectiveness of behavioral change interventions. The report also made a number of research recommendations for studies that could support and strengthen intervention trials. These recommendations included carrying out studies aimed at identifying the pathways through which social contexts affect disease pathogenesis and outcomes and performing cost-effectiveness analyses that could identify those interventions with the greatest potential to improve health at the least expense.
The general tenor of the report was very positive, because it emphasized the potential of intervention trials based on behavioral and social research and called for continuing work in the area. Less emphasis was placed on the fact that, up to that point in time, most interventions of this sort had either failed or were of only modest significance. There had been a few exceptions, such as the success in smoking cessation and in the use of seat belts and safety helmets, with these achievements attained by a combination of efforts on the individual, organizational, and legislative levels. Overall, however, the results had been disappointing.
The bottom line is that, as promising as intervention trials based on social and behavioral research may be, a number of challenges face anyone who would carry out such trials. In the rest of this chapter, we discuss these challenges, look at progress that has been made since the 2000 report in dealing with these challenges, and make some suggestions for future directions in dealing with such challenges.
CHALLENGES TO INTERVENTION TRIALS
During the committee’s work in preparing the 2000 report, it became clear that researchers who seek to carry out health intervention trials based
on social and behavioral research face a number of difficult challenges. In this section, we discuss some of the most commonly encountered ones.
Challenges to Randomized Controlled Trials
We begin by describing the difficulties associated with one of the most expensive and ambitious randomized and controlled clinical trials ever conducted: the Multiple Risk Factor Intervention Trial, or MRFIT (Multiple Risk Factor Intervention Trial Research Group, 1982). The challenges experienced by the researchers running this trial were similar to those experienced in many other clinical trials and thus serve as a case study in helping scientists think about the best approaches to conducting intervention trials in the public health field.
The trial was initiated in the 1970s at a time when it had been clearly established that the three most important risk factors for coronary heart disease were high levels of serum cholesterol, high blood pressure, and cigarette smoking. People who had all three of these risk factors had a six-fold increased risk of developing heart disease. Additionally, all three of these risk factors were clearly amenable to intervention.
The study team decided to focus on men who met two primary criteria: First, they had risk levels that were in the top 10 percent in the nation and, second, they were free of coronary heart disease at the time they enrolled in the study. Study power calculations indicated that 12,000 such men would need to be enrolled in the trial, half to the experimental intervention delivered in the MRFIT clinic and half to work with their own physicians as part of their usual care (i.e., the control arm). The decision was made to enroll only men in the study because including women would have approximately doubled the sample size.
To find these 12,000 men, almost a half million men were screened in 22 cities across the country. Following the initial study screening process, those still eligible were asked to attend two additional intensive screenings before they could be formally enrolled in the trial. Along the way, it was made clear to them that they should not participate if they had any reservations about the trial. They were told that if they were eligible they would be randomly assigned to work with an intervention team in the clinic or with their own physicians. They were warned that if they worked with the intervention team in the clinic, they would have to come to the clinic frequently at the beginning, sometimes with their family, and would be asked to change their diet, take medication to control their blood pressure, and stop smoking. They were also told that the trial would go on for six years. There was a behavioral team in every clinic to eliminate men thought likely to be poor participants over the long haul.
The study was carried out with great care and attention to detail. The clinics had large, well-trained staffs. Study participants were invited to come to the clinics with their families to observe demonstrations of low-fat cooking. The staff went to markets with them to show them how to read labels and visited them in their homes to show them how to use the foods that were already there. But after six years, when members of the study team were invited to Bethesda to learn the results of their work, they were told that there was no statistically significant difference between the “special care” and “usual care” groups.
The apparent reason became clear later when it was learned that a number of the men sent back to their own physicians became energized by being assigned to the usual-care arm. In interviews carried out after the study, a number of the men said, in essence, “First, I go through an exhaustive series of screening tests. Then you tell me I am in the top 10 percent risk group in the nation for developing coronary heart disease. Then you tell me to go back to my own doctor, which I could have done on my own, without having to go through all of those difficult screening exams. I’ll do it myself!” Others said that they had been invigorated by the substantial attention linked with the large amount of screening and assessment that they received throughout the course of the trial. In fact, subsequent trials have documented the positive effects that usual- or standard-care controls involving brief physician advice in combination with extensive screening and assessment activities can exert on health behaviors and study outcomes (Writing Group for Activity Counseling Trial Research Group, 2001).
To be sure, the MRFIT trial, despite the ambiguous nature of its original results, contributed decades’ worth of important scientific information that has advanced the understanding of cardiovascular risk factors and treatments. This classic study also helped to set the stage for improvements in randomized controlled trial (RCT) methods and design that have culminated in highly successful multisite trials such as the Diabetes Prevention Program. In that trial, lifestyle intervention aimed primarily at weight loss and increases in regular physical activity in a large study sample at high risk for diabetes led to a 58 percent lower diabetes incidence in this group relative to controls and a 39 percent lower diabetes incidence in the lifestyle group relative to the group taking metformin (Knowler et al., 2002).
But MRFIT also serves as a cautionary tale about the potential limitations of RCTs for advancing knowledge in the areas of treatment and prevention. One of the lessons from the MRFIT story is that the randomized controlled trial, while traditionally serving as the gold standard for answering a number of scientific questions related to intervention efficacy, may not be the best method for answering all intervention-relevant scientific questions, particularly as they pertain to intervention effectiveness in realworld conditions. Thus, while the RCT offers clear advantages in helping to
reduce confounding and other threats to internal validity, it also has specific limitations that need to be carefully considered.
One such limitation, for instance, is the problem of subject selection biases (e.g., the fact that samples typically are restricted to generally motivated individuals who are often healthier than other community members). While selecting a reasonably homogeneous group of research subjects can reduce the complexity of a trial, it also constrains study generalizability. For example, consider a trial aimed at evaluating the health consequences of an intervention to improve social support networks. The complexity of the intervention can be reduced by including only a specific segment of individuals who could potentially benefit from such an intervention, as opposed to broader representation from the larger community. Similarly, the often long list of eligibility criteria accompanying many RCTs often reduces sample heterogeneity, and, with it, external validity. The resulting study sample is often no longer representative of the populations targeted.
Furthermore, once a decision is made regarding the kinds of people who are to be invited to a trial, only some people will agree to join, further limiting representativeness. In addition, not all who agree to participate in the study remain in it to the end, and not all follow the study’s intervention advice.
There are other problems with the RCT as well. The nature of the intervention is usually specified with the hope that those who received the intervention can be compared with those who did not. Notably, however, the effects of an intervention can extend beyond what the investigators originally conceptualized or intended. In the Multiple Risk Factor Intervention Trial, the focus was on blood pressure, fat in the diet, and cigarette smoking. It turned out, however, that those who attempted to change their risk engaged in many other behaviors as well. Some became more physically active, some took yoga lessons, some went to church more often, some spent more time with their children, and so on. While scientists may focus on a defined range of “risk factors” being targeted by an intervention, study participants often focus on broader “life factors.”
General Challenges to Intervention Trials
In addition to the challenges specific to RCTs, there are a variety of challenges that face intervention trials in general, including both RCTs and other types of trials.
Choice of Outcomes
Many failures to successfully intervene may be due to the fact that the intervention is aimed at the wrong outcomes. This problem is illustrated
by a study that one of us (Syme) conducted with San Francisco bus drivers (Ragland et al., 1987). This study was initiated in response to the observation that this sample of bus drivers had a very high rate of hypertension. By the time drivers were 60 years old, the prevalence of hypertension among them was 90 percent. A research grant from the National Heart, Lung, and Blood Institute allowed us to conduct an epidemiologic study of hypertension risk factors that led to the development of an intervention project to reduce the rate of hypertension among the drivers.
Toward the end of the project, it was noticed that the drivers also had a very high rate of low back pain. We subsequently received a grant to study that problem as well and began to develop an intervention strategy to address it. We then learned that the drivers had high rates of respiratory difficulties, gastrointestinal problems, and alcohol issues, which led us, finally, to begin to rethink our project. While the research team was focusing on those issues that had come to our clinical attention, the real problem was the job itself. The job was creating a constellation of health problems and, thus, the attributes and parameters of the job itself should have been our focus. Without attending to this fundamental determining force, our interventions would have remained focused on treating individual clinical outcomes for this workforce without getting to the root cause of the many health issues associated with that occupation. Thus, it is important in intervention development to consider not only the type of risk factors being targeted but also the outcomes.
Selection of Risk Factors
Another challenge is the selection of risk factors that affect susceptibility. Although we, as experts, are good at selecting factors that we think are significant, it is important to realize that these priorities do not always coincide with the priorities of the people who we intend to help. This problem is illustrated by the experience of one of us (Syme) in a study of smoking cessation. In this study, the purpose was to reduce smoking rates in a community with a very high prevalence of smoking. In particular, the objective was to change the climate of opinion about smoking in the community, including challenging the social acceptability of smoking. In Richmond, California, the rate of smoking was more than 50 percent when the project started. It was a well-designed project that was later used as the basis for a nationwide series of smoking cessation interventions in 20 cities across the country (COMMIT Research Group, 1995).
After five years of rigorous work, however, the project failed to show any differences in smoking rates between Richmond and the two comparison communities, Oakland and San Francisco. The COMMIT study resulted in generally disappointing outcomes as well. Without going into
all the details of the Richmond failure, it is clear that the research team came to this community with the attitude that a 50 percent prevalence of smoking was unacceptably high. But Richmond is a very poor city with very high rates of crime, unemployment, high school dropouts, and drug use, as well as having few health services and few food markets. It later became clear that the smoking rate in Richmond was not high on the priority list of citizens of this community. The research team had never checked with residents about this, however, and in interviews with Richmond residents after the failure of the smoking cessation project, it became clear that residents were more concerned with crime, drugs, schooling, jobs, money, and safety than smoking. These and other examples have made clear that it is critical to take account of the social circumstances in which people live. An exclusive focus on risk factors alone may be inappropriate.
Choosing the Right Time Frame: Recognition of Life Course Events
Another issue that can compromise the ability to intervene successfully involves the recognition of life course effects. Berkman (2009) described two clinical trials that evaluated interventions directed at social support and depression among older adults. Both of the trials were well designed and carefully implemented, but neither showed significant results relative to controls.
One of these trials was the Enhancing Recovery in Coronary Heart Disease Patients (ENRICHD) trial (Berkman et al., 2003). It was designed to evaluate whether enhancing social support would lower the rate of depression among post-myocardial patients. The hope was that a lower rate of depression would reduce re-infarction and all-cause mortality among 2,481 cardiac patients. This ambitious trial involved 80 hospitals and 8 clinical centers.
The second clinical trial was called FIRST, for the Families in Recovery from Stroke Trial (Glass et al., 2004). It was designed to improve social networks with the hope that these networks would in turn increase the functional independence of stroke patients following their strokes. The study involved 291 patients recruited from 8 hospitals.
Neither study produced the desired results. Berkman has suggested that such intervention studies may need to be more mindful of the importance of life course issues and trajectories when targeting populations on which to intervene. If one intervenes on a problem such as depression or a way of coping that originated and stabilized early in life, or if the problem is the result of a cumulative exposure over the life course, it may make less sense to intervene on that problem much later in life. Berkman suggests being more mindful of these issues in planning interventions (Berkman, 2009).
A final problem to be considered concerns study reproducibility. All interventions rely on information that has been derived from prior research, and if that information is flawed, then it can affect subsequent investigations. Scientific inquiry has long emphasized the importance of determining the reproducibility of methods and results. In research, scientists are urged to use standard questions that have been used by others and to publish their methods so that others can repeat the work. We suggest, however, that this emphasis may have certain limitations. The goal of complete reproducibility may be achievable in basic laboratory research, but it is often unattainable in human research. Repeating a research project in an identical sample of people is virtually never possible. For instance, using a standardized questionnaire in two different groups of people drawn from the same population ignores the subtle differences of context that can make reproducibility difficult.
An alternative to strict reproducibility may be preferable. This alternative could involve maximizing differences to better test the generalizability of an intervention across different populations and circumstances. Continuing to observe the same patterns in all of this work indicates a finding of real importance, as seen in studies of social class and health. Social class has been measured in many different ways in different population groups, but the same result emerges in virtually all of these studies—those from reduced social circumstances have higher rates of virtually all diseases. The same result is found in studies of social support. A recent meta-analysis of 148 social support studies involving 308,000 people showed that people with better social support networks—defined and measured in many different ways—had better health related to a wide range of different diseases over a seven-year follow-up period (Holt-Lundstad, Smith, and Layton, 2010). As was noted earlier, this was true even after accounting for age, gender, and health at baseline. These social class and social support results are robust and transcend differences in research methods and populations. For that reason, they may be considered all the more important and useful.
WHAT HAS BEEN LEARNED SINCE THE 2000 IOM REPORT
Since the publication of the 2000 IOM report, basic understanding of interventions to change health behaviors has greatly improved. This growing knowledge, along with advances in intervention methods and design and an increasing understanding of the importance of contextual and environmental factors in influencing daily health decisions, has opened the way to more potent and sustainable interventions. Below we describe a few of
the areas in which significant improvements have been made in methods or understanding that should lead to improvements in intervention trials.
Appreciation of the Heterogeneity of Health Targets as Well as Populations
While a great deal of scientific attention has been focused on the demographic heterogeneity of the sample being enrolled in an RCT, disease and health targets themselves are often treated as reasonably uniform. This is often not the case, however. For example, health targets such as obesity typically represent a constellation of conditions with varying etiological, biobehavioral, and contextual features that need to be addressed when developing an intervention.
A growing appreciation of this fact has led some researchers to call for a targeted intervention approach aimed at meeting the needs of a particular population segment with common health features and circumstances (e.g., King et al., 2008). This targeting approach contrasts with the “one-size-fits-all” approach that has been pervasive in much of the intervention literature. The growing use of the targeted intervention approach has been accompanied by the development of statistical methods aimed at better understanding which subgroups of participants may have fared better or worse with a particular intervention (Kraemer et al., 2002). For example, as part of the ENRICHD trial described above, Schneiderman et al. (2004) found that while intervening in depression and low perceived social support within 28 days after myocardial infarction did not increase event-free survival rate, it seemed that white men—although not other subgroups—may have benefited from the ENRICHD intervention. Such exploratory findings are informative and suggest additional discovery work to further refine and improve interventions for additional population groups.
Advances in RCT Research Methodology
As researchers have come to understand the limitations of traditional RCT methods, particularly in the areas of external validity and translational efficiency, it has resulted in the development of various ways to increase the real-world applicability of RCT research. One such approach—the practical clinical trial—aims to extend design parameters to real-world contexts and to promote greater generalizability in the areas of intervention development, implementation, evaluation, and translation (e.g., through contextually relevant interventions, diverse populations and settings, and a broad range of outcomes). By using a structured framework to evaluate the impacts of the intervention in areas related directly to external validity
(e.g., the RE-AIM model), it is possible to achieve an improved translation of research into practice (Dzewaltowski et al., 2004).
Other innovative methods for enhancing the flexibility and real-world relevance of clinical trials research include adaptive designs that use a stepped-care approach to intervention delivery based on participant response and adaptive interventions that use prespecified decision rules based on tailoring variables to adjust intervention dose and related parameters (Collins et al., 2005). There have also been efforts to determine the most relevant control groups to use in clinical trials research depending upon a study’s aims, objectives, and resources (Mohr et al., 2009). Finally, given that changes are often required in multiple risk factors and behaviors that together influence disease endpoints, behavioral scientists and other researchers have begun to investigate the best methods for making multiple behavior changes (Prochaska, Spring, and Nigg, 2008). These efforts stand in contrast to the majority of the clinical trials that can be found in the literature, in which risk factor interventions have been combined in various ways with little regard to theory or evidence.
Employing “Stealth” Interventions That Harness People’s Intrinsic Life Values
Health represents only one area in people’s busy lives, and, in many cases, it may not be the most important or salient issue on a day-to-day basis. Therefore, it is important to explore other ways to engage people’s passions and motives, where health may not play an explicit role but may be benefited indirectly. Such motives can include family, faith, culture, recreation, leaving a reduced carbon footprint, or other pursuits. By identifying which values or motives might resonate most strongly with a population of interest, promoting health through harnessing these other motives becomes an attractive option. Published examples of such “stealth interventions” have included utilizing interests in social movements (e.g., environmental sustainability) to indirectly motivate dietary improvements (Hekler, Gardner, and Robinson, 2010) and providing meaningful public school volunteer opportunities for older adults that have been shown to improve their own physical health and cognitive and social engagement in addition to benefiting the school children (Fried et al., 2004).
SUGGESTIONS FOR FUTURE IMPROVEMENTS
Although various advances have been made since the release of the 2000 IOM report in the understanding of the importance of contextual and environmental factors as well as in intervention methods and design, much
remains to be done. In this section we offer a variety of suggestions for how intervention trials might be improved in the coming years.
Importance of Aiming for the Underlying Drivers of Health
One key fact about intervention trials is that they rarely attempt to intervene in the fundamental driving forces in society—both social and environmental—that are responsible for many of the health problems in the first place. These upstream social and environmental factors that influence health and disease include such factors as poverty, lack of education, and neighborhood incivilities that can negatively affect daily health-related decisions and behaviors regardless of how committed an individual or community may be to living a healthier life. Interventions aimed at making a discernible difference in promoting health and preventing or controlling disease need to reflect these multiple, interconnected levels of influence, including biological, behavioral, sociocultural, environmental, and institutional and policy levels (Institute of Medicine, 2002). Such ecological frameworks typically require more complex interventions than have often been evaluated in the RCT literature. However, given the importance of the “web of causation” described in ecological models for determining health outcomes, multilevel interventions aimed at these differing impacts deserve substantially greater attention.
Using a Two-Step Model Framework
Consider two puzzling facts rarely addressed. One is that the risk factors that are intended to explain the occurrence of specific diseases actually explain a relatively small proportion of the diseases they are intended to explain. The second is that many of the psychosocial risk factors of interest are related to a wide range of disease outcomes. For example, people in lower social class positions have higher rates of not just one or two diseases, but of virtually every disease studied. The same phenomenon exists for social support: In many studies, people with poor social connections have higher mortality rates than people with better social connections, and these observations hold after controlling for age, gender, and health status. This generalized disease finding is true for a number of other psychosocial risk factors as well.
Thus, on the one hand, a set of disease-specific risk factors do not fully explain the diseases they are intended to explain, and, on the other hand, a set of psychosocial risk factors are inexplicably related to virtually all diseases. Perhaps this puzzle can be explained by looking at disease causation in a different way, using a two-step model framework. The argument would be that psychosocial risk factors are related to host susceptibility.
The effect of harmful psychosocial risk factors would be to compromise the body’s immune system and increase individual vulnerability to disease, but they would not predict which diseases we get: These psychosocial risk factors would not necessarily increase the risk of any one disease but would make us vulnerable to disease more generally. The specific disease that one would contract would be attributable to the particular disease agents that one was exposed to—viruses, bacteria, tobacco smoke, high fat diets, air pollution, stress, and so on.
Colleagues at the University of California, San Francisco, recently presented some remarkable early data that are supportive of this way of thinking. In their two separate studies of 113 and 96 people of varying racial and ethnic backgrounds and from several social class groups, it was found that several important psychosocial variables were related to immune functioning in a very significant way (John-Henderson et al., 2011). Specifically, it was found that interleukin-6 responses varied in exactly the ways that would be expected with selected psychosocial factors. Interleukin-6 plays an important role in the immune functioning of the body and is related to a number of diseases, including coronary heart disease, depression, diabetes, prostate cancer, and rheumatoid arthritis. These and related results can further the understanding of such two-step models of disease etiology.
Those who study the epidemiology of infectious diseases have for many years approached their research in this way. These scholars know that disease agents must be considered in relation to both host susceptibility and environmental circumstances. Many researchers interested in psychosocial factors have not taken this into account in a satisfactory way. The implication of this work is the need to think about an additional set of risk factors—risk factors that affect people’s vulnerability and susceptibility to disease in general.
Better Defining the Needs of Different Population Groups Through Community-Based Participatory Research Methods
In recent years, there has been an increasing appreciation among public health researchers of the critical perspectives that community members can bring to the intervention research endeavor (Minkler and Wallerstein, 2008). Community-based participatory (CBPR) methods represent one such approach for doing so (Horowitz, Robinson, and Seifer, 2009). CBPR activities consider community stakeholders as active partners throughout the research decision-making and implementation process, and researchers who use a CBPR approach employ such techniques as team building, mutual exchanges of ideas, the sharing of resources, and joint decision making involving both community members and researchers. The approach requires that the interests of the researchers and the use of rigorous designs and
methods be balanced against the needs of the community and the larger aim of developing contextually valid and sustainable interventions (Horowitz, Robinson, and Seifer, 2009). Although such an approach may involve a greater upfront effort than more traditional approaches, by involving communities in the intervention development process from the beginning it is possible to avoid some of the problems discussed above (e.g., in the COMMIT study) and to improve the chances that the intervention will have positive effects on health.
Intervening Where People Live, Work, and Play
With the current explosion of electronic communication channels and devices, it has become easier to reach diverse groups of people in ways that can be personalized to meet their individual needs, preferences, and contexts. For example, eHealth advances provide previously unimagined opportunities to positively affect daily health behaviors and choices that can broadly influence the health of the population (King and Guralnik, 2010). They also offer innovative ways of capturing contextually rich information throughout an individual’s day that can help inform intervention development as well as evaluation. Systematic investigations of the efficacy of some of these technologies have demonstrated that they offer a viable alternative to human advisors for promoting sustained behavior change in health areas such as physical activity (King et al., 2007).
Concerns have also been expressed, however, about the potential impact of such technologies in increasing already significant health disparities in the United States by expanding the “digital divide.” This concern is fueled in part by socioeconomic and cultural differences related to computer access and literacy as well as to more general health literacy. Greater attention in the eHealth and mobile health fields needs to be paid to the types of communication technologies that can be tailored to a broad range of language, reading, and related user abilities and access issues. An example of one such technology is the embodied conversational agent, which requires minimal levels of reading and computer proficiency and health literacy to use (Bickmore et al., 2010) and could potentially be placed in community settings that residents frequently visit.
Putting the Public in the Driver’s Seat in Areas of Data Collection to Create Meaningful Community Change Related to Health
Some of the technology-based innovations noted above provide a unique opportunity to “deputize” community members to collect and share the types of local barriers to and enablers of a healthy lifestyle that are potentially amenable to change on an organizational or policy level. This
type of crowd-sourcing activity is currently being used in rudimentary ways by various municipalities and universities across the country to highlight neighborhood problems, such as litter or unsafe streets, that can be addressed by local decision makers. Such activities could also be aimed specifically at neighborhood barriers to healthy living and eating. For example, a recent pilot study found that a simple neighborhood audit tool, residing on an electronic tablet and adapted to the lower literacy skills and physical limitations of an older population, could be effectively used to gather photo, audio, and location data relating to the walkability of the local neighborhood environment that could, in turn, be shared with municipal decision makers (Buman et al., in press). Such applications can potentially empower legions of community residents to become actively engaged in contributing to the evidence base in ways that could directly impact program and policy decisions relevant to health in their communities.
Although there is great promise in health intervention trials based on social and behavioral research, researchers who seek to carry out such trials face a number of challenges, including choosing the most appropriate risk factors and outcomes and determining how to deal with a lower level of reproducibility than many researchers are accustomed to. Progress has been and is being made in the area, but much remains to be done. Researchers carrying out such intervention trials in the future should keep in mind the importance of aiming for the underlying drivers of health, of using community-based participatory research methods, of applying new digital communications technologies, and of including the public in data-gathering efforts. Researchers should also consider using a two-step model framework that interprets psychosocial risk factors as increasing a general susceptibility to poor health and disease.
Berkman, L.F. (2009). Social epidemiology: Social determinants of health in the United States: Are we losing ground? Annual Review of Public Health, 30, 27-41.
Berkman, L.F., Blumenthal, J., Burg, M., Carney, R.M., Catellier, D., Cowan, M.J., Czajkowski, S.M., DeBusk, R., Hosking, J., Jaffe, A., Kaufmann, P.G., Mitchell, P., Norman, J., Powell, L.H., Raczynski, J.M., and Schneiderman, N. (2003). Effects of treating depression and low perceived social support on clinical events after myocardial infarction: The enhancing recovery in coronary heart disease patients (ENRICHD) randomized trial. Journal of the American Medical Association, 289(23), 3,106-3,116.
Bickmore, T., Pfeifer, L.M., Byron, D., Forsythe, S., Henault, L.E., Jack, B.W., Silliman, R., and Paasche-Orlow, M.K. (2010). Usability of conversational agents by patients with inadequate health literacy: Evidence from two clinical trials. Journal of Health Communication, 15(Supplement 2), 197-210.
Buman, M.P., Winter, S.J., Baker, C., Otten, J., Hekler, E.B., and King, A.C. (in press). Policy to practice and back again: Engaging older adults in policy activities. Submitted to Translational Behavioral Medicine.
Collins, L.M., Murphy, S.A., Nair, V.N., and Strecher, V.J. (2005). A strategy for optimizing and evaluating behavioral interventions. Annals of Behavioral Medicine, 30, 65-73.
COMMIT Research Group. (1995). Community intervention trial for smoking cessation (COMMIT). I. Cohort results from a four-year community intervention. American Journal of Public Health, 85(2), 183-192.
Dzewaltowski, D.A., Glasgow, R.E., Klesge, L.M., Estabrooks, P.A., and Brock, E. (2004). RE-AIM: Evidence-based standards and a web resource to improve translation of research into practice. Annals of Behavioral Medicine, 28, 75-80.
Fried, L.P., Carlson, M.C., Freedman, M., Frick, K.D., Glass, T.A., Hill, J., McGill, S., Rebok, G.W., Seeman, T., Tielsch, J., Wasik, B.A., and Zeger, S. (2004). A social model for health promotion for an aging population: Initial evidence on the experience corps model. Journal of Urban Health, 81(1), 64-78.
Glass, T.A., Berkman, L.F., Hiltunen, E.F., Furie, K., Glymour, M.M., Fay, M.E., and Ware, J. (2004). The Families in Recovery from Stroke Trial (FIRST): Primary study results. Psychosomatic Medicine, 66(6), 889-897.
Hekler, E.B., Gardner, C.D., and Robinson, T.N. (2010). Effects of a college course about food and society on students’ eating behaviors. American Journal of Preventive Medicine, 38(5), 543-547.
Holt-Lundstad, T., Smith, T.B., and Layton, J.B. (2010). Social relationships and mortality risk: A meta-analytic review. PLoS Medicine, 7(7). Available: http://www.plosmedicine.org/article/info%3Adoi%2F10.1371%2Fjournal.pmed.1000316 [September 13, 2012].
Horowitz, C.R., Robinson, M., and Seifer, S. (2009). Community-based participatory research from the margin to the mainstream: Are researchers prepared? Circulation, 119(19), 2,633-2,642.
Institute of Medicine. (2000). Promoting Health: Intervention Strategies from Social and Behavioral Research. B.D. Smedley and S.L. Syme (Eds.). Washington, DC: National Academy Press.
Institute of Medicine. (2002). The Future of the Public’s Health in the 21st Century. Washington, DC: National Academy Press.
John-Henderson, N., Jacobs, E., Mendoza-Denton, R., and Francis, D.D. (2011). The Relationship Between Feeling Poor, Implicit Social Class Bias, and Health Outcomes. Paper presented at the annual meeting of the Society of Personality and Social Psychology, January, San Antonio, TX.
King, A.C., and Guralnik, J.M. (2010). Maximizing the potential of an aging population through shaping an active and vital old age. Journal of the American Medical Association, 304, 1,944-1,945.
King, A.C., Friedman, R.M., Marcus, B.H., Castro, C., Napolitano, M., Ahn, D., and Baker, L. (2007). Ongoing physical activity advice by humans versus computers: The Community Health Advice by Telephone (CHAT) trial. Health Psychology, 26, 718-727.
King, A.C., Ahn, D.F., Atienza, A.A., and Kraemer, H.C. (2008). Exploring refinements in targeted behavioral medicine intervention to advance public health. Annals of Behavioral Medicine, 35(3), 251-260.
Knowler, W.C., Barrett-Connor, E., Fowler, S.E., Hamman, R.F., Lachin, J.M., Walker, E.A., and Nathan, D.M. (2002). Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin. New England Journal of Medicine, 346(6), 393-403.
Kraemer, H.C., Wilson, G.T., Fairburn, C.G., and Agras, W.S. (2002). Mediators and moderators of treatment effects in randomized clinical trials. Archives of General Psychiatry, 59, 887-883.
Minkler, M., and Wallerstein, N. (Eds.). (2008). Community Based Participatory Research for Health: Process to Outcomes (2nd ed.). San Francisco: Jossey-Bass.
Mohr, D.C., Spring, B., Freedland, K.E., Beckner, V., Arean, P., Hollon, S.D., Ockene, J., and Kaplan, R. (2009). The selection and design of control conditions for randomized controlled trials of psychological interventions. Psychotherapy and Psychosomatics, 78(5), 275-284.
Multiple Risk Factor Intervention Trial Research Group. (1982). Multiple risk factor intervention trial. Risk factor changes and mortality results. Journal of the American Medical Association, 248(12), 1,465-1,477.
Prochaska, J.J., Spring, B., and Nigg, C.R. (2008). Multiple health behavior change research: An introduction and overview. Preventive Medicine, 46(3), 181-188.
Ragland, D.R., Winkleby, M.A., Schwalbe, J., Holman, B.L., Morse, L., Syme, S.L., and Fisher, J.M. (1987). Prevalence of hypertension in bus drivers. International Journal of Epidemiology, 16, 208-214.
Schneiderman, N., Saab, P.G., Catellier, D.J., Powell, L.H., DeBusk, R.F., Williams, R.B., Carney, R.M., Raczynski, J.M., Cowan, M.J., Berkman, L.F., and Kaufmann, P.G. (2004). Psychosocial treatment within sex by ethnicity subgroups in the enhancing recovery in coronary heart disease clinical trial. Psychosomatic Medicine, 66(4), 475-483.
Writing Group for Activity Counseling Trial Research Group. (2001). Effects of physical activity counseling in primary care: The activity counseling trial: A randomized controlled trial. Journal of the American Medical Asociation, 286, 677-687.
This page intentionally left blank.