Community research focuses on the impact of single and multiple interventions on the prevention and cessation of tobacco use among large groups of people. Community research examines various types of intervention strategies, such as educational programs, clinical services, media programs, and policy interventions in a wide range of settings, including schools, worksites, health care facilities, and community organizations. INTRODUCTION Ultimately, all tobacco control research should be applied at the state and community levels. Communities are the final "laboratories" where tobacco control interventions are tested to determine if and how they work in real-world settings. The need for such research on state and community interventions has never been greater. New tobacco control programs are underway in almost every state, and each program raises new research questions and opportunities to provide useful information that can improve the effectiveness of these interventions. To seize this opportunity, community and state research must use a transdisciplinary approach that draws on, applies, and assess the impact of research from many different research areas discussed in this report. Ideally, community research uses the results of more basic biological, behavioral, and treatment research and links their findings to policy research to develop and evaluate community tobacco control programs. Community interventions frequently include a component that promotes public health policies through coalitions and the mass media. Community research also draws extensively on epidemiology and surveillance methods to obtain critical information needed to assess the impact of tobacco control interventions. In other cases, community research takes advantage of what are, in effect, large "natural experiments" that occur when new programs or policies are implemented in a group of states or communities. State settlements with the tobacco industry are providing resources for tobacco control programs and new opportunities for natural experiments. The development of new measurement tools and surveillance systems will be needed to support this complex but vital research. During the last 15 years, the NCI has conducted a program of community intervention research that has developed and tested a large number of cancer control interventions. Trials of individual interventions ultimately led to the Community Intervention Trial for Smoking Cessation (COMMIT), which tested the impact of multiple interventions in community settings on adult smoking cessation rates. The combined interventions significantly increased smoking cessation rates among adults who smoked less than 25 cigarettes per day but not among heavier smokers. If this rate of success was expanded nationally, it would generate an additional 1.2 million individuals annually who successfully stop smoking. Currently, the NCI supports a much broader intervention program, the American Stop Smoking Intervention Study (ASSIST). Now in its final phase, ASSIST is testing the impact of a more comprehensive intervention on smoking rates among adults and youth in 17 states. Building on lessons learned from COMMIT, the intervention used in ASSIST states added a strong emphasis on public and private tobacco control policies. The evaluation of ASSIST now underway will detail the impact of the intervention on several different outcomes, including the prevalence of smoking and tobacco use. It also will describe the impact of different parts of the intervention on these outcome measures. A midpoint analysis has shown that cigarette consumption is significantly lower in ASSIST states than in other states. As a result, the Centers for Disease Control and Prevention will fund programs based on the ASSIST model in all 50 states beginning in 1999. Research-based programs like COMMIT and ASSIST have illustrated the potential of community and state tobacco control interventions to change the societal influences on tobacco use and reduce the tobacco-related cancer burden. As states continue to expand the implementation of such programs, we must seize this opportunity to expand research on the community and state levels to increase understanding of the most effective approaches to achieving broad reductions in tobacco use. RECOMMENDATIONS Research should be conducted to improve community and state tobacco control programs and to increase their effectiveness for populations at disproportionate risk. Although comprehensive state tobacco control programs have led to significant reductions in the tobacco use rates among very large populations, major questions remain about the relative importance of different components of these programs. Therefore, a high priority is to establish research protocols in state tobacco control programs to ascertain the impact of new and existing interventions and to determine how they and their individual components work to achieve their effects. This research should be an integral and ongoing part of state and local programs. We also need to conduct intervention trials to develop, test, and refine tobacco control interventions on an ongoing basis. One of the most important components of such trials should be assessment of mass media interventions. Counter-advertising campaigns funded by state governments are highly visible and costly. Research in this area should document effective components and guide future campaigns. Many other important research questions only can be addressed in communities and states that are conducting multiple interventions. For example, we need to conduct simultaneous testing of multiple interventions in different community settings to clarify the relative contribution of each intervention to reductions in tobacco use. We also need observational studies of interventions and policies whose implementation is beyond the control of the investigator to provide new information about how these initiatives influence tobacco-use behaviors. Such research will require the use of complex study designs that account for variation in the multiple parameters that are found in these very complex environments. Special emphasis should be given to high-risk populations in developing and evaluating community and state tobacco-control programs. These populations include ethnic and racial minorities, adolescents, children, people with low educational levels, and blue-collar workers. Of particular concern is research to address the needs of low-income Americans, who have consistently higher rates of tobacco use than the general population. Ethnographic studies and survey research, economic and behavioral-economic studies, studies of targeted tobacco marketing, and biobehavioral research to clarify the basic mechanisms contributing to higher tobacco use and nicotine addiction among low-income Americans and other high-risk populations could help to generate more effective prevention and cessation strategies and tobacco control policies for these groups. Research in this area also must consider the features of the population that influence the delivery, receipt, and impact of a particular intervention. Research should be conducted to identify mechanisms for optimal dissemination of proven prevention and treatment interventions at the community and state levels. As prevention and treatment interventions are found to be effective for particular groups, larger-scale dissemination and diffusion trials should be conducted to evaluate the best approaches for applying these programs at the state and community levels. For example, we know that few Americans currently have access to the most effective school-based prevention programs or physician-initiated and self-help smoking cessation programs that follow the guidelines of the Agency for Health Care Policy Research (AHCPR). Research is needed to increase scientific knowledge about policies and strategies that increase dissemination and application of these and other effective tobacco-control interventions. One promising treatment dissemination strategy makes nicotine addiction treatments with different levels of intensity available to large populations. This strategy uses a "stepped" approach to treatment and a matching of patients to appropriate therapies. It is inherently cost-effective because it assumes that people should be given the least intensive intervention that is effective. Research is also needed to develop new strategies for accomplishing the systemic change needed to institutionalize effective tobacco control interventions. Examples of promising approaches in health care settings include implementing systems in medical offices, clinics, and hospitals to identify patients' tobacco-use status; providing education and feedback to health care providers about tobacco cessation approaches; and providing reimbursement for smoking cessation treatment. THE IMPACT OF COMMUNITY AND STATE INTERVENTION RESEARCH The findings of community and state intervention research can have a tremendous impact on the tobacco control programs that are already underway in all 50 states as well as the new and expanding programs that are under development. The impact of existing programs on tobacco-use behaviors among large Populations already has been established. Expanding community and state research will increase our understanding of how to change societal on tobacco use, increase the effectiveness of these programs, and contribute to the development of the next generation of both targeted and broadly applicable tobacco-control programs, which includes much of the policy research described in the next section, is likely to have an enormous impact on the rapid reduction of tobacco-use rates and subsequent tobacco-related cancers.