Questions and Answers: ASSIST Evaluation

National Cancer Institute
Tuesday, 18 November 2003

1. What is ASSIST?

At the time of the study (1991-1999), the American Stop Smoking Intervention Study (ASSIST) was the largest government-funded demonstration project to help states develop effective strategies to reduce smoking. In 1991, the National Cancer Institute (NCI), part of the National Institutes of Health, funded 17 state health departments and formed a partnership with the American Cancer Society to undertake the study. Focusing on policy change, the goal of ASSIST was to alter states' social, cultural, economic, and environmental factors that promote smoking.

2. How were the intervention and outcomes measured?

The ASSIST evaluation compared changes in tobacco control policies, state per capita cigarette consumption, and adult smoking prevalence in ASSIST vs. non-ASSIST states and the District of Columbia. Smoking prevalence, or the number of people who smoke, was obtained from adults interviewed in the NCI-sponsored Tobacco Use Supplement to the U.S. Census Bureau's Current Population Survey in 1992-1993 and 1998-1999. Per capita cigarette consumption was calculated every two months for each state from sales data for the total number of cigarette packs moved from wholesale warehouses, divided by the state's adult population.

The focus of the ASSIST project was on policy change, which was assessed with a measure called Initial Outcomes Index (IOI). IOI was developed to serve as an indication of the intensity of states' tobacco control policies. It includes the percentage of smokers covered by 100 percent smoke-free work sites, total cigarette price, and legislative ratings. The authors also developed Strength of Tobacco Control Index (SOTC) to evaluate the ASSIST program. SOTC summarizes the multiple components of tobacco control efforts and provides information on which components of ASSIST or ASSIST-like programs might be related to lower smoking prevalence or cigarette consumption. SOTC is a multi-element measure that assesses the effects of three variables in each state: tobacco control resources (funding), capacity and infrastructure, and program efforts focused on policy and environmental change. Both IOI and SOTC are promising measures that can be used for future research in tobacco control and evaluation.

3. What were the key findings of the ASSIST evaluation?

The authors found that ASSIST states had a greater decrease in adult smoking prevalence than non-ASSIST states. The ASSIST evaluation also showed that states that experienced greater improvement in tobacco control policies had larger decreases in per capita cigarette consumption. States (not including the District of Columbia) with higher policy scores also had lower smoking prevalence. In addition, the authors found that states with greater "capacity," or ability to implement tobacco control activities - such as states with tobacco control infrastructure in the health department, staff experience, and strong interagency and statewide relationships - had lower per capita cigarette consumption. Lastly, there was evidence that policy interventions may be more effective at reducing women's smoking.

4. What do these results tell us?

The results from ASSIST are the latest evidence that investing in state tobacco control programs that focus on strong tobacco regulations and policies is an effective strategy for reducing tobacco use. The small but statistically significant differences in the reduction of adult smoking prevalence in ASSIST states, when applied on a population basis, could be expected to have a large impact on the public. If all 50 states and the District of Columbia had implemented ASSIST, there would be approximately 278,700 fewer smokers nationally.

The finding that states with a greater change in their tobacco control policies during ASSIST had larger decreases in per capita cigarette consumption suggests that interventions which result in tobacco control policy change can have a strong and sustained effect on the amount of cigarettes smoked. This conclusion adds to the body of similar research and expert reports that document the importance of a comprehensive approach to tobacco control. Although policy efforts take time, they can bring about major changes in social norms, including smoking behavior.

The finding that states with stronger infrastructure or capacity (ability to implement tobacco control activities) had lower per capita cigarette consumption is more evidence that when tobacco control programs are strong and well-supported, a decrease in the amount of smoking can be achieved. This is the first study to provide a method to quantify states' capacity to implement tobacco control programs.

5. What was the design of the study?

ASSIST was a large-scale experiment in which the observational unit was the state (i.e., its entire population and environment), and the goal was to change social, cultural, economic, and environmental factors in the state that promote smoking behavior. This goal was accomplished primarily through interventions in four policy areas: 1) promoting smoke-free environments, 2) countering tobacco advertising and promotion, 3) limiting youths' tobacco access and availability, and 4) increasing tobacco prices by raising excise taxes. The statewide tobacco control plans were carried out in the 17 ASSIST states by a network of state and local coalitions charged with developing and implementing interventions.

The ASSIST states implemented the project in two phases: a two-year planning phase (October 1991 through October 1993) and a six-year implementation phase (November 1993 through September 1999).

The ASSIST evaluation compared changes in tobacco control policies, state per capita cigarette consumption, and adult smoking prevalence in the 17 ASSIST states with those in the 33 non-ASSIST states and the District of Columbia. The authors also analyzed the effect of program components and tobacco control policies on smoking prevalence and per capita cigarette consumption.

6. What states participated in the study?

The 17 ASSIST states were: Colorado, Indiana, Maine, Massachusetts, Michigan, Minnesota, Missouri, New Jersey, New Mexico, New York, North Carolina, Rhode Island, South Carolina, Virginia, West Virginia, Washington and Wisconsin. However, the evaluation encompassed all states and the District of Columbia and includes comparisons between the 17 ASSIST states and the 33 non-ASSIST states and D.C.

The ASSIST states had a combined population of 91 million people, more than a third of the population of the United States. The ASSIST population included more than 10 million African Americans and 7 million people of Hispanic origin or other racial or ethnic minority groups.

7. How were the 17 ASSIST states chosen?

All 50 states and the District of Columbia were eligible to compete for the NCI ASSIST contracts; 35 states applied for the contracts and 23 of those 35 states were deemed eligible for funding based on published selection criteria. Only 17 states were awarded contracts due to budgetary constraints.

ASSIST was a demonstration project and not a randomized experiment. Therefore, the states chosen for ASSIST funding represented a wide range in terms of their ability and experience in developing and implementing tobacco control programs.

8. Did the evaluation include youth smokers, as well?

No. This study measured cigarette consumption and smoking prevalence in adults, defined as persons at least 18 years old.

9. How much or how often must a person smoke to be considered a smoker in this evaluation?

A smoker was defined as a person who used cigarettes on a current, everyday, or some-day basis and had smoked at least 100 cigarettes in his or her lifetime.

10. How much did the study cost?

NCI provided an average of $1.14 million per state per year during the six-year implementation phase (1993-1999), for a total of $128 million over the eight years of the program. Other additional funding and resources were available to the states through voluntary organizations and other non-federal sources.

11. What were the limitations of the study?

An evaluation of states restricts the number of observations to 51 (50 states plus the District of Columbia) and reduces the ability to detect small but statistically important changes, particularly for per capita cigarette consumption. Despite this lower ability to detect changes, the authors did find some statistically significant differences.

Complicating the ASSIST evaluation was the diffusion of materials and interventions from ASSIST states to non-ASSIST states, with no restriction on the free flow of knowledge and technical assistance. ASSIST served as an impetus for change, as by 1994, the Centers for Disease Control and Prevention supported tobacco control programs in all non-ASSIST states, and the Robert Wood Johnson Foundation's SmokeLess States program was implemented in many ASSIST and non-ASSIST states. In addition, during the 1990s, tobacco control activities and issues received more media attention than ever before in most states. These factors possibly decreased the relative difference in outcomes between ASSIST and non-ASSIST states.

Although the ASSIST researchers attempted to identify and address as many factors as possible that affect smoking behavior, some political, social, and economic factors outside the control of the ASSIST intervention could have caused the delivery of the intervention strategies to differ between states. For example, some ASSIST states might have been better able than others to build their infrastructure and focus on policy change as outlined in the ASSIST program plans.

In addition, during the period of the ASSIST project, the tobacco industry spent approximately $47 billion nationwide to market tobacco products (Federal Trade Commission, 1992-1999 spending) and previously confidential tobacco industry documents demonstrate that the tobacco industry acted to counter the ASSIST project and to oppose the policy measures it sought to implement at the state level. However, the authors of the ASSIST evaluation were not able to develop a quantitative measure of the tobacco industry's countervailing efforts and are unable to determine the degree to which the tobacco industry affected the success of the ASSIST project.

12. What is the National Cancer Institute doing now in the area of tobacco research?

NCI's goal in tobacco research is to understand the causes of tobacco use, addiction, and related cancers and to apply this knowledge to prevention and treatment. NCI is advancing tobacco research by conducting and supporting numerous research grants, partnerships, and special initiatives, including the Transdisciplinary Tobacco Use Research Centers and the Tobacco Research Initiative for State and Community Interventions, among many other projects. For detailed information about tobacco research at NCI, go to the Tobacco Control Research Branch's Web site,

For more information, or to contact National Cancer Institute, see their website at:

Email Article To A Friend Link to us!
Home » Medical Research » National Cancer Institute » Article 04838