Background: In the transtheoretical model of behavior change, “stages of change” are defined as Precontemplation (not even thinking about changing), Contemplation, Preparation, Action, and Maintenance (maintaining the behavior change). Marketing principles suggest that efforts should be targeted at persons most likely to “buy the product.”

Objectives: To examine the effect of intervening at different stages in populations of smokers, with various numbers of people in each “stage of change.” One type of intervention would increase by 10% the probability of a person moving to the next higher stage of change, such as from Precontemplation to Contemplation. The second type would decrease by 10% the probability of relapsing to the next lower stage, such as from Maintenance to Action, and also of changing from Never Smoker to Smoker. Nine hypothetical interventions were compared with the status quo, to determine which type of intervention would provide the most improvement in population smoking.

Methods: Three datasets were used to estimate the probability of moving among the stages of change for smoking. Those probabilities were used to create multi-state life tables, which yielded estimates of the expected number of years the population would spend in each stage of change starting at age 40. We estimated the effect of each hypothetical intervention, and compared the intervention effects. Several initial conditions, time horizons, and criteria for success were examined.

Results: A population of 40-year-olds in Precontemplation had a further life expectancy of 36 years, of which 26 would be spent in the Maintenance stage. In a population of former and current smokers, moving more persons from the Action to the Maintenance stage (a form of relapse prevention) decreased the number of years spent smoking more than the any other intervention. In a population of 40-year-olds that included Never Smokers, primary smoking prevention was the most effective. The results varied somewhat by the choice of criterion, the length of follow-up, the initial stage distribution, the data, and the sensitivity analyses.

Conclusions: In a population of 40-year-olds, smokers were likely to achieve Maintenance without an intervention. On the population basis, targeting quitters and never-smokers was more effective than targeting current smokers. This finding is supported by some principles of health marketing. Additional research should target younger ages as well as other health behaviors.


Health Services Research