OBJECTIVE: To obtain an early estimate of the effectiveness of the American Stop Smoking Intervention Study (ASSIST). DESIGN, SETTING, AND PARTICIPANTS: Seventeen American states funded through ASSIST are compared with 32 others regarding per capita cigarette consumption from 1989 to 1995. California, which already had an extensive tobacco control programme, was omitted. ASSIST states were selected competitively (not randomly) based on their proposals' merit, state smoking prevalence, and geographical distribution. INTERVENTIONS: Comprehensive tobacco control programmes, emphasising policy interventions, were implemented in the ASSIST states beginning in 1993. MAIN OUTCOME MEASURES: Trends in aggregated per capita cigarette consumption and inflation-adjusted average price/pack of cigarettes in the intervention states were compared. Percentage change in per capita consumption is also compared with percentage change in inflation-adjusted cigarette price by state in each group from 1992 to 1994. RESULTS: Per capita consumption and inflation-adjusted cigarette price were nearly identical in both groups of states before 1993, when full funding for the ASSIST interventions began. However, by 1996 smokers in the intervention states were consuming about 7% less cigarettes per capita (P<0.05, beginning in 1994), and in 1994 the average price was over $0.12/pack higher in the intervention states. All but three states (all intervention) showed decreases in cigarette price. Nonetheless, 76% of the intervention and 55% of the comparison states showed some decrease in consumption despite decreases in price. The relationship between changes in price and consumption was considerably diminished in the intervention group. CONCLUSIONS: These interim results suggest that the ASSIST programme is associated with a substantial difference in tobacco consumption in a third of the United States, and that increased price from taxation may not be the only programme influence.