Several randomized trials have demonstrated the efficacy of colorectal cancer screening by the fecal occult blood test in reducing colorectal cancer mortality, but an evaluation of population-based screening programs is still lacking. We compared the colorectal cancer mortality rates (both adjusted rates and 3-year moving average rates) during 1985-2006 for two geographic areas in the provinces of Florence and Prato in the Tuscany region of Italy that began active population-based screening for colorectal cancer at different times: the Empolese-Mugello district, in the early 1980s, and the rest of the Florence and Prato provinces, in early 2000. A log-linear Poisson model was used to estimate the annual percent change in mortality and to examine whether geographic area modified the effect of calendar year on it. The Empolese-Mugello district had a greater decrease in colorectal cancer mortality than the rest of the Florence and Prato provinces (estimated annual percent change in age-adjusted colorectal cancer mortality rate, 2.7% decrease per year [95% confidence interval {CI} = 1.7% to 3.7%] vs 1.3% decrease per year [95% CI = 0.8% to 1.7%], respectively). The interaction between calendar period and area was statistically significant (P < .001). Our results support the hypothesis that the implementation of colorectal cancer screening in the early 1980s in the Empolese-Mugello district, where approximately 17 500 people were tested each year with the fecal occult blood test, was associated with a larger reduction in colorectal cancer mortality than that observed in the rest of Florence and Prato provinces, where screening started 15-20 years later and where approximately 38 000 people were screened each year beginning in 2000.