Figure 1.

Effect of one additional biennial screening invitation on cumulative breast cancer mortality to different endpoints. The figure shows the estimated effect of getting access to the screening program in 1997 rather than 1995 on the number of breast cancer deaths from January 1995 to the end of each year indicated per 100,000 women who were aged 49–63 in January 1995. Estimates are scaled coefficients on the binary indicator of year of program entry in linear probability models with the dependent variable a binary indicator of having died from breast cancer between January 1995 and the end of the respective year. All models include yearly indicators of age in January 1995, that is, 49 (reference) and dummies for 50, 51, to 63. Error bars show 95% confidence intervals computed allowing for clustering at the municipality level.