Table 2.
Association between date of entry to screening program and cumulative breast cancer mortality 1995–2011
| Date of program access | Number of breast cancer deaths per 100,000 (95% CI)1 | Difference in deaths per 100,000 from number in reference (95% CI)2 | p value |
|---|---|---|---|
| Effects of delayed access to the screening program (6 month indicators; N = 256,712)3 | |||
| January to June 1995 (reference) | 1,138 (952 to 1,324) | ||
| July to December 1995 | 35 (−104 to 174) | 0.62 | |
| January to June 1996 | −44 (−161 to 74)) | 0.46 | |
| July to December 1996 | 10 (−137 to 157) | 0.90 | |
| January to June 1997 | 163 (21 to 306) | 0.03 | |
| July to December 1997 | 186 (33 to 339) | 0.02 | |
| Effect of one additional biennial screening invitation (1997 indicator; N = 148,920)4 | |||
| 1995 (reference) | 1,116 (901 to 1,332) | ||
| 1997 | 154 (40 to 267) | 0.009 |
Scaled model constant indicating estimated mortality in the reference group aged 49 and with access to the screening program from the period indicated in left‐hand column.
Scaled coefficient on indicator of period in which screening program started to operate in municipality.
Data for women aged 49–63 in January 1995 in 138 municipalities where the screening program was implemented between January 1995 and December 1997. Estimates are scaled coefficients from linear probability models with the dependent variable a binary indicator of having died from breast cancer between 1995 and 2011. The models include a (series of) binary indicator(s) of the period in which the screening program started to operate in the municipality of residence and yearly indicators of age in January 1995 (49 [reference], dummies for 50, 51, to 63). Statistical inference accounts for clustering among women in a municipality at its entry to the program.
Data for women aged 49–63 in January 1995 in 86 municipalities where the screening program was implemented between January and December 1995 and between January and December1997. Estimates are scaled coefficients from linear probability models with the dependent variable a binary indicator of having died from breast cancer between 1995 and 2011. The models include a (series of) binary indicator(s) of the period in which the screening program started to operate in the municipality of residence and indicators of age in January 1995 (49 [reference], dummies for 50, 51, to 63). Statistical inference accounts for clustering among women in a municipality at its entry to the program.