Table 2.
Results of secondary analyses of association of DCIS detection frequencies with subsequent invasive interval cancer incidence
| Poisson regression coefficient (95% CI) | p value | |
|---|---|---|
| Adjustment for small (<15 mm) invasive tumours | ||
| Frequency of screen-detected DCIS | −0·080 (−0·13 to −0·03) | 0·003 |
| Small screen-detected tumours | −0·0003 (−0·001 to 0·0004) | 0·38 |
| Adjustment for screen-detected grade 3 invasive tumours | ||
| Frequency of screen-detected DCIS | −0·089 (−0·14 to −0·03) | 0·001 |
| Grade 3 screen-detected tumours | 0·0003 (−0·0003 to 0·001) | 0·35 |
| Restriction to invasive interval cancers diagnosed within 24 months | ||
| Frequency of screen-detected DCIS | −0·089 (−0·16 to −0·016) | 0·016 |
| Repeat of primary analysis to include the 2003–04 data from Scotland | ||
| Frequency of screen-detected DCIS | −0·086 (−0·14 to −0·033) | 0·001 |
DCIS=ductal carcinoma in situ.