No overdiagnosis in the Norwegian Breast Cancer Screening Program estimated by combining record linkage and questionnaire information in the Norwegian Women and Cancer study

Abstract

<i>Background</i>: The Norwegian Breast Cancer Screening Program (NBCSP) was implemented across the country in 2005 and has been criticised for potential ‘overdiagnosis’, i.e. a breast cancer diagnosis that otherwise would not have been detected or treated in a woman’s lifetime. We aimed to estimate overdiagnosis in the NBCSP based on the Norwegian Women and Cancer (NOWAC) study using both questionnaire information and record linkage information from NBCSP.<p>

<p><i>Method</i>: For 124,978 women aged 49e79 years from the NOWAC study, information on <i>screened</i> women could be cross-validated from the NBCSP database. Based on information from the NOWAC questionnaire, <i>unscreened women</i> were further divided into those who had <i>mammograms taken only outside the NBCSP</i> and those who had <i>never had taken a mammogram</i>. Breast cancers diagnosed in 2005e2013 were identified through linkage to the Cancer Registry of Norway; in situ or DCIS 417; invasive 2845; combined 3262. Cumulative incidence rates (CIRs) for ages 49e79 years of breast cancer were compared using the log-rank test.<p>

<p><i>Results</i>: After exclusion of women with a family history of breast cancer, screened women had a CIR of 9.7% for combined breast cancer, non-significantly lower compared with unscreened women. Screened women had a 1.1% increased CIR or 13.0% increased relative risk of breast cancer diagnosis (significant) compared with women who had never had a mammogram, but for invasive breast cancer alone the difference was reduced to -0.2% (95% CI: -9.1; 8.8). Invasive breast cancers were significantly smaller (<2.5 cm) in screened versus unscreened women. There was a borderline significant decrease in lymph node positive cancer among screened (p Z 0.06).<p>

<p><i>Conclusion</i>: The findings of no significant overdiagnosis combined with smaller tumours and less lymph node metastases suggest that the prevailing view of overdiagnosis in the NBCSP should be challenged.<p>