This retrospective cohort study included women screened for breast cancer through organized mammography screening programmes in Denmark and Norway during 2016–2019 and in Spain during 2012–2015. Women aged 50–69 years were included in the study group if they had an abnormal screening mammogram recommending further assessment. Only the first recall during the study period was included. We excluded women with a diagnosis of breast cancer before the screening examination, and death or emigration within 1 month, where data were available (Denmark and Norway). The included women were followed through their diagnostic follow-up until a breast cancer diagnosis, or the mammographic findings were ruled out and the women returned to routine screening (max. two years).

Following the European guidelines for mammography screening [2] all three countries offer biennial population-based two-view digital mammography screening to women aged 50–69. The screening mammograms are classified into positive or negative based on the assessment consensus of two independent breast radiologists, disagreements assessed by a third radiologist [7].

The Danish programme was gradually implemented in 1991 in one municipality and became nationwide in 2010 [8]. According to Danish legislation, women should have the screening result within 10 working days and an appointment for diagnostic mammography within an additional 6 working days [7, 9]. Recent numbers show coverage of 79% and yearly participation rate of 83% [10].

In Norway, the national screening programme, BreastScreen Norway, started in 1996 as a pilot programme and became nationwide in 2005 [11]. The Cancer Registry of Norway administers the programme. Independent double reading with consensus is standard practice for initial interpretations. All examinations are given a score of 1–5 for each breast and cases with a score of 2 or higher by one or both radiologists are discussed at a consensus meeting where it is decided whether to recall the woman for further assessment, which usually takes place within 1–2 weeks [12]. The annual attendance rate is 75%, and coverage is about 84% [13].

In Spain, the first regional population-based mammography screening programme was launched in Navarre in 1990 and extended to the whole country in 2006. The programme is part of the National Health System strategy [14], however, managed in a decentralized manner structured by geographical regions. Two independent breast radiologists interpret the screening mammogram according to the Breast Imaging Reporting and Data System (BI-RADS) scale [15]. Independent double reading with consensus or arbitration by a third radiologist is standard practice. Those with abnormal mammography defined as BI-RADS scores 3, 4, 5, or 0 are recalled for additional assessments. Further assessment usually takes place within 2 weeks after the screening examination. Data from 2016 indicate that screening coverage in Spain was higher than 90% and that overall participation is 75.7% [16].

The Danish study population was identified using data from the National Breast Cancer Screening database [8]. Diagnostic procedures after screening and cancers were retrieved from the National Cancer Registry [17], The National Pathology Register [18], and the National Patient Registry [19]. Furthermore, information on death and emigration was retrieved from the Civil Registration System in Denmark [20] and different health registries merged using the individual identification number.

In Norway, it has been mandatory to report cancer cases since 1953. Information about assessment procedures and diagnosis among screening participants is stored in databases at the Cancer Registry of Norway. Data about death and emigration were extracted from the Population Registry. Data were merged using the 11-digit personal identification number assigned to all residents.

In Spain, data were obtained from nine different units of the Spanish Breast Cancer Screening Program. All screening examinations performed in these centres between 2012 and 2015 were included. The different centres gather information on screening mammograms, recalls, additional assessments, and diagnoses performed in their defined recruitment areas and they are considered representative of the general Spanish population [21, 22]

The assessment pathways were defined by including all relevant procedures after a positive screening examination and categorized based on the type of assessment and diagnostic procedures performed, i.e. whether additional (mammographic) imaging (including also tomosynthesis and breast MRI), ultrasound examinations or biopsies (core and fine-needle) were performed for a period of 6 months. Women who had no follow-up procedures registered within six months after the screening date were categorized as having ‘No follow-up/lost to follow-up’. The procedures had to be registered before or on the same day as the cancer diagnosis was registered, to be included as recall assessment and not part of pre-operative examinations and cancer treatment.

For Danish and Norwegian women, those cancers diagnosed within six months after the positive screening examination were considered SDC, and those diagnosed from six to 24 months after a positive screening examination or 0–24 months after a negative examination and before the next screening were considered IC. In Spain, all SDC are consistently recorded in the screening programmes databases. The information on interval cancers is actively obtained on a systematic basis using screening centre databases, hospital-based cancer registries, regional Minimum Data Sets, and population-based cancer registries.

SDC’s and IC’s were all defined as invasive carcinomas of the breast and ductal carcinoma in situ using either ICD codes (ICD-10: C50*, D05*).

Distributions of the diagnostic follow-up pathways were presented as absolute numbers with proportions of women referred to the particular assessment and follow-up out of all screened women to account for the varying referral rates in the populations. In Table 5, numbers are presented out of women referred for recall assessment.

Time to first follow-up was calculated in calendar days from the screening examination until the first registered diagnostic follow-up procedure and presented as cumulative proportions, by the final assessment outcome, i.e. false positive or SDC.

The positive predictive value of the diagnostic follow-up (PPV-1) was calculated as the number of SDC’s in each diagnostic pathway divided by the number of women referred to the specific follow-up pathway. In addition, we calculated the positive predictive value of a diagnostic assessment including at least one biopsy, i.e. PPV-2, as a biopsy is needed to have the cancer diagnosis confirmed in most cases. Cancer detection rates were calculated as the number of SDC’s among all screening examinations in the period. IC rates were presented as (1) the number of cancers diagnosed after a negative screening, but before the next screening or 24 months, whatever comes first, per 1000 screening examinations, i.e. ICrate, and (2) the number of cancers diagnosed after a false positive screening in each diagnostic follow-up pathway per 1000 false positive recalls, i.e. ICRpos.

The sensitivity of screening was defined as the number of SDC’s divided by the total number of cancers (SDC and IC), i.e. Sens-1 [6]. To diminish the potential effect of different referral rates in the populations, we also calculated the sensitivity of the diagnostic pathway by dividing the number of SDC by the IC diagnosed for two years in the referral population, i.e. Sens-2.