The Netherlands started screening for Breast Cancer 30 years ago. During this period, we learned a lot about the benefits and harms of Breast Cancer screening and therefore we constantly adjust the Dutch programme in order to optimise it as much as possible. This optimisation resulted in a mortality reduction of 50 per cent for women who do attend all screening rounds between the age of 50 and 75 in a biannual setting.
Due to our experiences, we developed a philosophy regarding the organisation and maintenance of a screening programme, based upon six pillars.
At first, we think of screening like a medical chain in which equipment, technicians and radiologists are the key players. This chain is as weak as the weakest link. Therefore, before starting a screening programme all three links have to be secured. In practice this means that an adequate educational programme as well as a system for quality control has to be implemented before starting the actual screening process itself. It is a misperception that starting a screening programme is equivalent to buying equipment.
The second important issue is to discriminate between clinical breast radiology and screening for breast cancer. The mindset in a screening environment differs from a clinical setting. In screening you only have to depict lesions with a high probability for breast cancer. All other lesions are not of interest in a screening setting. This requires additional training not only in a theoretical setting but even more it requires specific skills.
Therefore, in our opinion a radiologist is not a screening radiologist unless these specific skills have been trained. In the Netherlands, it is obligatory to pass an additional training before you are allowed to work in the Dutch national screening programme. The additional training is also obligatory for technicians. Apart from the certification of professionals all equipment has be certified as well. This is the third pillar of our system. Certifying new mammography equipment before installation, weekly calibration of every mammography unit and extensive testing every six months is part of our quality control system.
In order to maintain the highest quality possible and adjusting the individual performance of Technicians and Radiologist, auditing is essential and accounts for another important pillar of our philosophy. During these audits, which take place once every three years, we benchmark the results from screening units all over the country. This gives insight in the regional performance in comparison to the national performance. During these audits, we arrange peer to peer discussions between the auditing team and the screening unit, which is audited.
In this way we have created a system in which scientific discussion forms the basis of adjusting the programme instead of signing out of a list of items. The combination of benchmark and peer to peer discussions makes fine tuning a real option in the Dutch screening programme.
Another important element of our quality system is the recall rate. When we started screening 30 years ago, we thought that a recall rate of 1 per cent would be optimal for the Dutch programme. The advantages of this low recall rate were a high positive predictive value (this means that relatively few women were recalled for a benign lesion) while detecting a lot of breast cancers at an early stage. After evaluating this recall rate by means of ‘the optimisation study’, published in JCNI 2005, we changed our policy and went up to a recall rate of 2.4 per cent. Still very low compared to other countries, especially the U.S., but with the same breast cancer detection as in other countries (6,8 per 1,000). We pay a lot of attention to recall rate in the Netherlands. Recall rate reflects in essence the balance between benefits and harms of a screening programme. The more you recall, the more falls-positives (woman recalled from the screening programme but no malignancy after assessment) you will have.
False-positive recalls should be avoided as much as possible because it constitutes a serious drawback of screening without compensatory benefit for the affected subgroup of participants. In studies performed in the Netherlands, we discovered a drop in re-attendance after a positive recall as high as up to 30 per cent. Anxiety and discomfort for women as well as high costs for the community (additional imaging and image-guided interventions) makes false-positives a serious item to evaluate, educate and control. Defining a recall rate, teaching and training skills of radiologist forms the backbone to minimize harms in a screening programme.
The last pillar is about data. Data are essential in a screening programme. Not only information regarding incidence in relation to age, but also recall rate, detection of breast cancer, interval cancers (symptomatic cancers that appear between screening rou nds), stage of detection and in the long run figures in survival are essential to evaluate and adjust the program. In the Netherlands, we have central storage of all mammography images produced during screening. But we also have a unique custom-made reporting system. Besides that, all pathology reports from all over the country (not only breast) are stored centrally as well as all data regarding cancer (all cancer types). Coupling of these databases makes it possible to evaluate the Dutch screening programme on a national level. This gives us the opportunity to calculate other items like overdiagnosis.
Our philosophy that screening differs from clinical breast radiology also accounts for the women involved. Therefore, screening in the Netherlands is arranged completely outside the hospital setting. We have 58 trucks with mammography equipment driving throughout the country to facilitate women to have their mammography once every two years (paid for by the government) close to their home address avoiding a clinical setting. Apart from the 58 trucks, we have 20 stationary systems outside hospitals. This concept is accepted very well and is reflected in a constant high attendance rate of around 80 per cent. Only in case of recall and further assessment women are confronted with a hospital setting.
In conclusion, screening for Breast Cancer works but it requires a solid infrastructure consisting of education, monitoring auditing and continuous adjustment. The Dutch model reflects a comprehensive system resulting in a serious reduction of breast cancer mortality. The Dutch Expert Centre for Screening (LRCB) is frequently consulted by countries that want to start or set up a screening programme. Adjustment for local and cultural background is essential for making the programme successful. But regardless, the country or cultural background, education is the key to success.