Nadine Zielonke

Breast cancer is a major public health problem in Europe. It is by far the most frequently diagnosed neoplasm in European women as it accounts for nearly one third of all new cancers in women. At present, women in Europe have a 1:7 chance of developing breast cancer during their lifetime. Breast cancer remains the leading cause of death in European women. But encouragingly, breast cancer mortality has been declining in most of Europe since the 1980’s. This favourable trend in breast cancer mortality is essentially due to advancements in early diagnosis and improved treatment. However, great inequity persists in cancer incidence and mortality rates across Europe.

Breast cancer screening means checking a woman’s breasts for cancer before there are signs or symptoms of the disease. It aims to reduce morbidity associated with advanced stages of the disease and subsequently breast cancer mortality. At present, breast cancer screening programmes are well established in most European countries. Most of them adopted biennial screening for breast cancer in the minimum target age range (50–69 years, as recommended by the European Union). But disparities exist in terms of the status of implementation, the extent to which screening programmes are organized or coexist with opportunistic screening activity, as well as the invitation coverage and the attendance to screening.

With this thesis, we aimed to investigate the consequences of variations in breast cancer screening practices and potential ways to further optimize screening programs across Europe.

Part 1: The effectiveness of breast cancer screening
In Chapter 2, we performed a systematic review to summarize the current evidence on reducing breast cancer mortality with mammography screening in Europe. Our results strengthen previous findings that mammography screening reduces mortality from breast cancer, but highlight that this occurs at varying magnitudes. For this systematic review, we included randomized controlled trials (RCTs) as well as observational studies such as prospective and retrospective controlled cohort or case-control studies. The stringent use of widely acknowledged and transparent grading tools in order to appraise the quality of each included reference allowed us to highlight only those studies, which provide the most valid information. From this set of high quality studies we found that the impact of organized screening (target ages 50 to 69) on breast cancer specific mortality ranges from 12% to 58% in screening attenders versus non-attenders and from 4% to 31% in invited versus non-invited women.

In Chapter 3, we aimed to investigate the determinants of benefits and harms of different breast cancer screening approaches (mammography, ultrasonography, clinical breast examination and breast self-examination) and to summarize data from systematic reviews on those four screening approaches among the general population. Overall, we found that systematic reviews of breast cancer screening focus on mammography more than on the other screening approaches, and evaluate benefits of screening more frequently than harms. All included systematic reviews are consistent in certifying a reduction in breast cancer mortality among women aged 50–69 years. However, important nuances exist across studies and regions at variable levels of credibility. Results for overdiagnosis varied widely by type of original evidence and several methodological differences, e.g. the denominator, the duration of follow up or the accounting for ductal carcinoma in situ. Therefore, the magnitudes in effects of benefits and harms of mammography remain heterogeneous.

We estimated in Chapter 4, that breast cancer screening prevents nearly 21,700 breast cancer deaths in Europe each year. Cancer screening programs can only be effective in reduction in mortality from breast cancer if a high proportion of people within the target population make an informed choice to participate. Our aim in Chapter 4 was to investigate what the effect would be of an increased or even complete breast cancer screening coverage on breast cancer mortality for each European country. We found that introducing a hypothetical 100% coverage of screening in the advised target age group (meaning, all eligible women aged 50-69 are invited and follow this invitation), the number of breast cancer deaths of European women could be further reduced by almost 12,500 per year. The effect would be particularly notable in Eastern Europe.

Part 2: Modelling the impact of different interventions on the harms and benefits of breast cancer screening
In Part 2 of this thesis, we evaluated the impact of different screening interventions on the harms and benefits of breast cancer screening across different European countries, using the microsimulation model (MISCAN-Breast).

Breast cancer screening causes harms and benefits. The balance between the two varies by age. In Chapter 5, we used the MISCAN-Breast models of four European countries (the Netherlands, Finland, Italy and Slovenia) to assess how harm-to-benefit-ratios would vary if biennial breast cancer screening would be extended to younger and/or older age groups (45-69, 45-74, 50-74) when compared with the current strategy of screening women in the age group 50-69. We found that in all countries, adding screening between the ages 45 and 49 or 70 and 74 resulted in more life-years gained and more breast cancer deaths averted, but at the expense of increases in harms (i.e. the number of overdiagnoses and false-positive diagnoses). Adapting the age range of breast cancer screening is an option to improve harm-to-benefit ratios, but the prioritization of considered harms and benefits affects the interpretation of the results.

Despite the benefits of early detection, in practice, screening programs often fail to achieve their full potential. Therefore, in Chapter 6 we assessed how breast cancer screening programs can be further optimized. Using Italy as the example, we first identified barriers common to those identified in other countries, namely low adherence to screening and a combination of screening intervals due to opportunistic screening (leading to a screening interval < 2 years) and a lack of recourses (leading to a screening interval > 2 years). Subsequently, we modelled potential future changes (i.e. practical solutions) that can be initiated to the breast cancer screening programs to overcome these barriers and to improve screening. Our analysis shows that removing the most important barriers of the current Italian breast cancer screening programs could result in substantial improvements and could be cost effective or even cost saving. This Italian example illustrates a systematic approach and stepwise process that can be easily followed or adapted by other European countries or stakeholders.

Cost-effectiveness analyses are valuable tools to determine an optimal screening intervention by comparing it with alternative strategies. Breast cancer screening affects the length of life (mortality) as well as quality of life (morbidity). To date, the best way to combine mortality and morbidity in one single unit is using health-adjusted life years, an umbrella term for quality adjusted life years (QALY) and disability adjusted life years (DALY). To use cost-effectiveness analyses more effectively to inform health policy, in Chapter 7 we evaluated the impact of using DALYs versus QALYs on outcomes and optimal breast cancer screening strategies. We questioned if QALY and DALY were interchangeable measures of health. We found that using DALYs averted instead of QALYs gained to assess the effects on quality of life from breast cancer screening in the Netherlands yields differences in cost-effectiveness. However, these differences were only subtle, with nearly the same strategies on the efficient frontiers, and no effect on the conclusions of the cost-effectiveness analyses. Since the relationship between these two measures remained constant through all analyses, we conclude that both can be used to determine the optimal breast cancer screening strategy.

CONCLUSIONS

Based on the results of the studies described in this thesis, we derived the following conclusions:
- Several methodologically appropriate approaches exist that are able to capture the true beneficial effect of mammographic screening.
- The reduction in breast cancer mortality due to breast cancer screening in attenders versus non-attenders ranged between 12%-58% in Europe.
- Yearly, 21,680 breast cancer deaths have already been prevented due to mammography screening. The number of breast cancer deaths could be further reduced substantially (by more than 12,000 annually) in all European regions, if 100% of the eligible women aged 50-69 in Europe would be screened every two years.
- Starting screening women at 45 years, five years prior to the currently recommended age-range of 50-69 years, could improve the ratio of overdiagnosed breast cancer to breast cancer deaths averted.
- The major barriers to the breast cancer screening programs in (Southern) Italy could be overcome by initiating feasible changes leading to better long-term outcomes. The employed online tools can be an effective and reliable resource for European policymakers aiming for informed decision-making on cancer screening in their country.
- DALY and QALY can both be used to determine the optimal breast cancer screening strategy.