Dutch Breast Cancer Screening Programme Research Articles

League Tables of Breast Cancer Screening Units: Worst-case and Best-case Scenario Ratings Helped in Exposing Real Differences between Performance Ratings

Data on the performance of health boards, hospitals and medical specialists, etc., are being collected at various levels in the health-care system and are often presented as league tables. These tables ignore natural variation and/or confounders, and this introduces uncertainty about their interpretation. The purpose of this study was to devise and illustrate a method to expose the real difference between the ratings in league tables. Two values per rating were added to the league tables: the best-case scenario and the worst-case scenario. True performance will lie somewhere between these two values. The method is illustrated using data from the Dutch breast cancer screening programme. By focusing on one performance indicator and one confounder, it was possible to show shifts in the rating order of breast cancer screening units and thus expose the uncertainty about the true performance of each screening unit. The worst-case and best-case scenario ratings demonstrated the uncertainty within the ratings of a league table. League tables should therefore only be used with great caution and after providing the public with sufficient information.

Pathological and Biological Differences Between Screen-Detected and Interval Ductal Carcinoma in situ of the Breast

BackgroundThe incidence of ductal carcinoma in situ (DCIS) has risen dramatically with the introduction of screening mammography. The aim was to evaluate differences in pathological and biological characteristics between patients with screen-detected and interval DCIS.MethodsFrom January 1992 to December 2001, 128 consecutive patients had been treated for pure DCIS at our institute. From these, 102 had been attending the Dutch breast cancer screening program. Sufficient paraffin-embedded tissue was available in 74 out of the 102 cases to evaluate biological marker expression (Her2/neu, ER, PR, p53 and cyclin D1) on tissue microarrays (TMA group). Differences in clinicopathological characteristics and marker expression between screen-detected and interval patients were evaluated. Screen-detected DCIS was classified as DCIS detected by screening mammography, when the two-year earlier examination failed to reveal an abnormality. Interval patients were classified as patients with DCIS detected within the two-year interval between two subsequent screening rounds.ResultsScreen-detected DCIS was related with linear branching and coarse granular microcalcifications on mammography (p < .001) and with high-grade DCIS according to the Van Nuys classification (p = .025). In univariate analysis, screen-detected DCIS was related with Her2/neu overexpression (odds ratio [OR] = 6.5; 95%CI 1.3–31.0; p = .020), and interval DCIS was associated with low-grade (Van Nuys, OR = 7.3; 95% CI 1.6–33.3; p = .010) and PR positivity (OR = 0.3; 95%CI 0.1–1.0; p = .042). The multivariate analysis displayed an independent relation of Her2/neu overexpression with screen-detected DCIS (OR = 12.8; 95%CI 1.6–104.0; p = .018).ConclusionsThese findings suggest that screen-detected DCIS is biologically more aggressive than interval DCIS and should not be regarded as overdiagnosis.

Cost-effectiveness of different reading and referral strategies in mammography screening in the Netherlands

In mammography screening with double reading, different strategies can be used when the readers give discordant recommendations for referral. We investigated whether the results of the Dutch breast cancer screening programme can be optimised by replacing the standard referral strategy by consensus. Twenty-six screening radiologists independently and blinded to outcome read a test set consisting of previous screening mammograms of 250 cases (screen-detected and interval cancers) and 250 controls. Their referral recommendations were paired and, in case of discrepancy, re-read according to three referral strategies: (1) decision by one of the readers; (2) arbitration by a third reader; (3) referral if both readers agree (consensus). Data allowed studying other referral strategies, including referral if any reader suggests, as well. Double reading with referral if any reader suggests resulted in a 1.03 times higher sensitivity (76.6%) and a 1.31 times higher referral rate (1.26%) than double reading with consensus. To estimate the cost-effectiveness, the outcomes were used in a microsimulation model. Even if double reading with referral if any reader suggests results in four times as high referral rates and an accompanying increase of biopsies or other invasive procedures, the cost-effectiveness of 4,190 Euros per life-year gained may well be in the range of acceptable cost-effectiveness for Dutch health care programmes.

Methodological issues in international comparison of interval breast cancers

International comparisons of interval cancers (IC) are important to better understand the relationship between programmes' performance and screening practices. In this respect, differences in (i) definition, (ii) identification and (iii) quantification of IC have received little attention. To examine these 3 comparability issues and activities involving IC, an assessment was conducted among member countries of the International Breast Cancer Screening Network, and the impact of accuracy of identification and quantification practices was estimated using 1996-98 data from the Dutch breast cancer screening programme. Information was obtained from 19 screening programmes in 18 countries, 16 of which acknowledged the coexistence of opportunistic screening. IC data were collected to evaluate performance of the screening programme (100% of programmes) and the radiologists (89%); 53% of programmes had a designated review process for IC. Most programmes (84%) agreed with the European Guidelines definition of IC, but a case situation exercise evidenced substantial discrepancy in classification of cancers that occurred after a positive screen. Completeness of identification of IC appears to contribute most to international variation, and cannot be easily controlled for in methodologically rigorous comparisons. Statistically significant differences of about 4% were measured between quantification methods for IC. An operational definition of IC is proposed to enhance international comparability. Valid comparisons of IC are possible with careful attention to the definition but true differences in IC frequency across screening programmes should exceed 10% to be possibly indicative of real differences between programmes.

Seventy‐five years is an appropriate upper age limit for population‐based mammography screening

In this study, we assessed the results of the Dutch breast cancer screening programme for women aged 70-75, and discussed the current upper age limit of the women invited. We compared the main outcome parameters of the screening programme 1998-2000 for women aged 70-75, with those in women aged 50-69. Breast cancer detection rates were also compared with prediction from 2 variants of a simulation model of breast cancer screening, assuming the mean sojourn time, i.e., the duration of the preclinical detectable phase, to increase or not with age above the age of 65. The underlying idea is that an increase of the sojourn time with age will lead to a less favourable balance between screening benefits and harms from a certain age on. Of the 315,103 women (aged 70-75) invited, 65.6% participated. The attendance increased from 1998 to 2000. For women aged 74 and 75 years, this increase was almost 10%. As a result of the 187,207 screening examinations performed within this age group, 18.3 per 1,000 women were referred and 10.3 per 1,000 breast cancer were diagnosed. Detection rates in both initial and subsequent screens increased steadily with age and got close to those model-simulated rates, which assume a continuously increasing sojourn time with age. A major finding of this study is that the screening participation among elderly women is high. The outcomes of our study suggest a steadily increasing sojourn time of breast tumours beyond the age of 69, leading to a strong increase in detection of cancers, and therefore, disfavouring the balance with the benefits of screening. At present, 75 years of age can be regarded as an appropriate upper age limit for the Dutch programme.

Women's opinions about attending for breast cancer screening: Stability of cognitive determinants during three rounds of screening

To examine the stability of beliefs and intentions towards repeat attendance at breast cancer screening, using the Theory of Planned Behaviour. The aims of the study were to examine whether and how cognitions changed in the course of the programme, and whether intentions that were assessed proximally were better predictors of behaviour than those assessed distally. A total of 2,657 women filled out a baseline questionnaire (T(1)), 2 months after being invited for an initial mammogram in the Dutch Breast Cancer Screening Programme. Actual attendance data in the second and third screening rounds were subsequently collected and follow-up questionnaires were sent to parts of the sample at four points in time: shortly before (T(2)) and after (T(3)) the second screening round, and shortly before (T(4)) and after (T(5)) the third screening round. Only minor changes in beliefs and intentions were found. In the assessments shortly before screening, women were somewhat less positive about attending than in the assessments shortly after screening. Throughout the course of the programme, women's opinions about attending remained positive. In fact, women became somewhat more convinced that they were vulnerable to getting breast cancer, and that participating in screening was beneficial to them. Actual attendance in subsequent rounds of screening was higher than expected. Proximal beliefs and intentions were only slightly more predictive of actual behaviour than distal beliefs. In organized breast cancer screening, beliefs and intentions remain positive and rather stable. Although our results should be interpreted with caution, due to little variation in behaviour, they suggest that the gap between intentions and behaviour could not be substantially reduced by proximal assessment of determinants.

Decreased rates of advanced breast cancer due to mammography screening in The Netherlands.

The effect of the implementation of the Dutch breast cancer screening programme during 1990–1997 on the incidence rates of breast cancer, particularly advanced breast cancer, was analysed according to stage at diagnosis in seven regions, where no screening took place before 1990. The Netherlands Cancer Registry provided detailed data on breast cancer incidence in 1989–1997 by tumour stage, age and region. Annual age-adjusted incidence rates of all breast cancers and advanced cancers, defined as large tumours T2+ with lymph node and/or distant metastases, were compared with rates in 1989. In general, breast cancer incidence rose strongly in the early 1990s, especially in the age category 50–69 years (estimated annual percentage change (EAPC) 4.25; 95% CI 1.70, 6.86). The increase was mainly due to the increase in small T1 cancers and ductal carcinoma in situ. However, in women aged 50–69, advanced cancer incidence rates showed a significant decline by 12.1% in 1997 compared with 1989 (EAPC –2.14, 95% CI −3.47, −0.80), followed by a breast cancer mortality reduction of similar size after approximately 2 years. We confirm that breast cancer screening initially leads to a temporary strong increase in the breast cancer incidence, which is followed by a significant decrease in advanced diseases in the women invited for screening. It is evident that breast cancer screening contributes to a reduction in advanced breast cancers and breast cancer mortality.

Breast cancer risk after caloric restriction during the 1944-1945 Dutch famine.

Data from animal models suggest that caloric restriction may reduce the risk of breast cancer, although not all dietary regimens produce similar effects. We examined whether caloric restriction imposed as a consequence of the 1944-1945 Dutch famine reduced the risk of breast cancer in women participating in a Dutch breast cancer screening program. Between 1983 and 1986, approximately 15 000 women, aged 2 to 33 years during the 1944-1945 Dutch famine, responded to a questionnaire about their famine experience. A famine exposure score, graded as absent, moderate, or severe exposure, was derived from answers regarding hunger, cold, and weight loss. During follow-up until January 2000, 585 women with incident breast cancer were identified by the regional cancer registry. The relationship between famine exposure and breast cancer risk was assessed by weighted Cox regression models, in which person-years lived for the entire cohort was extrapolated from data from a random sample of 15% of the cohort. Women who were severely exposed to the famine had a statistically significantly higher risk of breast cancer (hazard ratio [HR] = 1.48, 95% confidence interval [CI] = 1.09 to 2.01) than women who were not exposed. Women who were moderately exposed to the famine had a 13% increased risk of breast cancer (HR = 1.13, 95% CI = 0.92 to 1.38; P(trend) = 0.016). The association between famine exposure and breast cancer risk was stronger for women who were exposed between the ages of 2 and 9 years (severely exposed versus unexposed: HR = 2.01, 95% CI = 0.92 to 4.41) than for women who were exposed at older ages. The risk of breast cancer was increased in women who were severely exposed to a short but severe famine decades earlier. This result is compatible with data from the few animal studies investigating effects of short-term, transient caloric restriction.

Prospective Study on the Determinants of Repeat Attendance and Attendance Patterns in Breast Cancer Screening Using the Theory of Planned Behaviour

This prospective study, using the Theory of Planned Behaviour (TPB) as a theoretical framework, was carried out to identify the determinants of repeat attendance and attendance patterns in organised breast screening. A group of 2657 women filled out a baseline questionnaire, approximately 8 weeks after having been invited for an initial screening in the Dutch Breast Cancer Screening Programme. Data on actual attendance in second and third screening round were subsequently collected. Personal variables such as family history of breast cancer, breast cancer in someone close and fear of breast cancer were not related to repeat attendance. The TPB variables could explain approximately 17% of the variance in attendance in the second and third round. Maintenance behaviour (consistent attendance vs. dropout) was related to the TPB variables, but the amount of variance explained was only 6%. Initiation of behaviour (consistent refusal vs. delayed attendance) also was related to TPB variables, with a substantially higher amount of explained variance than with maintenance behaviour. Results indicate that the TPB variables are more related to the initiation of screening behaviour than to the maintenance of screening behaviour. Implications of these findings for breast cancer screening are discussed.

Monitoring women's experiences during three rounds of breast cancer screening: results from a longitudinal study.

(a) To monitor experiences of women during three successive rounds of breast screening; (b) to examine the impact of previous experiences (obtained either immediately after the latest mammogram or shortly before the subsequent one) on reattendance; and (c) to examine which factors are associated with the experience of pain and distress during screening. The Dutch Breast Cancer Screening Programme 2657 women completed a baseline measurement (response rate 67%) about 8 weeks after they had been invited for an initial mammogram (T(1)). Actual participation data of these women in the second and third rounds of screening were collected. Follow up questionnaires were sent to subgroups of the sample at different times: shortly before the second screening (T(2); response rate 86%), shortly after the second screening (T(3); response rate 85%), shortly before the third screening (T(4); response rate 80%), and shortly after the third screening (T(5); response rate 78%). Most women were satisfied with the first screening round and remained positive about subsequent screens. Although pain and anxiety were not uncommon, only a few (10%-15%) experienced moderate or severe levels of distress or pain. Experiences were relatively stable: women who experienced pain in the first screen were more likely to experience pain in subsequent screens (r values from 0.39 to 0.50). Fear of breast cancer was associated with increased distress related to mammography and, to a lesser extent, with increased pain during the mammography. Evidence was found for a relief effect: women were more positive about their previous screen when asked shortly after this screen, than when asked just before the subsequent one. Previous experiences (obtained either proximally or distally) were only slightly predictive for future attendance. Experiences during mammography are fairly stable. Negative experiences were generally not a reason to drop out of the programme.

Does mammographic screening and a negative result affect attitudes towards future breast screening?

To investigate the impact of an experience of a benign mammographic result on intention to seek medical help immediately in the case of breast abnormalities, and on intentions and thoughts about future participation in screening. The Dutch Breast Cancer Screening Programme for women aged 50-69. Subjects were women who were invited for an initial breast examination: 223 women filled out a questionnaire about 10 days before and about 6 weeks after their initial breast examination. To be able to control for possible test effects, another group of 293 women filled out a questionnaire only after mammography. Changes in thoughts and intentions were examined. Most women were very satisfied with the course of their initial breast examination, although pain or discomfort was often mentioned. No clues to suggest false reassurance were found: more than 99% of the women would consider the possibility of breast cancer if they felt a lump in one of their breasts. In such a situation, most women intended to seek medical help within a week. These variables were not influenced by the experience of mammography with a benign result. In general, women were very positive about (repeat) participation, both before and after screening. After screening, the average woman perceived fewer costs in participating, and perceived her own ability to engage in future screening as higher. However, the experience of pain and anxiety during the initial screening did lead to reverse effects. Women who were less satisfied about their treatment by the staff were more likely to change their intentions to reparticipate in a negative way. As, in general, women became more positive about regular participation after they had attended breast cancer screening, efforts to improve first round attendance must be continued. At the same time, the screening organisations must continue to prioritise the high level of client friendliness throughout the screening. No evidence for detrimental effects of screening through false reassurance among participants was found.

A breast cancer screening programme operating in a liberal health care system: the Luxembourg Mammography Programme, 1992-1997.

The national breast cancer screening programme in Luxembourg, the Mammography Programme (MP), was launched in 1992. Its primary goal was to set up an organised breast cancer screening with biennial invitation-reinvitation of women 50-64 years, 2-view mammography, technical quality assurance, double-reading procedures and close monitoring of performance. In 1994, the decision was taken by health authorities to only reimburse screening mammograms done within the MP. We evaluated the performance of 3 screening rounds that took place in 1992-1997. In 1992, 4,815 women had opportunistic screening and 4,790 attended the MP. In 1997, 861 women had opportunistic screening and 8,603 attended the MP. In 1997, the participation rate in the MP reached 50%. The referral rate in the MP was 10% for initial screening and 5% for subsequent screening. Echographic examinations accounted for 71% of referrals. Per 1,000 screened women, biopsy and cancer rates were 18.0 and 8.0 for initial screening and 10.3 and 5.8 for subsequent screening. Twenty-one percent of the breast cancers diagnosed in screened women were interval cancers. The age-adjusted proportion of tumours >20 mm was 56.1% before 1992, 45.1% for women not (yet) screened by the MP, 27.0% for interval cancers, 26.6% for initial screening and 12.0% for subsequent screenings. Tumour size distribution was similar to that observed in the Dutch Breast Cancer Screening Programme, 1990-1993, except that in the latter programme, no difference was observed between initial and subsequent screening. The Luxembourg experience shows that in a liberal health care system, a policy of organising the screening process by reimbursing only those mammograms done within the context of the organised screening programme can drastically reduce opportunistic screening. Further efforts in the Mammography Programme will aim at increasing participation and look at ways to decrease referrals to echography.