1 Krebsregister des Kantons Zürich (formerly Krebsregister St Gallen-Appenzell), Zurich; 2 Centre de Coordination ASRT (Association Suisse des Registres des Tumeurs); 3 Registre Genevois des Tumeurs, Geneva; 4 Registro Ticinese dei Tumori, Locarno; 5 Kantonales Krebsregister Graubünden, Choir; 6 Krebsregister beider Basel; 7 Zentrale Informatikdienststelle Basel Stadt, Basel; 8 Registre Valaisan des Tumeurs, Sion, Switzerland
* Correspondence to: Dr J.-M. Lutz, Centre de Coordination de l'ASRT, c/o Registre Genevois des Tumeurs, 55 Boulevard de la Cluse, 1205 Geneva, Switzerland. Tel: +41-22-379-4950; Fax: +41-22-379-4971; E-mail: jean-michel.lutz{at}imsp.unige.ch
![]() |
Abstract |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Methods: Survival of 11 376 cases of primary invasive female breast cancer diagnosed between 1988 and 1997 and registered in seven Swiss cancer registries covering a population of 3.5 million was analysed.
Results: Comparing the two periods 19881992 and 19931997, age-standardized 5 year relative survival improved globally from 77% to 81%. Furthermore, multivariate analysis adjusting for age, tumour size and nodal involvement identified regional survival differences. Survival was lowest in the rural parts of German-speaking eastern Switzerland and highest in urbanised regions of the Latin- and German-speaking northwestern parts of the country.
Conclusions: This study confirms that survival differences are present even in a small and affluent, but culturally diverse, country like Switzerland, raising the issue of heterogeneity in access to care and quality of treatment.
Key words: breast cancer, cancer registry, population-based, survival differences, Switzerland
![]() |
Introduction |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Survival after diagnosis of cancer is one of the major outcome measurements and key criteria for assessing quality of cancer control related to both the preventive (early detection) and the therapeutic level. Using data from cancer registries allows population-based comparisons.
EUROCARE studies have been a milestone in population-based research on survival after cancer diagnosis in Europe. Major differences have been found in different countries and time periods within Europe [5,11
,12
]. Data from two Swiss areas, Basel and Geneva, were included in these studies. The survival rates estimated in these two regions were favourable compared with most other European regions. However, these two Swiss registries cover highly urbanised populations and therefore are not representative of the Swiss population overall.
There are marked cultural and geographical differences within Switzerland. In addition, organisation and provision of health care remains the responsibility of the cantons to a major extent and thus varies between them. This is also reflected in the cantonal differences of per capita health care expenditure. Earlier analyses have shown clear differences in incidence (Association of Swiss Cancer Registries: www.asrt.ch) and mortality for many cancers between the various regions of Switzerland [13]. Furthermore, a previous exploratory analysis of breast cancer in two Swiss German-speaking cantons suggested the existence of survival differences [14
]. In the present study we extended the investigation to other regions of the country by including data from patients diagnosed between 1988 and 1997 in seven population-based cancer registries.
![]() |
Materials and methods |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
The data were merged into a central database after a number of quality controls for validity and accuracy. These quality controls included routine plausibility checks on diagnosis, morphology, topography, age, sex, dates and checks on completeness, as well as controls on compatibility of the variables used for staging. All participating registries contribute their data to the International Agency of Research on Cancer (IARC) and fulfil their quality controls. In cases with multiple tumours, only the first diagnosed malignant tumour was included. Bilateral synchronous breast cancers were counted as a single tumour. Cases known from their death certificate only (DCO) and cases detected at autopsy were excluded.
Six of the seven registries have collected information on stages since 1993, all but one of those by active collection of information through mailing of questionnaires to the treating physicians or consultation of patient records. All used UICC rules (fourth or fifth edition). Stage data were grouped into five EUROCARE categories : stage 1 = T1, N0, M0; stage 2 = T23, N0, M0; stage 3 = T13, N1, M0; stage 4 = T4, any N, M0; stage 5 = any T, any N, M1; stage 6 = unspecified stage. Data on the number of lymph nodes examined were available since 1995 from all registries. The accuracy of such information was checked by local staff and, in addition, a standard procedure for checking validity and concordance was conducted centrally for all data.
The Zürich registry was unable to provide data on metastatic status at diagnosis. For this reason, Cox modelling analyses for inter-cantonal comparisons of survival were adjusted for age (continuous), tumour size (three categories: T01, T24, unknown), nodal involvement (three categories: N0, N1, unknown) and number of nodes examined (five categories : zero, 14, 59, 1014, 15+), but not for complete stage including metastatic status. However, we used the EUROCARE five-category staging to compare frequencies of stage distribution by age across all registries except Zürich.
Since information about the specific cause of death was not available for all registries, the effects of mortality from competing causes were taken into account by computing relative survival rates, a net survival measure representing survival in the absence of other causes of death. This is the preferred method for analysing the survival of cancer patients in population-based studies. This survival probability is calculated using general mortality tables related to each study period and each canton (one table per year and per canton) provided by the Swiss Federal Statistical Office.
Age-standardised relative survival rates were calculated taking the age distribution of the EUROCARE-2 Study breast cancer population as the reference [12]. The confidence intervals (CIs) were calculated on the basis of likelihood ratio statistics.
In all participating registry regions there was at least one specialised radiotherapy service unit in operation in the mid-1990s. Oncology services were available in all regions except Valais, where the unit came into operation only after the period of this study. The nearest available hospital-based oncology service was about 100150 km distant in Lausanne. However, there were three free practising oncologists available in the canton of Valais during the study period. The centres in some of the more rural cantons (GraubundenGlarus, St GallAppenzell, Valais, Ticino) are widespread, and it may take longer to reach a centre, even if available, than in urban regions (Geneva, Basel, Zürich).
![]() |
Results |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Age at diagnosis was not equally distributed between the registries (P < 0.01 for the whole period). During the first period (19881992) the differences were statistically significant for the youngest (1544 years) and oldest (65 years) age groups only, but not for women aged 4564 years. During the second period (19931997/8), the differences were statistically significant for all age groups (P < 0.05) except for 1544 years. The cantons of St GallAppenzell (28%), GraubundenGlarus (24%), Ticino (24%) and Basel (23%) had the highest proportions of cases aged >75 years (Table 1).
|
|
Table 3 shows the geographical distribution of T stage for all age groups combined, and Table 4 shows the stage distribution according to EUROCARE stage. Overall, stages were not equally distributed between the registries during the period 19931997 (P < 0.05). For instance, Valais (12.4%) and Basel (8.4%) had more frequent stage 4 than GraubundenGlarus (5.6%) and St GallAppenzell (5.8%) (P < 0.01). This heterogeneity was also observed in age-specific figures. For example, in the age groups 5564 years and 6574 years, stage 1 was less frequent in St GallAppenzell (19%, 18%), Valais (27%, 25%) and GraubundenGlarus (31%, 25%) than in Basel (37%, 29%), Ticino (37%, 28%) and Geneva (40%, 39%) (P < 0.01), possibly reflecting differences in screening activities. In the age group 75 years, stages 2, 4 and 5 differed statistically significantly between the six registries.
|
|
|
|
|
|
|
|
![]() |
Discussion |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Our observation is in line with previous data from Europe [24
] and with data from other countries, i.e. Italy, where major cultural differences also exist [4
]. A difference in survival was observed between the south and other parts of the country for many tumours, including breast cancer. Additional analyses of time trends by region also showed that there was less improvement in survival in southern parts of the country. In the EUROCARE study, lower survival rates had been found in Denmark than in Sweden, which are neighbouring Nordic countries. Jensen et al. [15
] compared survival of breast cancer patients from Aarhus in Denmark and Malmö in southern Sweden in a population-based historical follow-up study of patients diagnosed between 1983 and 1989. Even after adjustment for several possible explanatory factors (age, tumour size, intensity of lymph node examination, regional and distant spread), as well as use of mammography screening, a clear difference in breast cancer survival between these two areas persisted. Differences in health care and treatment modalities may explain some of these remaining differences in two otherwise quite comparable areas. In a later comparison between Danish and Swedish patient cohorts diagnosed between 1996 and 1997 the between-country difference tended to disappear [16
]. Survival of breast cancer patients in populations from 99 health authority regions in England was compared in a study based on cancer registry data [17
]. Even after taking socio-economic differences between the regions into account, a statistically significant variation in breast cancer survival between health authorities in England remained. No tumour-specific prognostic factors were taken into account in this study. Differences in survival by district (health boards) were also investigated in patients from Scotland diagnosed in 1987, and the authors found statistically significantly different survival (based on disease-specific deaths) in some districts even after adjustment for clinical factors such as tumour stage [18
].
Many factors may influence survival, including socio-economic status, degree of urbanisation, cultural differences, patient behaviour, physicians' attitudes, treatment, and health care system [1922
]. While the available data do not allow identification of specific reasons for the observed survival differences, the observed tendency of a lower proportion of early-stage cases in regions with less favourable survival rates implicates differences in secondary prevention across cantons as one potential source.
Populations from various cultural backgrounds as well as from urban and rural regions are included in this study. Zürich and Basel are highly urbanised mainly German-speaking areas, St GallAppenzell and GraubundenGlarus are rural and German-speaking (in Graubunden a large part of the population is bicultural, Rumantsch and German, and a minority are Italian-speaking). Geneva is highly urbanised and French-speaking, Ticino is partly urban with an Italian background and Valais is rural with a mixed cultural background, but is predominantly French-speaking. These cultural differences are known to have an influence on the attitude to health problems in the population as well as in health care professionals (e.g. towards early detection) [23]. There were no mammography screening programmes during the study period. However, spontaneous screening regularly increased and official programmes started in 1998 in the French-speaking part of Switzerland, whereas no such programmes have been instituted so far in the German part. The prevalence of women ever having undergone a mammography differs substantially between the Latin- and German-speaking parts of Switzerland [23
]. These differences in behaviour could also explain why breast cancer incidence is higher in the western cantons than in the eastern cantons of Switzerland: age-standardised incidence (European) between 85 per 100 000 in St GallAppenzell and 127 per 100 000 in Geneva in the period 19931997. Trends in mortality (European standardised rates) over the past 20 years are also clearly different in these two cantons: from 29 to 27 per 100 000 in St GallAppenzell, and from 40 per 100 000 to 24 per 100 000 in Geneva.
The strength of this study is its population-based case ascertainment through cancer registries, minimising the potential for selection bias as observed for hospital-based survival data and for clinical studies. The results of our study are representative of the whole population. Methodological biases and possible confounders affecting population-based survival comparisons have been discussed in depth elsewhere[11,12
,24
], and some of these factors are briefly mentioned below.
Incomplete registration or coding mistakes, which can bias the survival estimates due to patient selection, is not a likely source of bias in this study. Incidence data of all participating registries have been approved by the IARC. The proportion of cases known through death certificate only is <2%, and the proportion of microscopically verified cases is 97% in all registries. In addition, cases known to registries through death certificates only (DCO) were excluded from the analysis.
Another source of bias is incomplete follow-up. All registries routinely used active follow-up, and specific central checking for follow-up has been carried out for this study. Relative survival in Geneva may have been slightly overestimated because of the migration bias described elsewhere [25], but is not likely to explain the size of the observed survival differences.
As survival continuously improved with time in all cantons, there could be an earlier or faster downstaging in some cantons. Although proportions of cases with negative lymph nodes at diagnosis did not differ much during the two periods (49% compared with 48%, overall), frequency of early stage did, especially for registries with poor staging during the first period. For instance, frequency of stage 1 (all ages combined) changed from 9% to 20% in St GallAppenzell and from 14% to 27% in Valais, but only from 32% to 35% in Geneva and from 24% to 28% in Basel, and was unchanged in GraubundenGlarus (28%). During the same period the proportion of unknown stage decreased from 57% to 12% in St GallAppenzell and from 6% to 5% in Geneva, and was unchanged in Basel (6%). Therefore the simple hypothesis of a true and rapid downstaging cannot be confirmed; we only observe a higher quality of casenotes with a better systematic staging.
Another possible explanation is the quality of staging assessment (e.g. metastasis status was not available in Zürich). Unfortunately, the group of cantons with large sampling (Basel, St GallAppenzell, Zürich and GraubundenGlarus) have opposite results in survival with Basel and Zürich different from St GallAppenzell and GraubundenGlarus. In addition, Cox modelling clearly showed that, even after adjustment for tumour size, nodal involvement and number of nodes sampled, the differences still persisted.
As mentioned above, many factors may influence survival including differences in patient behaviour due to different cultural background or socioeconomic status, differences in physicians' attitudes, treatments and access to health care. The intercantonal differences in survival could be related to a combination of such factors. The respective contributions of these factors cannot currently be measured by data routinely collected by all registries in Switzerland, but is the focus of ongoing participation by Switzerland in the EUROCARE 4 and CONCORD projects. (http://www.eurocare.it/ and http://www.lshtm.ac.uk/ncdeu/cancersurvival/concord/index.htm).
In conclusion, we found a global improvement of survival after diagnosis of breast cancer over a relatively short time period. We also found differences in survival in different regions of Switzerland, with a more favourable outcome in urbanised regions and the western part than in the more rural regions of the eastern part.
![]() |
Acknowledgements |
---|
Received for publication March 14, 2005. Revision received June 24, 2005. Accepted for publication August 2, 2005.
![]() |
References |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
2. Sant M, Allemani C, Capocaccia R et al. Stage at diagnosis is a key explanation of differences in breast cancer survival across Europe. Int J Cancer 2003; 106: 416422.[CrossRef][ISI][Medline]
3. Levi F, Lucchini F, Negri E, La Vecchia C. The decline in cancer mortality in the European Union, 19881996. Eur J Cancer 2000; 36: 19651968.[CrossRef][ISI][Medline]
4. Survival of cancer patients in Italy in the nineties: figures from the cancer registries. Epidemiol Prev 2001; 25: S1S375.
5. Sant M, Capocaccia R, Verdecchia A et al. Survival of women with breast cancer in Europe: variation with age, year of diagnosis and country. The EUROCARE Working Group. Int J Cancer 1998; 77: 679683.[CrossRef][ISI][Medline]
6. Coleman MP, Babb P, Damiecki P et al. Cancer Survival Trends in England and Wales, 19711995: Deprivation and NHS Region. London: Stationery Office 1999.
7. Sant M, Allemani C, Berrino F. Breast carcinoma survival in Europe and the United States. Cancer 2004; 100: 715722.[CrossRef][ISI][Medline]
8. Jones AL. Reduction in mortality from breast cancer, screening and increased use of adjuvants are responsible, adjuvants more so. BMJ 2005; 330: 205206.
9. Overgaard M. Radiotherapy as part of a multidisciplinary treatment strategy in early breast cancer. Eur J Cancer 2001; 37 (Suppl 7): S3343.[ISI][Medline]
10. Kingsmore DB, Ssemwogerere A, Hole DJ et al. Increased mortality from breast cancer and inadequate axillary treatment. Breast 2003; 12: 3641.[CrossRef][ISI][Medline]
11. Berrino F, Sant M, Verdecchia A et al. Survival of Cancer Patients in Europe. The EUROCARE Study. IARC Scientific Publication 132. Lyon: International Agency for Research on Cancer 1995.
12. Berrino F, Capocaccia R, Estève J et al. Survival of Cancer Patients in Europe. The EUROCARE-2 Study. IARC Scientific Publication 151. Lyon: International Agency for Research on Cancer 1999.
13. Schüler G, Bopp M. Atlas der Krebsmortalität in der Schweiz 19701990. Basel: Birkhäuser Verlag 1997.
14. Fisch T, Torhorst J, Perret E et al. Results from two cancer registries in Switzerland (Basel and St Gall-Appenzell). Bull Suisse Cancer 2002; 2200: 1517.
15. Jensen AR, Garne JP, Storm HH et al. Does stage at diagnosis explain the difference in survival after breast cancer in Denmark and Sweden? Acta Oncol 2004; 43: 719726.[CrossRef][ISI][Medline]
16. Jensen AR, Garne JP, Storm HH et al. Stage and survival in breast cancer patients in screened and non-screened Danish and Swedish populations. Acta Oncol 2003; 42: 701709.[CrossRef][ISI][Medline]
17. Mullee MA, De Stavola B, Romanengo M et al. Geographical variation in breast cancer survival rates for women diagnosed in England between 1992 and 1994. Br J Cancer 2004; 90: 21532156.[ISI][Medline]
18. Twelves CJ, Thomson CS, Dewar JA et al. Variation of survival in women with breast cancer: Helath board remains a factor at 10 years. Br J Cancer 2001; 85: 637640.[CrossRef][ISI][Medline]
19. Rosso S, Faggiano F, Zanetti R, Costa G. Social class and cancer survival in Turin, Italy. J Epidemiol Community Health 1997; 51: 3034.[Abstract]
20. Coleman MP. Opinion: why the variation in breast cancer survival in Europe? Breast Cancer Res 1999; 1: 2226.[CrossRef][Medline]
21. Schrijvers CTM, Mackenbach JP, Lutz JM et al. Deprivation and survival from breast cancer. Br J Cancer 1995; 71: 738743.[ISI][Medline]
22. Bouchardy C, Rapiti E, Fioretta G. et al. Undertreatment strongly decreases prognosis of breast cancer in elderly women. J Clin Oncol 2003; 21: 35803587.
23. Keller B, Zemp Stutz E, Tibblin M et al. Screening mammographies in Switzerland: what makes female and male physicians prescribe them? Swiss Med Wkly 2001; 131: 311319.[ISI][Medline]
24. Berrino F. The EUROCARE-3 study: strengths, limitations and perspectives of population based, comparative survival studies. Ann Oncol 2003; 14 (Suppl 5): 913.
25. Raymond L, Fischer B, Fioretta G, Bouchardy C. Migration bias in cancer survival rates. J Epidemiol Biostat 1996; 1: 167173.
|