1 Supportive Research Unit and 3 Oncology Department, Virgen del Rocío Hospital, Seville; 2 Benacazón Health Center, Seville; 4 Department of Social Sciences, University of Seville, Andalucia, Spain
Received 13 May 2003; revised 31 October 2003; accepted 2 January 2004
![]() |
ABSTRACT |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
The aim of this paper is to analyse breast cancer mortality in Andalucia (Spain) between 1975 and 1999 based on ageperiodcohort.
Patients and methods:
Mortality data were obtained from the Mortality Registry of Andalucia. Deaths and population were divided into 13 age groups and five 5-year periods. From this, age-specific mortality rates for 17 birth cohorts were computed. These were plotted and fitted to Poisson regression models to assess age, period and cohort effects.
Results:
The best fit was found for the complete model, which simultaneously considered the effects of ageperiodcohort. Cohort effects were found to be more important than period effects in terms of model fit.
Conclusion:
These effects were manifest as a seemingly consistent increase in the relative risk of breast cancer mortality with a three-fold increase in women born in the 1950s relative to those born in the 1890s.
Key words: Andalucia, breast cancer, epidemiology, mortality
![]() |
Introduction |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Although overall age-standardized breast cancer mortality rates declined by 7% in Europe between 1988 and 1996 [2], breast cancer is the leading cause of death from cancer in women (17%) [3].
In Spain there is no evidence of a decline or levelling off of mortality in recent birth cohorts or in recent years [1, 4]. In a previous report we analysed trends in mortality from breast cancer in Andalucia (Spain) during the period 19751992, showing that overall age-standardized breast cancer mortality rates increased (2% per year) [5]. In this study the trend of breast cancer mortality could be explained by changes in factors that act around the time of death (period effect) and by risk factors that are present in early life (cohort effect). To provide updated information about breast cancer mortality in Andalucia, we have now considered death rates between 1975 and 1999 and applied an ageperiodcohort model in order to separate the effects of age, period at death and cohort of birth on breast cancer mortality.
![]() |
Patients and methods |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Parallel population groups were constructed using mid-year official population estimates. From these data, age-specific mortality rates were calculated for each birth cohort and plotted for the purposes of graphical presentation. Age-adjusted breast cancer mortality rates were calculated by the direct method, using the world standard population as the standard [6].
Evaluation of the effects of age, birth cohort and period of death was performed by means of a log-linear Poisson model, fitted using the General Linear Interactive Modelling (GLIM) software with appropriate macros. The predictive variables (age, cohort and period) were sequentially included in the model. Age was considered first since breast cancer mortality increases markedly with age. In the next step, three two-factor age + period, age + cohort and cohort + period models were considered. In the third model, values for cohort and period were derived after estimating the age effect alone. Age effects are expressed in terms of rate per 100 000 population. Cohort of birth and period of death are expressed as relative values compared to their weighted average (i.e. as estimators of relative risk for each cohort and period). The final model was defined by minimizing the sum of the Euclidean distances between the three two-factor models [79]. The resulting adjusted values for this model were then plotted for presentation. Goodness-of-fit was evaluated by comparing the deviance for each model with that for the age model.
![]() |
Results |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
|
|
Figure 3 depicts the graphs for the age, period and cohort effects. The age effect grows exponentially, but then dips around the age of 50 years before continuing to rise once more. Risk by birth cohort shows an increase from the 1890 generation up to the 1960 cohort and then a decline in women born after 1960. Period of death values also tended to decline.
|
![]() |
Discussion |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
The age effect grows exponentially, but then dips around the age of 50 years before continuing to rise once more. This phenomenon, known as the Clemmensen hook, has been observed in different countries with reference to both incidence and mortality, and is interpreted as the overlapping of two curves corresponding to pre- and post-menopausal tumours, respectively.
The period effect shows a levelling off and decline in breast cancer mortality in recent years. This suggests that the decline might be due to mass phenomena that affected the Andalucian population and exerted their effects with a short latency time. Most prominent among the possible candidates for these phenomena are changes in the health-care system and quality of mortality statistics. Changes in how cause of death is registered and coded can influence the cause-specific mortality rates that are reported. The quality of the Spanish cause of death statistics is similar to that of the USA and some European countries [10].
Health-care services reform in Andalucia has been accompanied by major improvements in the health-care system. There has been an increase in the number of physicians per 100 000 population and in the number of primary health-care centres. The use of mammography increased sharply in the mid-1980s, and a breast cancer screening programme was introduced in 1995. All women aged 5065 years are invited for screening every 2 years. Evidence from the programme itself indicates that the increase in activity was gradual (the screening programme covered 58.7% of the target population in 1999) and that the prevalent round of screening will not be completed until the year 2002.
Breast cancer mortality began falling in Andalucia around 1993, before the breast screening programme was introduced. In Andalucian women in the target age group (5065 years), breast cancer age-adjusted mortality rates fell by 4.3% during the periods 19901994 and 19951999. This may be the result of screening, although greater falls in mortality also took place in younger women (mortality in 2549 year olds fell by 14.3% between the periods 19901994 and 19951999). It is suggested that primary medicine could play a fundamental role in making women more aware of breast cancer, regardless of early detection campaigns [11]. The impact of mammographic screening during the present study is thus likely to be small.
There has been considerable recent debate over whether the reduction in mortality from breast cancer can be attributed to improvements in treatment or screening. Our research indicates that for the period 19931999, improvements in treatment have probably played a major role in reducing breast cancer mortality in Andalucia. The fact that the decline in mortality occurred (in the 1990s) later than the increase in survival (reported in the 1980s) [12] reflects the relatively long life expectancy of women dying of breast cancer, who are known to survive on average 78 years after diagnosis [13].
Risk by birth cohort shows an increase from the 1890 generation to the 1960 cohort and then declines in women born after 1960. This finding must be interpreted carefully. However, cohort values for the most recent periods are based on fewer age-specific rates and fewer deaths (this is based solely on two points) and are therefore less reliable than central cohort values. Trends in birth cohort effects usually reflect risk factor trends, so this decline is surprising because trends in most known or suspected risk factors would seem to predict an increasing breast cancer risk in these cohorts.
![]() |
FOOTNOTES |
---|
![]() |
REFERENCES |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
2. 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]
3. Bray F, Sankila R, Ferlay J, Parkin DM. Estimates of cancer incidence and mortality in Europe in 1995. Eur J Cancer 2002; 38: 99166.[CrossRef][ISI][Medline]
4. La Vecchia C, Negri E, Levi F et al. Cancer mortality in Europe: effects of age, cohort of birth and period of death. Eur J Cancer 1998; 34: 118141.[CrossRef][ISI][Medline]
5. Cayuela A, Rodríguez S. Análisis de la mortalidad en Andalucía 19751992. Sevilla: Instituto de Estadística de Andalucía 2000.
6. Waterhouse J, Muir CS, Shanmugaratnam K, Powell J. Cancer Incidence in Five Continents, Vol. IV. IARC Scientific Publication no. 42. Lyon: IARC 1982.
7. Decarli A, La Vecchia C. Age, period and cohort models: review of knowledge and implementation in GLIM. Rev Stat App 1987; 20: 397409.
8. Osmond C, Gardner MJ. Age, period and cohort models applied to cancer mortality rates. Stat Med 1982; 1: 245259.[Medline]
9. Gardner MJ, Osmond C. Interpretation of time trends in disease rates in the presence of generation effects. Stat Med 1984; 3: 113130.[ISI][Medline]
10. Fernandez E, Gonzalez JR, Borras JM et al. Recent decline in cancer mortality in Catalonia (Spain). A joinpoint regression analysis. Eur J Cancer 2001; 37: 22222228.[CrossRef][ISI][Medline]
11. Prieto F, Serra Ll, La Vecchia C et al. Mortality trends and past and current dietary factors of breast cancer in Spain. Eur J Epidemiol 1996; 12: 141148.[ISI][Medline]
12. García F, Bolumar F, Peris R. Quality of death certificates in Valencia, Spain. Am J Public Health 1989; 79: 13521354.[Abstract]
13. Villar A, Candel MT, Delgado F et al. Extensión del cáncer de mama en el momento del diagnostico inicial. Med Clin (Barc) 1992; 98: 285289.[ISI][Medline]
14. Sant M, Capocaccia R, Verdecchia A et al. Survival of women with breast cancer in Europe: variation with age, year of diagnosis and country. Int J Cancer 1998; 77: 679683.[CrossRef][ISI][Medline]
15. Sant M, Capocaccia R, Coleman MP et al. Cancer survival increases in Europe, but international differences remain wide. Eur J Cancer 2001; 37: 16591667.[CrossRef][ISI][Medline]