Modelling, interpretation and forecasting of cancer incidence and mortality in Europe
Background Previous instances of this project have been monitoring cancer mortality in Europe Since 1992 and have been extremely productive, as shown by the huge number of relevant and highly cited publications. Overall, cancer mortality in Europe (and in Italy) has been declining and is likely to further decrease in the near future. There are, however, cancer sites showing unfavourable trends, and areas of Europe where rates remain comparatively high, requiring further quantitative attention. Hypothesis Monitoring past and predicting (short term) future trends in cancer mortality provides important public health information, and helps measuring success of implemented preventive strategies and identifying priorities for cancer prevention and management. It can also generate hypothesis to be tested in analytical studies, and provide support, or lack thereof, of hypotheses generated elsewhere. Aims To maintain and improve the integrated system of monitoring of cancer mortality in Europe and other areas of the world. Within the project¿s main aim, in 2014-2017 cancer mortality data from recent years will be used in order to promptly identify and interpret early signals of evolution. To further develop and test modelling techniques for interpreting past trends (i.e. age-period-cohort models), and for predicting short term trends, which are of key interest for cancer management on a public health scale. Experimental design Data on number of cancer deaths, incidence and resident populations are obtained from official public access databases. Regular updating of the project¿s core database will be performed. Age-specific and age-standardized rates will be computed and carefully examined, in tabular and graphical form. Joinpoint regression models will be applied and (average) annual percent change computed to help interpret the trends. To disentangle the effects of age, birth cohort and period of death on cancer mortality, an age-period-cohort model will be used, using the bootstrap technique to obtain confidence intervals. A short term prediction model based on logarithmic Poisson count data joinpoint models was also developed. We plan to further test the models and develop new ones to exploit more local and spatially dependent data. Expected results i)age-period-cohort analysis of cancer mortality in Europe for the period 1970-2009, i.e. for the cohorts born between 1885 and 1970; ii) annual prediction of cancer mortality in the EU in the current year; iv) monitoring of the tobacco-related cancer epidemic in Europe, with specific focus on oesophageal and laryngeal cancer; v) developing methods for evaluating the impact of cancer screening on mortality; vi) analysis of cancer burden and trends in older Europeans; vii) in depths analyses for specific cancer sites viii) Further expand reports on specific regional geographical areas with higher risk exposures. A major characteristic and advantage of this project is its extreme flexibility. Besides the planned tasks, high priority will be given to issues arising from the scientific literature or from interactions with colleagues.
- maria domenica CAPPELLINI
- Ente finanziatore:
- MINISTERO DELL'ISTRUZIONE, DELL'UNIVERSITA' E DELLA RICERCA
- SIR2014 - Bando Scientific Independence of young Researchers 2014
- UNIVERSITA' DEGLI STUDI DI MILANO