Project

Back to overview

Development of statistical methods for analysing geographical disparities, trends and projections in gender-related cancers in Switzerland

English title Development of statistical methods for analysing geographical disparities, trends and projections in gender-related cancers in Switzerland
Applicant Vounatsou Penelope
Number 135769
Funding scheme Project funding (Div. I-III)
Research institution Swiss Tropical and Public Health Institute
Institution of higher education University of Basel - BS
Main discipline Methods of Epidemiology and Preventive Medicine
Start/End 01.09.2011 - 31.08.2014
Approved amount 332'735.87
Show all

All Disciplines (2)

Discipline
Methods of Epidemiology and Preventive Medicine
Cancer

Keywords (8)

Prostate Cancer; Breast cancer; Spatio-temporal models; Mortality; Age-period-cohort models; Incidence; Bayesian methods; Markov chain Monte Carlo

Lay Summary (English)

Lead
Lay summary

Background In Switzerland, breast and prostate cancers are the two most frequent diagnosed cancers in women and men, respectively. Moreover, breast cancer is the leading cause of cancer-related death and the leading single cause of premature mortality in Swiss women. Prostate cancer comes second to lung cancer mortality in men. Both cancers, share other characteristics e.g. they can be detected at an early stage with screening procedures. Important regional disparities in early diagnosis and in patterns of care in breast cancer have been recently described. There is also evidence for an impact of socioeconomic status on prostate cancer diagnosis, treatment and prognosis.. With the aim of reducing disparities quality assured breast screening programs have been started or are in planning in several cantons. Maps of geographical patterns and trends of incidence, mortality and management of gender-related cancers (i.e. breast, ovarian, cervix and uterus in women and prostate and testis in men) as well as estimates of the future disease burden at different regional scales will be important for the design, implementation and evaluation of cancer programs. Over the last fifteen years, Bayesian spatio-temporal models are the state of art methodology in disease mapping, however they have not yet been applied to Swiss cancer data. These data are available from several independent sources: Federal Statistical Office (mortality data from death certificates, diagnosis and procedures of hospitalised patients, national health survey and census) and regional Cancer Registries (cancer incidence and survival). To our knowledge the wealth of information provided in these databases have not been fully explored to assess space-time patterns and trends of breast, prostate and other gender-related cancers at different administrative levels, for control and health planning purposes. These analyses require development of statistical methods to take into account data characteristics such as lack of annual population data at small area level (i.e. municipality)  and incomplete registration of cancer cases.

Objectives and Goals  The main objectives of this research are to (i) assess spatio-temporal patterns and produce smoothed maps of age-specific patterns of  gender-related cancer mortality since 1969 (ii) explore differences in cancer mortality rates between linguistic regions, urbanisation, screening patterns and affluence iii) identify regional differences and trends in cancer management especially surgical procedures and explore their relation to possible socioeconomic disparities and screening patterns (iv) project gender-related cancer mortality by municipality for 2012-2021 (v) assess spatio-temporal patterns and project gender-related cancer incidence for the next ten years.

Methods of investigation We propose to accomplish these objectives by employing and further developing Bayesian Poisson and negative binomial (a) spatio-temporal models to describe disparities in mortality, incidence and care patterns (b) age-period-cohort models with spatial and temporal random effects to forecast geographical patterns and trends of mortality and incidence (c) back-calculation models to estimate incidence from mortality data and (d) time-series models which project and estimate inter-censual small area populations. The models will be fitted using Markov chain Monte Carlo (MCMC) simulation algorithms and they will be used to analyse cancer (a) mortality data extracted from the Federal Statistical Office mortality database; (b) incidence data extracted from the databases of Cancer Registries (c) hospital-based data with national coverage from the medical hospital statistics for case management procedures (d) socioeconomic indicators for the small area region extracted from the Federal Statistical Office census database and (e) self reported use of screening procedures by region from the national health survey database.

Time scale September 2011 to August 2014

Significance This research will contribute with (i) Bayesian statistical methodology for studying spatio-temporal patterns and projections of cancer incidence and mortality driven by data availability and characteristics in Switzerland. These models will enable (a) estimation of incidence rates of the disease in the parts of the population which are not covered by cancer registries and (b) inter-censual small area population estimates and projections that are needed for the calculation of incidence and mortality rates; (ii) smoothed maps of age specific patterns of gender-related cancer mortality and morbidity over time. These maps will identify discrepancies of disease burden and assist implementation and evaluation of cancer programs; (iii) a better insight into the differences in cancer mortality rates between linguistic regions, urbanisation, affluence and cancer management procedures and (iv) estimates of the geographical patterns gender-related cancer mortality and morbidity for the next 10 years: This information will be useful for planning future demand and resource allocation for early detection, diagnosis and treatment.
Direct link to Lay Summary Last update: 21.02.2013

Responsible applicant and co-applicants

Employees

Publications

Publication
Screening is associated with lower mastectomy rates in Switzerland
Herrmann C., Ess S., Walser E., Frick H., Thürlimann B., Probst-Hensch N., Rothermundt C., Mousavi M., Morant R., Vounatsou P. (2019), Screening is associated with lower mastectomy rates in Switzerland, in The Breast, 44, S47-S47.
Impact of mammography screening programmes on breast cancer mortality in Switzerland, a country with different regional screening policies
Herrmann C, Vounatsou P, Thürlimann B, Probst-Hensch N, Rothermundt C, Ess S (2018), Impact of mammography screening programmes on breast cancer mortality in Switzerland, a country with different regional screening policies, in BMJ Open, 8(3):e017806, 1-7.
Mortality atlas of the main causes of death in Switzerland, 2008-2012.
Chammartin Frédérique, Probst-Hensch Nicole, Utzinger Jürg, Vounatsou Penelope (2016), Mortality atlas of the main causes of death in Switzerland, 2008-2012., in Swiss medical weekly, 146, 14280-14280.
Using lung cancer mortality to indirectly approximate smoking patterns in space.
Jürgens Verena, Ess Silvia, Schwenkglenks Matthias, Cerny Thomas, Vounatsou Penelope (2015), Using lung cancer mortality to indirectly approximate smoking patterns in space., in Spatial and spatio-temporal epidemiology, 14-15, 23-31.
40 years of progress in female cancer death risk: a Bayesian spatio-temporal mapping analysis in Switzerland.
Herrmann Christian, Ess Silvia, Thürlimann Beat, Probst-Hensch Nicole, Vounatsou Penelope (2015), 40 years of progress in female cancer death risk: a Bayesian spatio-temporal mapping analysis in Switzerland., in BMC cancer, 15, 666-666.
A Bayesian generalized age-period-cohort power model for cancer projections
Jürgens Verena, Ess Silvia, Cerny Thomas, Vounatsou Penelope (2014), A Bayesian generalized age-period-cohort power model for cancer projections, in Statistics in Medicine, 33(26), 4627-4636.
Cancer survivors in Switzerland: a rapidly growing population to care for
Herrmann C, Cerny T, Savidan A, Vounatsou P, Konzelmann I, Bouchardy C, Frick H, Ess S (2013), Cancer survivors in Switzerland: a rapidly growing population to care for, in BMC Cancer, (13:287), 1-7.

Collaboration

Group / person Country
Types of collaboration
Department of Statistics, University of Florence, Italy Italy (Europe)
- in-depth/constructive exchanges on approaches, methods or results
Cancer Registry Grison-Glaris Switzerland (Europe)
- Publication
Duke University Institute of Statistics and Decision Sciences United States of America (North America)
- in-depth/constructive exchanges on approaches, methods or results

Communication with the public

Communication Title Media Place Year

Associated projects

Number Title Start Funding scheme
118379 Development of spatial statistical methodology for the analysis of health demographic surveillance system (DSS) data 01.10.2007 Project funding (Div. I-III)
118379 Development of spatial statistical methodology for the analysis of health demographic surveillance system (DSS) data 01.10.2007 Project funding (Div. I-III)

Abstract

2.1. Summary of the research planBackground: In Switzerland, breast and prostate cancers are the two most frequent diagnosed cancers in women and men, respectively. Moreover, breast cancer is the leading cause of cancer-related death and the leading single cause of premature mortality in Swiss women. Prostate cancer comes second to lung cancer mortality in men. Both cancers, share other characteristics e.g. they can be detected at an early stage with screening procedures. Important regional disparities in early diagnosis and in patterns of care in breast cancer have been recently described. There is also evidence for an impact of socioeconomic status on prostate cancer diagnosis, treatment and prognosis.. With the aim of reducing disparities quality assured breast screening programs have been started or are in planning in several cantons. Maps of geographical patterns and trends of incidence, mortality and management of gender-related cancers (i.e. breast, ovarian, cervix and uterus in women and prostate and testis in men) as well as estimates of the future disease burden at different regional scales will be important for the design, implementation and evaluation of cancer programs. Over the last fifteen years, Bayesian spatio-temporal models are the state of art methodology in disease mapping, however they have not yet been applied to Swiss cancer data. These data are available from several independent sources: Federal Statistical Office (mortality data from death certificates, diagnosis and procedures of hospitalised patients, national health survey and census) and regional Cancer Registries (cancer incidence and survival). To our knowledge the wealth of information provided in these databases have not been fully explored to assess space-time patterns and trends of breast, prostate and other gender-related cancers at different administrative levels, for control and health planning purposes. These analyses require development of statistical methods to take into account data characteristics such as lack of annual population data at small area level (i.e. municipality) and incomplete registration of cancer cases.Objectives and Goals: The main objectives of this research are to (i) assess spatio-temporal patterns and produce smoothed maps of age-specific patterns of gender-related cancer mortality since 1969 (ii) explore differences in cancer mortality rates between linguistic regions, urbanisation, screening patterns and affluence iii) identify regional differences and trends in cancer management especially surgical procedures and explore their relation to possible socioeconomic disparities and screening patterns (iv) project gender-related cancer mortality by municipality for 2012-2021 (v) assess spatio-temporal patterns and project gender-related cancer incidence for the next ten years. Methods of investigation: We propose to accomplish these objectives by employing and further developing Bayesian Poisson and negative binomial (a) spatio-temporal models to describe disparities in mortality, incidence and care patterns (b) age-period-cohort models with spatial and temporal random effects to forecast geographical patterns and trends of mortality and incidence (c) back-calculation models to estimate incidence from mortality data and (d) time-series models which project and estimate inter-censual small area populations. The models will be fitted using Markov chain Monte Carlo (MCMC) simulation algorithms and they will be used to analyse cancer (a) mortality data extracted from the Federal Statistical Office mortality database; (b) incidence data extracted from the databases of Cancer Registries (c) hospital-based data with national coverage from the medical hospital statistics for case management procedures (d) socioeconomic indicators for the small area region extracted from the Federal Statistical Office census database and (e) self reported use of screening procedures by region from the national health survey database.Time scale: April 2011 to March 2014Significance: This research will contribute with (i) Bayesian statistical methodology for studying spatio-temporal patterns and projections of cancer incidence and mortality driven by data availability and characteristics in Switzerland. These models will enable (a) estimation of incidence rates of the disease in the parts of the population which are not covered by cancer registries and (b) inter-censual small area population estimates and projections that are needed for the calculation of incidence and mortality rates; (ii) smoothed maps of age specific patterns of gender-related cancer mortality and morbidity over time. These maps will identify discrepancies of disease burden and assist implementation and evaluation of cancer programs; (iii) a better insight into the differences in cancer mortality rates between linguistic regions, urbanisation, affluence and cancer management procedures and (iv) estimates of the geographical patterns gender-related cancer mortality and morbidity for the next 10 years: This information will be useful for planning future demand and resource allocation for early detection, diagnosis and treatment.
-