Expanding the Role of Cancer Registration: from Cancer Incidence and Mortality Monitoring to Evaluation of Cancer Care Delivery

(1)

Thesis

Reference

Expanding the Role of Cancer Registration: from Cancer Incidence and Mortality Monitoring to Evaluation of Cancer Care Delivery

RAPITI AYLWARD, Elisabetta

Abstract

The 20th century saw the establishment of the first population-based cancer registries with the purpose of generating data to describe cancer patterns and trends, and with the ultimate goal of guiding national cancer control planning. In recent years, the roles and activities of cancer registries have been steadily expanding. This dissertation argues that these new roles are now becoming as important and valuable to clinicians, policy makers and public health authorities as the traditional ones. For some issues, these new roles played by cancer registries can address knowledge gaps that cannot be assessed or evaluated using other approaches. The dissertation presents ten papers conducted using the data of the Geneva Cancer Registry that explore these new directions. Together these papers present a strong case for ensuring that national and local authorities build on recent experience to systematically invest in and expand the roles and activities of cancer registries.

RAPITI AYLWARD, Elisabetta. Expanding the Role of Cancer Registration: from Cancer Incidence and Mortality Monitoring to Evaluation of Cancer Care Delivery. Thèse de privat-docent : Univ. Genève, 2014

DOI : 10.13097/archive-ouverte/unige:46937

Available at:

http://archive-ouverte.unige.ch/unige:46937

Disclaimer: layout of this document may differ from the published version.

1 / 1

(2)

Department: Médecine et santé communautaires Service : Registre genevois des tumeurs

EXPANDING THE ROLE OF CANCER REGISTRATION:

FROM CANCER INCIDENCE AND MORTALITY MONITORING TO EVALUATION OF CANCER CARE DELIVERY

Thesis submitted to the Medical School of the University of Geneva

for the degree of Privat-docent by

Elisabetta RAPITI

Geneva, 2014

(3)

(4)

1

SUMMARY

The 20^th century saw the establishment of the first population-based cancer registries with the purpose of generating data to describe cancer patterns and trends and with the ultimate goal of guiding national cancer control planning. Today at least 290 population-based cancer registries exist worldwide. The Geneva Cancer Registry, established in 1970, was one of the first population-based cancer registries in Europe. Since its creation, the Geneva Cancer Registry has given particular importance to expanding and improving the quality of the data collected.

In recent years, the roles and activities of cancer registries have been steadily expanding.

This dissertation argues that these new roles are now becoming as important and valuable to clinicians, policy makers and public health authorities as the traditional ones. The dissertation presents ten papers conducted using the data of the Geneva Cancer Registry that explore these new directions. The first three papers are used to illustrate and discuss three of the main traditional roles played by cancer registries such as monitor cancer trends and burden, project future cancer burden to help plan services, and propose potential risk factors for specific cancers. The subsequent seven papers then explore in more detail the validity and importance of three new roles that cancer registries can play such as monitoring of the effectiveness of cancer treatments at a population level, examining the quality of cancer care, and exploring potential long-term adverse effects of cancer treatments.

The recent expansion of the important roles played by cancer registries has been driven by a number of factors, ranging from the development and availability of improved tools for data collection, management and analysis, through enhanced national and international collaboration, and data sharing between cancer registries and other relevant bodies. In addition to discussing these factors, this dissertation recognizes and explores the limitations to the use of registry data. The reasons for caution when inferring broader and/or definitive implications from some types of studies conducted using cancer registry data include

(7)

4

incomplete information and the presence of inherent biases, both of which are illustrated along with some of the approaches that are used to minimize the impact of such limitations.

While there are limitations to the use of these data, as is the case with most data, these are outweighed by the value of building on recent experience to close gaps in cancer knowledge that are difficult if not impossible to address through other approaches. For example, cancer registries and their data can be of particular utility to explore situations where experimentation may be inappropriate or inadequate, to monitor the long-term impact of interventions in the general population and in sub-groups that are not included in trials, and to evaluate procedures that are highly dependent on the provider, setting and patient. Fully exploiting the value of cancer registries, however, requires addressing a combination of practical and policy challenges, from providing sustainable financing to ensuring appropriate data access and security.

Since the introduction of cancer registration over 50 years ago, its fundamental importance to health planners and society is evidenced by the first set of papers presented here. The second set of papers included in this dissertation demonstrate that the value of cancer registries has only increased since then, as the result of new data sources, the availability of new methodologies for data analysis, improved tools for data collection and management, and enhanced capacity to link data sources. Together these papers present a strong case for ensuring that national and local authorities build on recent experience to systematically invest in and expand the roles and activities of cancer registries.

(8)

5

1. INTRODUCTION

(9)

6 1.1. Background

Cancer is a devastating disease accounting for an increasingly high proportion of premature deaths. In the last World Cancer Report (2014) it was estimated that over 14 million new cases of cancer were diagnosed worldwide in 2012, 8.2 million of whom died. ¹ Despite progress in preventive and treatment measures, cancer will continue to increase in the coming years: a predicted global burden of more than 20 million new cancer cases is anticipated by 2025.¹ Increasing life-expectancy and the ageing of populations are the main drivers of this growth, but exposures to old and new risk factors also impact this trend. ¹ Nations, in both developed and developing regions, will face growing needs in terms of diagnosis, treatment and care of cancer patients and consequently growing costs. National governments, public health and health care systems in countries at all stages of economic development will have to plan cancer control programs to reduce the incidence of and mortality from cancer and to improve the quality of life of patients by implementing policies and actions for prevention, early detection, treatment and palliation.

In this context, the role of a population-based cancer registry is pivotal. Epidemiological data obtained from cancer registration are indispensable to document trends and patient numbers; ascertain the need for prevention, screening and therapeutic measures; allocate targeted resources for research; document the quality and efficiency of cancer treatments;

and conduct causal research.²

Traditionally, the principal function of a cancer registry was to provide data on the burden of the disease on the basis of which a cancer control program could be developed. Data on cancer incidence and mortality were essential to evaluate the current situation, to set objectives, to define priorities, and to assess the future evolution of the disease burden.³ However, cancer registries have progressively developed their activities to include data on time trends and geographical variations and survival from cancer. This information allows monitoring of progress in implementing cancer control activities, and evaluation of prevention, early detection, screening and treatment interventions.

(10)

7

Furthermore, using information derived from cancer registries it has been possible to conduct descriptive studies, so-called “hypothesis generating”, about aetiological factors and to support analytical studies results.

Only in recent years, however, has consideration been given to broadening the number of variables collected and exploiting cancer registry databases for other purposes, such as to evaluate the quality of the clinical care provided to cancer patients, including the degree of implementation of treatment recommendations, disparities in the provision of care, and the treatment side effects at a population level.⁴

1.2. Objectives of this work

The purpose of this work is to demonstrate that the role of a modern cancer registry goes well beyond the core functions of surveillance of cancer trends and patterns and guidance for the planning and evaluation of cancer control efforts by using examples of studies performed at the Geneva Cancer Registry, Switzerland.

The first part of this dissertation presents studies which reaffirm the traditional functions of cancer registries, including the monitoring of incidence trends and survival, projections of future burden and etiology hypothesis generating. These activities are mostly directed to public health, epidemiology and health services specialists and policy makers.

The second part of the dissertation shows how through the use of more sophisticated analytical methods and expanded data collection it is possible to evaluate the effectiveness of treatments, assess the patterns and quality of cancer care, and monitor the long term side effects of treatments at a population level. Clinicians, to whom the results of such evaluations are addressed, can then use them in their daily practice to improve patient outcomes and quality of life.

The final section of this dissertation summarizes the major findings and discusses the relevance and the value of the new roles for cancer registries, highlights the practical and policy challenges cancer registries face, and closing suggests future potential roles for cancer registries that are rapidly evolving using new data sources, e.g. biobanks.

(11)

8 1.3. History of cancer registration

The first known systematic collection of information on cancer is London’s “General Census of Cancer” in 1728. But the first real attempts to collect statistical information on cancer in terms of morbidity, mortality and prevalence of the different forms of cancer were made at the end of the 1800s. The chosen method was to perform general surveys of doctors on cancer occurrence. This type of investigation was undertaken in several European countries.

The results, however, were rather scarce and eventually they were completely abandoned. It is only in 1926 that the first permanent system to register cancer cases in a defined population was established in Hamburg, Germany.^3,5 An after-care organization obtained the status of patient follow-up care service and registered all the cancer cases from hospitals and practitioners. Once a week the central index of cases was compared with the official death certificates.⁶ Shortly after, in 1935, in Connecticut the first population-based cancer registry was created in the United States while the first national cancer registry was created in Denmark in 1942.⁷ The Danish registry, established under the auspices of the Danish Cancer Society, received case reports from physicians on a voluntary basis with the support of the Danish Medical Association. The National Board of Health, on the other hand, provided full access to death certificates and mortality data. The purpose of data collection was to: (a) allow individual patient follow-up; (b) have reliable morbidity statistics with a view to an accurate estimate of therapeutic results; and (c) evaluate variations in the incidence of malignant neoplasms (e.g. secular, geographical, occupational).⁸ In the late 1940s other registries in the USA, in Canada and a number of European countries were started.

At an international meeting of 12 leading experts in the field of cancer control in Copenhagen in 1946 the worldwide establishment of cancer registries was recommended to the Interim Commission for the World Health Organization.^9,10 The arguments for such a recommendation were:

“ (a) great benefit would follow the collection of data about cancer patients from as many different countries as possible;

(b) such data should be recorded on an agreed plan so as to be comparable;

(12)

9

(c) each nation should have a central registry to arrange for the recording and collection of such data;

(d) there should be an international body whose duty it should be to correlate the data and statistics obtained in each country. ”^9,10

Thereafter, in 1950 the World Health Organization established a subcommittee on the registration of cancer cases and their statistical presentations followed soon after by the Committee on Geographical Pathology which was established by the International Union Against Cancer (UICC). In 1965 the World Health Organization established the International Agency for Research on Cancer (IARC) as a specialized cancer research centre, and in 1966 the International Association of Cancer Registries (IACR) was formed.¹⁰ The IACR serves as a membership organization dedicated to fostering the aims and activities of cancer registries worldwide. The association collaborates closely with IARC.

Currently at least 290 population-based cancer registries exist in various parts of the world, most of them in more developed regions, which contribute to the publication of the Cancer Incidence in 5 Continents report every 5 years.¹¹ In addition, there are approximately 205 registries that cover only the registration of specific age groups or cancer sites (e.g., childhood tumours; gastrointestinal or gynaecological cancers only) plus a large number of hospital-based cancer registries.

These cancer registries have been founded and supported by a wide range of organizations, including local governments, health departments (at city, provincial or regional level), non- governmental organizations (especially anticancer societies) and universities.⁷

1.4 Utilization of cancer registry data

As outlined in the previous section and stated by WHO “the primary function of a cancer registry is the maintenance of a file or register of all cancer cases occurring in a defined population in which the personal details of the patients and the clinical and pathological characteristics of the cancers, collected continuously and systematically from various data sources, are documented.”¹⁰

(13)

10

There exist different types of registries: hospital-based, cancer-specific or age-specific, and population-based.

Hospital-based cancer registries maintain data on all patients diagnosed and/or treated for cancer at a particular facility. These kinds of registries are useful for administrative purposes and for reviewing clinical performance.

There are two sub-categories under hospital-based registries: single hospital registries and multi-institution registries. The primary goal of the single hospital (institution) registry is to improve patient care by medical audit-type evaluation of outcomes. Cancer data collected by the single hospital (institution) registry are also used for physician education, as a source of data for some types of research, and for some facility utilization assessment.

The primary goal of collective registries (multi-institution registries) is to improve patient care by supporting institutional registries with common standards and pooled data. They often seek to establish the natural history and aetiology of the reported cancers.

Another type of registry is the pathology-based cancer registry which collects information from one or more laboratories on histologically diagnosed cancers. Usually this registry supports the need for laboratory based services and serves as a quick “snapshot” of the cancer profile. However, their data set is constructed from laboratory-based surveillance only, therefore on an incomplete and biased sample of the population. The cancer profile is determined by cancers for which tumour tissue investigations were undertaken.¹²

Some cancer registries record only specific types of cancer (e.g. Registre Bourguignon des Cancers Digestifs, France; Registre d’hémopathies malignes in Côte D’Ôr, France) and/or cancer for a specific age-group (e.g., Registre national des tumeurs solides de l'enfant, France). These registries can be national, subnational or can cover only limited regions of a country. They are often part of or used for specific research projects or patient-care initiatives and are operated by a health or academic centre involved in studying or treating those with the relevant disease.

Population-based registries record all new cancer cases in a defined population, most frequently a geographical area such as a province, region or state. K. Shanmugaratnam in

“Cancer registration: Principles and Methods” emphasized that a registry “analyses and

(14)

11

interprets such data periodically and provides information on the incidence and characteristics of specific cancers in various segments of the resident population and on temporal variations in incidence. Such information is the primary resource not only for epidemiological research on cancer determinants but also for planning and evaluating health services for the prevention, diagnosis and treatment of the disease.”¹⁰

A National Cancer Registry is often thought of as an ideal tool but may not always be feasible due to the scale or cost of implementation. Most of the requirements for planning and monitoring can be achieved through registration of a well-described subset (sample) from the national population, such as from a specific region. The subset sample can then be extrapolated to the national level.

There are differences in the number of variables and the detail of information collected by each cancer registry but in “Cancer Registration: Principles and Methods”, IARC defined a set of 10–11 essential variables (Table 1). As IARC considers that no cancer registry can function with less than this, these variables might be considered the minimum data set. ¹⁰ However, a reasonable list of essential variables is more substantial than this. Table 2 is based on the recommendations of the European Network of Cancer Registries (ENCR)

(http://www.encr.eu/images/docs/recommendations/recommendations.pdf).

Table 1. Essential variables to be collected by a population-based cancer registry according to IARC (Minimum data set) (From Cancer Registration: Principles and methods)¹⁰

Item Comments

The person

Personal identification^a

Name According to local usage

Sex Date of birth or age Estimate if not known Demographic

Address Usual residence

Ethnic group^b When population consists of two or more groups

The tumor Incidence date

Most valid basis of diagnosis

Topography (site) Primary tumor

Morphology (Histology) Behaviour

Source of information For example, hospital record number, name of physician

a The minimum information collected is that which ensures that if the same individuals are reported again to the registry, they will be recognized as being the same person. This could also be a unique personal identification number.

b Ethnic group is included here because it is important for most registries, especially in developing countries.

Source: MacLennan (¹⁰).

(15)

12

Table 2. Essential variables to be collected by a population cancer registry according to ENCR

The information collected by cancer registries can be placed into four categories: patient demographics, tumor (cancer) identification, treatment, and outcome. Demographic data consists of the personal information about a patient such as the patient's name, age, gender, race, ethnicity, birthplace residence. This information individually identifies the cancer patient. Without individual identifiers to check for duplicate registrations, the data would be inaccurate and unsuitable for analysis.

The cancer information comes from the diagnostic findings for a patient. It includes the primary site of the malignancy, its cell type, and the extent of disease. Dates and results of procedures used to diagnose cancer are also recorded.

Data collection continues after a patient is diagnosed with cancer. Information regarding cancer treatment is recorded by many, but not all, registries (surgery, radiation therapy,

(16)

13

chemotherapy, hormone, immunotherapy, and other). Registries usually record such information for at least the first 6 months after diagnosis.

Cancer registries also continue to gather data after the cancer patient has received treatment. These data consist of information concerning the outcome of the treatment.

Patient status is updated regularly to maintain accurate surveillance information. Lifetime follow-up data from patients permit registries to record information about patient survival.

Cancer data are highly confidential. Improper disclosure of these data could result in emotional, psychological, and financial harm to patients and their families. Therefore, one of the most important responsibilities of cancer registry professionals is to protect the confidentiality of cancer patient information.

In every country legislative efforts have been made to improve and protect the confidentiality of cancer data. Specific regulations and laws determine to whom cancer information may be reported, how cancer information is reported, and what procedures should be taken to access cancer information. For example, information identifying a patient is removed when the data are reported. In addition, researchers who need access to cancer data must receive special permission from a designated authority. Each cancer registry may have its own strict policy regarding the way in which files and documents containing confidential information are handled. As a result of such policies and procedures, the privacy of cancer patients is protected.

1.4.1 Traditional uses of population-based cancer registry data

In contrast to hospital-based registries, population-based registries are designed to quantify the burden of cancer among various populations and monitor its trends over time, to make projections and guide planning services; to monitor screening or prevention programs and evaluate cancer control efforts; and to suggest aetiological hypothesis.

The scale and profile of cancer can be evaluated in terms of incidence and mortality, but other dimensions are often considered, including prevalence, person-years of life lost, and quality- or disability-adjusted life years. Cancer registry databases are the principal source of such statistics. Such data can be disseminated via cancer incidence reports, cancer registry websites, research articles, and press releases, and through direct communication with

(17)

14

clinicians, health authorities, the media, or other data users. Most cancer registries are publishing new data regularly. The cancer incidence report is the routine and baseline means of presenting registry data. These reports contain information on all reportable cancers and represent the main deliverable of a cancer registry, providing feedback to the stakeholders and the data providers. Using data from cancer registries IARC is able to estimate the global cancer burden. Its last release was GLOBOCAN 2012, an online database resource¹³ providing the most recent estimates for 28 types of cancer in 184 countries worldwide.¹ According to GLOBOCAN 2012, an estimated 14.1 million new cancer cases and 8.2 million cancer-related deaths occurred in 2012 and a rise in incidence was observed in most regions of the world for both men and women.¹

An appraisal of the current situation provides a framework for action and cancer control planning, and their monitoring over time allows evaluation of the success (or otherwise) of interventions. At a regional, sub regional and local level, health professionals and health planners need to have an accurate picture of cancer in their population in order to target interventions and ensure rational and equitable provision of services and access to care.

Data on incidence rates and temporal trends, together with changes in diagnostic procedures and prevention measures, are also indispensable to estimate the future course of the disease. Cancer predictions, obtained using different mathematical models – the easiest being just the projection of recent trends into the future - have been published for several countries to help prioritize health resource allocation. ^14-16

Cancer registry data are vital to monitor the impact and effectiveness of policy implementation, and screening or prevention programs, and to evaluate cancer control efforts. For example, cancer registry and mortality data have been used to investigate the effectiveness of cervical and breast screening programmes in several countries.^17-20 To be able to perform these kinds of evaluations cancer registries need to have comprehensive and accurate data on the stage of cancers at diagnosis.

Variations in the patterns of cancer incidence are also used to generate hypotheses about the aetiology of cancers. Differences in cancer incidence worldwide have prompted many hypotheses linking diet and cancer, ^21-23 supplying the initial evidence that environmental aflatoxin causes cancer in human beings,²⁴ and providing the first reasonably certain evidence that Pap-smear screening reduces the incidence of cervical cancer.²⁵

(18)

15

Cancer registries are also a primary source of data for unbiased population-based case control studies, the end points for cohort studies and clinical trials and, perhaps most importantly, the starting point for survival analysis.

1.4.2 New roles for cancer registration

Monitor the effectiveness of treatments at a population level

In those situations where there is not, or it is not possible to carry out, a randomized controlled trial, observational studies using population-based registry data can and are increasingly being used to fill the knowledge gap. Randomized controlled trials are clearly the gold standard for comparing treatment efficacy but they are often not feasible, either because of expense, ethical concerns, or patient acceptance. In addition, they often have restrictive enrollment criteria so that the participants do not resemble patients in practice, particularly with respect to clinical characteristics such as comorbidity, age, and medications or sociodemographic characteristics such as race, ethnicity, and socioeconomic status.

Studies conducted with registry data can confirm (or not) the results of randomized studies at the population level, or suggest clues which lead, in turn, to hypothesis-directed clinical trials.^26,27 The use of registry data also allows for the analysis of treatment benefits, costs, and toxicity in specific subsets of patients. These types of comparative effectiveness studies are particularly helpful, given that one can analyze treatment effects in populations that would not normally be included in randomized trials.^28-30 However, the random allocation of patients to treatment, which balances the effect of factors that may influence the exposure, is not possible in such studies. Therefore, observational studies such as those conducted using registry data are affected by selection or susceptibility bias.³¹ A second issue that must be considered in population-based registry studies is confounding. Confounding occurs when an extraneous factor is associated with both exposure and outcome, often leading to biased results.

Researchers using observational data currently have a number of advanced statistical methodologies at their disposal to perform risk adjustment and to limit selection bias and confounding. Among those used most frequently are cohort selection, matching, propensity score analysis and instrumental variable analysis.³² Careful framing, appropriate study

(19)

16

design, and application of sophisticated analytic techniques can improve the accuracy of nonrandomized studies.

Evaluate the quality of cancer care across providers and populations

The EUROCARE Study, a European Union project to assemble survival data from population- based cancer registries and analyse them according to standard procedures, was the first large international comparison of population-based survival probabilities.³³ Initially, survival was measured without clinical detail by age, sex and cancer site and was proposed as an indicator of care provided by the oncology system. Publication of the first reports from the project in 1995 demonstrated substantial international variations and had a strong impact on policy-making in some countries (e.g. in the United Kingdom and Denmark). Subsequently the methods have been refined, and in “high resolution” studies detailed information on stage at diagnosis and treatment received were collected from the hospital records for specific cancer sites and limited numbers of patients.^34,35 Currently, the project has been expanded to include almost 100 registries that together cover 13 million patients

(www.eurocare.it). One of its focuses is the standardization of reporting of population-based survival data.

After the publication of the first EUROCARE reports some cancer registries expanded their role to include the evaluation of the quality of healthcare performance in terms of structure, process and outcomes, as first suggested by Donabedian.³⁶ For this purpose some registries expanded the number of variables collected for each case to include the stage, the extent of disease, the type of treatments, follow-up and some included even comorbidity, recurrence and metastasis data.

Patterns of care (POC) studies evaluating variations in the process of care have been conducted on samples of patients for whom additional detailed information have been retrieved.^37,38 Different indicators for the process of clinical care and for outcome measures have been proposed. Among the first ones were the health sector of care (public, private), the place of treatment (type of hospital), the caseload of the institute/therapist, the specialization of the therapist, the diagnostic procedures, and the waiting times (between diagnosis and first treatment). Among the outcome measures, disease-free survival, overall or specific survival, and some quality of life indicators- between treatment and death in

(20)

17

terms of disability or side effects of treatment – have been collected. Sometimes economic considerations have been also included.

Monitor and evaluate side effects of cancer treatments

As the result of advances in screening, early detection and treatment, the survival of cancer patients has significantly improved in recent decades, leading to increasing numbers of individuals who are either cured of their cancer or are living with it as a chronic disease.³⁹ However, the multimodal therapies that are frequently used to achieve cure may be responsible for late health consequences. For most treatments, there is better information on the acute and intermediate effects of therapy, within the first 5 years of treatment, than possible late effects. In addition, cancer survivors are often at risk for second malignancies, either related to the biology of the primary cancer (e.g. breast cancer and colorectal cancer) or secondary to previous cancer treatments. To study various health outcomes in cancer survivors, it is important to use a defined population to make accurate inferences about the effects of treatments. There is a paucity of longitudinal cohort studies to link specific treatment regimens to late physical and psychosocial effects, and thus it is not possible to describe the natural history of these events for future patients and their health care providers. Population-based cancer registries that have accurate data on treatments offer the best opportunity for this type of research. A cohort of patients for study can be identified and counselling about interventions and risk reduction offered.⁴⁰

1.5 The Geneva Cancer Registry

The Geneva Cancer Registry, the first registry in Switzerland, was created in 1970 following a decision of the Council of State. The registry was attached to the Center of Cytology and Cancer Screening (Centre de cytology et de dépistage) that was supported by the Cantonal Department of Social Welfare and Public Health (Direction générale de la santé DGS). The first assessments for creating a cancer registry in Geneva were performed at the same time as the first screening campaigns against cervical cancer at the beginning of the 1960s.

However, practical obstacles prevented implementation of the project for several years.

The main purpose of the Registry was to have statistical indicators on cancer to predict the needs in terms of hospital equipment. A secondary purpose, according to the first published

(21)

18

report of the Registry, was to contribute to research on the aetiological factors for cancer.⁴¹ One of the first periodical reports of the Registry stated its tasks⁴²:

a) Collect information on all neoplastic diseases detected among the resident population in the canton, in order to obtain epidemiological statistics (calculation of the incidence and prevalence, calculation of survival, etc.), establish the level of impact of these conditions and measure the stage of the lesion at the time of its discovery in the general population or in specific groups. The identification of high risk groups or those for whom an early diagnosis is relatively infrequent serves both the preventive and the aetiological research functions.

b) follow-up the cases reported for the evaluation of survival from major tumour types and for calculating the intervals between significant events in the history of the lesion (overall evaluation of preventive and therapeutic measures).

The early objectives of the registry also included the organization of a system for the regular exchange of information with the physicians responsible for monitoring cancer patients.^41,42

During the first two years of registration 1112 cases of cancers among males and 1227 among females were recorded in the Canton (in 1970 Geneva had 331 599 inhabitants in an area of 282 km²). Forty years later, at the end of 2012, these figures had become 66,600 and 71,200 for males and females, respectively (by 2012 the number of inhabitants in the Canton was 470 512). During these 40 years a lot of changes and improvements occurred in the organization and management of the registry as well in its functions. Not only had the number of cases that were registered increased substantially but the number of variables recorded for each case went from 38 to the current 150.

Among the variables added to the collection were those relevant to the patient, like the place of birth and nationality, the profession at the moment of diagnosis, and the marital status; some cancer risk factors such as weight, height, BMI, family history of cancer, smoking status; many variables related to the tumor including method of discovery, previous screening with date, stage, extension, grade, histology; a number of prognostic factors; exams used for diagnosis confirmation; all treatments received during the first 6 months after diagnosis, and the health sector providing the care (e.g. public or private).

Recently, for breast and prostate cancer patients the Registry is also collecting data on

(22)

19

comorbidities. Since 2012 data were collected on treatments received during the first 12 months after diagnosis.

In 1999 the Registry became part of the Social and Preventive Medicine Institute (IMSP) of the Faculty of Medicine of the University of Geneva. In 2013, the IMSP became the Global Health Institute.

Since 1994, the Registry was granted a general approval to collect cancer data of residents in the Geneva canton and use this data for public health and clinical research purposes (art. 321 bis CP) by the Federal Expert Commission on Professional Secrecy in Medical Research of the Federal Office of Public Health. This authorization allows doctors to transmit patient data which are not anonymized to the registry, but does not require them to declare. (http://www.bag.admin.ch/themen/gesundheitspolitik/10374/10375/index.html)

All information reported to the Registry is confidential, and strict procedures are in place to protect patient privacy. All employees are trained in handling confidential information. All the research performed with cancer registry data are done with files from which personal identification have been stripped. When publishing the results of such research studies, patient data are disseminated only as aggregate information and summary results. The identity of individuals remains completely undisclosed.

In 1999, the Geneva Cancer Registry set up the first Swiss Familial Breast Cancer Registry, by extending its data set to include the detailed family history of cancer for all women diagnosed with invasive breast cancer in the Geneva population.⁴³ For breast cancer patients diagnosed between 1990 and 1999, family history was collected retrospectively, using information from medical records from the public university hospitals and private physicians. For 90% of the breast cancer patients, information on family history was obtained and the accuracy of this retrospectively retrieved information has been validated (Verkooijen et al, 2004).⁴⁴ Since January 2000, family history on >= 3 generations is prospectively collected by sending standard questionnaires to health-care providers. This registry currently contains the complete, validated cancer family histories for more than 8000 women diagnosed with in situ and invasive breast cancer in the Canton of Geneva between 1990 and 2011. This registry has allowed a multidisciplinary team of experts to plan and perform studies that have contributed to a better knowledge

(23)

20

of the factors that affect cancer susceptibility, the interaction with environmental and life-style factors, prognosis, second tumor occurrence and the heritability of prognosis of patients with a family history. Many publications have resulted from the Geneva Familial Breast Cancer Registry.

The Geneva Cancer Registry regularly links to the Cantonal Breast Cancer Screening Programme dataset for the evaluation of the programme. The registry also shares data in national and international collaborative projects.

Currently in Switzerland there are 15 population-based registries that cover 19 cantons (Basel Stadt and Land, Fribourg, Geneva, Graubunden and Glarus, Jura, Lucerne, Neuchatel, St. Gall and Appenzell Ausserrhoden and Appenzell Innerrhoden, Ticino, Valais, Vaud, Zug and Zurich). Since 2007 all of these registries have been sending their data to the National Institute for Cancer Epidemiology and Registration (NICER) which compiles and aggregates the data, and promotes and supports population-based cancer registration and epidemiological cancer research in Switzerland. (www.nicer.org)

A new law which would make cancer registration compulsory for all Swiss cantons is under consideration.

(24)

21

2. TRADITIONAL USES OF CANCER

REGISTRATION

(25)

22

2.1 . Monitor trends in cancer frequency, mortality and survival 2.1.2. Recent increase of breast cancer incidence among women under the age of forty.

Bouchardy C, Fioretta G, Verkooijen HM, Vlastos G, Schaefer P, Delaloye JF, Neyroud-Caspar I, Balmer Majno S, Wespi Y, Forni M, Chappuis P, Sappino AP, Rapiti E.

Br J Cancer. 2007 Jun 4;96(11):1743-6.

(See appendix for the full article)

ABSTRACT

Using data from the Geneva Cancer Registry, we found that in 2002–2004, breast cancer incidence in women aged 25–39 years increased by 46.7% per year (95% CI: 7.1–74.0, P=0.015), which surveillance or detection bias may not fully explain.

(26)

23

2.2. Project the future of cancer burden and plan health services 2.2.1. Planning for the future: cancer incidence projections in Switzerland up to 2019 Rapiti E, Guarnori S, Pastoors B, Miralbell R, Usel M.

BMC Public Health. 2014 Feb 1;14:102.

ABSTRACT

Background: Projections of the national burden of cancer play a key role in planning cancer control programmes and investments. We present projections of cancer incidence rates and cases for the period up to 2015-2019 in Switzerland.

Methods: Projections were based on cancer incidence data estimated from cancer registries for the 1989-2009 periods and demographic projections of the Federal Statistical Office.

Age-specific incidence rates were modelled as a function of age, period-birth cohort using NORDPRED.

Results: Up to 2019 the incidence of all cancers combined is expected to decrease slightly for both sexes. Nevertheless, the overall number of cases is predicted to increase. The number of male cancer cases will increase by 30%, from 20005 in 2005-2009 to 25910/year in 2015- 2019. For females the number will increase by 20%, from 16913 to 20359/year in 2015-2019.

Changes in the population size and structure will be responsible for most of the increase.

Among men, the largest increase is observed for melanoma (+54%), thyroid (+45%), non- Hodgkin lymphoma (+43%), and prostate (+37%). Prostate cancer will contribute with 8083 cases, colorectal cancer with 2908 and lung cancer with 2791. For women, cases of lung and oral cavity cancers will increase by +48% and +38%, respectively; those of thyroid by +45%

and non-Hodgkin lymphoma by +36%. The sites with the most cancer predicted are breast (5870), colorectal and lung (over 2000 each), melanoma (1341) and corpus uteri (1040).

The overall annual cancer burden predicted for 2015-19 is of 46269 new cases in Switzerland.

Conclusions: Substantial investments appear to be needed in Switzerland cancer services to meet and fill absolute increased demand driven by aging population.

(27)

24 2.3. Suggest aetiological hypothesis

2.3.1. Risk of second breast cancer according to estrogen receptor status and family history.

Bouchardy C, Benhamou S, Fioretta G, Verkooijen HM, Chappuis PO, Neyroud-Caspar I, Castiglione M, Vinh-Hung V, Vlastos G, Rapiti E.

Breast Cancer Res Treat. 2011 May;127(1):233-41.

ABSTRACT

A recent study reported an increased risk of contralateral estrogen-negative breast cancer after a first primary estrogen-negative breast cancer. Our study aims to confirm this result and to evaluate how the risk of second breast cancer occurrence is affected by family history of breast cancer and anti-estrogen treatment. We included all 4,152 women diagnosed with breast cancer between 1995 and 2007, using data from the population-based Geneva Cancer Registry. We compared the incidence of second breast cancer among patients according to estrogen receptor (ER) status with that expected in the general population by age-period Standardized Incidence Ratios (SIRs). Among the cohort, 63 women developed second breast cancer. Patients with ER-positive first tumors had a decreased risk of second breast cancer occurrence (SIR: 0.67, 95% CI: 0.48-0.90), whereas patients with ER-negative primary tumors had an increased risk (SIR: 1.98, 95% CI: 1.19-3.09) limited to ER-negative second tumors (SIR: 7.94, 95% CI: 3.81-14.60). Patients with positive family history had a tenfold (SIR: 9.74, 95% CI: 3.57-21.12) higher risk of ER-negative second tumor which increased to nearly 50- fold (SIR: 46.18, 95% CI: 12.58-118.22) when the first tumor was ER-negative. Treatment with anti-estrogen decreased the risk of second ER-positive tumors but not ER-negative tumors. The risk of second ER-negative breast cancer is very high after a first ER-negative tumor, in particular among women with strong family history. Surveillance and prevention of second cancer occurrence should consider both ER status of the first tumor and family history.

(28)

25 2.4 Discussion

The papers in this chapter demonstrate the basic cancer registry roles in providing unbiased profiles of the cancer burden in a population and if/how that burden changes over time.

They also show the role of a registry in planning and evaluating health services. These papers also show that a registry is at the service of the public to respond to cancer inquires and concerns and guide policy planning.

The study “Recent increase of breast cancer incidence among women under the age of forty” was carried out following the concern of local physicians about an “epidemic” of young women with breast cancer. Analyzing our registry data we confirmed that their impression was right.

Right after the publication of our article the Vaud Cancer Registry submitted to the same journal an analysis of their data in a letter to the Editor showing that in their canton the same phenomenon was not observed.⁴⁵ The authors suggested that the small numbers and consequent great variability were the main reason for our results. Shortly afterwards, however, several studies in different parts of the world also showed an increase in breast cancer incidence in their youngest cohorts. In particular, two European collaborative studies confirmed an increase in the incidence of breast cancer in European women in their 20s and 30s.^46,47 In the first study breast cancer incidence data were pooled for 17 European Cancer Registries for the calendar period 1995–2006. The mean annual change in the incidence rate was 1.032 in women aged 20–29 (95 % CI = 1.019–1.045), and 1.014 in women 30–39 years old at diagnosis (95 % CI = 1.010–1.018).⁴⁶ The second study included data for women aged less than 40 years from 1990 to 2008 from 37 European population-based cancer registries in Belgium, Bulgaria, France, Italy, Portugal, Spain and Switzerland. The overall incidence rate of breast cancer in the area covered increased linearly during the study period by 1.19%

(0.93; 1.46) on average per year. The rise in incidence was greater for women under the age of 35 years and for ductal carcinomas.⁴⁷ In their conclusions, both studies emphasized the importance of surveillance of young women and further studies to understand the risk factors responsible for this phenomenon. The Geneva Cancer Registry participated in both collaborative studies.

(29)

26

The paper on projections of future cancer cases (“Planning for the future: Cancer incidence projections in Switzerland up to 2019”), was conceived to help clinical facilities plans. By providing with data on the number of oncologic cases expected in the canton during the following 10 years we helped predict the treatment capacity and radiation therapy equipment needed. We applied well accepted mathematical models to calculate the projected number of cases in Geneva in 2019 as well as the radiotherapy needs for the canton. We then decided to expand the estimation to the entire country of Switzerland, basing our models on larger samples thus obtaining more reliable figures.

The paper “Risk of second breast cancer according to estrogen receptor status and family history” falls into the category of use of registry data to suggest aetiological hypotheses/risk factors. It is well known that a positive family history for breast cancer is a strong risk factor for developing breast cancer.⁴⁸ Family history has also been associated with a higher risk of a second breast cancer.^49,50 Recently it has been demonstrated that the receptor status of breast cancer influences the risk and the receptor status of a second breast cancer .⁵¹ With our study we wanted to assess the risk of subsequent ER-positive and ER-negative contralateral tumors in breast cancer patients in Geneva. In addition, we evaluated whether the ER status of the first tumor, a family history of breast and/or ovarian cancer, and the use of anti-estrogens modified the association. This study, as with many others in the area of breast cancer, was conducted using data from the Familial Breast Cancer Registry (see page 19 for details) which includes information on the family history of breast cancer for all patients diagnosed since 1990. Our study confirmed some of the results showed from other authors about an increased risk of contralateral ER-negative breast cancer if the first breast cancer was ER-negative.⁵¹ In addition, we demonstrated that this risk is modified by family history, showing that a strong family history of breast and/or ovarian cancer further increases the risk of developing a second ER-negative tumor. In particular, patients with both ER-negative tumors and a strong family history presented a very high risk of developing a second ER-negative tumor. These data allow clinicians to establish the correct follow-up for breast cancer patients. Surveillance for a second cancer occurrence should be adjusted according to both the ER status of the primary breast cancer and the family history of the patient. In this particular case we suggested that specific preventive interventions such as chemoprevention or prophylactic surgery should be considered for women with both positive family history and ER-negative first tumors.

(30)

27

3. NEW ROLES FOR CANCER

REGISTRATION

(31)

28

3.1 Monitor the effectiveness of cancer treatments

3.1.1. Omission of excisional therapy is associated with an increased risk of invasive cervical cancer after cervical intraepithelial neoplasia III.

Rapiti E, Usel M, Neyroud-Caspar I, Merglen A, Verkooijen HM, Vlastos AT, Pache JC, Kumar N, Bouchardy C.

Eur J Cancer. 2012 Apr;48(6):845-52

ABSTRACT

Background: Using data from the population-based Geneva Cancer Registry we evaluated the risk of invasive cervical cancer following carcinoma in situ (CIS) or cervical intraepithelial neoplasia (CIN) III according to type of treatment.

Methods: Included in the study were all women diagnosed with CIS/CIN III in Geneva (Switzerland) between 1970-2002 (n=2,658) and followed for invasive cervical cancer occurrence until December 31st 2008. We calculated age and period standardised incidence ratios (SIR) and multiadjusted hazard ratios (HR) of invasive cervical cancer by treatment groups.

Results: During follow-up, 17 women developed invasive cervical cancer, conferring a SIR of 5.1 (95% Confidence Intervals [CI] 3.0-8.1). The risk of cervical cancer was significantly increased until 10 years after diagnosis. The risk was highest for women ≥50 years (SIR= 7.3, 95% CI: 2.7-15.8) and for women who did not undergo excisional treatment (SIR=25, 95% CI:

12.0-46.0). The multiadjusted HR of invasive cervical cancer for women who did not undergo surgical excisional treatment was 9.4 (95% CI: 2.8-32.2) compared with women who did.

Conclusion: Women diagnosed with CIS/CIN III are at increased risk of developing invasive cervical cancer. This risk is particularly high for women who did not have excision of cervical lesions.

(32)

29

3.1.2. Complete excision of primary breast tumour improves survival of patients with metastatic breast cancer at diagnosis.

Rapiti E, Verkooijen HM, Vlastos G, Fioretta G, Neyroud-Caspar I, Sappino AP, Chappuis PO, Bouchardy C.

J Clin Oncol. 2006 Jun 20;24(18):2743-9.

ABSTRACT

Purpose: Surgery of the primary tumour is usually not advised for patients with metastatic breast cancer at diagnosis as the disease is considered incurable. In this population-based study we evaluate the impact of local surgery on survival of patients with metastatic breast cancer at diagnosis.

Methods: We included all 300 metastatic breast cancer patients recorded at the Geneva Cancer Registry between 1977-1996. We compared mortality risks from breast cancer between patients who had surgery of the primary breast tumour to those who had not and adjusted these risks for other prognostic factors.

Results: Women who had complete excision of the primary breast tumour with negative surgical margins had a 40% reduced risk to die of breast cancer (multi-adjusted Hazard Ratio [HR] =0.6, 95% Confidence Interval [95% CI] =0.4-1.0) compared to non-operated women (p=0.049). This mortality reduction was not significantly different among patients with different sites of metastasis, but in the stratified analysis the effect was particularly evident for women with bone metastasis only (HR=0.2, 95% CI: 0.1-0.4, p=0•001). This mortality reduction was particularly evident for women with bone metastasis only (HR=0.2, 95% CI:

0.1-0.4, p=0•001). Survival of women who had surgery with positive surgical margins was not different from that of non-operated women.

Conclusion: Complete surgical excision of the primary tumour improves survival of patients with metastatic breast cancer at diagnosis, particularly among women with only bone metastases.

(33)

30

3.2 Evaluate the quality of cancer care across providers and populations

3.2.1. Breast cancer management and outcome according to surgeon's affiliation: a population-based comparison adjusted for patient's selection bias.

Taban F, Rapiti E, Fioretta G, Wespi Y, Weintraub D, Hugli A, Schubert H, Vlastos G, Castiglione M, Bouchardy C.

Ann Oncol. 2013 Jan;24(1):116-25.

ABSTRACT

Background: Studies have reported that breast cancer (BC) units could increase quality of care but none has evaluated the efficacy of alternative options such as private BC networks, which is our study objective.

Patients and methods: We included all 1’404 BC patients operated in the public Unit or the private Network and recorded at the Geneva Cancer Registry between 2000 and 2005. We compared quality indicators of care between public BC Unit and private BC Network by logistic regression and evaluated the effect of surgeon’s affiliation on BC-specific mortality by Cox model adjusting for propensity score.

Results: Both groups had high care quality scores. For invasive cancer, histological assessment before surgery and axillary lymph node dissection when indicated were less frequent in the public sector (adjusted odds ratio (OR): 0.4, 95% confidence interval (CI): 0.3- 0.7, and OR: 0.4, 95%CI: 0.2-0.8, respectively), while radiation therapy after breast- conserving surgery was more frequent (OR: 2.5, 95%CI: 1.4-4.8). Surgeon affiliation had no significant effect on BC-specific mortality (adjusted hazard ratio (HR): 0.8, 95%CI: 0.5-1.4).

Conclusions: This study suggests that private BC networks could be an alternative to public BC units with both structures presenting high quality indicators of BC care and similar BC- specific mortality.

(34)

31

3.2.2 Impact of socioeconomic status on prostate cancer diagnosis, treatment, and prognosis.

Rapiti E, Fioretta G, Schaffar R, Neyroud-Caspar I, Verkooijen HM, Schmidlin F, Miralbell R, Zanetti R, Bouchardy C.

Cancer. 2009 Dec 1;115(23):5556-65.

ABSTRACT

Background: This study aims to evaluate the impact of socioeconomic disparities on prostate cancer presentation, treatment, and prognosis in Geneva, Switzerland, where healthcare costs, medical coverage, and life expectancy are among the highest in the world.

Methods: This population-based study included all patients diagnosed with invasive prostate cancer among the resident population between 1995 and 2005. Patients were divided into three socioeconomic groups according to their last known occupation. We compared patient and tumour characteristics, and treatment patterns between socioeconomic groups. We used Cox multivariate regression analysis to assess and explain socioeconomic inequalities in prostate cancer-specific mortality.

Results: Compared to patients of high social class, those of low social class were more often foreigners, had less frequently screen-detected cancer, more advanced stage at diagnosis and less often information on disease characteristics and staging. Complete diagnostic assessment. These patients had less often prostatectomy and were more often managed with watchful waiting. The risk of dying as a result of prostate cancer (Hazard Ratio [HR]) in patients with low vs. high socioeconomic status was 2.0 -fold increased (95% Confidence Interval [IC]: 1.5-2.6,). After adjustment for patient and tumour characteristics and treatment, the mortality risk was no longer significantly increased (HR 1.2, 95%CI: 0.8-1.6).

Conclusion: Patients of low social class are at increased risk of dying as a result of their prostate cancer. This over-mortality is largely attributable to delayed diagnosis, poor diagnostic work-up, and less invasive treatments.

(35)

32

3.2.3. Impact of obesity on diagnosis and treatment of breast cancer.

Deglise C, Bouchardy C, Burri M, Usel M, Neyroud-Caspar I, Vlastos G, Chappuis PO, Ceschi M, Ess S, Castiglione M, Rapiti E, Verkooijen HM

Breast Cancer Res Treat. 2010 Feb;120(1):185-93.

ABSTRACT

In this population-based study, we evaluated the impact of obesity on presentation, diagnosis and treatment of breast cancer. Among all women diagnosed with invasive breast cancer in the canton Geneva (Switzerland) between 2003 and 2005, we identified those with information on body mass index (BMI) and categorized them into normal/underweight (BMI

<25 kg/m(2)), overweight (BMI > or =-<30 kg/m(2)) and obese (BMI > or =30 kg/m(2)) women. Using multivariate logistic regression, we compared tumour, diagnosis and treatment characteristics between groups. Obese women presented significantly more often with stage III-IV disease (adjusted odds ratio [OR(adj)]: 1.8, 95% CI: 1.0-3.3). Tumours > or =1 cm and pN2-N3 lymph nodes were significantly more often impalpable in obese than in normal/underweight patients (OR(adj) 2.4, [1.1-5.3] and OR(adj) 5.1, [1.0-25.4], respectively). Obese women were less likely to have undergone ultrasound (OR(adj) 0.5, [0.3-0.9]) and MRI (OR(adj) 0.3, [0.1-0.6]) and were at increased risk of prolonged hospital stay (OR(adj) 4.7, [2.0-10.9]). This study finds important diagnostic and therapeutic differences between obese and lean women, which may impair survival of obese women with breast cancer. Specific strategies are needed to optimize the care of obese women with or at risk of breast cancer.

(36)

33

3.3 Monitor and evaluate side effects of cancer treatments

3.3.1. Increased risk of colon cancer after external radiation therapy for prostate cancer.

Rapiti E, Fioretta G, Verkooijen HM, Zanetti R, Schmidlin F, Shubert H, Merglen A, Miralbell R, Bouchardy C.

Int J Cancer. 2008 Sep 1; 123(5):1141-5.

ABSTRACT

Background: Radiotherapy can induce second cancers. Controversies still exist regarding the risk of second malignancies after irradiation for prostate cancer. We evaluated the risk of developing colon and rectum cancers after prostate cancer in irradiated and non-irradiated patients.

Methods: Using data from the population-based Geneva cancer registry, we included in the study all men with prostate cancer diagnosed between 1980 and 1998 who survived at least five years after diagnosis. Of the 1134 patients, 264 were treated with external radiotherapy.

Patients were followed for occurrence of colorectal cancer up to 31 December 2003. We calculated Standardized Incidence Ratios (SIR) using incidence rates for the general population to obtain the expected cancer incidence.

Results: The cohort yielded to 3798 person-years. At the end of follow-up, 19 patients had developed a colorectal cancer. Among irradiated patients the SIR for colorectal cancer was 3.4 (95% Confidence Intervals [CI] 1.7-6.0). Compared to the general population, the risk was significantly higher for colon cancer (SIR=4.0, 95% CI: 1.8-7.6), but not for rectal cancer (SIR=2.0, 95% CI: 0.2-7.2). The risk of colon cancer was increased in the period 5-9 years after diagnosis (SIR=4.7, 95% CI: 2.0-9.2). The overall SIR of secondary cancer in patients treated with radiotherapy was 1.35 (p=0.056). Non-irradiated patients did not have any increased risk of rectal or colon cancer.

Conclusion: Colon cancer has increased. This study shows a significant increase of colon but not rectum cancer after radiotherapy for prostate cancer. The risk of second cancer after irradiation, although probably small, needs nevertheless to be carefully monitored.

(37)

34

3.3.2. Excess of cardiovascular mortality among node-negative breast cancer patients irradiated for inner-quadrant tumors.

Bouchardy C, Rapiti E, Usel M, Majno SB, Vlastos G, Benhamou S, Miralbell R, Neyroud- Caspar I, Verkooijen HM, Vinh-Hung V.

Ann Oncol. 2010 Mar;21(3):459-65.

ABSTRACT

BACKGROUND: Radiotherapy of the left breast is associated with higher cardiovascular mortality linked to cardiotoxic effect of irradiation. Radiotherapy of inner quadrants can be associated with greater heart irradiation, but no study has evaluated the effect of inner- quadrant irradiation on cardiovascular mortality.

PATIENTS AND METHODS: We identified 1245 women, the majority with breast-conserving surgery, irradiated for primary node-negative breast cancer from 1980 to 2004 registered at the Geneva Cancer Registry. We compared breast cancer-specific and cardiovascular mortality between inner-quadrant (n = 393) versus outer-quadrant tumors (n = 852) by multivariate Cox regression analysis.

RESULTS: After a mean follow-up of 7.7 years, 28 women died of cardiovascular disease and 91 of breast cancer. Patients with inner-quadrant tumors had a more than doubled risk of cardiovascular mortality compared with patients with outer-quadrant tumors (adjusted hazard ratio 2.5; 95% confidence interval 1.1-5.4). Risk was particularly increased in the period with higher boost irradiation. Patients with left-sided breast cancer had no excess of cardiovascular mortality compared with patients with right-sided tumors.

CONCLUSIONS: Radiotherapy of inner-quadrant breast cancer is associated with an important increase of cardiovascular mortality, a possible result of higher irradiation of the heart. For patients with inner-quadrant tumors, the heart should be radioprotected.

Expanding the Role of Cancer Registration: from Cancer Incidence and Mortality Monitoring to Evaluation of Cancer Care Delivery

Thesis

Reference

Expanding the Role of Cancer Registration: from Cancer Incidence and Mortality Monitoring to Evaluation of Cancer Care Delivery

EXPANDING THE ROLE OF CANCER REGISTRATION:

FROM CANCER INCIDENCE AND MORTALITY MONITORING TO EVALUATION OF CANCER CARE DELIVERY

Elisabetta RAPITI

TABLE OF CONTENTS

SUMMARY

1. INTRODUCTION

2. TRADITIONAL USES OF CANCER

REGISTRATION

3. NEW ROLES FOR CANCER

REGISTRATION