The Cancer Registry: Its Purpose and Uses

by Donald Maxwell Parkin ¹

Patients waiting to be seen in the outpatient clinic of the Cancer Institute (W.I.A.) in Chennai, India. All cancer patients seen at the center are included in both the institutional registry and the population-based registry for Chennai.

Cancer Registration

Disease registers – inventories of newly diagnosed cases – have been used as a means of keeping track of the importance of several different diseases in the community, and how their distribution varies between different places, different subgroups of the population, and over time. Disease registers can supplement and extend the information we can get from vital statistics – records of deaths from specific diseases – as the latter deal only with fatal events, and death may be the outcome of a disease only in a minority of cases – even for many cancers. Registration of cancer has probably been the most successful and widely used form of disease registration. Cancer lends itself to the process, since it is a serious disease that almost always comes to medical attention, sooner or later, has a relatively clear definition, and generally occurs only once in the same individual. It is much more difficult to count cardiovascular events (such as myocardial infarctions or strokes), many episodes of which occur in sub-clinical form, with multiple events of the same disease in the same person, or congenital malformations, where definition is a problem, since most are trivial and may not even be noticed, let alone reported.

Cancer registration is a continuing process of systematic collection of information on individuals developing new cancers. The cancer registry (or rather, its staff) are responsible for collecting, processing, storing, analysing, interpreting and reporting data on the characteristics of the cancers and of the individuals affected. Hospital-based cancer registries are concerned only with the recording of information on the patients with cancer seen in one specific hospital. Their purpose is to contribute to patient care by providing readily accessible information on the patients seen at the institution over defined time periods, the treatment they received and its outcome. They also serve an administrative purpose and are useful in reviewing clinical performance of staff and deciding on present and future institutional needs. By contrast, a population-based cancer registry (PBCR) collects data on every subject with cancer in a defined population - usually, the residents of a particular geographical region (city, province or country). The cooperation of the medical profession and health care services within the region is vital to the success of cancer registration.

Since population registration involves head counts of new cancer cases occurring in a defined time period and relates this to a population of known size, it is possible to calculate the incidence rates of specific cancers for different subgroups of the population. Ideally, the registered cases can be followed up after diagnosis to find out which of them die from their disease, so that the survival rate for different cancers, at specified intervals after diagnosis, can be calculated.

The main sources of information for population-based registries include: (1) information from treatment facilities, such as cancer centers, hospitals, private clinics and hospices, (2) information from diagnostic services, especially pathology laboratories (3) death certificates from the vital statistics registration system. The way the data are collected invariably depends upon the local conditions; more and more it relies upon the transfer of computer files of patients from hospital administration systems, laboratory databases and population registers to the cancer registry. The amount of information on each cancer case collected by a registry is dictated by the purpose for which the registry has been established, but, in general, PBCRs collect only a limited amount of data for each patient. The basic minimum dataset is shown in Table 1.

Essential variables Personal identification: names (in full) AND/OR unique personal identification numbers Sex Date of birth: day, month, year Address: usual residence (coded) Incidence date (at least month and year) Most valid basis of diagnosis Topography (site) of primary cancer (ICD-O coded) Morphology (histology, ICD-O coded) Behavior (e.g., benign, malignant, in-situ, ICD-O coded) Source of information Recommended variables Date of last contact (at least month and year) Status at last contact (at least dead or alive) Stage or extent of disease Initial treatment

Table 1. Items of information collected by registries (from Jensen et al., 1991).

It is essential to be able to identify individuals with cancer, not just the cancer itself (a few patients may have more than one cancer). Name, sex, date of birth (age) and address are essential identification items. Address is required for establishing the residence status, and for follow-up. Age is of great importance in the description of cancer incidence. The incidence date (date the cancer was detected) must be a definite, consistent and reliable point in time which can be verified from records. The most valid basis of diagnosis is of great interest in assessing the reliability of incidence rates; the minimum requirement for evaluation is differentiation between microscopic and non microscopic confirmation. The International Classification of Disease for Oncology (ICD-O) coding scheme (Fritz et al, 2000) is used to describe the topography (site of primary tumor) and morphology (histological type). The topography of a tumor is the most important data item recorded and provides the main basis of tabulation of registry data. The fifth digit in the ICD-O morphology code describes the behavior of the tumor (benign, borderline, in situ, malignant). The optional (recommended) items in Table 1: clinical extent of disease (stage) before treatment, details of initial treatment and outcome in terms of follow up, are collected by most registries.

A PBCR must document its findings and conclusions in the form of cancer incidence reports and other special reports and articles in scientific journals that ensure dissemination of the information to users. A registry that simply collects and stores information serves no useful purpose – the data are only valuable if used for information, research, planning and evaluation, as described later. Registries with a wide portfolio of activities will tend to improve the quality of their routine statistics simply by utilizing the collected data, as well as by activating interest among collaborators (clinicians and researchers) in the daily registration procedures.

The History and Development of Cancer Registration

The idea of recording information on all cancer cases in defined communities dates from the middle of the 20th century, and there has been a steady growth in the number of such cancer registries since. Originally, they were concerned primarily with describing cancer patterns and trends, but provision of information on other aspects of cancer occurrence and on the control of the disease has developed progressively. This resulted initially from the need for information on the survival after cancer at the population level, and later to study the effects (value) of various aspects of services for prevention, early diagnosis, treatment and care. This template has been applied to a greater or lesser extent in various world regions, and the steady increase in the number of cancer registries attests to their value in cancer research and control. Figure 1 illustrates the growth in the number of cancer registries that are members of the International Association of Cancer Registries (IACR) between 1979 and 2006. These cancer registries cover around 21% of the population of the world, albeit with a rather uneven spread among the continents. Some entire national populations are covered; in smaller countries (e.g. Singapore, the Gulf States, the Nordic countries) this is possible with a single cancer registry, but larger populations pose considerable technical and logistic problems, and there are few registries covering populations in excess of 10-15 million. National registries for the larger countries therefore rely upon input from independent regional registries (as in UK, Australia and Canada). In most countries, one or more cancer registries provide coverage of a sample of the population, although this is by no means random. Registry associations have grown up to deal with issues such as cross-notifications (of cancer patients resident in one area but treated in another), common definitions and coding, quality control procedures and staff training. There are many national as well as other groupings on a regional or linguistic basis (Table 2). Most registry associations hold training courses, sponsor joint research projects and hold scientific workshops and meetings. Almost all PBCRs are members of the IACR.

Figure 1. International Association of Cancer Registries - Membership*.

Cancer Registry Uses: Research

The original function of the cancer registry was to calculate rates of incidence, so that the risk of various cancers in different populations could be compared. Although this still remains their most basic role, the activities of cancer registries have developed far beyond this, to include studies of cancer cause and prevention. The registry may be a source of information on the importance of any risk factors relevant to the local population. At the very least, local researchers should be aware of the registry and the potential for using its database in approved research projects.

Descriptive Studies

Descriptive studies use information from the registry database to examine differences in the incidence (or survival) of cancer, according to variables associated with place (of residence or of birth), time and personal characteristics (sex, ethnic group, social status, etc). Classically, such descriptive studies are said to be “hypothesis generating” – providing clues to possible causes, to be followed up in studies that focus on specific risk factors.

Fig 2 shows time trends in cancer of the colon in two Asian populations, to illustrate the rapidly increasing risk of this disease, where previously it was relatively rare, compared to the incidence in North America and Europe.

Figure 2. Changes in incidence of colon cancer in men over a 35-year period.

Studies of Cause

Cancer registries are not involved in the majority of epidemiological studies of cause – except as a source of incidence rates for studies correlating incidence in different populations with prevalence of exposure to suspected risk factors (so-called “ecological” studies). Cancer registries have, however, been extensively used to follow up defined groups of individuals (“cohorts”) to detect the occurrence of new cases of cancer. They may involve linkage of pre-existing databases with the cancer registry (e.g. registers of specific occupations or of HIV-AIDS). A special example is the study of the risk of second cancers (in relation to the initial cancer and its treatment) – the interest here is in detecting commonality of risk factors (or susceptibility to them) or the adverse effects of treatment.

International Association of Cancer Registries http://www.iacr.com.fr Association of Nordic Cancer Registries http://ncu.cancer.dk Australasian Association of Cancer Registries http://www.aihw.gov.au/cancer/aacr/index.html European Network of Cancer Registries http://www.encr.com.fr Groupe de Coordination pour l’epidemiologie et l’enregistrement du cancer dans les pays de Langue Latine http://www.grellnet.org Gulf Center for Cancer Registration http://www.gccr.org/main.html North American Association of Central Cancer Registries http://www.naaccr.org The Japanese Association of Cancer Registries http://home.att.ne.jp/grape/jacr

Table 2. Some cancer registry associations.

Effectiveness of Interventions in Prevention, Early Diagnosis/Screening and Therapy

The effectiveness of new treatments for cancer has to be demonstrated by means of a randomised controlled trial (RCT), before they are implemented in clinical practice. One might think that this is a reasonable condition to meet before other costly (or potentially harmful) interventions that aim to prevent cancer, or to detect it at an early stage (for example, by screening) are introduced. However, although sometimes the results of RCTs were available before widespread acceptance (e.g. vaccination against HPV in preventing precursors of cervical cancer, or mammographic screening for breast cancer), very often the benefit of a procedure was thought to be obvious, and no RCT evidence was available before widespread implementation (examples are most anti-smoking interventions, cytological screening for cervix cancer and PSA screening for prostate cancer).

The cancer registry has had a limited role in most RCTs. On the other hand, post hoc evaluation of already implemented programs has made wide use of cancer registration. This sort of research is the same as that involved in monitoring a prevention or screening program that was implemented after a proper appraisal of effectiveness, where we are concerned with the efficacy of the program in practice, rather than its theoretical effectiveness in the context of an RCT. Examples of use of cancer registries for the evaluation of programs that have already been implemented are given below.

Cancer registry uses: Evaluation and Monitoring

Primary Prevention

The effectiveness of preventive interventions against cancer has rarely been evaluated by randomized controlled trials. We therefore have to judge their effectiveness by observation of established programs. This may involve comparing the incidence rates actually observed versus those expected (allowing for a time lag for the effects to emerge), where the “expected” rates are based on a prediction model of some kind. This approach has been widely used to evaluate the success of anti-smoking programs. An even more striking example is the dramatic effect on incidence of hepatocellular carcinoma recorded by the cancer registry in Taiwan, following the introduction of vaccination against hepatitis B in the 1980s, first to neonates born of HBsAg positive mothers, then, in 1984, for all newborns. By 1994, it was possible to compare liver cancer incidence in children aged 6-9 born before vaccination was introduced, and those born after (Chang et al, 1997). There was a fourfold difference (Fig 3).

Figure 3. Incidence of liver cancer in children in Taiwan, born before and after introduction of vaccination against Hepatitis B.

Another approach has been to compare incidence rates in areas with or without preventive programs, or with different intensities of intervention.

Screening and Early Detection

Cancer registry data have been widely used for evaluation and monitoring of screening programs. The effectiveness of screening can only be correctly judged by the extent to which the objective of reduced mortality (or reduced incidence, for cancer of the cervix) is achieved. Thus, screening for cervix cancer aims to reduce incidence of invasive cancer; this is the aim of oral cancer detection programs also. Other screening programs (e.g., breast, large bowel), which aim to detect cancers early, do not reduce incidence. Their objective is to decrease the number of deaths.

Ideally, information on the screening status of individuals from a suitable database can be linked to that of the cancer registry in order to study outcome for individuals with different intensities of screening exposure. The simplest study design is the cohort study, in which occurrence of disease (incidence or mortality, as appropriate) is compared in screened vs. unscreened individuals with the objective of deciding the decrease in risk which might plausibly be ascribed to screening. Studies such as these were used to study the potential benefit of cytological screening in preventing invasive cancer of the cervix or mortality from gastric cancer, programs whose effectiveness has never been demonstrated in an RCT. The problem with these studies is that individuals who agree to take part in screening programs are not at the same risk of disease as those who do not, so a lower incidence in those screened may be due to the intervention itself, or to the type of person who chooses to have it.

If screening is known to be able to reduce mortality from cancer (for example, mammography screening for breast cancer), the so-called “intermediate endpoints” - such as tumor size, stage and survival- may be used to monitor the screening process. The cancer registry is essential here, providing information on the screened population and also on the ‘expected’ distribution of cancers by stage or incidence rates (overall, by age group and/or by stage), that would have been expected in the absence of screening. Linkage of the cancer registry and the screening program allows calculation of incidence rates of interval cancers (detected between screenings) and incidence of advanced cancers. The incidence of interval cancers (detected between screenings) is useful in making decisions about the appropriate intervals between tests, and incidence of advanced cancers provides an indication of the probable benefit in terms of mortality.

More usually, there is no information on the screening status of individuals, and population-level analyses are used. The simplest are time trends in incidence, for cancers where screening should prevent invasive cancer, such as in the cervix. Fig 4 shows perhaps the most celebrated example – the trends in incidence in the Nordic countries following the introduction of population screening (Iceland, Finland, Sweden), its partial implementation (Denmark) and postponement (Norway) - Hakama et al, 1991.

Figure 4. Trends in incidence of cancer of the cervix in the five Nordic countries.

Detection of cancers in an early stage by screening will lead to improved survival, whether this results in reduction in mortality (the goal of screening), or is simply due to advancing the date of diagnosis (lead time bias) or differential detection of slow-growing tumors (length bias). But, improved survival can also result from patients presenting earlier with their disease. During 1996-1999, health personnel in Sarawak, Malaysia, were trained to improve their skills in early detection of cancers of the nasopharynx, breast and cervix, and, at the same time, a public education program was carried out to raise awareness of these diseases, and their early signs and symptoms (Devi et al, 2007). Fig 5 shows the results; for both breast and cervix cancer, the percentage of cases presenting in late stages (III and IV) showed marked declines.

Devi B. et al. Ann Oncol 2007 18:1172-1176; doi:10.1093/annonc/mdm105

Figure 5. Percentage of patients presenting at late stage for nasopharyngeal cancer (NPC), breast and cervix in the Department of Radiotherapy and Oncology (DRO), Sarawak General Hospital (SGH) (1991-1999). Reprint with permission of the Annals of Oncology and the authors (from Devi et al., 2007).

Survival and Quality of Life

While cancer control aims to reduce deaths from cancer, the effect of cancer control activities on mortality will often be quite delayed. Changes in survival from cancer will usually be evident much sooner. Differences in survival rates (between populations or over time), as measured by cancer registries are the result of many factors other than the effectiveness of treatment. The methods of data collection and analysis, tumor characteristics (histology, stage, method of detection) and patient characteristics (age, sex, ethnicity, socio-economic status and lifestyle) also affect survival at the population level, as measured by cancer registries. The study of survival gives an indication of the possible role of the process of care, and not simply the effectiveness of a specific treatment. International comparisons of survival, such as those of EUROCARE group1 (Fig 6) have influenced policy making for cancer treatment services in several countries, e.g. the U.K. and Denmark.

Figure 6. Age-adjusted 5-year relative survival from cancer of the colon & rectum (males & females), by country, with area-weighted mean European relative survival. Horizontal bars show 95% confidence intervals (from Berrino et al., 2007).

Clinical care

In the past, research on the quality of clinical cancer care, leading to the development of management guidelines, has been based on the study of clinical case series, often in randomized trials. But series such as these are not at all typical of cancer patients in clinical practice, who are older, with other coincident diseases. Thus, in recent years, increasing attention has been paid to examining variations in the process of care for the cancer patient between different providers or different groups of patients (clients), and cancer registries have proved very useful in this role. For example, they can be used to study the location (e.g. type of hospital) of treatment for specific cancers, and the specialty and case load of individual clinicians treating cancer patients. Although these are indirect indicators of quality of care, they have been shown to be important in determining outcome. Simple measures such as delay in diagnosis, or in receiving appropriate treatment, provide information on equity and access, as well as potentially influencing outcome. Auditing the nature of the therapeutic procedures actually performed may also be possible – for example, in monitoring the proportion of early breast cancer cases receiving breast-conserving surgery followed by radiation therapy (Edwards et al, 2005).

Conclusions

The policy of the World Health Organization is to encourage as many as possible of its member states to develop comprehensive national cancer control programs (NCCPs), as the most efficient method of reducing their cancer burden (WHO, 2002). Planning and evaluation of cancer control needs a surveillance system, and, although mortality statistics (from vital registration systems) can provide much important information, the scope and utility of surveillance is greatly extended by the presence of one or more population-based cancer registries (PBCRs). For planning purposes, a PBCR does not need to cover the entire national population, although for many epidemiological studies this is highly desirable. Coverage of a sample of the population is adequate for many purposes. Ideally, the sample would be a reasonably representative one (as was done with the Surveillance, Epidemiology and End-results (SEER) program in the United States, which now currently covers about 26% of the national population), but even partial, non-random coverage can provide adequate estimates of the national situation, although the sometimes marked heterogeneity of populations in developing countries may necessitate careful planning as to the siting of cancer registries, and significant differences in different populations may be found.

The resources required to operate a PBCR for a limited population (for example, a city or province) are quite modest, comprising, as well as a medical director (part-time with their usual duties), one or two trained data clerks, a personal computer, expenses for office supplies and transport. The benefits, in terms of providing essential information in formulating the cancer control plan, as well as in monitoring its success, means that all countries introducing a cancer control program should attempt to include a PBCR as part of the cancer surveillance program of their NCCP.

¹ Clinical Trials Service Unit & Epidemiological Studies Unit, University of Oxford, Oxford, UK

References

Berrino F, De Angelis R, Sant M, Rosso S, Bielska-Lasota M, Coebergh JW, Santaquilani M; EUROCARE Working group. Survival for eight major cancers and all cancers combined for European adults diagnosed in 1995-99: results of the EUROCARE-4 study. Lancet Oncol. 2007;8:773-83.
Chang M.H., Chen C.J., Lai M.S., Hsu H.M., Wu T.C., Kong M.S., Liang D.C., Shau W.Y. & Chen D.S. Universal hepatitis B vaccination in Taiwan and the incidence of hepatocellular carcinoma in children. Taiwan Childhood Hepatoma Study Group. N Engl J Med, 336, 1855-1859(1997).

Devi BC, Tang TS, Corbex M. Reducing by half the percentage of late-stage presentation for breast and cervix cancer over 4 years: a pilot study of clinical downstaging in Sarawak, Malaysia. Ann Oncol. 2007;18:1172-6.

Edwards BK, Brown ML, Wingo PA, Howe HL, Ward E, Ries LA, Schrag D, Jamison PM, Jemal A, Wu XC, Friedman C, Harlan L, Warren J, Anderson RN, Pickle LW. Annual report to the nation on the status of cancer, 1975-2002, featuring population-based trends in cancer treatment. J Natl Cancer Inst. 2005;97:1407-27.

Fritz A, Percy C, Jack A, Shanmugaratnam K, Sobin L, Parkin DM & Whelan S. International Classification of Diseases for Oncology: Third Edition. Geneva: World Health Organization, 2000.

Hakama M., Magnus K., Pettersson F., Storm H., Tulinius H. (1991) Effect of organized screening on the risk of cervical cancer in the Nordic countries. In: Miller A., Chamberlain J., Day N., Hakama M., Prorok P. (eds) Cancer Screening. Cambridge University Press, pp153-162.

Jensen O.M., Parkin D.M., MacLennan R., Muir C.S. & Skeet R., [eds] Cancer Registration, Principles and Methods (IARC Scientific Publications No. 95) International Agency for Research on Cancer, Lyon (1991).

World Health Organization. National Cancer Control Programmes. Policies and Managerial Guidelines. 2nd Edition. 2002, WHO, Geneva.