IARC TECHNICAL PUBLICATIONS - WHO

PrefacePopulation-based cancer registries have provided decisive contributions to cancer epidemiology andcancer control, spanning three quarters of a century. Cancer registrationbegan in earnest in the 1930s and1940s, at the same time that modern epidemiology began to seek thecauses of chronic diseases. Cancerregistration progressively expandedduring the subsequent decades, andcancer registries have now becomedefinitive and unique resources formeasuring the cancer burden in thecommunity (still today, no comparable data system is available forother major diseases). Registrieshave contributed in a number of important ways across the spectrum ofcancer control, from determining theburden and geographical variationin cancer, and thereby aiding understanding of its causes, through topopulation-based survival analysesand assessments of the quality ofdiagnosis and care received by cancer patients.The accumulation and expansion of registry data have enabledgeographical and time trends ofincidence, mortality, survival, andprevalence to flourish. The individual data sets collected havealso fed into a very large numberof analytical epidemiological stud-viiiPrefaceies. More recent developments include research based on registrylinkages with clinical databanksand biological sample repositories.Although these achievements arebecoming standard practice in registries in industrialized countries,much work still remains to ensurea similar development in low- andmiddle-income countries (LMICs).Registry coverage with high-qualitydata remains well below 10% in Africa, Asia, and Latin America, andthere is an urgent need to supportthe initiation, expansion, and development of registries in many LMICs.The approach relies upon the synergy between local resources andwillingness on the one hand, and international cooperation on the other.It is in this context that the International Association of Cancer Registries, an organization with memberregistries across all continents, willbe pleased to link activities and future plans with the ongoing development of the IARC Regional Hubs forCancer Registration, as part of theGlobal Initiative for Cancer RegistryDevelopment (GICR).A key requirement for the development of population-based cancerregistries is resources to supportthe delivery of training. Needs vary,from detailed how-to guides for cancer registrars to instruction in statistical methodologies for the analysisof registry data sets. This guidancedocument provides an overview ofthe key concepts in cancer registration, covering the steps involved inplanning a registry, the sources ofinformation a registry will need toaccess, methods for ensuring dataquality, and how registry resultsshould be reported. As such, it willbe of value to those who are seekingto establish a registry or are in theearly stages of developing a registry. It covers the major componentsthat need to be thought about whensetting up a registry and ensuringthat it provides the necessary information for its main stakeholders –especially those involved in cancercontrol planning.Roberto Zanetti, MD, PhDPresident, InternationalAssociation of Cancer RegistriesDavid Forman, PhDHead, Section of CancerInformation, InternationalAgency for Research on CancerExecutive Secretary,International Associationof Cancer Registries

Executive summaryMore than 20 million new casesof cancer are predicted worldwide in2025, with four fifths of the burdenfalling on low- and middle-incomecountries (LMICs). To understandthe local cancer situation and tacklethe increasing incidence, there is apressing need for planners to haverelevant and unbiased data on thecancer burden in their communities.Population-based cancer registries(PBCRs) provide such informationand are a standard requirement forcancer control planning and evaluation in every country of the world.They are especially valuable inLMICs, where few other populationbased data on cancer occurrenceand outcome are available.In planning a PBCR, there aremany elements to consider, including the definition of the population,the personnel required, the physicallocation of the registry, the necessary equipment and office space,adequate financing, ensuring thatlegal aspects and confidentiality areappropriately addressed, and – lastbut not least – the appointment ofan advisory committee to overseethe activities of the registry. Most ofthe requirements for planning andmonitoring can be achieved throughregistration of a subset (sample) ofthe national population, using oneregional PBCR or a series of regional PBCRs. The political will and support of the key stakeholders are veryimportant at the outset to ensure thesustainability of the PBCR. Successalso depends on the collaboration ofclinicians, pathologists, and staff inadministration in ensuring access totheir data.PBCRs rely on the use of multiple sources of information on cancer cases in the target population.These sources can be grouped intothree broad categories: hospitals,laboratories, and death certificates.Registry procedures allow identification of the same cancer case fromdifferent sources (while avoiding duplicate registrations). The minimumdata set is the list of variables for agiven case that is essential for anycancer registry to collect. Several ofthe variables require coding, to facilitate analysis. Standard, international coding schemes are availablefor some variables, and cancer registries should use them so that comparison of results between registriesis possible. The most important arethe coding of the tumour (site, histology, behaviour, basis of diagnosis),using the International Classificationof Diseases for Oncology (ICD-O),and the coding of stage, using thetumour–node–metastasis (TNM)staging system.As well as collating the data,PBCRs are responsible for analysing and reporting. Cancer incidencereports contain information on all reportable cancers and represent themain deliverable of a cancer registry, providing feedback to the stakeholders and the data providers. Themain components of the report arebackground information, evaluation and presentation of the results,and the tabular section. All PBCRsshould be able to provide some objective indication of the quality of thereported data. The methods available are described in the contextof lower-income settings and coverthe four dimensions of data quality:comparability, validity, timeliness,and completeness.To support the local planningand development of PBCRs in countries within defined regions, a seriesof IARC Regional Hubs for CancerRegistration in Africa, Asia, and Latin America have been established.A tailored set of recommended localactivities involving training, technical guidance, research capacitybuilding, and advocacy are providedto increase the data quality, coverage, and utility of PBCRs in servingcancer control purposes.Executive summary. Planning and developing population-based cancer registration in low- and middle-income settingsix

AbbreviationsASRage-standardized rate (World Standard) per 100 000CDCCenters for Disease Control and PreventionCI5Cancer Incidence in Five ContinentsCTcomputed tomographyDCOdeath certificate onlyDCO%percentage of cases for which the only information came from a death certificateENCREuropean Network of Cancer RegistriesGICRGlobal Initiative for Cancer Registry DevelopmentHBCRshospital-based cancer registriesHDIHuman Development IndexHPVhuman papillomavirusIACRInternational Association of Cancer RegistriesIARCInternational Agency for Research on CancerIARCcrgToolsIARC–IACR Cancer Registry ToolsICD-10International Statistical Classification of Diseases and Related Health Problems, 10th RevisionICD-OInternational Classification of Diseases for OncologyICD-O-3International Classification of Diseases for Oncology, 3rd EditionLMICslow- and middle-income countriesM:Imortality-to-incidence ratioM:I%percentage mortality-to-incidence ratioMRImagnetic resonance imagingMV%percentage of cases with a morphologically verified diagnosisNCDsnoncommunicable diseasesNCRPNational Cancer Registry ProgramPBCRspopulation-based cancer registriesSEERSurveillance, Epidemiology, and End ResultsTNMtumour–node–metastasisUICCUnion for International Cancer ControlWHOWorld Health OrganizationxAbbreviations

CHAPTER 1chapter 1.IntroductionChanging fertility rates, increasing longevity, and changing lifestyleshave led to an increasing burden fromnoncommunicable diseases (NCDs)worldwide: of the estimated 55 million deaths occurring globally in 2011,almost two thirds were deaths fromNCDs (WHO, 2011). The morbidity and mortality from NCDs are setto further increase over the next fewdecades, and for cancer, more than20 million new cases are anticipatedworldwide in 2025, with four fifths ofthe burden falling on low- and middle-income countries (LMICs) (Bray,2014). According to World Health Organization (WHO) estimates in 2011,cancer is now the leading cause ofdeath worldwide (Fig. 1.1).In recognition of the rising burden, WHO Member States during the65th World Health Assembly agreedto adopt a global target of a 25% reduction in premature mortality fromFig. 1.1. The 10 leading causes of death worldwide in 2011, for all ages and bothsexes. COPD, chronic obstructive pulmonary disease. Source: Data compiledfrom the Global Health Observatory Data Repository (http://apps.who.int/gho/data/).Chapter 1. Introduction1

cancer and the other major NCDs(cardiovascular diseases, respiratorydiseases, and diabetes) by 2025. Anaction plan and its monitoring framework have subsequently been adopted to achieve the target (WHO, 2013).NCD surveillance is critical toproviding the information needed forpolicy and programme development,and to support the monitoring andevaluation of the progress made inimplementing NCD policies and programmes. Cancer registries are theonly disease-specific registries thatare in use for NCDs and are therefore of pivotal importance not only inassessing the cancer burden but alsoin measuring the impact of interventions in cancer prevention and control.Population-based cancer registries(PBCRs) are thus a unique sourceof information for research and public health programme monitoring. Inimplementing the NCD action plan,WHO is mandated to report back onprogress towards achieving the nineglobal targets in 2015, 2020, and2025. The monitoring of indicatorsin the WHO Medium-Term StrategicPlan is linked directly to the indicatorsand targets agreed by Member Statesat the 66th World Health 32/B132 27-en.pdf).To provide information on progress in the implementation of the action plan, Member States agreed tothe collection of 25 indicators, including “cancer incidence”. More specifically, the request is for governmentsto collect data on “cancer incidence,by type of cancer, per 100 000 population”, and thus commit to developingand sustaining PBCRs. The PBCRis unique in that it systematically collects and classifies information on allreportable cancers occurring in a geographically defined population frommultiple sources, including hospitals,diagnostic laboratories, and vital statistics departments. As well as collating the data, PBCRs are responsiblefor analysing and reporting. The routine calculation of rates “per 100 000”by PBCRs provides information onhow the cancer patterns are affectingtheir communities and how the trendsin different cancers are evolving.PBCRs provide the solid basis for theplanning, establishment, monitoring,and evaluation of cancer control programmes and the dimension of cancercare services required.This guidance document consistsof six chapters that provide technicaladvice to planners and health specialists in LMICs wishing to implement anddevelop PBCRs as information systems that inform cancer control policy.This first chapter has placed the needfor cancer registration in the contextof the rapidly increasing burden fromthe disease seen worldwide. Chapter2 describes the characteristics of thedifferent types of cancer registry andthe unique functions of PBCRs andtheir present status worldwide. Chapters 3 and 4 outline the critical stepsin planning and developing a PBCRin lower-resource settings, includingdiscussion of the key sources of information required and the minimal standard set of data items that the PBCRshould collect. Aspects in the set-upthat will help ensure the sustainabilityof the registry are emphasized, including comments on infrastructure andresource requirements as well as thecommitment of stakeholders. Chapter 5 describes the main techniquesto evaluate and further enhance thedata quality at the PBCR. Chapter 6provides some advice on reportingthe results to the community at largein support of cancer control and thuspromoting the increasing utility of theregistry.Key points Along with an increasing NCD burden, more than 20 million new cases of cancer are predicted worldwide in2025, with four fifths of the burden falling on LMICs. As a response, WHO Member States have agreed to adopt a global NCD target of a 25% reduction in prematuremortality from the four major NCDs by 2025 and to collect data on cancer incidence by type to provideinformation on progress. PBCRs are critical for collecting and collating such incidence data so as to assess how cancer patterns areaffecting their populations and how trends in different cancers are evolving. They provide the solid basis for theestablishment, monitoring, and evaluation of cancer control programmes.2

The role and statusof population-basedcancer registrationPopulation-based cancer registration represents the gold standard forthe provision of information on cancer incidence in a defined population;PBCRs can serve to identify possiblecauses of cancer in the communityand to assess the impact of cancercontrol activities. A functioning healthcare system is, however, of critical importance to achieve full case ascertainment and an unbiased picture ofthe true cancer burden.Fig. 2.1 compares present levelsof the national Human DevelopmentIndex (HDI) (Fig. 2.1A) versus available sources of cancer incidence(Fig. 2.1B) and mortality (Fig. 2.1C)data. Cancer incidence and mortality data are more commonly available in countries that have attainedhigh or very high levels of HDI; suchcountries have a longer history of reasonably complete national mortalitystatistics, and many have developedeither a national PBCR or one or moreregional PBCRs over the past decades, notably during the 1970s and1980s. In such settings, the activitiesof PBCRs have developed far beyondthe basic role of estimating rates andcomparing cancer profiles in differentpopulations (Parkin, 2006). Registriesexpand their range of activities asthey develop, undertaking studies ofcancer causes and prevention, andproviding the information required inplanning and evaluating cancer control programmes (see Box 2.1).The situation is different for countries presently categorized as havinglow or medium HDI. In many of theseareas of the world, most notably insub-Saharan Africa and South Asia,both vital registration and PBCR systems of reasonable quality have beenslower to emerge.1. What is the status ofpopulation-based cancerregistration worldwide?The concept of population-basedcancer registration has been in existence for well over half a century,with the first PBCR founded in the1930s. Currently, there are more than700 PBCRs worldwide, although theirpace of development has been muchslower in LMICs than in high-incomecountries. This reflects a shortageof human and financial resourcesin LMICs, rather than a lack ofawareness of need. The CancerIncidence in Five Continents (CI5)series (http://ci5.iarc.fr) published byIARC and the International Association of Cancer Registries (IACR) isregarded as the definitive source ofhigh-quality cancer incidence data,and in Volume X published incidencedata for the period circa 2003–2007.Chapter 2. The role and status of population-based cancer registration3CHAPTER 2chapter 2.

Fig. 2.1. Global maps depicting (A) the development level of individual countries, according to the four-level Human Development Index (HDI), based onquartiles, for 2012; (B) status of population-based cancer registries (PBCRs),as of mid-2013; (C) status of vital registration systems, as of mid-2013. Sourcefor A, B: Data compiled from the United Nations Development Programme.ALow HDIMedium HDIHigh HDIVery high HDIBI High-quality PBCR (national)II High-quality PBCR (regional)III PBCR (national or regional)IV Registration activityV No data/status unknownC1 High-quality complete vital registration2 Medium-quality complete vital registration3 Low-quality complete vital registration4 Incomplete or sample vital registration5 Other sources (cancer registries verbal autopsy surveys, etc.)6 No data4Although there is a substantial disparity between high-HDI countriesand low- or middle-HDI countries incoverage in CI5 (e.g. almost completecoverage in North America compared with 10% in South America,Asia, and Africa), the circumstancesare less bleak when it is consideredwhether cancer information is available and can be built upon in a givencountry in these regions.A series of IARC Regional Hubsfor Cancer Registration in Africa,Asia, and Latin America have beenestablished and will be the first pointof call for countries within the respective regions. The Hubs in liaison withIARC develop specific tools in supportof registries to: assess cancer registry quality, publication, and presentation of data assess capacity for registry development and evaluate the quality ofexisting registries m onitor overall progress in expandingcoverage of cancer registration coordinate the development, follow-up, and evaluation of formalagreements between individual cancer registries and IARC, coveringspecific activities consistent with theneeds and recommendations and atime-limited plan coordinate research projects, including the development of monographsin collaboration with IARC, including continental reports based on allavailable registry data in the Hubregion.Currently, the registration status ofthe 138 countries within the six Hubscan be placed into one of five categories, as indicated in Fig. 2.1B:Grade I. High-quality PBCRs (included in CI5 Volume X) and nationally representative (registries withcoverage of 50% of the country’spopulation)Grade II. High-quality PBCRs(included in CI5 Volume X) and regionally representative ( 50% coverage)Grade III. Nat

iarc technical publications planning and developing population-based cancer registration in low- and middle-income settings freddie bray, ariana znaor, patricia cueva, anne korir, rajaraman swaminathan, andreas ullrich, susan a. wang, and donald maxwell