|The NORDCAN database (Version 7.3, 07.2016)|
1. Warning: Coding of cancer incidence and cancer mortality might differ between countries, especially in earlier years. We have tried to convert data from the national cancer and mortality registries to datasets according to international rules, as described in some of the notes below. Some differences due to variations in coding between countries may still exist. Some of the known problems are also described below.
2. The incidence data was delivered by the national Cancer Registries. In
collaboration with the Registries, data was first recoded to ICD-O-3. The
IARCcrgTools software was used in the process of recoding to ICD-O-3
and afterwards to ICD-10, and to count multiple cancers in the same site as one
only (following the IARC/IACR
rules, 2004). At last the data was converted into the NORDCAN
“entity” schema. In consequence, the numbers presented in NORDCAN may
not correspond exactly to those published by the National Cancer
Registries or in international studies such as
Cancer Incidence in Five Continents.
The incidence of bladder tumours may not be comparable between the Nordic countries due to varying coding practice over time concerning non-invasive tumours. From version 7.0 (December 2014) and for Finland from version 7.1 additional codes (ICD-10 D30.1, D41.1-3, D41.5-9) have been included in the definition. For Finland the ICD-10 codes D09.0+D41.4 were also new from version 7.0.
The incidence of brain and central nervous system tumours may not be comparable between Nordic countries because of the inclusion of benign or unspecified tumours. From version 7.0 (December 2014) additional codes (ICD-10 D35.2-4, D44.3-5) have been included in the definition.
The incidence of skin, non-melanoma before 1978 is not comparable between Denmark and the other Nordic countries because Denmark was the only country which included basal cell carcinoma in the period 1943-1977.
The all cancer group including non-melanoma skin is not comparable between countries before 1978 due to the Danish problems with inclusion of basal cell skin cancer in non-melanoma skin before 1978.
In Denmark a change in prostate cancer incidence level took place from 2004 due to inclusion of registrations solely based on pathology register information.
For Lip, oral cavity and pharynx new subgroups based on the topography classification used in the TNM-manuals are available from 1978 but in cancer incidence maps already from 1971.
For leukaemia seven subgroups are available from 1978, but not yet for Sweden and the Nordic countries together.
Two new groups, “Other specified cancers” and “Unknown and ill-defined”, are shown from 1978 and in maps from 1971. These cancers were previously only shown as part of the all cancer groups.
The Swedish cancer registry does not search for and perform follow-back in other registers for cancer cases identified from death certificates, as the other Nordic cancer registries do. This may lower Swedish incidence rates, especially for older ages and for cancers types with a high lethality, and thus result in too optimistic survival, since incident cases with short survival may not be included. Lung and pancreatic cancer are examples of cancer types where Swedish incidence rates are not directly comparable to rates from other Nordic countries and the use of mortality rates for comparisons is recommended.
3. The mortality data was first converted to ICD-10 and then into the NORDCAN
“entity” schema. The data is based on the national mortality registers. In
Finland the mortality data was based on the cancer register record indicating
whether or not the patient died from the registered cancer.
Numbers of Danish cancer deaths from 2007 are too low; 3-5% of all Danish death certificates from 2007-2014 are missing cause of death. In some of them the actual cause-of-death might have been cancer.
Gallbladder mortality is missing, and here set to 0 up till 1960 for Denmark and Sweden, 1957 for Iceland, and 1968 for Norway.
Pleura mortality is missing, and here set to 0 up till 1960 for Denmark and Sweden and until 1968 for Norway.
|4. Prevalence is calculated for each country from follow-up of each cancer patient for death and emigration through linkage of the cancer register and the population register.|
5. The survival data originated from the project “The survival of cancer patients diagnosed 1964-2003 in the Nordic countries” published in Acta Oncologica 2010 issue 5, now updated and supplemented with 1-year survival and calculations for newer periods. Links to the papers can be found at www.ancr.nu/cancer-data/cancer-survival/.
For Finland, patients were only followed-up for death or emigration to the end of 2014. The other countries followed patients to the end of 2015.
From version 7.1 Finnish survival is 1-2% point higher for each site and all years (back to the 1960s) than in previous versions, due to a higher number now being excluded from the survival calculations as tumours identified from death certificate alone (DCO) or as incidental finding at autopsy.
|6. In the cancer incidence maps the new additional codes for “Bladder etc.” (ICD-10 D30.1, D41.1-3, D41.5-9) and “Brain, Central nervous system” (ICD-10 D35.2-4, D44.3-5) have not yet been included for any country and the incidence levels might thus be slightly too low.|
|7. The database is organised by country, data type (incidence, mortality, prevalence and survival), year or period (survival data), cancer, sex and age-group. The NORDCAN database does not contain individual records.|
|8. Data from the Faroe Islands and Greenland are new and should be considered preliminary. They are therefore not included in the "Nordic countries" group.|
|9. Regional information might not add up to the national level since a few persons do not have information on region.|
The data files have been converted at the Secretariat for NORDCAN. The final database for the Internet access has been build up by the Section of Cancer Surveillance (CSU) at IARC.
Different coding classification systems have been used over time. Jumps in the incidence and mortality rates might occur at the points of change due to ICD-versions not being completely comparable (i.e. liver cancer) and/or conversion problems. The tables below show for each country the time periods where a specific ICD version has been used in the coding of diagnoses for incidence and for mortality.
Queries about the data process should be directed to the Secretariat for the NORDCAN product, Danish Cancer Society, Dept. of Documentation & Quality, Strandboulevarden 49, DK-2100 Copenhagen O. Att. G. Engholm (firstname.lastname@example.org) or H.H. Storm (email@example.com) or to the national representatives.
|NORDCAN © 2011 Association of the Nordic Cancer Registries - All Rights Reserved.|