Limitations

Overall limitations. The quality and completeness of data transmitted to the National Health Information System (Czech National Cancer Registry, Death Certificate Information System) may affect outputs on the portal. The quality and completeness of the data have been gradually increasing, and the CNCR data are benchmarked against the administrative data of health insurance companies within the NHIS. Nevertheless, due to the use of specific diagnostic systems not reflected in ICD-10, there is a real risk of underreporting for some diagnoses. These are mainly a group of haematological diagnoses, specifically the leukaemia group (C91–C95), CNS tumours, specific tumours of the gastrointestinal tract (neuroendocrine tumours), as well as GIST (gastrointestinal stromal tumours). However, a substantial proportion of haematological diagnoses are adequately described by the ICD-10 system (in particular Hodgkin's lymphoma [C81], non-Hodgkin's lymphoma [C82–C86] and multiple myeloma [C90]) and, from 2019 onwards, the current WHO classification of lymphoid and haematopoietic neoplasms is fully implemented in the ICD-O-3.2 in use.

Limitation of mortality data. There are differences between the cancer-specific mortality values from the population-based Death Certificate Information System and the cancer-specific mortality values from the CNCR. These differences are most pronounced for diagnoses of malignant neoplasm of the oesophagus (C15), malignant neoplasm of the liver and intrahepatic bile ducts (C22), malignant neoplasm of the pancreas (C25), malignant neoplasm of connective and soft tissues and peripheral nerves (C47, C49) and malignant neoplasm of the brain, spinal cord and other parts of the central nervous system (C70-C72). The most significant difference is seen in malignant neoplasms of the liver and intrahepatic bile ducts (C22) and in malignant neoplasms of the brain, spinal cord and other parts of the central nervous system (C70-C72), with mortality rates according to the population-wide database exceeding those found in the CNCR data. The same problems can also be observed in extensive international data, suggesting that the international methodology for identifying a single leading cause of death for statistical purposes is not straightforward. In this case, a significant proportion of deaths from tumours metastasising to the liver, spinal cord or brain appear to be miscoded as deaths from primary liver, spinal cord or brain tumours.

Limitations of data on clinical stage of disease. Staging data for newly diagnosed neoplasms were only recorded in the CNCR until 2018. From 2019 onwards, these data can also be supplemented by data from the National Registry of Reimbursed Health Services according to reported DRG markers. Objective reasons for not reporting the stage of the disease are findings only based on the death certificate or at autopsy, very early death of the patient (within 30 days), patients not treated due to contraindications to cancer treatment or due to refusal of treatment by the patient. If the failure to indicate the stage is not explained, the record is considered erroneously incomplete.