Comparison of Hungarian Central Statistical Office's causes of death data with the database of the Hungarian National Cancer Registry / A Központi Statisztikai Hivatal halálozási adatainak összevetése a Nemzeti Rákregiszter adatbázisával. Egy adat-összekapcsolás tanulságai

ABSTRACT

Introduction: In international comparison, Hungary is in the forefront of cancer incidence and mortality statistics. Based on paper-based death certificates, mortality statistics are compiled by the Hungarian Central Statistical Office, while population-based measures of cancer incidences are performed by the Hungarian National Cancer Registry. Objective: Linking the records of these two independent databases can highlight their weaknesses and provide an opportunity to reconcile and verify collected data, which may emphasize the need to expand current data exchange protocols. Method: Based on the Hungarian unique health care insurance ID, the mortality database of the Hungarian Central Statistical Office between 2012 and 2020 was compared with the data of the Hungarian National Cancer Registry from 2001 to 2020. Deaths in 2018, in particular those related to lung cancer, were examined in more depth to demonstrate the biases resulting from erroneous data collection. Results: The mortality database of the Hungarian Central Statistical Office contained 32 586 cases with an underlying cause of death of malignant neoplasm for 2018, of which 29 970 were identified in the Hungarian National Cancer Registry. Out of the 8716 deaths coded to lung cancer, 7957 corresponding individuals were also found in the Registry. From the matches, 7381 cases were marked with lung cancer in the Hungarian National Cancer Registry. For the remaining 576 cases, the Registry recorded different types of cancers, of which in 69 cases with lung metastasis. Discussion: The differences between the two databases may be caused by methodological differences in data collection, incomplete, inaccurate reporting and differences in processing algorithms. Nevertheless, the majority of the data in the examined databases were found to be appropriate for epidemiological studies. Conclusion: Based on the outcomes of the present analysis, a revision of the data transfer between the two institutions is in order. The introduction of electronic Death certificate recording and validity checks are expected to improve the reliability of ID numbers and may shorten data processing times.