Loading
The registry checked its own work by sending people to 39 hospitals and laboratories to count the cancers by hand, then seeing how many were in the database. About one in sixteen was not.
The Israel National Cancer Registry was established in 1960 and notification has been compulsory since 1982. Reportable disease covers all invasive and in-situ malignancies and neoplasms of uncertain behaviour, and benign as well as malignant tumours of the brain and central nervous system, but excludes basal and squamous cell carcinomas of the skin, which is why Israel's non-melanoma skin cancer figures should not be read as a count.
To measure its own completeness the registry sent abstractors into the medical records departments, pathology and cytology laboratories and oncology and haematology institutes of 39 Israeli medical facilities to identify every reportable case diagnosed or treated in 2005, then linked those cases to the registry database by national identity number. Completeness was the proportion of independently identified reportable cases that the registry already held; timeliness was the proportion of 2005 cases in the database by 31 December 2007.
Completeness was 93.7 percent for all reportable disease, 96.8 percent for invasive solid tumours and 88.0 percent for haematopoietic tumours. Cases diagnosed in the index year were less likely to be in the database than older ones, which is what a reporting lag looks like. The authors judge both measures to meet international guidelines and argue for fully automated reporting.
It is the number to quote when using Israeli registry data, and the reason to treat the most recent year in any registry report as provisional. It also explains why haematological malignancies are the weakest part of the count, and why Israeli non-melanoma skin cancer statistics are not comparable with countries that register it.