Objectives: to assess whether the data source of cancer exemption ticket (code 048) correctly estimate the cancer incidence produced by Cancer registries (CR).
Design: comparison between incidence estimates produced by cancer exemptions ticket and cases registered by CR.
Setting and participants: six CRs provided incidence data for one year in the five-year period from 2007 to 2011 and for the previous five years, the exemptions provided for the same year and for the previous five years.
Main outcome measures: incidence distribution by gender, age and tumour site, exemptions 048/incident cancers ratio, and trend estimates.
Results: out of 14,586 patients with 048 exemption, a first group was present in the CR database in the same reference year (No. 8,015) and a second group in the previous 6 months (No. 1,696). Of the remaining 4,875, only 2,771 were prevalent cases and 2,104 were manually re-valued: 514 non-cancer; 710 non-malignant/noninfiltrating tumours, 250 non-residents, 532 unknown, and 98 lost at CR. The exemption/ tumours ratio was 32%in males and 37% in females. Out of 27,632 cancer patients in CR, only 29% had a 048 exemption. Among linked cases, there is a case-mix problem: the exemptions overestimated the weight of some cancer sites (breast, prostate), but underestimate the weight of other sites (stomach, liver, lung) and the burden of tumours in the elderly.The trend estimated from the exemptions underestimates the true incidence of tumours and presents fluctuations, because of local administrative and organisational issues.
Conclusions: the 048 codes are an accessory source for CRs, but when used as single flow they are not able to estimate the true incidence of tumours and, therefore, do not provide useful information on cancer trends.