Estimates of thyroid cancer incidence at district level using cancer registries data and linkage of two sources of medico-administrative data, France, 2007-2011

Objectives - In France, cancer registries cover 20% of the population. The objective of this study was to provide estimations of thyroid cancer incidence at the administrative district level ( 'departements') over the 2007-2011 period in metropolitan France, using registries data and medico-administrative data. Methods - A medico-administrative indicator 'HUL' [Hospital union LLD] combining Hospital discharge data and health insurance data on Long Duration Diseases (LDD) was constructed. It counts the number of people with a new LDD or hospitalized for thyroid cancer. The principle of the estimation consists in adjusting HUL data at the district level by the Incidence/HUL ratio of the registry area. The accuracy of the estimations was first evaluated in the districts covered by registries by comparing estimated to observed incidence. Results - The preliminary phase of the assessment confirmed that the HUL/Incidence ratio were sufficiently accurate to provide district level estimates of thyroid cancer incidence over the whole territory. The district variability of the estimations for the 2007-2011 period was important: the 5. and 95. percentiles of the distribution of standardized incidence rates across districts were 2.8 and 7.1 for 100,000 in men, and 8.3 and 21.2 for 100,000 in women. Areas with significant over-incidence were observed in the South-East, and on the South-West coast of France compared to the national level. Conclusion - This study confirms the magnitude of geographical variability of thyroid cancer incidence between French administrative districts. It also confirms the value of cross-referencing medico-administrative data to estimate the incidence at an administrative district level. (authors)