Data Sources

The evaluation is based on data submitted annually to the InVS by the managing entities using extraction modules integrated into their specialized software for database management. Two types of data are submitted:

• data aggregated at the departmental level, used to calculate program participation,

• and anonymized and standardized individual data, known as evaluation data, which enable the calculation of evaluation indicators.

This data is validated by the administrative bodies and by the InVS.

The format of the reported data and the indicators are described in a guide, the first version of which was developed in 2009 by a working group comprising members of the involved institutions (DGS, Health Insurance, INCa, InVS), coordinating physicians from cancer screening management organizations, gastroenterologists, and anatomic cytopathologists. The latest updated version of this guide is available on the InVS website.

Aggregated data are provided in the form of cross-tabulations (by age and sex) and provide information on the number of people who underwent a screening test, the number of people excluded from screening for medical reasons, and the number of people who tested positive during the evaluation period.

Individual evaluation data are extracted as "text" files containing one line per person who underwent a screening test during the evaluation period and 58 variables for each individual.

Program Participation and Evaluation Indicators

Program participation indicators are calculated for rolling two-year periods (2009–2010, 2010–2011, 2011–2012, 2012–2013, 2013–2014…) that reflect the biennial frequency of program invitations (the entire target population of each department must be invited over a two-year period). The reference population for calculating participation rates is derived from population projections by the National Institute of Statistics and Economic Studies (INSEE) (2007–2042, central scenario) to ensure consistency in estimates across departments. To calculate the participation rate, individuals excluded from screening for medical reasons during the evaluation period (permanent or temporary exclusions) are subtracted from the target population aged 50–74. Information on the number of people excluded from the program for medical reasons also allows for the calculation of program exclusion rates. These rates are standardized using a direct standardization method based on the age and sex structure of the French population aged 50–74 in 2009 (INSEE population projections, central scenario 2007–2042).

Indicators of the quality of testing, examinations, follow-up, and lesion detection are estimated by department (where sample sizes permit) and for France as a whole. They are calculated according to sex, age, and screening stage (first screening versus subsequent screenings). They are also calculated over two-year periods to reflect the duration of an invitation campaign. The rates of positive tests, detection of advanced adenomas, and detection of colorectal cancers are standardized as described above.

To calculate participation indicators for the 2013–2014 period across the 99 departments, aggregated data were extracted by the administrative units in January 2014, and individual 2011–2012 screening data were transferred to the InVS between April 1 and October 14, 2014.