Objectives To describe optimization of a nationwide newborn screening program for cystic fibrosis (CF) that combines an immunoreactive trypsinogen (IRT) assay and DNA mutation analysis in dried blood samples at day 3. Study design Data from regional screening laboratories and CF care centers were centralized and periodically analyzed to allow adaptation, thus limiting the number of false-positive cases. Results A total of 2 717 905 infants were screened between 2002 and 2005. Flow chart protocol was modified twice. First, the IRT d3 cutoff value increased from 60 to 65 μg/L, thus decreasing the percentage of samples requiring mutation analysis from 0.82% to 0.64%. Second, for infants with no mutations using the screening panel, a recall for IRT was performed only if IRT d3 was > 100 μg/L; the percentage of recalls decreased from 0.51% to 0.12%, and the percentage of infants requiring a sweat test decreased from 0.14% to 0.01%. No significant change in the CF detection rate was observed after these 2 modifications. A total of 625 CF cases were detected, and 22 false-negative findings (3.4%) were observed, most of them inevitable, with a low initial IRT. Conclusions The centralized data analysis led to changes in the screening strategy to optimise the newborn screening program.