A surveillance system for severe influenza cases requiring intensive care was established in 2009. It is coordinated at the regional level during each flu season by the CIREs in collaboration with intensive care units. After several seasons, variations in reported case rates were observed depending on the predominant virus and the region. The objective of this study was to assess the completeness of surveillance, both overall and at the regional level, and to identify factors that may contribute to these variations. The approach involved comparing cases reported through the surveillance system with an independent data source: hospital stays recorded in the PMSI. The cases selected were patients admitted to intensive care with a diagnostic code for influenza. Cross-referencing the cases identified in both sources allowed for the estimation of the total number of cases using a capture-recapture method. A description of the various regional surveillance methods was compiled through a survey of the CIREs. For the 17 former metropolitan regions that participated (out of 22), the total number of cases was estimated at 4,435 [4,352–4,519] for the 2009–2010 through 2012–2013 seasons. Surveillance coverage was estimated overall at 47% [47–48%], with regional variations ranging from 28% to 74%. The practice of certain CIREs of sending a weekly report to clinicians was associated with higher coverage (63% vs. 46%, p=0.02). The completeness of influenza coding in the PMSI was estimated at 73% [72–74%]. The estimated overall incidence varied by season from 14 to 30 cases per million inhabitants. In each of the seasons studied, a geographic gradient was observed, with higher incidences in eastern France. The study identified areas for improvement, including the routine use of PMSI data—consolidated at the end of the season—to supplement surveillance data. Streamlining the management of the current network is under consideration, with a shift toward a sentinel system and optimization of regional surveillance protocols.