The SAGES bariatric surgery outcome initiative

Background: The recent initiative for identifying centers of excellence in bariatric surgery calls for documentation of surgical outcomes. The SAGES Outcomes Initiative is a national database introduced in 1999 as a method for surgeons to accumulate and compare their data with summary national data. A bariatric-specific dataset was established later in 2001. The aim of this study was to compare the outcomes of bariatric surgery from the Society of American Gastrointestinal Endoscopic Surgeons' (SAGES) bariatric database with data derived from a national administrative database of academic centers. Methods: Between 2001 and 2004, 24 surgeons with 1,954 patients participated in the SAGES Bariatric Outcome Initiative, and 97 institutions with 42,847 patients participated in the University HealthSystem Consortium (UHC) database. Only 7 of the 24 surgeons participating in the SAGES Bariatric Outcome Initiative submitted more than 50 cases. The main outcome measures included demographics, comorbidities, type of bariatric procedure, operative time, length of hospital stay, short- and long-term complications, mortality, and weight loss. Results: Both datasets were comparable for gender. Roux-en-Y gastric bypass had been performed for 88% of the patients in the SAGES database and 96% of the patients in the UHC database. Associated comorbidities were similar between the two groups except for a higher rate of hyperlipidemia for the patients in the SAGES database. The SAGES database contains more bariatric-specific information such as body mass index, operative time, blood loss, bariatric-specific complications, long-term complications, and weight loss data than the UHC database. According to the available data, no statistically significant differences exist between the two datasets in terms of perioperative complications and mortality. Conclusions: The SAGES Bariatric Outcome Initiative provides valuable bariatric-specific data not currently available in an administrative database that may be useful for benchmarking purposes. However, this database is currently underutilized.