Regionalization of complex visceral surgery across the US has followed identification of a volume to outcomes association. However, a simultaneous trend toward improved surgical outcomes might have attenuated this relationship. We hypothesize that the difference in adjusted postoperative mortality between low- (LV), medium- (MV), and high-volume (HV) hospitals has decreased over time.

The National Inpatient Sample (NIS) was used to identify patients undergoing bladder, esophageal, pancreatic, liver, lung, and rectal surgery from 2003 to 2011. Hospitals were divided into LV (<33^rd centile), MV (34^th to 66^th), and HV (>67^th centile) groups. Annual organ-specific adjusted in-hospital mortality (AIHM) for each volume strata was calculated and the difference in AIHM between volume strata was plotted over time.

The proportion of hospitals classified as HV was 6% for lung; 5% for rectal; 3% for esophageal, pancreatic, and bladder; and 2% for liver surgery patients. The AIHM after operation was higher in LV compared with HV hospitals in 2003 to 2005 for all visceral resections except liver surgery. The difference in AIHM between LV, MV, and HV hospitals showed a decreasing trend from 2003 to 2005 to 2009 to 2011 for pancreatic, esophageal, bladder, and lung surgery. For patients undergoing rectal resections, the difference in AIHM was low and stable, and increased for liver resections only.

A reduction in the differences in AIHM among LV, MV, and HV hospitals for 5 of 6 organs studied suggests attenuation of the volume to outcomes relationship with time. This is likely due to system-wide improvements in surgical care.