Accuracy of administrative database estimates of national surgical volume: Solid organ transplantation in the National Inpatient Sample.

Abstract

Little is known about the accuracy of procedural coding in the National Inpatient Sample, in part because it is challenging to validate population-level estimates. We evaluated the accuracy of the National Inpatient Sample by comparing estimates of solid organ transplantation to known national transplant volumes from the Organ Procurement and Transplant Network. The mean deviation of National Inpatient Sample point estimates from true transplant volume for the study period was 17.5±20.8%. The mean deviation of point estimates from 2005 to 2011 was 26.4±22.8% compared to 4.9±6.3% from 2012 to 2016 (P<.001). Although future National Inpatient Sample transplantation research may be limited by the inability to subgroup procedures by donor type, surgical procedure coding of solid organ transplantation within the National Inpatient Sample appears to be accurate and reliable for generating national estimates, particularly after the National Inpatient Sample redesign in 2012.