Retrospective inpatient study shows gradual increase in upcoding for gastrointestinal surgeries
Findings from a retrospective study conducted by researchers at Johns Hopkins University School of Medicine in Baltimore show that the number of upcoded cases for patients who underwent gastrointestinal surgeries increased over a ten-year period. Upcoded cases were also significantly correlated with higher inpatient costs.
Researchers collected data from the National Inpatient Sample (NIS)—the largest inpatient administrative database in the U.S., which includes demographic and clinical data on inpatient admissions from approximately 35 million patients annually.
Data from the NIS was used to identify 1,344,152 inpatients over the age of 18 who underwent one of six gastrointestinal operations (i.e., colectomy, rectal resection, gastrectomy, esophagectomy, pancreatectomy, hepatectomy) between 2001 and 2011.
Researchers then recorded the total number of ICD-9-CM diagnosis codes reported per hospitalization and compared the values to the average number of codes reported for all hospitalizations included in the study. Hospital admissions with greater than the average number of diagnosis codes were classified as upcoded admissions, while those with less than the average number of diagnosis codes were classified as appropriately coded admissions.
Researchers also used recorded hospital charges to estimate the total in-hospital costs for each inpatient admission.
Findings showed that over the ten-year period, the proportion of upcoded admissions (i.e., over nine ICD-9-CM codes reported per admission) increased from 14.1% to 32.9%. The trend was observed for all six gastrointestinal operations but was greatest for hepatectomy procedures.
Additionally, adjusted analyses showed that upcoded admissions were independently associated with a $13,754 greater in-hospital cost.
These findings were presented at the 13th Annual Academic Surgical Congress of the Association for Academic in February and recently published in Elsevier’s Surgery.