Abstract

BACKGROUND: Efforts to improve healthcare value (quality/cost) have become a priority in the US. While many seek to increase quality by reducing variability in adverse outcomes, less is known about variability in costs. In conjunction with the AAST Healthcare Economics Committee, the objective of this study was to examine the extent of variability in total hospital costs for two common procedures: laparoscopic appendectomy (LA) and cholecystectomy (LC).METHODS: Nationally-weighted data for adults =18y was obtained for patients undergoing each operation in the 2014 and 2016 National Inpatient Sample. Data were aggregated at the hospital-level to attain hospital-specific median index hospital costs in 2019 USD and corresponding annual procedure volumes. Cost variation was assessed using caterpillar-plots and risk-standardized observed/expected cost-ratios. Correlation analysis, variance decomposition, and regression analysis explored costs' association with volume.RESULTS: In 2016, 1563 hospitals representing 86,170 LA and 2276 hospitals representing 230,120 LC met inclusion criteria. In 2014, the numbers were similar (1602, 2259 hospitals). Compared to a mean of $10,202, LA median costs ranged from $2,850 to $33,381. LC median costs ranged from $4,406 to $40,585 with a mean of $12,567. Differences in cost strongly associated with procedure volume. Volume accounted for 9.9% (LA) and 12.4% (LC) of variation between hospitals, after controlling for the influence of other hospital (8.2%, 5.0%) and patient (6.3%, 3.7%) characteristics and in-hospital complications (0.8%, 0.4%). Counterfactual modeling suggests that were all hospitals to have performed at or below their expected median cost, one would see a national cost savings >$301.9 million/year (95%CI: $280.6 to $325.5 million).CONCLUSIONS: Marked variability of median hospital costs for common operations exists. Differences remained consistent across changing coding structures and database years and were strongly associated with volume. Taken together, the findings suggest room for improvement in EGS and a need to address large discrepancies in an often-overlooked aspect of value.LEVEL OF EVIDENCE: Epidemiological, III.

View details for DOI 10.1097/TA.0000000000002611

View details for PubMedID 32039972