Hospital Markup and Head and Neck Cancer Surgery Outcomes in the United States

Presentation: AHNS-QS-099
Topic: Value / Quality
Type: Quickshot
Date: Thursday, May 2, 2019
Session: 5:30 PM - 6:00 PM
Authors: Warren C Swegal, MD1, Peter Vosler, MDPhD1, Carole Fakhry, MD, MPH1, David W Eisele, MD1, Kevin D Frick, PhD2, Christine G Gourin, MD1
Institution(s): 1Johns Hopkins Medicine, 2Johns Hopkins Carey Business School

Context: Health care spending continues to grow at a rate that is predicted to exceed growth in the US economy, with marked variation in hospital costs and payments a target for healthcare reform efforts. Limited data exists to explain variability in prices for head and neck surgical procedures, and if variations in surgical price are associated with outcomes.

Objective: To characterize variations in hospital price markup for head and neck cancer surgical procedures, and examine associations with postoperative complications and in-hospital mortality.

Design, Setting, and Participants:  The Nationwide Inpatient Sample was used to identify 157,464 patients who underwent head and neck cancer surgery at 4,833 hospitals for a malignant upper aerodigestive tract neoplasm in 2001-2011. Markup ratio (charges to costs) was modeled as a continuous and categorical variable. Hospital volume was modeled as a categorical variable. Hospital market concentration was evaluated using a variable-radius Herfindahl-Hirschman Index from the 2001, 2003, 2006, and 2009 Hospital Market Structure files.

Main Outcomes and Measures: Cross tabulations and multivariable regression was used to evaluate associations between markup ratio, hospital and patient variables, markup ratio, postoperative complications and in-hospital mortality.

Results:  Hospital markup ratios ranged from 0.8-8.7, with a mean markup ratio of 2.8 ( 95% CI 2.7-2.8). Hospitals in the lowest markup ratio quartile (low markup) had a mean markup ratio of 1.82 (1.76-1.87), while hospitals in the top markup ratio quartile (extreme markup) had a mean markup ratio of 4.02 (3.9-4.1).  Compared to low markup hospitals, extreme markup hospitals were more often large (78.4% vs. 68.3%), urban (99.1% vs. 89.4%), private for-profit hospitals (18.7% vs 1.5%) and were less likely to be high-volume hospitals (12.6% vs. 19.2%) or in competitive markets (64.5% vs 85.1%). Postoperative complications occurred more often in extreme markup hospitals (22.8% vs. 17.0%). Overall in-hospital mortality was 1.0% and did not vary by markup ratio. On multivariate analysis, a significantly higher markup ratio was associated with private, for-profit hospitals (50.6% [36.6-66.1%], P<0.001), urban hospitals (31.0% [25.4-36.9%]; P<0.001), hospitals located in the West (27.2%; 14.6-41.2%]; P<0.001), and advanced comorbidity (3.5% [1.0%-6.0%]; P=0.0006). The extent of surgery, hospital volume, market share, morbidity, and mortality were not associated with significant differences in markup ratios.

Conclusions and Relevance: There is wide variation in hospital markup for head and neck cancer surgery, with a four-fold increase in charges relative to costs in 25% of hospitals. Variations in surgical price are not associated with the extent of surgery, hospital volumes, or outcomes. These data suggest that greater transparency is needed to address disparities in hospital pricing in an era of cost containment.