High-tech versus high-touch: Components of Hospital Costs and Charges Vary Widely By Clinical Condition

Background & Study Objective:  Information on hospital prices has historically lacked transparency.  In 2013, the Centers for Medicare and Medicaid Services released average charges and payment data for 100 Medicare-covered services based on billing data for over 3,000 hospitals.  The release of the data sparked a national dialogue regarding the variability in hospital prices across hospitals and the seemingly lack of correlation between hospital prices and actual payment.  In order to begin to understand hospital prices, it is important to first understand the underlying costs of providing a service that drive hospital prices.  In this study, we examine five common MS-DRGs to address the following questions:  1) What are the components that drive the overall cost and price of a hospital service? and 2) How do these cost components differ by type of service, hospital characteristics, and over time. 

Data and Methods:  Hospital level data representing 4,785 unique hospitals come from the Medical Provider and Analysis Review (MEDPAR) files for fiscal years 2008 and 2012.  We investigate cost and charge components for five different Medicare Severity-Diagnosis Related Groups (MS-DRGs): Total Joint Replacement, Septicemia, Acute Myocardial Infarction (AMI), Unstable Angina and Appendectomy.  These MS-DRGs were selected for various reasons.  Total Joint Replacement and Septicemia represent the most common surgical and medical MS-DRGs among Medicare discharges, respectively.  AMI and unstable angina are heart conditions commonly investigated in the health services research literatures, and appendectomies represent a routine, non-elective procedure.

Results:  Our preliminary results suggest that there is considerable variation in distribution of cost components (e.g. nursing, medical supplies, labs, radiology, operating room) depending on the clinical condition.  For example, nursing care represents less than a quarter of the total costs for joint replacement but over half of total costs for septicemia.  Medical supplies, which includes the cost of medical devices account for over 40 percent of the total costs for joint replacement but are only 5 percent of costs for septicemia.  Total average costs and payment by condition were similar among government, not-for-profit and for-profit hospitals; however, charges were significantly higher in for-profit hospitals.  The cost-to-charge ratio for for-profit hospitals is 0.265 compared to 0.381 in not-for-profit and 0.53 in government-owned hospitals.  On average, costs increased 11 percent between 2008 and 2012 while charges increased twice that rate at almost 25 percent.

Conclusions and policy implications:  Cost components that drive total cost vary substantially by MS-DRG.  This suggests that strategies to reduce costs may be most effective if tailored to the clinical condition.  Moreover, new payment designs that are based on the expected costs of treating an episode of care, such as bundled payments or other risk-sharing arrangements, must account for variation in key cost drivers and market conditions that may affect input prices over time.  Finally, understanding the variation in cost components and charge structure by medical condition and hospital characteristics will help inform current discussions around hospital prices.