What Accounts for the Markup in Hospital Charges?
Our analysis aimed to (i) quantify the difference between what hospitals charge for inpatient care and the amount Medicare reimburses for that care, and (ii) identify the drivers of that difference.
For the first aim, we performed log-linear regression on the CMS Inpatient Charge Data from Fiscal Year 2011 for each provider that provided care for more than 11 Diagnosis Related Groups (example, first figure). The residuals after regression were normally distributed. To look at those data in the aggregate, we created a histogram to show the distribution of ratios of hospital charges to Medicare reimbursements (slopes from log-linear regression, second figure). No statistical test is needed to see that the first moment of the distribution of markup ratios is greater than one because no element of the distribution crosses the vertical line that denotes unity.
For the second aim, we used the HCAPS Survey of Patientâ€™s Hospital Experiences, the Hospital Outcome of Care Measures, and the Hospital Value-Based Purchasing Performance Scores to create scatter plots (final four figures). Scatter plots identify relationships between two variables, regardless of the form of those relationships.
We created a static visualization to help the public better understand drivers of healthcare costs. We use an analogy with car sales to explain our data sources and the measure we created, the markup ratio. Our visualization helps the public better understand what drives hospitals to mark up prices by showing that expected drivers of healthcare costs do not drive the markup.