Skip Navigation Archive: U.S. Department of Health and Human Services U.S. Department of Health and Human Services
Archive: Agency for Healthcare Research Quality
Archival print banner

This information is for reference purposes only. It was current when produced and may now be outdated. Archive material is no longer maintained, and some links may not work. Persons with disabilities having difficulty accessing this information should contact us at: Let us know the nature of the problem, the Web address of what you want, and your contact information.

Please go to for current information.

CareScience Risk Assessment Model - Hospital Performance Measurement

Presentations from a November 2008 meeting to discuss issues related to mortality measures.

Appendix A—Calculating Costs

In the hospital setting medical "cost" does not have a common, single definition, in part because of the diverse reasons for tracking costs. Record keeping at the patient level should ideally distinguish at least three different aspects of medical cost and its components. First, the most straightforward patient-oriented view of cost is the value of all resources directly included in the provision of individual patient care. Examples of these variable costs include nursing, technical, and physician labor, as well as any tests/procedures performed or supplies consumed. These costs are often the basis for clinical profiling, because they are presumed to be under the control of treating physicians. A second broader measure of total hospital cost includes, in addition to the aforementioned variable costs, all overhead and administrative costs (fixed costs). These costs are not limited to patient care directly, but rather encompass the costs of running the hospital and financing its physical plant and equipment. Most of these components are included in Medicare's Hospital Cost Report Information System (HCRIS) and are grouped by hospital departments. Third are cost reimbursements, the dollar amount paid to the provider for each case. Reimbursements are not necessarily the true treatment costs, but rather the amount that the payer has determined to be typical for the specific case mix and care requirements, in this case Medicare's RBRVS (Resource Based Relative Value Scale) computations. These, too, are available from the Center for Medicare and Medicaid Services (CMS) and are grouped by DRGs.

One other dimension to tracking costs and charges deserves note. Aggregate or departmental dollars can be assembled for the entire facility or restricted to the inpatient population only. Facility level cost-to-charge ratios are often much greater than those limited to inpatients. There are two reasons: (1) outpatient costs are much closer to billings; and (2) facility level aggregations more naturally allow the inclusion of overhead (fixed) costs.

A number of published studies have assessed the measurement of the costs of providing medical care in hospital settings. When comparing costs derived from departmental cost-to-charge ratios to costs derived from RVUs14, CCR-calculated costs were shown to be a reliable methodology for estimating hospital costs and for comparing average costs for a collection of patients in a given DRG among hospitals.15 Hospital-level CCRs, however, are not a reliable way to assess costs for individual patients when evaluated by the RVU "gold standard." Nonetheless, when large representative samples of patients are the basis for comparison, hospital-level CCRs can be useful for overall cross-hospital profiling.

The accuracy of cost-to-charge ratios is influenced by a number of hospital factors, including nursing and physician labor, teaching status, patient risk and demographics, insurance status, and even location. An alternative to using RVUs to assess the reliability of CCRs is to compare hospital charges to costs that are derived from internal cost accounting systems, for those hospitals that have them.

Some medical centers, for example, have adopted a cost accounting and management information system developed by Transition Systems, Inc. (TSI). This system relies on DRGs to allocate costs and involves the following process: First, all billable procedures in the hospital are grouped into intermediate products which are then mapped into departments. Within each department, each intermediate product is assigned a relative weight for up to nine cost categories—variable labor,16 variable supplies, variable other, fixed direct labor,17 fixed direct equipment, fixed direct facilities, fixed direct other, variable indirect, and fixed indirect. This relative weight is then multiplied by the annual volume of the product and then summed over all products per department to determine the total relative weight units expended by the department. These costs are broken down into the nine cost categories. One main problem with this system is that hospitals differ in how they define intermediate products and in how they map them into departments.18

The CareScience data model explored a number of alternatives and after extensive testing found that the best approach was to exploit the Medicare cost and charge information from the minimum data set hospital cost report (HCRIS) to create cost-to-charge ratios (CCR) by department. The cost and charge information is for the entire facility, organized by departments, and focuses on ancillary services without distinguishing variable and fixed costs. The application of these ratios to client data poses challenges, because the CareScience Data Manager tracks charges at the patient level rather than by department. Some studies have shown that the overall hospital CCR, in contrast to departmental CCRs, is not a reliable basis for comparing costs across hospitals. These computed CCRs contain some extreme values, which require trimming. After trimming (excluding the top and bottom 1 percentile) the average ratio is 0.62 with a standard deviation of 0.23. These total facility ratios are somewhat smaller than the average (unweighted) departmental ratios with which they correlate at about 86 percent.

Our ultimate goal is to accurately account for costs and resources attributable to an individual patient's care. Meeting it would require detailed information from our clients both in terms of charge breakdowns by department and complete cost accounts. This kind of information could help us build a model of CCR-based costs. CCR derived costs have been shown to be reliable when comparing the relative cost of patients in a DRG in one hospital to the average cost of patients in the same DRG in a group of hospitals.19 This could be accomplished if we obtained charge break downs on a departmental level. These charges could then be multiplied by departmental CCRs obtained either from the hospitals own total annual departmental costs and charges or those calculated from the CMS minimum data sets.

14 RVU, relative value units, describes a procedure's intensity of resource usage. The details about RVU can be found in the section of Clinical Knowledge Base.
15 Shwartz M, Young DW, Siegrist R. The Ratio of Costs to Charges: How good a basis for estimating costs? Inquiry 32:476-81. 1995.
16 Variable labor is direct patient care or services.
17 Fixed labor includes strictly administrative staff and caregivers when they perform administrative duties.
18 Shwartz M, Young DW, Siegrist R. The Ratio of Costs to Charges: How good a basis for estimating costs? Inquiry 32:476-81. 1995.
19 Shwartz M, Young DW, Siegrist R. The Ratio of Costs to Charges: How good a basis for estimating costs? Inquiry 32:476-81. 1995.

Return to Article Contents
Proceed to Appendix B


Page last reviewed March 2009
Internet Citation: CareScience Risk Assessment Model - Hospital Performance Measurement. March 2009. Agency for Healthcare Research and Quality, Rockville, MD.


The information on this page is archived and provided for reference purposes only.


AHRQ Advancing Excellence in Health Care