Average Cost Method
[ Learn more about "What's new" with HERC's Average Cost Data ]
[ Learn more about registering to use HERC's Average Cost Data ]
[ Learn more about HERC's Average Cost Data
We are using VA cost and utilization data to estimate the cost of VA patient care encounters. Cost data consist of department-level costs. Utilization data are from the VA Patient Treatment File (PTF) and the National Patient Care Database Outpatient Procedures File. Information from non-VA sources is used to estimate the relative cost of each VA healthcare encounter; the relative cost can be adjusted with data from the Cost Distribution Report (FY98-03) and the Decision Support System (FY04-present) to estimate the actual cost of each encounter.
We call this the "average cost" method; it assumes that every healthcare encounter has a cost that is the average cost of all encounters that share the same characteristics. (Although the U.S. Panel of Cost Effectiveness and Medicine calls this a "gross cost method", we feel our name is more descriptive).
Since the accuracy of the estimate depends on the level of detail, we are using the greatest level of detail found in centralized VA datasets for which we can determine a measure of relative value. The method also assumes that relative value weights from the non-VA sector accurately represent VA costs. While these assumptions limit the accuracy of our cost estimates, they permit us to create a comprehensive set of encounter-level estimates of all VA patient care cost. This method is useful to find the cost of care that is not directly affected by the intervention under study (see Swindle articles in Medical Care Supplement).
HERC is using this method to estimate the cost of every VA inpatient stay and outpatient visit for FY98 and subsequent years. We are creating two different cost estimates for each encounter: a local cost based on the costs reported for that facility, and a national average cost, based on costs reported by all facilities. These data have been made available via the Austin Automation Center and are documented by a handbook that describes their limitations and appropriate use. Link to more information on HERC Average Cost databases available at the Austin Automation Center.
To find the cost of outpatient visits, we use the relative values of all Current Procedures and Terminology (CPT) codes assigned to the visit. We use the relative values from the Resource Based Relative Value System (RBRVS), which is used to reimburse providers for services provided to Medicare patients. We assign every VA visit to one of 12 different categories of outpatient care. For each category, we find a specific factor to convert the relative value to a VA cost estimate. We assume that the resources used to provide VA outpatient care are proportionate to the relative values assigned in the Medicare reimbursement. For FY98-03 costs, see the Excel spreadsheet in the downloads section to view the VA clinic stops (now called DSS identifiers) associated with each Cost Distribution Account.
A centralized database on outpatient prescriptions filled by VA is not available. For this reason, the HERC estimates cannot reflect the cost of prescriptions provided to individual outpatients. The DSS national extracts include estimates of the cost of all prescriptions dispensed to an individual on a given day.
To find the cost of inpatient stays in rehabilitation, domiciliary, psychiatric, substance abuse, and intermediate medicine treatment units, we find the average cost of a day of stay, and apply it to estimate the cost of care. This makes the assumption that every day of stay has the same cost, that is, that costs are proportionate to the length of stay.
Long-term care: For FY98-00, we found the cost of long-term stays by incorporating the relative values for resource utilization from Resource Utilization Groups (RUGs). VA undertakes a biannual assessment in order to assign every long-term care patients to a RUG. We used these data to assess cost, making the assumption that the cost of long-term care is proportionate to the RUG relative value. For FY01-present, we calculated costs based on an average daily rate. VA changed from using RUG II to RUG III, but these data are not yet ready to include in the average cost method (see Technical Report #11).
To find the cost of acute hospital care, we use relative value units (RVUs) from the non-VA sector (Medicare). Our RVUs are based on the relationship between cost-adjusted charges, diagnosis, and length of stay. The RVUs reflect the effect of diagnosis on the relative quantity of resources used in a hospital stay. We used these RVUs so that our cost estimates would reflect the effect of diagnosis on resource use. The method we employed makes the following assumptions: (1) the relative value units (e.g., DRG weights and length of stay), based on non-VA costs, reflect the relative costs of VA hospital stays and (2) all stays with the same characteristics have the same cost.
