HERC Inpatient Average Cost Data

HERC estimates encounter-level costs of VA inpatient care. These estimates are known as the HERC Average Cost data.

To create HERC Average Cost data, we distribute aggregate-level VA costs to encounter-level VA utilization using non-VA relative value weights. We call this the "average cost" method because it assumes that all encounters with the same observed characteristics have the same average cost.

Data Overview

The HERC inpatient estimates represent the national average cost of a hospital stay given its Diagnosis Related Group (DRG), overall length of stay, and days in intensive care. The inpatient estimates are based on analysis of cost-adjusted charges in Medicare funded stays of veterans in non-VA hospitals. Estimates are adjusted so that the estimates tally to actual national VA expenditures for that type of care.

• HERC inpatient average cost data include estimates for 11 types of inpatient care.
• Data include cost estimates for each facility (a local cost) and estimates at the national level (a national cost). We recommend that most researchers use the national cost.

return to top

Data Availability

HERC Inpatient Average Cost data are available for FY98-FY24. The dataset is updated annually, with updated expected to occur approximately July 1st each year.

Data are available on VINCI to VA research and operations data users. Researchers can request access to HERC data through DART (Data Access Request Tracker). Visit the VHA Data Portal to access DART and learn more about the data request process for both research and operations (http://vaww.vhadataportal.med.va.gov/Data-Sources/HERC-Cost-Data; VA intranet only).

Data location: Data are available on RB03 at VINCI_HERC and on the SAS Grid at /data/prod/HERC.

return to top

Resources for Using the Data

Interested in proc contents for earlier years? Contact herc@va.gov.

Guidebook: HERC's Average Cost Datasets for VA Inpatient Care

This guidebook provides a detailed description of the methods used to prepare the HERC Inpatient Average Cost Data estimates and a user's guide for working with the data.

return to top

Limitations of HERC Average Cost Data

A limitation with this method is precision. Because all encounters that share the same characteristics receive the same average cost estimate, the HERC Average Cost data may not be ideal if a researcher needs very precise data, such as when comparing close substitutes that may share many of the same characteristics.

More information about choosing a cost data source is available in the Data FAQ: "Which Data Source Should I Use: HERC or MCA?

return to top 

---

Methods

Inpatient Non-Medical/Surgical Care

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 assumes that every day of stay has the same cost, that is, that costs are proportionate to the length of stay.

Long-Term Care

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; see Technical Report #11 - Intranet only - https://vaww.herc.research.va.gov/files/RPRT_185.pdf). For FY01-present, we calculated costs based on an average daily rate.

Inpatient Medical/Surgical Care

To find the cost of medical/surgical care, we use relative value units (RVUs) from the non-VA sector (Medicare). Our RVUs are based on a regression model that uses DRG weight, ICU days, length of stay and patient characteristics to estimate cost-adjusted charges.

Bedsection/Treating Specialty Codes to HERC Category of Care Assignments

For inpatient care, the VA does not distinguish between medical/surgical care and non medical/surgical stays. The VA keeps track of bedsections. Bedsection is a VA-specific term analogous to a hospital ward; MCA uses the term treating specialty, but the values are the same as the bedsection. Because a patient can get transferred among bedsections multiple times within a single medical/surgical hospital stay, keeping track of bedsections provides us with a great amount of detail.

Notes

Number 4 is surgery, but this becomes combined with category 0. The skip in number from 3 to 5 maintains consistency with average daily rates calculated before FY98.

*Stations with an approved PRRTP program: 459 463 501 504 515 516 518 523 528 541 546 549 554 555 556 561 568 573 586 589 590 595 598 620 622 631 632 635 637 640 645 653 656 658 662 663 666 676 678 687 689

return to top

FAQs

What are the average daily costs of VA inpatient care?

The tables below present the average daily cost of stays in VA care by fiscal year. The first table displays the local average cost; the second table displays the national average cost. Both tables were produced using the HERC Average Cost Dataset. See the notes for more information about the local and national cost estimates. The data are not adjusted for inflation.

Notes

This table includes overhead and physician costs.

National cost estimates reflect VA national expenditures in each category of care, while local cost estimates reflect VA expenditures at a particular medical center. The sum of the local cost estimates for visits in each category of care will approximately equal the total national expenditures for each category. Visit HERC's Average Cost Datasets for VA Inpatient Care  or HERC's Outpatient Average Cost Dataset for VA Care guidebooks to learn more about the difference between local and national cost estimates. Because the 2019 estimates start to diverge from prior trends, we’ve included both estimates beginning 2017 here. For earlier years of national average cost estimates, please contact herc@va.gov.

*In 2023 we excluded sta3n 673. 

**In 2018, VA discontinued a range of inpatient treating specialty codes (also known as bedsections) for substance use disorder (SUD) care. VA left intact one treating specialty code for intensive alcohol treatment and intensive drug treatment. Most facilities that provided inpatient SUD care diverted patients to the domiciliary—a unit where patients can live while they get outpatient services. Because domiciliary care is less expensive than traditional inpatient care, this allowed some VA facilities to expand access.

For example, the Baltimore VA medical center provided 11,662 days of inpatient substance use treatment and 6710 days of domiciliary care in 2018. In 2019, it provided 2 days of intensive inpatient SUD care and 31,458 days of domiciliary care.

The 2018 policy change affects the HERC average cost inpatient data. There was a large increase in average daily costs for patients in intensive SUD treatment: $952 in 2018 to $2,650 in 2019. None of these changes are unexpected, but they could affect studies that are tracking substance use care and costs.

return to top

How do I obtain the DRG value weights from Medicare?

The Centers for Medicare and Medicaid Services (CMS) maintains a list of relative value units (RVUs) for inpatient hospital care. These RVUs are also known as the Diagnosis-Related Group (DRG) weight. DRGs are used to determine how much Medicare reimburses a hospital for providing care, as patients within each DRG category are expected to use a similar amount of hospital resources. A DRG is assigned by a hospital based on the diagnoses and procedures noted by physicians in a patient’s medical record.

Under the conventional DRG system, hospitals were reimbursed the same amount for treating a patient within that DRG, whether that patient was extremely sick or relatively uncomplicated. Recognizing that it was more expensive for hospitals to treat sicker patients, in 2008, CMS changed from DRGs to MS-DRGs, which stand for Medicare-Severity Diagnosis Related Groups. The MS-DRG reflects varying resource intensity to treat a condition through its classification of inpatient admissions into three mutually-exclusive categories: a base MS-DRG; a MS-DRG with care complications or comorbidities (CC); or a MS-DRG with major care complications or major comorbidities (MCC). Note that while these MS-DRGs all reference the same underlying condition, each of the three MS-DRGs will have its own MS-DRG number. A three-digit code is used to represent DRGs and MS-DRGs, but there is no way to match DRGs in the old system to the MS-DRGs, as a given DRG may have been split into two or more MS-DRGs. There were approximately 475 DRGs under the old system, and there are currently approximately 750 MS-DRGs.

The latest DRG relative value units, along with information on the average length of stay for that MS-DRG, can be found on the CMS website, in Table 5 entitled, “Proposed MS-DRGs, Relative Weighting Factors and Geometric and Arthritic Mean Length of Stay.” MS-DRG's weights can be merged to a utilization file using the MS-DRG number. Therefore, in order to be able to use the MS-DRG weights, your dataset must include MS-DRG code numbers. If you are using these data to determine length of stay for your population of interest, you must choose between the average length of stay and geometric mean length of stay. The average length of stay (ALOS) is the simple arithmetic mean, and is calculated by adding up the length of stay for all patients with that MS-DRG and dividing by the number of patients. The geometric length of stay (GLOS) is calculated by multiplying all of the lengths of stay for all patients with that MS-DRG and taking the nth root of that number, where n is the number of patients. The GLOS has the effect of reducing the influence of outlier values, or patients with very high or very low length of stay.

An example of the type of data available from CMS can be seen in the DRG weight files for 1983-2007 and the MS-DRG weight files for 2008-2013. However, please refer to the files on the CMS website for the most up-to-date information. It is important to make sure that the DRG or MS-DRG weight file is the same year as the utilization data you are using. HERC also has historical DRG files since 1983; you can email herc@va.gov to obtain these.

Last updated: May 1, 2025