Abstract
Introduction: The Veterans Health Administration (VHA) historically has spent a larger percentage of total medical expenditures on primary care than other US payers, but more recent estimates are unknown. Further, no studies have identified whether geographic differences in primary care spending exist within the VHA’s 18 regional Veteran Integrated Service Networks (VISNs). Our objective was to characterize the percent and per-Veteran absolute spending on primary care nationally and by VISN in the VHA in 2022.
Methods: We calculated primary care and total VHA spending in fiscal year 2022 nationally and by VISN. We then calculated the per-Veteran absolute spending and percent of total national expenditures attributed to primary care.
Results: In 2022, VHA spent $848 per-Veteran on primary care (9.2% of total medical expenditures). Across the 18 VISNs, per-Veteran spending ranged from $641 to $1079, with the 25th and 75th percentiles ranging from $802 to $881. Primary care percent spending ranged from 7.1% to 11.2%, with the 25th and 75th percentile ranging from 8.5% to 9.7%. Only 6 of 18 VISNs were in the same quartile for primary care spending in each metric.
Discussion: Primary care spending in the VHA in 2022 as a percentage of total medical expenditures was stable compared with prior estimates. In addition, our results show discordance in each VISN’s spending quartile based on the metric utilized. This finding suggests that each metric may capture a different element of primary care spending.
Introduction
On a percentage basis, the US spends less on primary care than other high-income countries. Primary care spending—most frequently measured as a percent of total medical expenditures—has also been decreasing—from 5.4% nationally in 2012 to 4.6% in 2022.1,2 In contrast, other high income countries spent on average 13% of medical expenditures on primary care in 2021.1,3 In light of this underinvestment, many states have launched primary care investment initiatives that set spending goals using percentage-based spending metrics.4 For example, in 2010, Rhode Island required that commercial health insurers increase primary care spending by 1% of total medical expenditures per year between 2010 and 2014.5 In 2022, the Washington State Legislature set a goal for 12% of all state medical expenditures be spent on primary care.6 One step further, Oregon passed legislation in 2022 that required health insurers and Medicaid coordinated care organizations to spend at least 12% of medical expenditures on primary care by 2023.7
These are important primary care spending goals, as they capture the relative investment in primary care compared with other categories of medical spending. However, primary care percentage metrics may not fully capture primary care spending, as they do not provide information on the absolute resources primary care has to care for patients. As a complement, primary care spending per-person is a less frequently utilized metric of primary care spending, and provides information about the absolute resources provided to primary care.8 Together, these two metrics may create a comprehensive understanding of the relative primary care orientation and absolute resources provided to primary care in health systems.
The Veterans Health Administration (VHA) is one of the nation’s largest integrated health systems, and is divided into 18 regional systems of care delivery called Veteran Integrated Service Networks (VISNs). Each VISN serves as the main funding unit and system for medical care in their respective region.9 This health system and funding structure provides an opportunity to study relative and absolute primary care spending nationally and regionally.
Historically, the VHA has spent a larger percentage of total medical expenditures on primary care than other US payers, spending 9.3% of total medical expenditures in 2014 and 7.8% in 2018.10 Previous high VHA spending on primary care was in the context of a nationwide effort to launch the Patient Aligned Care Team model, a patient-centered medical home initiative beginning in 2010 that included nearly $2 billion in new primary care spending in the first 2 years of implementation.11,12 In the PACT model, care is delivered by a team consisting of a primary care provider (either a physician, nurse practitioner, or physician’s assistant), a registered nurse, a licensed practical nurse, and a clerk. Together, this team shares responsibility for proactively supporting patient health, facilitating smooth transitions across care settings, serving as first point of contact for care throughout the VHA, and coordinating with all care team members in the expanded medical neighborhood.11
Still, more recent estimates on primary care spending at the VHA, either in per-person or percent spending, are not known. Further, no studies have identified whether geographic differences in primary care spending exist within the VHA’s 18 VISNs, or whether insights on geographic spending variation would differ based on whether percent or per-person spending metrics are used. This national and regional information would enable greater understanding of relative and absolute primary care spending in the VHA, and where additional resources might be needed to achieve desired primary care policy goals. Therefore, the objective of this study was to characterize the percent spending and per-Veteran absolute spending on primary care nationally and by VISN in the VHA in 2022.
Methods
We conducted a cross-sectional analysis of medical expenditures for care delivered in VHA facilities. Data sources included the VHA administrative files and the Health Economics Resource Center (HERC) Average Cost files.12,13 VHA administrative data contain comprehensive records describing characteristics of VA enrollees and all encounters nationally. Enrollee characteristics include dates of birth and enrollment, residence addresses, and level of service-connected disability. Health care encounter data include diagnosis, service procedure and place of service codes, dates of service and rendering provider identifiers. Encounter-level costs were based on national HERC estimates that are proportional to Medicare reimbursement rates for similar encounters. These HERC estimated costs are derived from private sector relative value units billed to Medicare, which incorporate any variation in professional fees, overhead costs, and professional liability.14 The national HERC estimates do not incorporate differences in costs to deliver services across different VHA medical facilities. We calculated total VHA spending in fiscal year 2022 by 4 categories: primary care and integrated behavioral health expenditures, nonprimary care outpatient expenditures (eg, specialty care, outpatient general surgery, diagnostic and radiology), inpatient care expenditures (eg, hospital, postacute, or emergency care), and pharmacy expenditures (eg, drug fees and dispensing costs). Outpatient care categories were derived from VHA-specific stop code groups that reflect the type of services provided at a clinic location (eg, primary care clinic, cardiology clinic, emergency department).15,16 Primary care services were defined using a previously published algorithm that considers all outpatient encounters (including face-to-face and telehealth visits as well as procedures) occurring in designated stop codes that are considered to be primary care.12 Total Veterans served per VISN were derived from previously reported 2022 estimates.17 We summed costs nationally for all encounters within each category, and calculated the per-Veteran absolute spending and percent of total national expenditures attributed to primary care. We then repeated this calculation for primary care spending by VISN, and categorized each VISN by quartile according to each primary care spend metric. This evaluation was designated as nonresearch quality improvement by the VHA Office of Primary Care and was not human subjects research and therefore did not require institutional review board approval.
Results
In 2022, total VHA spending was $77.6 billion (increase from $50.9 billion in 2014 and $58.5 billion in 2018) for 8.4 million enrollees (increase from approximately 6 million enrollees in 2014 and 2018), equating to $9264 per-Veteran enrollee.10,17 National VHA primary care spending was $7.1 billion (increase from $4.7 billion in 2014 and $4.6 billion in 2018), which equated to $848 per-Veteran or 9.2% of total medical expenditures.10 Across the 18 VISNs, per-Veteran absolute spending ranged from $641 (VISN 20) to $1079 (VISN 8), with the 25th and 75th percentiles ranging from $802 to $881 (Figure 1a). Primary care percent spending ranged from 7.1% (VISN 2) to 11.2% (VISN 17), with the 25th and 75th percentile ranging from 8.5% to 9.7% (Figure 1b). Only six of 18 VISNs were in the same quartile for primary care spending in each metric. Of the 12 VISNs whose quartile differed across metrics, six were in a higher quartile in the per-Veteran spending metric, and six were in a higher quartile in the percent spending metric (Table 1).
Veteran Integrated Service Network-level spending on primary care, by quartile, 2022. (A) Absolute per veteran spending, 2022 (B) Percent of total medical expenditures, 2022.
Primary Care Spending, by VISN, 2022
Discussion
Primary care spending in the VHA in 2022 as a percentage of total medical expenditures was stable compared with prior 2014 and 2018 estimates.10 Although primary care percent spend was higher in the VHA than in the US overall, it was also lower than health systems outside the US.18 Prior research has demonstrated that patients of the VHA may be on average sicker than the general population, which may in part contribute to the higher relative primary care spend to the US overall.19 However, patients of the VHA have been shown to be similar in their health characteristics to Medicare beneficiaries, which is estimated to have spent 3.4% of total medical expenditures on primary care in 2022.2,19
Our estimates show geographic variation in primary care spend across VISNs in both percent spending and absolute per-Veteran spending, demonstrating that primary care spending varies by region even within a nationally integrated health system. In addition, our results show discordance in each VISN’s primary care spending quartile based on the metric utilized. This finding suggests that each spending metric may capture a different element of primary care spending. Therefore, each metric should be selectively utilized for primary care spend target setting and surveillance depending on its relevance to the target policy goal. For example, the primary care percentage spend metric may be useful for policy goals aimed at emphasizing the balance between primary care spend relative to other service lines (eg, specialty or hospital care) in a particular region or health system. However, this metric has limitations in its ability to inform data-driven primary care spending goals. For example, a region with low primary care and total per-Veteran health spending may still have a high percent spending on primary care, thereby mischaracterizing a region’s primary care infrastructure as comparatively well-resourced. Northwest Network (VISN 20), for instance, has one of the highest percent primary care spend metrics (10.1%, quartile 4), but lowest absolute spending metrics ($641 per Veteran, quartile 1). Therefore, a version of the per-Veteran spending metric (eg, age-adjusted per patient spending) may be more useful as a complement to percent spend metrics for policy goals oriented toward tracking the absolute resources dedicated to primary care, or for tracking the resources needed for primary care to achieve desired population health outcomes.
Future research can identify whether regional spending variation is associated with certain quality (eg, cancer detection rates, emergency department utilization), population health (chronic disease burden), or workforce (clinician satisfaction, retention) related outcomes. Based on such research, policy leaders can direct primary care investment nationally and within individual regions to achieve absolute and relative primary care spending levels most likely to improve outcomes. This work also provides a current frame of reference for primary care spending by which future policy efforts can compare against. Additional future research can focus on explaining the drivers of observed geographic differences in primary care spending after adjusting for age and comorbidity differences in respective populations.
Ultimately, our study demonstrates regional variation in primary care spending across a nationally integrated health system, and differing regional insights depending on the primary care spend metric used. This study is limited by the fact that Veterans may receive primary and specialty care outside the VHA for a variety of reasons, and such spending can be funded by commercial, other public payers (eg, Medicare, Medicaid), or even by the VHA (also known as “community care”). If unmeasured non-VHA care had a different distribution of primary and nonprimary care services (eg, subspecialty outpatient, hospitalizations) than VHA administered services, such a finding could bias our estimate of primary care spending as a proportion of medical expenditures. However, the focus of this study was on spending for services delivered within the VHA’s internal care delivery system and not those delivered via community care. Analyses of primary care inclusive of VHA and community care delivered services can be an area of future work. Second, our analysis did not adjust across regions for VHA patient mix or regional cost of living. However, our results suggest national and targeted regional opportunities to increase spending in primary care. Third, some types of care provided by primary care clinicians (eg, ordering of diagnostic tests, medication prescriptions) contributes to utilization and spending that was categorized as nonprimary care spending in this analysis. Since we were unable to identify which diagnostic tests or medications were prescribed by primary care versus other clinician types, these costs were not attributed to primary care unless they occurred at a primary care stop code location. This approach limits this analysis’s ability to estimate the totality of medical expenditures that may be influenced by primary care. Fourth, the methodology used in this study may differ from approaches used to estimate primary care spending in the US or in other countries, limiting the ability to make direct comparisons between VHA and non-VHA primary care spending. Indeed, between 2010 and 2021, at least 67 separate efforts have been identified to approximate primary care spending by individual states, policy organizations, the VHA, or other groups using variable approaches.20 However, we used the same approach utilized in a prior published study of primary care spending in the VHA,10 making this study an appropriate comparator to prior VHA estimates. Fifth, our methodology is a model-based approach that uses Medicare rates to impute costs onto primary care and other encounters. This methodology may potentially under- or overestimate primary care spending if the operating costs for primary care and other services delivery in the VHA substantively differ from the operating costs for Medicare beneficiaries in civilian care settings.
Notes
This article was externally peer reviewed.
Conflict of interest: None.
Funding: This work was supported by the VHA Primary Care Analytics Team (PCAT), funded by the Office of Primary Care. Data for this evaluation were developed by PCAT. The views expressed are those of the authors and do not necessarily reflect the position or policy of the Department of Veterans Affairs or the U.S. Government.
- Received for publication December 30, 2024.
- Revision received March 11, 2025.
- Accepted for publication March 24, 2025.
Jump to section
Related Articles
Cited By...
- No citing articles found.