Programming note: This piece is co-published with , a newsletter run by the Center for Land Economics. The Center for Land Economics creates tools and research to promote equitable assessments and sustainable development.

A federal program designed to subsidize safety-net healthcare is quietly shifting tax burden onto the very communities it is meant to serve. The 340B Drug Pricing Program has become a major engine for hospital consolidation, incentivizing nonprofit systems to acquire private practices and hospitals to capture their drug revenue. Because these acquired properties typically fall off the tax rolls upon conversion to nonprofit status, the resulting revenue shortfall is redistributed as a silent surcharge on local taxpayers.

For those unfamiliar, the 340B Program lets qualified nonprofit hospitals purchase drugs from manufacturers at a steep discount; the hospital then charges patients (or their insurance) full price, pocketing the difference. This buy-low-sell-high setup is incredibly lucrative for hospitals and has caused a gold rush to join the program by any means necessary. As a result, 340B has expanded well beyond its original mandate and is now creating unintended consequences throughout the wider economy.

One of those consequences is increased property taxes. The basic causal chain goes like this:

• Step 1: The 340B Program incentivizes nonprofit health systems to acquire or roll up for-profit hospitals, doctors’ offices, and oncology practices in order to access their patient bases.

• Step 2: Once acquired, the physical property of the acquired for-profit entity becomes tax-exempt, significantly reducing or eliminating its property bill.

• Step 3: In many jurisdictions, other people’s property taxes automatically increase to fill the gap left by the formerly-tax-paying property.

This isn’t happening on a massive scale, but it’s a big enough deal to significantly impact the communities where it does happen. Because property taxes are the largest source of revenue for local governments - and hospitals are often among the largest and most valuable commercial properties - the fiscal impact of exempting them can be huge. Further, it’s a great example of a surprising second-order effect – one that almost certainly wasn’t intended when 340B was created.

So, let’s walk through each link in the causal chain above, using a case study to demonstrate the on-the-ground impact.

1. 340B promotes consolidation

A long line of studies has shown that the 340B Program incentivizes market consolidation among providers. Most of those studies focus on vertical integration, which is when 340B hospitals acquire private practices, infusion centers, etc. to make into child sites. The hospitals then inherit the patients of the acquired child sites, giving them more opportunities to purchase (and administer) 340B-discounted drugs and thus increase their drug revenue.

However, the 340B Program also seems to incentivize horizontal integration among hospitals, which is when a 340B hospital purchases a non-340B hospital with the intention of enrolling it in the program. For example, for-profit hospitals cannot enroll in 340B. However, some for-profit hospitals still meet the underlying requirements for the program (usually via DSH percentage). So, if a nonprofit hospital acquires an otherwise-qualified for-profit hospital and converts it to a nonprofit, the acquirer gets not only a 340B-qualified hospital, but also all its potential child sites and patients.

This buy-then-convert pattern turned out to be surprisingly common. Using searches of Becker’s, CHOW, and hospital finance websites, I found 25 cases of for-profit hospitals getting acquired by nonprofits in the last decade. In over half those cases, the newly acquired hospital almost immediately (within a year) enrolled in the 340B Program. The remainder did not enroll because they were not 340B-eligible even after conversion (e.g., they did not meet eligibility criteria such as the DSH threshold). Here’s the full timeline of the acquisitions I found:

It’s impossible to know if 340B caused these acquisitions, but I think it’s fair to say that it increased their financial attractiveness and probably made some of them pencil out. To check this, I pulled each hospital’s financial performance as reported in their CMS Cost Reports.

The plot below shows each hospital’s net patient revenue (NPR) before and after acquisition. Hospitals enrolled in 340B generally seemed to have a sharp takeoff in NPR and total drug charges (not shown, but they track together), hinting that at least some portion of their post-acquisition growth was fueled by 340B.

Some caveats here: the dip around the acquisition date of each provider is mostly (though not entirely) an artifact of hospitals switching/truncating their fiscal years when reporting. Further, while I tried to statistically quantify the difference between 340B-enrolled and non-enrolled hospitals, the groups here are qualitatively different enough that it’s not straightforward. Most of the non-340B, acquired hospitals (Marina Del Rey, Columbia Medical Center) are associated with a “premier” local health system or are dedicated cancer centers (City of Hope).

2. Acquired for-profit property falls off the tax rolls (usually)

The downstream consequence of all this integration and consolidation is a property tax shift. As private practices, infusion centers, and hospitals get acquired by nonprofit systems, they also get removed from local property tax rolls.

How nonprofit property is treated in practice is messy, piecemeal, and inconsistent. Different states and localities have different laws around which properties are deserving of property tax exemptions. Most require some kind of exclusive use – the property must be used solely for charitable, religious, or scientific use. However, what counts as “exclusive” differs widely, and many states carve out special rules for hospitals.

This is complicated further by the complex nature of property ownership and the various commercial agreements many hospitals maintain. For example, a hospital may lease part of its property to a commercial vendor (e.g. a cafe) or use it for a non-qualified use (e.g. fundraising), blurring the line between exempt and non-exempt use. Further, exemptions for a big property like a hospital typically involve a lengthy application process. That process can sometimes be contested by outside parties or result in a payment in lieu of taxes (PILOT).

All this complexity makes it hard to measure the actual property tax impact of programs like 340B. I initially tried to determine the property taxes and exemption value of each of the hospitals in the timeline above. I wanted to put a total dollar figure on the property tax cost of the 340B Program. This proved essentially impossible. Most states have very little historical property tax data available online, and fewer still have data about tax-exempt properties.

However, here’s what I can say given the available data – all these hospitals are property tax exempt as of 2026. That’s according to records from local assessors’ websites, property tax payment portals, and scouring financial records. There’s some nuance here - many properties are partially exempt or have a PILOT, others have closed or have questionable status - but all of them eventually became exempt after acquisition.

And here’s what I can’t say with the data – I don’t know when each hospital first became exempt, how much their exemption is worth, how much they’re saving in property taxes, or how much their exemption is costing others. All that information is locked away in the historical records of ten different counties.

That said, at least some counties do maintain good enough open records to check all those things and quantify 340B’s impact. So, since a national analysis isn’t possible, I thought it would be instructive to look at a case study.

MacNeal Hospital was owned by Tenet Healthcare, one of the largest for-profit health systems in the U.S. In 2018, Loyola Medicine acquired MacNeal and all its affiliated operations for $270M. Shortly thereafter, MacNeal was granted a nonprofit property exemption by the Illinois Department of Revenue.

Not all of Loyola’s parcels qualified for the exemption; many were owned by the hospital but dedicated to commercial use (parking lots, offices, etc.). As such, MacNeal’s property tax burden fell, but not all the way to $0. The taxable value (EAV) of its properties fell from $23M in 2019 to $8M in 2020, resulting in a $2.5M decrease in its annual property tax bill. Here’s what that looks like graphically:

Note that I’m not saying that giving MacNeal an exemption is bad, per se. Nonprofit hospitals get these tax breaks with the expectation that they’ll provide community benefits, and most do. In MacNeal’s case, they now get a tax break of around $2.5M per year, but they also drastically increased their charity care post-acquisition - from $2.6M in 2017 to $8.4M in 2024 - according to data from the Illinois Health Facilities & Services Review Board.

However, in Illinois and most other states, property taxes are zero-sum and therefore have inherent tradeoffs. Decreasing one person’s property taxes means increasing the property taxes of others. Removing a large hospital like MacNeal from the tax rolls means shifting its substantial tax burden to nearby properties. Let’s see how that actually plays out in practice.

3. Property taxes rise to make up the shortfall

There are essentially three common scenarios when a large property is removed from the tax rolls:

1. Property taxes rise on surrounding properties to make up for the shortfall (as described above). This is typical for most jurisdictions in the United States.

2. Property taxes stay the same on surrounding properties and local governments have a revenue shortfall. This basically only happens in California and Oregon.

3. Local governments arrange a Payment in Lieu of Taxes (PILOT) to prevent 1 or 2 and set a fixed payment for the property via legislation or a regulatory body.

MacNeal followed the first scenario. In Cook County, Illinois, where MacNeal is located, all local governments set a yearly revenue target called a levy. Property assessments (the base) are then used to set a tax rate which determines each parcel’s share of the overall levy burden. If one property’s assessment falls dramatically, then the tax rate increases to compensate, which increases everyone else’s bills. This system ensures stable revenue for local governments, but it also means that big tax breaks or assessment changes on neighboring properties can affect your tax bill. Here’s what that looks like as an equation, where MacNeal’s taxable value fell by $15M:

Figuring out how one property’s tax break affects its neighbors in a levy-based system can be tricky. Fortunately, Cook County has a public property tax simulator meant for exactly this purpose (disclaimer: I made this). To quantify the impact of the MacNeal acquisition, I used their tool to calculate counterfactual property tax bills for all surrounding properties, then compared them to actual tax bills. The difference between the two bills is essentially the cost that each property is paying due to MacNeal’s tax exemption.

Around 15,000 properties in Berwyn, where MacNeal is located, all saw increased property taxes due to the MacNeal exemption. The average increase was around $100 annually, but a long tail of properties saw much higher increases as well. Further, the effects weren’t limited to just Berwyn. Technically, all property tax bills in Cook County rose due to MacNeal’s exemption, though spread across the much larger county tax base (1.8M parcels) the increase was just a couple cents per property.

It’s not obvious that a hospital acquisition would increase your property tax bill. There’s no line item on your tax bill that says “MacNeal exemption surcharge.” The increase just gets quietly baked into your tax rate, indistinguishable from any other year-over-year change. Most people in Berwyn probably have no idea that Loyola’s acquisition of MacNeal is partly responsible for their higher bills. The practical point here isn’t that any one household is being bankrupted by these changes; it’s that a policy lever meant to subsidize safety-net care quietly created a mechanism for redistributing local tax burden in ways most taxpayers (or policymakers) would never suspect.

Wrapping up

MacNeal isn’t a one-off. The same basic causal chain applies to most of the nonprofit acquisitions in the timeline above: when a for-profit hospital (or a network of affiliated sites) gets converted into a nonprofit, some portion of that real estate is very likely to fall off the tax rolls.

Even if the per-property effect is small, these shifts stack: multiple acquisitions over time, plus the ongoing roll-up of private practices and infusion centers into tax-exempt systems, gradually erode the commercial tax base in the exact places where hospitals are the largest and most valuable properties.

While the actual tax impact is important, the deeper lesson here is about how complex systems route incentives in unexpected directions. A federal drug discount intended to support safety-net care became a catalyst for consolidation; that consolidation converted taxable property to exempt property; and local property tax systems turned that conversion into a distributed surcharge on everyone else’s bill. None of those steps are crazy in isolation, but chained together they produce an outcome that’s both unintuitive and hard for the public to even notice.

Nonprofit exemptions and 340B are both, functionally, public subsidies. They’re justified on the theory that the hospital will return value via charity care and broader community benefits. But those benefits (and the gains from 340B) are often diffuse, hard to value, and accrue over time. Meanwhile, the subsidy costs are hidden and nearly impossible to discern even for experts.

To preserve the intent of 340B and nonprofit exemptions without turning them into a hidden surcharge, the minimum starting point is transparency. Jurisdictions should be able to put a dollar figure on each exemption, show who absorbs the shifted levy, and compare that cost to measurable community benefits so homeowners aren’t unknowingly underwriting federal policy.

Rural hospitals are in trouble. Over the past decade, over 100 rural hospitals have closed, and those remaining face new and stronger headwinds. Massive cuts to Medicaid from the OBBB, more scrutiny on federal rural reimbursement programs, lower reimbursement and patient volume from Medicare Advantage, reduced 340B savings, and increased labor costs all point to an increasingly uncertain future for rural hospitals.

In light of such obstacles, policymakers and providers alike are interested in boosting the income of rural hospitals. Congress, for its part, has focused its effort on strengthening public payers and programs (e.g. Medicare, Medicaid, 340B, CAH, REH, etc.). However, that singular focus has left a huge gap in knowledge about the private, commercial side of the ledger.

How much are rural hospitals actually paid by insurers? How do those payments compare to urban hospitals? What factors drive differences in urban-rural reimbursement? Recent studies are scant and limited, leaving policymakers to try to answer these questions themselves. This study aims to fill such knowledge gaps and provide a new data point for discussing rural hospital finances.

Using an unprecedentedly large (93M+ rates) sample of price transparency data from over 5,000 U.S. hospitals, I show that rural hospitals are reimbursed significantly less than urban hospitals by commercial insurers. Further, I find that the price gaps between inpatient/outpatient care and independent/health system hospitals widen in rural areas. These trends are robust to different definitions of rurality, baskets of procedures, and weighting methods.

The size of this urban-rural reimbursement gap - and how it changes over time - has huge potential consequences. Small fluctuations in income can push rural hospitals’ operating margins into the red, possibly forcing them to merge with larger, more urban health systems, eliminate service lines such as obstetrics, or close altogether. The result is higher healthcare costs, expanding care deserts, worse health outcomes, and local economic disruption.

Rural hospitals receive less

Rural hospitals are reimbursed roughly 20 percentage points (pp) less than urban ones, relative to baseline Medicare rates. In other words, the median commercial rate at a rural hospital is 232% of Medicare, while the median commercial rate at an urban hospital is 253% of Medicare. Rural hospitals also have slightly more variation in their prices than urban hospitals, with IQRs of 77 pp and 73 pp, respectively. Here’s the difference visualized nationally:

It’s worth emphasizing the massive scale of the data included in this simple plot. It includes rates from 5,073 hospitals and 51 major health insurance companies. It includes all MS-DRGs, all APR-DRGs, and the top 5,000 Medicare-payable HCPCS (excluding drug, device, and some specialty codes).

In all, it includes 93,319,856 cleaned, validated, and normalized rates spanning dozens of contracting methods and edge cases. The data here is many orders of magnitude larger and more complex than the price transparency data used in previous urban-rural studies. And it is more up-to-date (as of October 2025) than studies that rely on claims data.

The scale of the data also lets us break the high-level urban-rural aggregation into finer geographies and classifications. Where previous studies mostly used states as the geographic unit of study, this new data is dense enough to use counties or even Census tracts (previous studies achieved this but were sparse/incomplete). Here’s a county-level map of commercial reimbursement relative to Medicare across the entire United States:

Each county shows the weighted median of rates from providers inside its borders. Some counties don’t have any providers, and in those cases, a 1-hour driving time catchment is used to aggregate providers from surrounding counties (exclusively for this map, not the rest of the analysis).

The patterns here match fairly well with previous state-level studies examining commercial-to-Medicare prices, but the county-level map lets us see details that are masked when aggregating to the state level. For example, the highest reimbursement in Georgia is concentrated around Atlanta and its surrounding counties, while the eastern part of the state has much lower reimbursement rates. Washington and Oregon likewise have their highest rates concentrated in the western, more urban parts of each state.

However, the real advantage of the finer, county-level granularity is that we can more precisely classify areas as urban or rural. This will let us expand the simple urban-rural dichotomy into a spectrum. Here’s the map again, now colored by each county’s rurality using the National Center for Health Statistics (NCHS) classification system:

The NCHS uses a six-level scheme to classify counties by rurality. Orange counties are rural (or nonmetropolitan), while green counties are urban (metropolitan). Each color is segmented out further by population density and metro type/size.

This classification scheme isn’t perfect. It doesn’t account for within-county variation in rurality, which matters a lot for larger counties. For example, Coconino County in Arizona is classified as a “Small metro” county because it contains Flagstaff, but the vast majority of its land area is desert, reservations, and parks.

Still, combining this classification system with county-level aggregate rates lets us move beyond the simple urban-rural divide and see more interesting trends. Here’s the urban-rural boxplot from before, now broken out by the six levels from the map:

The six-level scheme reveals some previously hidden trends. First, there’s huge variation within the rural classification. Hospitals in the most rural U.S. counties (Noncore) are reimbursed roughly 33 percentage points less than urban hospitals, and 23 percentage points less than the Micropolitan counties that are also considered rural. Second, there’s not much variation within the urban classification – medians are around 250% of Medicare across the board.

The overall takeaway here is that the most rural places receive the lowest reimbursement, while prices everywhere else are fairly similar. We call this the urban-rural reimbursement gap.

Subscribe now

Unpacking the urban-rural reimbursement gap

There are dozens, possibly hundreds, of reasons why rural hospitals might be paid less than urban hospitals. A full causal analysis or examination of all possible confounding variables is beyond the scope of this piece. However, we can at least unpack the gap further by looking at some common drivers of reimbursement variation.

The cleanest cut is by site of care. There’s a huge gap between inpatient and outpatient reimbursement in general, but the gap gets much wider for rural hospitals. Across the four urban NCHS classifications, the gap between inpatient and outpatient rate medians hovers between 55 to 65 percentage points. However, that gap jumps to roughly 95 percentage points for rural hospitals. This jump is driven largely by inpatient reimbursement rates dropping rapidly as hospitals become more rural, while outpatient rates remain (relatively) stable.

Low rural inpatient reimbursement could help explain why rural hospitals are increasingly choosing to cut inpatient service lines like obstetrics or convert to outpatient-only Rural Emergency Hospitals (REHs).

Health system affiliation is another interesting cut. Independent hospitals tend to be reimbursed at lower rates than hospitals affiliated with a health system, and rural hospitals are more likely to be independent. In practice, that means that aggregated rural reimbursement is dragged down by the high proportion of independent hospitals in rural areas, even though the gap between independent and health system-affiliated hospitals is larger in urban areas.

The lower reimbursement of independent rural hospitals mostly stems from their reduced market power. Such hospitals just don’t seem to have much negotiating leverage with commercial payers, even if they’re the only hospital in a given county or area. Clinically integrated networks are one way independent rural hospitals are starting to push back and regain negotiating power.

Reimbursement also varies by hospital ownership structure. Government-owned hospitals tend to be reimbursed less by private insurers, and rural hospitals are more likely to be government-owned and less likely to be for-profit. Here’s hospital ownership by rurality; note that the most rural (i.e. noncore) hospitals are excluded since they’re almost entirely missing ownership data:

Government-owned rural hospitals have the lowest median reimbursement of any group, while private rural hospitals have the highest. Curiously, the most urban government-owned hospitals also have very low reimbursement.

We can also break out urban and rural hospitals by service line. Here the familiar trend repeats; urban reimbursement is generally higher than rural reimbursement. However, there are some interesting exceptions. Lab work and radiology both receive higher reimbursement at rural hospitals. I’m not entirely certain what to make of this. It could be the result of rural hospitals attempting to cross-subsidize costly care using high volume, well-reimbursed care. Or it could simply be denominator-driven, i.e. Medicare reimbursing those service lines at slightly lower relative rates in rural areas (thus making their commercial-to-Medicare ratios higher).

There are a lot of other ways to slice this data, but most of them don’t make for very interesting plots. Here are some of the other variables I looked at and how they compared across urban and rural hospitals:

• Payer mix - Rural hospitals treat more patients on Medicare than urban hospitals. Since Medicare reimbursement is (usually) lower than commercial reimbursement, rural hospitals may cross-subsidize care by hiking commercial rates. I looked for evidence of this trend, or any relationship between payer mix, rurality, and commercial prices, but didn’t find anything compelling.

• Patient volume - Rural hospitals typically have lower patient volume due to the low population density of rural areas. I hypothesized that low patient volume might require higher commercial rates to offset the high fixed costs associated with each patient, but if anything, the opposite seems to be true. Larger, higher-volume, non-government rural hospitals had significantly higher reimbursement than small, low-volume rural hospitals.

• Procedure volume - This is a tricky one. My theory here is that rural hospitals may negotiate higher rates for their highest volume procedures and take low rates on everything else. This would muddle the aggregations in the plots above, since those use state-level rather than provider-level utilization weighting, and individual providers might have very different utilization profiles. Unfortunately, utilization data at the provider-level comes from claims, which are incomplete, lagged, and sparse for rural hospitals.

• Quality - Most studies find little correlation between hospital quality and prices. To confirm this, I compared reimbursement rates across CMS star ratings and levels of rurality. I similarly found no relationship between quality, rurality, and commercial reimbursement.

• Market concentration (provider) - Quantifying market concentration for rural hospitals is complex. The Herfindahl-Hirschman Index (HHI), the most commonly-used measure of hospital market concentration, is typically calculated using CBSAs or MSAs as proxies for competitive markets. However, rural hospitals are typically outside such geographies and thus need to be grouped into custom markets or peer groups.
  
  Further, any analysis of market concentration in rural areas has to disentangle two competing price effects. Small, independent rural hospitals start with almost no negotiating power (no health system, low patient volume, etc.), so payers will give them low reimbursement. However, such hospitals may have exclusive control over a region, increasing their negotiating power with payers who want to build a complete network.

  Quantifying rural market concentration and price effects would be a massive undertaking and is beyond the scope of this study. Health system-affiliation (used in the plot above) is the closest we’ll get for now.

• Market concentration (payer) - States or regions that are dominated by a single insurer tend to have lower reimbursement rates. I used state-level payer market share to construct an HHI, then compared rates across high- and low-market-concentration states. I found that while high-market-concentration states do have lower rates, there’s no clear relationship between rurality, payer market share, and prices. In other words, prices move in tandem across urban and rural hospitals as payer market share changes.

• Program status (CAH, REH, RRC, LVH, etc.) - Federal and state governments run dozens of special payment designations and programs to support rural hospitals. Many of these change the way rural hospitals are reimbursed (e.g. CAH and REH). I didn’t find any relationship between special payment designations and commercial rates. See the Methods section for further complications related to Medicare rates.

There are probably dozens of other variables that drive the urban-rural reimbursement gap. If you have suggestions for others I should look at, feel free to drop a comment or email me.

Closing the gap

Rural hospitals are systematically reimbursed less than their urban counterparts by commercial insurers, with the gap reaching 20 percentage points nationally and widening to 33 percentage points for the most rural counties. While there’s significant variation in urban/rural prices and some rural hospitals certainly receive higher rates than equivalent urban ones, the overall national trend is unambiguous – rural hospitals simply receive less on average.

The granularity of this new dataset, spanning over 5,000 hospitals and 93 million rates, also exposes other previously invisible patterns: rural inpatient reimbursement drops sharply compared to outpatient reimbursement, independent rural hospitals lack the negotiating leverage of their health system-affiliated peers, and government-owned rural hospitals are reimbursed much lower than private and nonprofit facilities.

For policymakers focused primarily on Medicare and Medicaid, these findings should serve as a wake-up call about the commercial side of the ledger. Small shifts in commercial reimbursement can mean the difference between operating in the black or red for rural hospitals already operating on thin margins. To stabilize commercial reimbursement, states should consider using money from the recently announced Rural Health Transformation Program (RHTP) to help rural hospitals improve the technical parts of their revenue cycle (e.g. increasing automation, getting better data for commercial negotiations, and integrating with revenue recovery software). States could also use RHTP funds to help rural hospitals form Clinically Integrated Networks, groups that allow them to pool resources and collectively negotiate rates.

Without addressing the commercial reimbursement gap, we risk accelerating the closure of rural hospitals, the elimination of essential service lines like obstetrics, and the continued consolidation of rural providers into larger health systems – all of which will ultimately lead to worse health and economic outcomes for rural areas.

Subscribe now

Methods

This sort of price analysis can be sensitive to the assumptions and methods used during cleaning, classification, and aggregation. To ensure that the results are robust to such choices, I re-ran the analysis with a few major changes. In other words, I did a sensitivity analysis of the most important variables and assumptions. The checks here are not exhaustive, see the GitHub repository for the public code and a full methods breakdown.

Weighting and sampling

This analysis uses weighting to aggregate individual negotiated rates at the payer-provider-code level all the way up to the county level. Weights are applied to three different aggregations:

1. Starting from the payer-provider-code level, rates are aggregated to the payer-provider level by taking the mean, weighted by each procedure code’s state-level utilization (sourced from claims volume).

2. From the payer-provider level, rates are aggregated to the provider level by again taking the mean, this time weighted by each payer’s state-level market share (sourced from Policy Reporter).

3. Finally, provider-level means are aggregated to the county-level by taking the median, weighted by each provider’s total bed count (as a proxy for hospital size/throughput).

The goal of all this weighting is to downweight codes, providers, and payers that aren’t actually used very much and get the final aggregated commercial-to-Medicare ratio for each county closer to what insurers actually pay. However, this amount of weighting adds a lot of complexity and makes the results slightly suspect. As such, I re-ran the aggregation without any weights by substituting the unweighted mean or median in the steps above.

The results were somewhat surprising. The gap between urban and rural hospital reimbursement grows even larger using unweighted rates. Further, the entire distribution of commercial-to-Medicare ratios shifts downward i.e. things get cheaper relative to Medicare. You can see these effects in the middle boxes below, while the left-most boxes show the original, weighted distributions:

I suspect that this result is due to the composition of rates. Outpatient codes (HCPCSs, technically CPTs) tend to have higher utilization and higher reimbursement than inpatient codes (MS-DRGs). As such, weighting by code utilization will move the mean toward higher-priced HCPCS, raising the overall commercial-to-Medicare ratio.

To test this effect, and to make sure that my results weren’t otherwise dependent on the mix of codes, I also ran the aggregations using a sub-sample of common procedures. The results for the sub-sample are shown in the right-most boxes above. They are slightly higher than the unweighted results, but still below the weighted ones. The gap between urban and rural providers is roughly the same. This aligns with my expectations given the sub-sample’s mix of codes, as it contains a higher proportion of inpatient rates than the main sample (and would therefore be slightly lower).

Overall, the results do not seem sensitive to the weighting or composition of the sample – a significant urban-rural reimbursement gap persists basically regardless of how one aggregates or (reasonably) subsets the data.

Rural classification

The analysis relies on a pre-made, county-level rurality classification system created by the National Center for Health Statistics. This system is based on the 2023 Office of Management and Budget (OMB) delineation of metropolitan (urban) and nonmetropolitan (rural) counties.

Given the centrality of urban-rural geographic classification to the analysis, I wanted to make sure that the results were not sensitive to the choice of rurality classification system. To test this, I re-ran the analysis using the urban-rural definitions from other classification systems, including Rural-Urban Continuum Codes (RUCCs), Rural-Urban Commuting Area Codes (RUCAs), and Urban-Influence Codes (UICs). To my surprise, RUCCs, UICs, and the NCHS system all shared an identical urban-rural split, despite differences in their underlying codes. It seems that the vast majority of urban-rural cutoffs used by the federal government are derived from OMB’s metro/nonmetro definitions. This is also true of CMS (ignoring dual classification) and HRSA, which use OMB’s definitions to geographically classify hospitals.

RUCAs do not share the same split as the other systems since they’re based not on counties, but on ZIP codes/Census tracts. Nevertheless, they have an urban-rural reimbursement gap that’s basically identical to the other code systems.

Though the classification systems share the same OMB-based urban-rural split, they look very different when broken out to their underlying codes. To check that the general trend visible in the NCHS codes is also present in the other code sets, I plotted all the code sets together:

Some errant boxes aside, the trend here is unmistakable. Median commercial-to-Medicare ratios start high on the left (urban) side of the plot, then fall as you move right, toward the more rural classifications. The additional levels of rurality (compared to the two in the NCHS system) show that this drop in reimbursement is gradual and nearly monotonic – more rurality almost equals less reimbursement. Indeed, the most rural classification in each system always has the lowest reimbursement rates. The strength and clarity of this trend makes it very unlikely that the results are sensitive to the chosen classification system.

Medicare rates

In this analysis, Medicare rates are used as the denominator of our variable of interest: commercial-to-Medicare ratios. As such, it’s important that these rates are calculated correctly and don’t contain any systematic bias or errors that would confound our results. The Medicare rates in this analysis are calculated by Turquoise Health using IPPS/OPPS rules. Nearly all rates are provider-adjusted (i.e. they include adjustments for geography, DSH, etc.).

To check for systemic issues, I compared Medicare rates across urban and rural counties. I was specifically looking for Medicare rates that were erroneously high in rural areas, as these would lower commercial-to-Medicare ratios and (falsely) create the result we found.

Like the commercial-to-Medicare ratios, the Medicare rates also have a clear trend: rates decrease monotonically as you move from urban to rural counties. This intuitively makes sense, as CMS adjusts Medicare reimbursement to account for local average wages, and urban areas are more expensive/have higher wages than rural areas. But there are also other, more subtle things that would create this trend. For example, urban hospitals might have a procedure mix that would increase their reimbursement because they offer more complex, higher-priced care.

Ultimately, I found it difficult to disaggregate which parts of the Medicare rate trend were real (caused by expected sources like wages, DSH adjustments, and procedure mix) vs fake (caused by some kind of systemic error). None of the statistical tests or aggregations I performed indicated an obvious issue, but the rates are so inherently complex that I can’t rule out all potential problems.

One more wrinkle with Medicare rates is the previously mentioned special payment designations. These designations can increase Medicare reimbursement or even replace fee-for-service reimbursement entirely. For example, Critical Access Hospitals (CAHs) forgo IPPS/OPPS and are reimbursed at-cost.

Given the number and complexity of these special payment programs, they are not accounted for in the Turquoise-calculated Medicare rates. Instead, CAHs and other hospitals receive the counterfactual rates that they would have received under IPPS/OPPS. This is defensible as using counterfactual rates actually decreases the magnitude/size of the urban-rural reimbursement gap. Conversely, as Medicare reimbursement goes up under special payment programs, the commercial-to-Medicare ratio falls, widening the gap. In other words, the urban-rural reimbursement gap would probably be even larger if CAH and other special rural programs were accounted for.

For more information on methods and to view the public code used to make this article, see the associated GitHub repository.

The loophole, known as BlueCard underpayments, involves arbitraging prices between different BCBSA-affiliated insurers located in the same state. The process is obscure and works in just a few states under specific conditions. However, it can be incredibly lucrative for both vendors and providers, netting them thousands per patient encounter with minimal effort. Here’s how it works.

Most states have just a single Blue insurer – think BCBSIL or Florida Blue. However, some states have multiple Blue insurers (“Blues”). For example, California has both Blue Shield of CA and Anthem Blue Cross.

When someone with Blue insurance travels to a different state and receives medical care, that care is covered and reimbursed by the “local” Blue via the BlueCard program. BlueCard lets providers bill the local Blue instead of the Blue from the patient’s home state, AKA the “home” Blue. This decreases the complexity of billing for providers while making the patient's travel experience more seamless.

The trouble starts when a Blue-insured patient travels to a state with multiple Blues. In that case, the patient is covered by multiple insurers at the same time. Those Blue insurers may have different prices, even for the same service. After the patient receives care, providers typically bill one of the insurers at random, irrespective of price.

Whenever that happens, there’s an arbitrage opportunity for vendors. If the provider bills the cheaper of the two Blue insurers, a vendor can void or cancel the original claim, rebill the more expensive Blue, then take a cut of the difference in prices. Here’s the whole process visualized using real negotiated prices from California:

Providers are in on the process too. They contract with vendors to search for BlueCard underpayments in their recent claims, usually under the auspices of “underpayments recovery.” For providers, every rebilled patient encounter means more revenue. However, that extra revenue has to come from somewhere. Usually it comes from the higher-reimbursing Blue insurer, but patients can also get stuck with a higher bill depending on their cost share.

For vendors, this entire operation is a closely guarded secret – the equivalent of Jane Street’s India options trade, just at a much smaller scale. They’ve found a tiny gap in the system, a free money faucet to exploit for as long as possible before the Blues shut it off or the competition gets stronger. But the exploit isn’t really free; the cost is ultimately borne by patients, who end up with higher bills and premiums without even knowing they were used.

So today, let’s expose the whole thing. We’ll take a look at where BlueCard underpayment vendors are operating and what sort of procedures they’re targeting, then briefly talk about how the industry got here and why it’s likely to continue to face similar issues.

Where BlueCard underpayments actually happen

BlueCard underpayments are rare, even relative to the number of normal rebilled underpayments, and can only happen under specific conditions.

First, they can only occur in states with two or more Blue insurers, of which there are seven. Of those seven, only three are serious targets for BlueCard underpayment vendors:

Additionally, in states with multiple Blues, each Blue may control its own region of the state. Below is a map of providers with negotiated rates from two or more Blues. Areas with many dots are regions where Blues overlap, while blank areas represent regions controlled by a single Blue. BlueCard underpayments can only occur in regions where Blues overlap.

The regional overlaps get extremely messy. Here’s a map showing how just two Blues split coverage of Washington state. The light blue and green areas show the overlapping service area, while the other colors show permutations of coverage from different single Blues. You can see the light blue area roughly mirrors the area with many dots from the map above.

In other states, the situation is even more complex. New York and Pennsylvania both have four Blue insurers. Those insurers mostly don’t overlap, except in specific regions like eastern PA. Some states, like Virginia and Missouri, have just a tiny area where Blues overlap.

Further, whether or not BlueCard underpayment claims are even accepted is up to the local Blue insurer. Some don’t allow the sort of claim voiding and rebilling that makes the BlueCard scheme possible.

Finally, providers need to be in on the game. Underpayment vendors need access to providers’ detailed claims, contracts, and medical records in order to do their work. Providers grant that access with the expectation of getting additional, risk-free revenue from their closed or recent claims. In the BlueCard case, providers may tell their general underpayments recovery vendor to search for BlueCard claims, or they may contract with a BlueCard-specific vendor.

To summarize, BlueCard underpayments can only happen when:

1. A state has multiple Blue insurers.

2. Those insurers have overlapping coverage regions in the state.

3. A patient travels from out-of-state to one of those overlapping regions and receives care.

4. The local Blues allow BlueCard-style underpayment rebilling.

5. A provider in an overlapping coverage region contracts a vendor to find underpayments.

Although these conditions seem highly restrictive, several common scenarios make meeting them surprisingly likely. First, remote work has drastically increased the chance of someone being covered by an out-of-state Blue plan. Second, almost all of the multi-Blue states are economic or tourism powerhouses, making them likely destinations for travel or remote work.

Third, many large companies offer a Blue plan from their headquartered state to employees spread throughout the country. In fact, the BCBSA requires employers to purchase coverage from the licensee in their home state1, regardless of where their employees are located. For example, Target has group insurance from Blue Cross Blue Shield of Minnesota and more than 400K employees in the United States. Here are the top 20 companies with a Blue plan by employee count:

Note, however, that not all these employees will have a Blue plan, as many large employers offer plans from multiple insurers (e.g. Target employees in California can get Kaiser). Further, some employer Blue plans require providers to submit claims to a specific local Blue. For example, in Washington state, Boeing (headquartered in VA) has an agreement that requires all BlueCard claims to flow through Regence, negating the possibility of underpayments for their employees.

Still, when you add up all multi-state employers with Blue plans, the scale is surprising. If even 1% of their employees meet the five criteria listed above, they’d likely generate hundreds of thousands of BlueCard claims per year, all subject to underpayments chicanery.

Vendors target the highest bills and largest rate gaps

But vendors using this scheme don’t make money through volume (i.e. rebilling lots of small claims). Instead, they seek out expensive claims with big rate gaps between Blues. Those claims tend to be for things like long inpatient stays, surgeries, or complex treatments. For example, here’s the gap in median rates across Blues for MS-DRG 871 (sepsis with MCC):

Finding a sepsis rate gap like this might make a vendor a few hundred dollars. Not bad, but not huge money. Sepsis is also common, so pricing it is easier and the rate gaps for it are likely to disappear over time.

What vendors really want are the big, yet-to-be-closed rate gaps for obscure, expensive care. The plot below shows the distribution of rate gaps between Blues by service line in the three major BlueCard underpayment states.

Here you can see that services like oncology, surgery, and infectious disease treatment have the largest rate differences by far. For those services, the 90th percentile difference is around $45K, meaning a lucky vendor could make $9K on contingency (at 20%) for rebilling a single claim. And these gaps are just for single billing codes. Real bills could include multiple codes, professional fees, carveouts, and other complexities, all of which could have their own rate gaps.

In fact, the higher the complexity of the bill, the more likely it is to be scrutinized by underpayments vendors. The absolute juiciest targets in the BlueCard underpayments industry are stop-loss claims. These are claims where the cost of care exceeds a threshold amount and triggers a new payment methodology (usually percent of charges).

Stop-loss clauses are added to contracts to protect providers from the unknown cost of expensive, high-intensity care like traumatic injuries, major complications, and pre-term births. However, stop-loss payment methods/percentages can vary significantly by payer, creating huge opportunities for BlueCard arbitrage. For example, here are some real reimbursement rates and stop-loss clauses for a premature birth without major problems (MS-DRG 792) from West Valley Medical Center in Idaho:

In a normal case, Regence reimburses a flat $38,513.19 compared to Blue Cross’s $51,568.93, as seen in the Negotiated rate field above. The difference in those negotiated rates is big enough to arbitrage, but the payout wouldn’t be huge. However, in costly outlier cases, Regence has a much less generous stop-loss clause, meaning they would reimburse far less than Blue Cross and create a large rate gap.

For example, if a 2-week hospital stay generates $500K in gross charges, Regence would reimburse ((($500,000 gross charges - [$38,513.19 rate * 4]) * 38% of charges) + $38,513.19 rate) $169,973 compared to Blue Cross’s ($500,000 gross charges * 46.5% of charges) $232,500. That’s a rate gap of $62,527, which would net a BlueCard vendor around $12.5K in revenue, assuming a 20% contingency take.

The gap gets even larger for more extreme cases. Given a 4-week hospital stay that generates $1.5M in gross charges, Regence would reimburse ($1,500,000 gross charges * 17% of charges) $255,000 compared to Blue Cross’s ($1,500,000 gross charges * 46.5% of charges) $697,500. Here the gap balloons to $442,500, which yields $88.5K in revenue at 20% contingency.

And the gaps can be even bigger for super complex care. Burns, for example, typically have their own negotiated reimbursement algorithm and a stop-loss clause. UCI Health in Orange, CA is one of about 125 burn centers in the United States. Here are their reimbursement algorithms for burns:

Not exactly straightforward. In the simplest possible case, if we take the California median gross charge of $1,771,782 for this billing code and assume we hit the max allowable % of charges for each payer, then the gap between them would be a massive $652,016. That’s $130K in revenue for a BlueCard vendor at 20% contingency (and $522K extra for UCI Health).

Realistically, accurately pricing these large claims requires genuine expertise and effort – you need to not only perfectly follow each payment algorithm and stop-loss formula, but also know each payer’s policies around what will be covered and reimbursed. Priced correctly, these claim gaps would probably be much smaller. Indeed, the largest gap I’ve actually heard of was around $300K on a long NICU stay.

That said, the juice is worth the squeeze. Find just a few of these big claims and you’re set for the year. Stack enough of them together and you end up with revenue upwards of $10M/year from BlueCard underpayments alone.

How did we get here?

BlueCard underpayments are a symptom of the complexity of U.S. healthcare. Their existence stems from the tangled history of the Blues, the accumulated complexity of the healthcare revenue cycle, and the constant need for providers to find revenue.

Historically, Blue Cross and Blue Shield were separate entities, each with their own purpose and national association. In 1982, the separate Blue Cross and Blue Shield associations merged, forming the modern BCBSA and licensing model (which permits one Blue per state). Over time, most separate state-level Blues merged or dissolved.

The remaining states with multiple Blues are artifacts of history. They either had separate Blue Cross and Blue Shield companies that never merged (California, Idaho, Washington), or had multiple regional plans that remained distinct (Pennsylvania, New York, Missouri, etc.). When BlueCard arrived in 1994, those unmerged Blues became seeds for the underpayment arbitrage that sprang up 20+ years later.

Over those same decades, the healthcare industry grew at a staggering rate, and administrative complexity grew with it. New regulations, new codes, CMS guidance, the ACA, the No Surprises Act, the OBBB, and more have all congealed into a thick regulatory soup. Payers and providers are locked in an everlasting war, writing increasingly complicated contracts designed to give themselves the upper hand.

The result is a system without design, full of independent, rational actors that can't stop to consider the whole. Each payer negotiates its own fee schedules, case rates, per diems, carve-outs, and percent-of-charges clauses. They adopt different bundling rules, clinical policies, stop-loss formulas, and claims processes. Before federal price transparency rules, even the Blues were effectively completely independent – they couldn’t see each other’s negotiated rates. Now that prices are public, the gaps and inefficiencies are clear, embarrassing, and easier than ever to exploit.

Meanwhile, most providers are just trying to survive. Public rates are tight and commercial rates are constantly pressed down by payers. Labor and capital costs are ever-increasing. And the federal government seems intent on cuts and policy changes. In this environment, providers use whatever means they have to eke out revenue and staunch the bleeding. If a vendor promises them essentially free, no-risk extra money, then they’re going to take it, ethics of rebilling be damned.

Put all this together and it’s easy to see why BlueCard underpayment vendors exist. The complexity of this ecosystem, grown over decades and thick with regulations and self-interested big players, creates seams and gaps which get filled by opportunists, eager to capitalize on that complexity. BlueCard vendors are hardly alone in this. Hundreds of other small players now inhabit this world, searching for any niche where they can make money.

The problem is that all those players have a vested interest in the continuation of the system that feeds them. For most vendors, complexity isn’t a problem to be solved; it’s their business model, the thing that gives them both the opportunity and reason to exist. A simpler system would both lower the barriers to entry (e.g. domain knowledge) and remove niches for them to fill, so most vendors will advocate for the status quo.

BlueCard vendors fit this mold perfectly. They’re the bottom-feeders of the healthcare ecosystem, getting fat on the detritus of the Blue whales above them. They’ll continue to thrive until someone actually puts in the work to simplify contracting and close the gaps between Blues.

What’s next?

Fortunately, change does seem to be happening, albeit slowly and incrementally. Last October, the BCBSA settled a massive, $2.8B anti-trust case focused on the BlueCard program and competition across states. Under the settlement, the BCBSA agreed to implement a long list of reforms to BlueCard focused on transparency, timeliness, and administrative costs. There’s nothing specifically about BlueCard underpayments, but many of the changes will probably indirectly reduce the attack surface for such arbitrage.

Further, some BlueCard vendors are improving their own processes and trying to reduce complexity. Specifically, they’re moving from retrospectively rebilling closed claims to proactively monitoring charges in real-time. That means providers are now just billing the higher payer up front, no rebilling needed.

As for the broader healthcare system, things are at the very least becoming more transparent. The 2021 Price Transparency rule and follow-up Transparency in Coverage mandate made healthcare prices publicly available for the first time. That pricing data is the source of this post, and it’s incredibly useful for shining a light on things like BlueCard underpayments. But illuminating and fixing aren’t the same thing. Price transparency lets us see the problems, but the immense complexity of the U.S. healthcare system continues to foster opportunities for difficult-to-detect exploitation and arbitrage. The hard work of simplifying it - closing loopholes and replacing systems - is just beginning.

Subscribe now

FAQs

What is the scale of this industry?

It’s hard to say. There’s essentially no information about BlueCard underpayments online, and I couldn’t get anyone from one of the Blues to talk to me about it (and not for lack of trying). Backing out the exact amount of money being made here (or a count of claims rebilled) is nearly impossible, as the BlueCard rebilling process varies from state to state and isn’t associated with any specific CARC/RARC code.

Why don’t the Blues do something about this?

Some have. At least a few have disallowed this sort of BlueCard rebilling. Others, like Blue Shield of California, have claims routing tools seemingly intended to prevent BlueCard issues. I’m guessing that for most Blues, this issue is just small enough relative to their usual business that it doesn’t warrant special attention.

Why don’t providers just bill the higher price up front?

Due to the sheer complexity of medical billing/coding/contracting, providers often don’t know the exact dollar amount they’ll be reimbursed for a complex claim. That means they also don’t know which Blue will pay out more. Sometimes providers may have a general sense of which Blue reimburses more for a specific service line (e.g. inpatient surgery gets higher rates from Blue Shield) and bill accordingly. All that said, the BlueCard vendors themselves are increasingly moving to a proactive approach, where they help providers pick the higher Blue payer before filing a claim.

Is the cost of a BlueCard underpayment really passed on to the patient?

It depends. If the patient has insurance with high cost sharing (high deductible, coinsurance, out-of-pocket max, etc.), then they absolutely could be saddled with the extra cost of changing to the higher-reimbursing payer. In practice however, many BlueCard underpayment claims are for high-cost care, meaning the patient is likely to have already met their cost share obligations. Additionally, some providers will write off the difference in the patient’s cost share when switching to the higher-reimbursing payer.

Who are the BlueCard vendors?

Although it would be funny to call them out, I don’t like naming names for a piece like this. Plus, some of them were nice enough to chat with me for this piece.

How do I know if I’m affected?

If you travel to or live in a state with two Blues and have Blue insurance from a different state, you could be affected by this. As a patient, you’ll probably never know this even happened, you’ll just get a larger-than-the-counterfactual EOB/bill in the mail.

The One Big Beautiful Bill (OBBB) Act may disqualify over 300 hospitals from the 340B program – about 12% of the ~2,500 providers currently enrolled. Disqualified hospitals would lose access to 340B’s discounted drug prices, which would severely impact their financial health and possibly even force some to close.

Subscribe now

The disqualifications stem from the OBBB’s effect on hospitals’ Disproportionate Share Adjustment percent (DSH %), a key metric used to determine 340B eligibility. DSH percent is derived from the Disproportionate Share Patient Percent (DPP), which quantifies the proportion of low-income Medicare and Medicaid patients a hospital serves:

Holding all else constant, the OBBB’s substantial Medicaid spending cuts will lower the number of Medicaid patient days (bolded) in the DPP formula above. As a result, DPP (and therefore DSH percent) will decrease, shifting many providers below the thresholds required by 340B:

Using state-level estimates of the OBBB’s impact on Medicaid spending, I recalculated the DSH percent of each 340B hospital based on 2023 CMS cost reports. In a worst-case scenario, 314 hospitals would lose 340B eligibility due to the OBBB cuts. This assumes that hospitals do nothing and simply allow their DSH percentages to drift downward. Here’s what the fallout looks like by state:

And by hospital type:

To summarize, the OBBB mainly impacts the 340B eligibility of hospitals that are:

• In states with substantial OBBB Medicaid spending cuts (see the KFF map)

• Right above the 340B DSH threshold relevant to their covered entity type

• Serving a large number of Medicaid patients specifically impacted by work requirements

• Rural or mid-sized urban DSH hospitals (using the CMS definition of rurality)

In short, the OBBB is likely to impact already-vulnerable safety-net hospitals the most, cutting their Medicaid revenue and removing the cushion of 340B savings in one blow. That double hit could leave many rural and mid-sized urban providers underwater, undoing the very support the 340B program was meant to provide.

Subscribe now

Notes

Some notes and assumptions about this analysis:

• For a full list of affected 340B hospitals, including their current and counterfactual DSH percent, see the CSV here.

• This analysis assumes that percent cuts in Medicaid spending translate to equivalent percent cuts in patient days. This is imperfect, but a good-enough proxy until all state-level estimates of the OBBB’s effects on Medicaid enrollment become available.

• This analysis assumes ceteris paribus, i.e. all other factors affecting DSH percent are held constant over time. This isn’t likely to hold, as historically hospitals are fairly scrappy and will probably find a way to move their DSH percent if needed.

• How quickly Medicaid enrollment decreases will affect DSH percentages is a bit ambiguous. States must implement the OBBB’s work requirements by December 2026, but stringency and implementation speed will likely vary significantly by state.

• Critical Access Hospitals aren’t affected since they don’t have a DSH percent requirement.

• The code needed to replicate this analysis is available on GitHub. See this PR for the relevant bits.

1. The 340B Drug Pricing Program isn’t working as intended. What started as a way to subsidize safety-net hospitals has become a critical way for many hospitals to generate income.

2. As a result, 340B program participation in Illinois has grown by 229% over the last 15 years. Over half of Illinois hospitals are now 340B covered entities.

3. In Illinois, 340B disproportionately benefits large regional hospitals and academic medical centers such as UChicago Medicine. Such providers earn an increasingly large proportion of their net patient revenue from outpatient drugs.

4. Due to the program’s opacity, it’s difficult to measure how 340B proceeds are actually used to serve low-income patients. There’s only a minor difference in charity care between 340B and non-340B hospitals in Illinois.

5. Most providers probably aren’t abusing 340B; they’re just responding to the incentives created by bad policy. Providers and states are trapped in a bad, FOMO-driven equilibrium.

Subscribe now

If the road to hell is paved with good intentions, then the 340B Drug Pricing Program must cover a mile-long stretch.

What started as a way to subsidize safety-net hospitals and help low-income patients has morphed into a critical way for hospitals to generate income. The program lets hospitals buy drugs at a discount, give them to patients, and then get reimbursed at a much higher price, pocketing the difference.

In theory, those profits are supposed to be passed on to patients – used to provide uncompensated care, offer community benefits, expand care access, and/or subsidize otherwise unprofitable lines of business. However, the program doesn’t require hospitals to report their 340B-derived spending, or even how much they make from the sales.

The unchecked nature of the program has resulted in numerous scandals. And its rapid growth in the past few years has started to garner attention. Mark Cuban tweeted about it, the Senate majority released a massive report on its problems, and industry leaders are calling it out.

Meanwhile, drug companies, hospitals, and states are in an all-out legislative war. In 2020, drug companies began adding new restrictions and requirements to try to limit the scale of the 340B program. Hospitals responded by lobbying states to pass 340B contract pharmacy protection laws. Eighteen states to date have done so, and at least nine additional states had pending legislation in 2025 designed to preserve 340B profits for hospitals.

One of those states was Illinois, where legislators were considering SB2385 and HB3350. This pair of bills would have effectively stopped drugmakers from restricting 340B drug discounts, netting Illinois hospitals hundreds of millions of dollars. Yet the bills got curiously little attention outside of the business press, and there’s been essentially no public analysis or due diligence covering them.

So let’s fix that and take a hard look at 340B in Illinois. This post is a state-level effort to quantify the extreme growth of the program, identify winners and losers, and ultimately catalog 340B’s many unintended consequences.

It’s also a rebuttal to some of 340B’s critics. The program may be flawed and opaque, but many providers genuinely depend on it for survival. With already thin margins and no better alternatives, hospitals treat 340B as a necessary subsidy. Participation isn’t about opportunism; it’s often a rational response to structural underfunding and a broken system of incentives.

To understand how we got here - and why so many providers feel stuck - we’ll start with some background on the 340B program, then look at Illinois specifically, and finally take a look toward 340B’s future.

What is 340B?

The 340B Drug Pricing Program lets qualified health care providers, known as covered entities, buy drugs from manufacturers at a deep discount. Covered entities then sell those drugs to patients or their insurers at a higher price – usually the existing cash or negotiated rate.

The stated goal of the 340B statute is to enable covered entities, “to stretch scarce Federal resources as far as possible, reaching more eligible patients and providing more comprehensive services.” But the actual intent is somewhat ambiguous. Are covered entities supposed to pass on their discounts directly to patients? Provide free care for uninsured patients? Invest in otherwise unprofitable lines of business? The law doesn’t specify, so providers are free to do whatever they want with their 340B savings.

To participate in 340B, covered entities must be a federal grantee or one of six hospital types. The vast majority of the program is made up of just two hospital types: Disproportionate Share Hospitals (DSH) and Critical Access Hospitals (CAH). DSHs must be nonprofit or government-run and have a certain percentage (11.75%) of their total inpatient days come from low-income patients. CAHs must be nonprofit or government-run but don’t have the same DSH percentage requirement.

Qualified hospitals can also register off-site outpatient clinics as “child sites” of the main parent hospital. These child sites also qualify for 340B and can purchase and administer discounted outpatient drugs. Child sites must appear as reimbursable outpatient departments on the parent hospital’s Medicare cost report, but otherwise there are few restrictions on what qualifies. Most child sites are things like oncology centers, specialty clinics, and family medicine centers.

In addition to child sites, 340B covered entities can also dispense discounted drugs at both in-house and contract (external) pharmacies. Up until 2010, each hospital was limited to just a single contract pharmacy, but guidance changes after the passage of the ACA lifted that restriction, allowing an unlimited number. Contract pharmacies typically take a small cut of each 340B-eligible prescription and/or charge an administrative fee.

340B discounts don’t extend to every drug. Inpatient-administered drugs are entirely excluded, as are vaccines, OTC drugs, and some orphan drugs. 340B-purchased drugs must be separated (virtually or physically) from non-340B drugs of the same type. Calculating the 340B discount amount is complex, and the resulting discounted prices aren’t public. CMS estimates the average discount is somewhere around 35%, but others put it much higher.

Since the 340B discount rate is (usually) a fixed percentage, higher-cost drugs yield more savings for covered entities. The Health Resources and Services Administration (HRSA), which oversees 340B, publishes aggregate sales figures showing that the largest 340B purchases are for high-cost specialty drugs for conditions like cancer and HIV.

Finally, to participate in 340B, providers must prevent duplicate discounts and diversion. Duplicate discounts occur when a manufacturer provides a 340B discount and Medicaid rebate for the same prescription. Diversion occurs when a provider sells 340B-discounted drugs to someone who doesn't qualify as their patient. Drug manufacturers claim both these mechanisms are ripe for abuse.

How 340B pricing actually works

Before diving further into the 340B program in Illinois, let’s walk through a quick example to see how 340B actually benefits a covered entity.

Below are real prices for the cancer drug Keytruda at Javon Bea Hospital in Rockford, IL. In the standard (non-340B) case, Javon Bea likely practices what’s called buy-and-bill:

1. They buy the drug from a specialty distributor for somewhere close to the Average Sale Price (ASP). In this case, Keytruda is $11,198 for a 200 mg dose.

2. Next they administer the Keytruda to a patient. The patient has commercial insurance, which reimburses the hospital at a negotiated rate of $18,520 for the dose.

3. The hospital makes $7,322 ($18,520 - $11,198). Not all of that is profit. There are costs associated with the handling, administration, and storage of each dose.

In the end, the hospital gets something close to $7,322 in profit from a single Keytruda dose when reimbursed via standard buy-and-bill. Now let’s see what happens in the 340B case:

1. Javon Bea, a 340B covered entity, buys the same 200 mg Keytruda dose at a discount. Here the discount is estimated at -23.1%, but it’s likely much higher. In this case, they save at least $2,586 (original price - discounted price) via 340B pricing.

2. They administer the drug and bill the patient’s insurance. The negotiated rate stays the same at $18,520 for the dose.

3. The hospital now makes $9,908, with the $2,586 in acquisition cost savings getting added directly to their profits.

So, under 340B pricing, one dose of one drug for one patient netted Javon Bea an extra $2,586 in profit. Add that up across thousands of doses, and you start to glimpse the importance of 340B for many hospitals.

But 340B covered entities aren’t limited to giving 340B drugs to patients with commercial insurance, they can also give them to publicly-insured patients under Medicare (and Medicaid, but it’s more complex). Let’s see how 340B impacts Medicare Part B patients:

1. Once again, Javon Bea gets the 340B-discounted Keytruda price of $8,612.

2. This time however, they administer Keytruda to a Medicare patient. Medicare reimburses at ASP ($11,198) plus 6%, so $11,870.

Since the reimbursement is so much lower than commercial insurance, the hospital makes much less ($3,258) despite the 340B discount. However, without the discount they would’ve made only $672, so 340B almost 5x’d their profit.

When serving Medicare patients, 340B pricing can be the difference between a hospital essentially breaking even and making a sustainable profit. In the past, CMS tried to lower the Medicare reimbursement rate for 340B drugs to ASP - 22.5%, roughly matching the minimum 340B discount. However, 340B covered entities objected strongly, and later won a Supreme Court case forcing CMS to reverse its decision in 2022.

Now, with some background on 340B and knowledge of how it works, let’s see how 340B has played out in Illinois.

340B is huge in Illinois

340B has grown rapidly in Illinois, from 34 participating hospitals in 2010 to 112 in 2025, an over 200% increase. Over half the hospitals in the state are now 340B covered entities.

Of the six possible hospital types, Critical Access (CAH) and Disproportionate Share (DSH) are the most common, with CAH spread throughout rural parts of the state and DSH concentrated around Chicago. Broadly speaking, 340B hospitals tend to be located in less wealthy areas. However, their satellite clinics (i.e. child entities) often extend into wealthier areas to capture greater returns from commercially-insured clients.

The Affordable Care Act (ACA) expanded the 340B program to include four new hospital types: cancer hospitals, Rural Referral Centers (RRC), Critical Access Hospitals (CAH), and Sole Community Hospitals (SCH). The ACA also expanded state Medicaid programs, including in Illinois, which indirectly boosted 340B enrollment by raising hospital DSH percentages. Accordingly, the vast majority of Illinois hospitals joined after 2010:

340B covered entities aren’t limited in size, so many of the largest hospitals in Illinois are now part of the program. Northwestern Memorial, Rush University Medical Center, and UChicago Medicine are all 340B covered entities. Each had more net patient revenue than all 50 CAHs in Illinois combined.

In order to qualify for 340B, a certain percentage of a hospital’s inpatient days must serve low-income patients. For DSH providers, the DSH percentage threshold is 11.75%. For RRCs and SCHs, the threshold is lower at 8%. Many Illinois 340B hospitals sit right above their respective DSH thresholds, indicating that they may take on just the required number of low-income patient days, but no more. Edit: Note that at least some of the hospitals just above the threshold have their DSH percentage adjustment capped at 12% by formula.

Other hospitals get themselves reclassified to qualify. In 2020, Northwestern Memorial was classified as a Rural Referral Center, even though it’s in the middle of downtown Chicago. Despite the name, Rural Referral Centers are not required to serve rural patients, only to have 275 inpatient beds. The lower RRC threshold of 8% allows Northwestern to qualify for 340B despite a DSH percentage of around 10%.

Illinois hospitals clearly see the value of the 340B program – the vast majority of those that qualify have now joined. And while the post-ACA growth in 340B hospital enrollment has tapered off, a new avenue for 340B growth has emerged.

340B child entities and contract pharmacies are everywhere

Each 340B hospital can have an unlimited number of child entities (CE) and contract pharmacies (CP), and both types of entities have grown immensely in Illinois.

Fifteen years ago, Illinois had around 25 total child entities. Now that number is close to 1,250, a 50x increase. The vast majority of those entities belong to large regional hospitals and DSH providers in the Chicago area, including Carle Foundation Hospital (110 CE), Javon Bea Hospital (92 CE), Loyola University Medical Center (93 CE), and Rush University Medical Center (75 CE).

Each child entity can administer and prescribe 340B drugs, expanding the parent hospital’s opportunities for savings. This has led to concerns about health system consolidation, as hospitals move to buy up community practices to expand their 340B networks and patient pools.

There are minimal limits (up to 35 miles, but sometimes more) on how far child entities can be from the parent hospital. As a result, a common approach to maximizing 340B profits involves opening satellite clinics in distant, well-off areas, using the parent hospital’s DSH status to qualify.

Contract pharmacies have grown rapidly as well. HRSA’s guidance change in the wake of the ACA had a profound effect on Illinois’ contract pharmacy market. In 2010, there were just 5 unique contract pharmacies in the entire state. By 2011, there were 155. And in early 2025, there were 1,243.

Each pharmacy can contract with more than one 340B hospital, so it’s useful to quantify contract pharmacy growth by the number of unique provider-pharmacy relationships. Here’s what that looks like for Illinois:

Since patients are largely unaware of and unaffected by the 340B status of their drugs, hospitals are incentivized to expand their contract pharmacy networks as much as possible. The wider the network, the more likely patients are to use it, which means more profits captured for the hospital. The three Illinois hospitals with the largest number of contract pharmacies are Rush University Medical Center (515 CP), Loyola University Medical Center (350 CP), and UChicago Medicine (290 CP).

Pharmacies themselves are incentivized to participate because they typically get a share of the 340B profits or a per-transaction dispensing fee. Many big-name pharmacies also run their own 340B third-party administrator (TPA), separate entities that manage 340B inventory on behalf of covered entities. For example, CVS owns Wellpartner, one of the largest 340B TPAs.

The growth of the contract pharmacy industry has prompted pushback from drugmakers, who have placed restrictions on 340B covered entities with the goal of limiting their discounts. Some manufacturers require 340B covered entities to designate a single contract pharmacy. Others require covered entities to submit detailed claims data, ostensibly to prevent duplicate discounts. These restrictions have generally been upheld in federal court, prompting states to create laws to protect the access of covered entities (such as Illinois SB2385 and HB3350).

Throughout the U.S. 340B contract pharmacy relationships are increasingly dominated by a few big companies. That’s also true in Illinois, where CVS and Walgreens are by far the largest players:

The reduction in contract pharmacy relationships since 2023 seems to be the result of pharmacy closures, mostly among Walgreens and grocery-store pharmacies like Kroger (Mariano’s) and Jewel-Osco.

Who benefits most from 340B in Illinois?

So, 340B has grown massively in Illinois. Now that it’s big, who benefits the most?

To find out, we’d ideally want to compare the actual 340B drug profits of each Illinois hospital. However, those numbers aren’t public and they’re nearly impossible to estimate using public data.

Instead, we can estimate the overall outpatient drug profits of each hospital, a close proxy for 340B profits. The calculations are a little bit complex and imperfect (see here for details), but the relative results are robust and clear – big academic medical centers (AMCs) and regional hospitals simply make way more money from outpatient drugs than smaller hospitals.

Part of that is simply scale, large hospitals just have more patients, larger outpatient networks, and more revenue than smaller ones. But part of it is also a focus on more complex care, which in the case of 340B, returns higher profits.

The plot below shows the range of likely outpatient drug profits per hospital. The lower bound of the range is based on the hospital’s overall conversion rate from their CMS cost report. The upper bound of the range is based on the hospital’s average gross-to-net ratio for common drug HCPCS.

Looking at Illinois, you can see that many AMCs and regional hospitals like OSF Saint Francis have higher outpatient drug profits than every CAH in the state combined. Whether or not you think this is bad depends on your beliefs about the intent of the 340B program.

A big part of public policy is fairly allocating limited resources. In the case of 340B in Illinois, a disproportionate and growing share of the benefits are going to large, well-capitalized health systems, not to the small, scrappy safety-net hospitals the program (probably) intended to support.

On the one hand, hospitals like OSF Saint Francis and UChicago Medicine really do serve a high number of low-income patients (the latter has one of the highest DSH percentages in Illinois). On the other hand, these large hospitals likely have the patient volume and finances to be sustainable without the subsidy of 340B. Further, 340B participation changes their incentives, pushing them to acquire outpatient practices and making them increasingly reliant on the profits from 340B drugs (see below).

Plus, the 340B program isn’t costless. It’s essentially a direct transfer from drug companies to hospitals. If pharma companies can’t make money on certain 340B drugs, then they’ll find other ways to generate profits. It’s like squeezing a balloon: the harder the collective 340B program is squeezing on one part, the more prices are likely to go up elsewhere – and having big hospitals participate in the program adds a lot of squeeze.

Looking at the plot above, you might get the impression that only big Chicago hospitals benefit from 340B, but that’s not the case. Here’s each hospital represented by a circle scaled to the midpoint of its estimated outpatient drug profits:

Regional hospitals in Peoria, Rockford, and Urbana-Champaign also participate in 340B. Insofar as these cities are regional hubs for Illinois and destinations for advanced care, it may be beneficial to subsidize their hospitals, even though they’re quite large.

This adds some nuance to the question of which hospitals the 340B program should help. Should 340B subsidize regional hospitals but not AMCs and big-city hospitals? Or should 340B only allow small, truly critical hospitals to participate? Such questions are critical but unanswered. HRSA’s deference to statute and unwillingness to look critically at 340B qualification criteria lets big hospitals run away with the ball.

The result has been massive growth of the 340B program and, perhaps more dangerously, an increasing reliance on outpatient drugs as a source of income. The plot below shows the average proportion of net patient revenue stemming from outpatient drugs by hospital type:

The DSH average is clearly pulling away from the others, mostly driven by big providers in Chicago. Some hospitals drew close to 20% of their net patient revenue from outpatient drugs alone.

The exact cause of the increased drug share among DSH providers is hard to determine. Drug negotiated rates could be rising faster than negotiated rates for other types of care. Or perhaps prescription volume went up while negotiated rates remained static. Or maybe providers bought up existing outpatient practices and made them into child entities.

Whatever the reason, the 340B program seems to further increase hospitals’ reliance on outpatient drugs as a source of income. This increases their risk exposure to changes in the law and to restrictions from drugmakers. So far, hospitals have managed to insulate themselves from such changes via litigation and state-level laws, but a major shakeup of 340B would likely be disastrous for many of them.

What does 340B mean for Illinois patients?

To recap, 340B has grown a lot in Illinois and most of that growth has benefited large hospitals. But 340B is ostensibly about benefiting low-income patients, so presumably some of the savings get passed on to them via lower prices or free care, right?

No, not really. At least not in a way that’s publicly visible. Illinois patients (or their insurers) pay basically identical drug prices across 340B and non-340B hospitals. That’s true state-wide and across many other cuts of the same pricing data (e.g. broken out by hospital size, drug indication, etc.). If 340B discounts are getting passed along to patients, it's through indirect means like medication assistance programs or free vaccinations, not through reduced rates.

The same story is true for uncompensated care. On average, Illinois 340B hospitals treat only slightly more charity care cases than non-340B hospitals. They also don’t provide significantly more expensive charity care – their costs as a percentage of net revenue are just 0.3 percentage points higher.

There are some exceptions to this trend. Some 340B CAHs, such as Hammond-Henry Hospital, do seem to provide a significant amount of uncompensated care. Stroger Hospital is another unique case. It’s a publicly-owned DSH hospital with an extremely high disproportionate share percentage (~34%), a high proportion of charity care patients, and a drug cost-to-charge ratio (CCR) of around 0.64.

It’s very possible that 340B hospitals are using their savings to benefit patients in ways that are mostly invisible in public data. This could include subsidizing otherwise unprofitable service lines like obstetrics, creating new community health or outreach programs, or preventing a hospital closure.

However, none of these activities are obvious because the 340B program has no reporting requirements whatsoever. Hospitals don’t need to share what they did with their 340B profits, or even report how much 340B saved them. This lack of transparency has obscured 340B’s potentially good effects while cracking the door for attacks from drugmakers.

In sum, evidence of 340B’s benefit to patients is ambiguous at best. More data and transparency is needed to enable meaningful study of the program.

All aboard?

Many providers do use 340B savings for their intended purpose. However, the design of the 340B program itself encourages growth and invites malfeasance. Congress created an open-ended discount with no reporting requirements, and providers responded rationally by treating it as a general subsidy rather than a charity fund.

Now everyone is stuck with 340B. The growth of the program has made it self-sustaining, a FOMO-fueled money train running rampant through the U.S. healthcare system. Providers that are already on the train have a strong incentive to preserve its inertia. Many of them probably even depend on it to survive. Providers that aren’t yet on board know they’re leaving free money on the table.

Meanwhile, drug manufacturers are trying to hit the brakes. For the past 5 years, they’ve come up with a litany of restrictions to slow the program down. Most recently, they pulled some of the few levers left to them by limiting contract pharmacies to a single location and by switching to rebates instead of upfront discounts.

States have responded by legislating to protect 340B providers. In Illinois, SB2385 and HB3350 would prevent drugmakers from limiting contract pharmacies or collecting any data related to 340B. Other states have passed similar bills – here’s the legislative landscape as of June 2025:

For states, 340B is a way to bring out-of-state money into the state at no cost to taxpayers. Letting pharma companies restrict access is akin to leaving free money on the table, so states have responded rationally by blocking them.

All of this is to say that providers and states are largely trapped. Their most morally righteous move - to not engage with a program that has clearly grown beyond its original intent - is against their self-interest and probably against the interests of their patients. Forgoing 340B would mean forgoing millions in subsidies all while other hospitals and drug companies profit. It’s no wonder that providers do everything they can to leverage the program.

If anyone is to blame for the current state of 340B, it’s Congress and HRSA. They started the train, got it moving, then took their hands off the controls – abdicating their responsibility to tweak the program. Worse still, they turbocharged many of the program’s problems: adding unlimited contract pharmacies, weakening enrollment requirements, and ignoring the need for transparency. 340B may be broken, but it’s broken by design – everyone is acting in their own self-interest in the system of incentives Congress created.

What’s next for 340B?

The 340B program has gone off the rails. Its uncontrolled expansion has critically warped hospital incentives and finances, creating issues that will be difficult to undo without legislative intervention. In Illinois, outpatient drug revenue now makes up an increasing portion of DSH hospital revenue, and the number of child entities continues to multiply. This excessive dependence on 340B leaves hospitals vulnerable; a sudden federal shift, such as the One Big Beautiful Bill (OBBB), could severely impact their margins or even their existence – more on that soon.

Congress and HRSA should take responsibility for the 340B program. A sensible reform plan would involve tightening eligibility criteria, demanding clear reporting on how savings benefit patients, and limiting contract pharmacy arrangements. These actions would require little federal spending and would start to rebuild the link between discounted drugs and the benefits the program was meant to deliver.

A few weeks ago, some residents of Orange County, California received a pair of worrying letters. The first came from Hoag, one of Orange County’s major hospital systems. The second came from Blue Shield of California, one of the state’s largest health insurers.

Both letters said the same thing: we are fighting with each other, and if we don’t stop fighting by July 1st July 15th, then Blue Shield will stop covering Hoag and you’ll need to find a new doctor.

The letters are the result of deadlocked contract negotiations between Hoag and Blue Shield. Each party is trying to set their preferred rates for the care Hoag provides: Hoag likely wants higher rates, while Blue Shield wants lower rates. Such negotiations usually happen quickly and quietly, but when they do spill into public, like in this case, it's often the result of an irreconcilable gap between the two parties.

Whatever the gap is between Hoag and Blue Shield, it isn’t detailed in their respective letters. The only hint at it comes from Blue Shield, which says, “Currently, Hoag hospitals are some of the highest cost facilities for our members in the region.” No further context is provided. No numbers or timeline for resolution.

This type of public payer-provider fight is increasingly common in the United States – 2024 saw a 54% increase in the number of public disputes compared to 2023. From Florida to Connecticut, health systems and insurers are duking it out in the court of public opinion like the cable companies of the 2010s. Both sides stoke outrage in the hopes of gaining a better deal, relying on the obfuscation and complexity of prices to win the vibe war. Patients are left waiting in the dark, hoping they don’t have to upend their medical care.

But this opacity isn’t inevitable. Thanks to price transparency data, it’s now possible to attach some numbers to these disputes, see where the parties actually stand, and ultimately develop an informed opinion. Hopefully, doing so will reduce payers and providers’ ability to appeal to emotions in the absence of hard data. At the very least, publishing these numbers will give patients the context they deserve.

So, as a proof-of-concept, let’s dive in and look at Hoag and Blue Shield. We’ll start with each party’s background and market position, then do some data detective work to figure out the most likely sticking point in their negotiations.

Hoag

Hoag is a big player in the Orange County healthcare market. They have two major hospitals, an inpatient orthopedic institute, and a huge array of clinics, ASCs, and urgent care centers. All told, Hoag has ~620 inpatient beds and over 8,000 employees.

Hoag’s marketing heavily emphasizes care quality by citing awards, CMS ratings, and patient stories. They claim the title of “Orange County’s best health system” on their website. They seem to be positioning themselves as the premium (AKA highest quality) provider in the market.

Hoag’s main competition seems to be UC Irvine Medical Center, the other large (400+ bed), high acuity (CMI > 1.5) provider in Orange County. Hoag also competes with medium-sized, urban hospital networks like Providence and MemorialCare, whose facilities have similar bed counts, care acuity, and payer mixes.

However, Hoag is much more expensive than its competition. The map below shows the median negotiated rate for major hospitals in Los Angeles, relative to Medicare benchmark rates:

Hoag prices are high, even relative to large academic medical centers like UCI and USC. The two Hoag hospitals shown here are the Newport Beach and Irvine locations, and both hospitals share the same set of high negotiated rates. The only hospitals more expensive than Hoag are a few Adventist outliers and St. Mary’s in Long Beach. Cedars-Sinai, perhaps LA’s most well-known hospital, is just below Hoag.

This map has some caveats. It doesn’t show every hospital (some are excluded for lack of data, Kaiser is excluded because it’s an HMO). It relies on a (carefully chosen) sample of rates for 20 representative procedures. And it only includes rates from 5 major insurers. But the conclusion here is robust, changing the particulars doesn’t change the overall takeaway: Hoag’s premium care comes at a premium cost.

But what about Blue Shield? Let’s take a look at the other party in this fight.

Blue Shield

Blue Shield of California is the third largest health insurer in California (~15% commercial market share), behind only Kaiser (at 37%) and Anthem/Elevance (at 24%). Based on aggregated claims data, I estimate Blue Shield has slightly higher market share in Orange County (around 18%).

Blue Shield has a history of publicly feuding with California hospitals. In 2024, it settled a (much spicier) contract negotiation with Providence Health, followed by a similar negotiation with Stanford Health in Northern California. It is currently negotiating with both Hoag and the UC health system, with both contracts set to expire on July 1st.

In Orange County, Blue Shield’s main competition (besides Kaiser) is for-profit insurers with similarly large networks. Here’s how Blue Shield stacks up:

Basically, Blue Shield is weird. Relative to other major insurers in Orange County, Blue Shield has the lowest inpatient rates and the highest outpatient rates. Their rates are particularly high for cheap(er) procedures like CT scans and colonoscopies. From the outside, it’s hard to know the reason for this split. It’s possible that Blue Shield is offering higher outpatient rates as a tactic to make providers take a cut on their (typically higher cost) inpatient rates.

Where’s the beef?

So, Hoag is expensive relative to other hospitals. And Blue Shield is a bit idiosyncratic but is probably in the middle or top end of payers. If they’re both near the top of their respective markets, why are they still fighting?

To return to Blue Shield’s hint from their press release, “Currently, Hoag hospitals are some of the highest cost facilities for our members in the region.” Let’s see if that’s true:

So Blue Shield isn’t lying. Hoag really is an outlier among their Orange County providers. They have the highest outpatient rates (by a large margin) and are behind only UCI for inpatient rates. I think the story here is actually pretty simple – Hoag is just really expensive for Blue Shield.

To flesh this out and speculate a bit, I think each party is looking at a different set of comparable providers.

Hoag positions itself as the premier hospital in Orange County, offering the highest-quality care. They’re likely comparing themselves to other top-end providers like UCI, USC, and Cedars-Sinai and demanding the same compensation. Hoag is banking on their brand and betting that Blue Shield needs them to maintain a strong network in the area.

Meanwhile, Blue Shield is comparing Hoag to the other regional hospitals in Orange County and finding that it’s one of the most expensive providers on their books. Blue Shield is probably trying to bring Hoag’s rates in line with systems like Providence and MemorialCare, and is betting they can strongarm Hoag using their market share.

Of course, there are probably lots of other sticking points here, overall cost is just the big, obvious one. Hoag’s Medicare cost-to-charge ratios are high for inpatient care, indicating that they’re not making much money on the inpatient side. It’s possible that they’re pushing for higher inpatient rates or more favorable (i.e. flexible) contract types.

Ultimately though, the dispute between Hoag and Blue Shield likely boils down to a fundamental disagreement about the value of Hoag's care, with Hoag emphasizing its quality and Blue Shield focusing on cost relative to other regional providers.

So what?

Blue Shield and Hoag aren’t the only players having this kind of fight. The media covers dozens of disputes like this each year, and almost none of that coverage includes hard numbers or data.

That opacity is strategic. Healthcare is emotionally charged, and when negotiations go public, it’s easier for both sides to appeal to emotion than to defend their financial demands. Providers invoke loyalty and quality. Insurers warn about affordability and access. Patients, meanwhile, are left vulnerable, anxious, and uninformed – forced to choose sides in a game with no clear score.

But that can change. Price transparency data lets us peek behind the curtain. It offers a concrete foundation for patients to form their own opinions and avoid being used as leverage. Patients deserve visibility into the processes that affect their healthcare – and we’re here to provide it. Subscribe for future articles that bring clarity, context, and data to the complex world of healthcare.

Subscribe now

The price of childbirth varies significantly by geography, complexity of care, insurer, and lots of other factors. It can be as little as $5K (for an uncomplicated vaginal delivery in rural Tennessee) or as much as $70K (for a C-section with major complications in the Hamptons). Nationwide, the average uncomplicated vaginal delivery costs right around $10K, which makes the U.S. one of the most expensive places in the world to give birth.

However, most patients won’t actually pay that much. These are negotiated rates; they’re what insurers pay out to hospitals, not what patients typically pay out of pocket. Depending on your deductible, co-insurance, and out-of-pocket max, you’ll likely pay a fraction of that $10K, maybe even nothing.

So why look at these prices if they aren’t what patients actually pay? Because the prices are paid by insurers, and patterns within them can tell you a lot about a hospital or insurer’s financials, market power, and strategy. Negotiated rates also let you compare across geographies and procedures without the complexity or lag of claims data. You get a real-time, nationwide picture of the true cost of care, even if not all of that cost is experienced by patients.

So, what does this nationwide picture look like for childbirth? Here's the state-by-state breakdown of negotiated commercial rates for an uncomplicated vaginal delivery (MS-DRG 807). Note that these prices only cover the hospital/facility fee, there may be additional costs for consultations, an epidural, etc.:

Broadly speaking, delivery prices are closely correlated with cost of living. Expensive states in the Northeast and Pacific Northwest have the highest prices, while states in the South have the lowest. Alaska is a major outlier, with high prices that likely stem from its low labor supply, low population density, and limited market competition. Similarly, Florida is dealing with a nurse shortage, limited competition, and an aging (i.e. more expensive) population. Hawaii is an outlier in the opposite direction; it has low delivery prices relative to its high cost of living.

But a map like this doesn’t actually reflect the way patients shop. Some states are huge; people aren’t going to drive from Houston to El Paso to deliver their baby. And people on the edge of a state don’t just stop at the border. Let’s try a smaller geography.

Here’s the same underlying data as the state-level map, but now at the ZIP code level. Each ZIP price is the weighted median of all providers within a 3-hour drive, regardless of state boundaries. Larger and closer hospitals get more weight, while smaller and distant hospitals are basically ignored.

Now we can see submarkets and trends within states: the rural northeast part of Maine is dominated by a single (expensive) healthcare system, while Portland and Bangor enjoy significantly lower delivery prices; New York City has some of the most expensive ZIP codes in the country, but most of New York state is close to the national median; and in California, the Bay Area and Central Valley top the charts for delivery cost, while southern California fares much better.

Some rural areas stand out as well: northeast Nevada, southwest Texas, and Nebraska all have clusters of high-priced ZIP codes. Some of those clusters are just noise - the result of errors in the MRFs of small, understaffed hospitals - but some of them are real. At least a few of the clusters stem from large providers cross-subsidizing their Medicaid patients via sky-high commercial rates. Others, like those in Nevada and Alaska, demonstrate how rurality and small patient populations can drive up the cost of care.

Other patterns in the map are hidden and only become apparent when you plot them. For example, here’s delivery price broken out by rurality:

Despite some high-priced rural outliers, prices in urban counties are nearly 50% higher than in rural counties and have a much wider range (whiskers represent the 10th and 90th percentile). Curiously, the 75th percentile price in central metro counties is actually lower than in fringe metro counties like Palm Beach County, FL, or Marin County, CA, perhaps due to higher cost of living in many exurbs.

We can get a closer look at individual markets by using CBSAs (core-based statistical areas), which are the healthcare industry’s preferred unit of analysis. Here are the top 20 largest CBSAs with at least 15 hospitals in the data:

Once again, New York and San Francisco lead the pack, with vaginal delivery prices ~60% higher than the national median. Conversely, low-cost-of-living cities in the Midwest are actually well below the national median (and don’t even show up on the ZIP code map). C-sections are shown on the right, and are generally about 50% more expensive than vaginal deliveries.

As you might expect, cost increases even further with complications and complexity. Here’s the national distribution of all delivery prices, broken out by severity and type of care:

Complications and comorbidities (CC/MCC) increase costs a lot, particularly for C-sections. They also increase variance. The standard deviation of prices for a vaginal delivery with no complication (MS-DRG 807) is ~$4.3K, while a complicated C-section (MS-DRG 786) is 3x higher at ~$13K.

All these plots paint a clear, if predictable, picture – higher acuity care costs more, and delivering a baby in San Francisco costs more than in rural Iowa. But this picture is incomplete. Prices vary significantly based on which insurer is paying, the market share of the hospital, and even the specific type of contract negotiated. How do narrow network plans affect these prices? How can hospitals in the same market have such vastly different prices? How is healthcare in Hawaii so surprisingly affordable?

Answering such questions - understanding the real nuance and complexity of the U.S. healthcare market - is crucial to achieving the transparency and efficiency CMS originally envisioned when mandating this data's release. In the coming months, Price Points will tackle these issues with open-source code, methods, and data. Subscribe to join me on the next deep dive exploring the complexities of healthcare pricing.

Subscribe now

It’s now 2025, and compliance with the CMS mandate is… surprisingly good. Hospitals were initially slow to respond, but CMS stepped up fines + enforcement and created a standardized schema, so most hospitals (~90%) now post some kind of machine-readable file (MRF) containing their prices. But compliance ≠ results, and the actual impact of dumping 5,000+ MRFs onto the internet is a little less clear.

Subscribe now

For instance, no one can currently say whether the CMS mandate actually increased competition. Prices are moving - descriptive statistics suggest they are converging - but no large studies to date have quantified a causal effect (studies of the entire U.S. healthcare system are difficult, to say the least).

The impact on research and policymaking has likewise been muted. Collating every hospital’s MRF is a Herculean task, and once combined, the data is massive and requires substantial care/domain knowledge to use. As such, the flurry of whitepapers and reports that CMS probably hoped for hasn’t materialized. Most studies that use price transparency data are small and focused on specific markets and/or procedures. Few have looked at the national picture or dug deep into healthcare pricing and contracting.

However, one result of the CMS mandate is clear. There’s now a huge amount of valuable new healthcare data to sort through, including negotiated rates for every common procedure and drug across almost every hospital and insurer in America. This data is a vast ocean of prices that few have deeply explored, so starting today, I’m going to dive into it.

Once every few weeks, I’ll publish novel, quantitative analyses using price transparency data from Turquoise Health. Turquoise is paying me to do this as an FTE, but doesn’t tell me what to publish – or what not to publish, other than insisting I remain relatively neutral. The goals here are simple: unpack the data, fill research gaps, and explore price transparency’s impact – and do it all in public to help others with the same goals.

To that end, I’ve created a public GitHub repo which will contain the code and methods for each analysis. I’ll also include any datasets necessary for replication, as long as licensing allows (not all data used by Turquoise is theirs to share, e.g., claims data). My hope is that sharing code, data, and domain knowledge will improve the price transparency ecosystem, make research easier, foster more collaboration, and ultimately drive down healthcare prices.

So with that, let’s dive in, starting with a question: