Healthcare Administrative Costs: Where the $250 Billion Goes

Administrative spend consumes roughly 25 percent of U.S. healthcare expenditure, and published estimates of wasteful administrative spend cluster around $250 billion per year. Breaking down where that money goes is the first step toward deciding what to automate.

Every few years, a new analysis lands on the same uncomfortable fact: the United States spends far more on healthcare administration than any peer country, and a substantial fraction of that spend is wasteful. The JAMA analysis by Shrank and colleagues estimated administrative waste at $265 billion annually. The Commonwealth Fund has consistently reported U.S. administrative spend at 8 percent of total healthcare spending versus 1 to 3 percent in most peer OECD countries. The CAQH Index tracks transaction-level administrative cost year over year and shows the number growing, not shrinking.

That $250 billion figure gets repeated in boardrooms, policy papers, and earnings calls. It's less often broken down. Knowing the total doesn't help an operator decide what to do. Knowing where the $250 billion flows does.

This breakdown is for the CEOs, COOs, and CFOs trying to decide which administrative cost buckets are realistically addressable through automation in the next 12 to 24 months, and which are structural problems that require policy or payer-side change.

How big is the healthcare administrative cost problem?

The top-line numbers frame the scale.

Total U.S. healthcare spending in 2024 crossed $4.9 trillion according to CMS National Health Expenditure data. Administrative spending, variously defined, runs between $900 billion and $1 trillion annually across providers, payers, and government programs. Of that, roughly $250 billion is estimated to be waste, meaning administrative cost that could be reduced without reducing clinical output or care access.

For a point of comparison, $250 billion is larger than the total annual spending on cancer care, larger than the annual Medicare budget for physician services, and roughly equivalent to what the U.S. spends on all of prescription drugs in a year. It's not a trivial problem.

On a per-provider basis, the MGMA benchmarks show primary care practices running administrative cost as 27 to 32 percent of total practice expense. Specialty practices run 22 to 28 percent. For a 10-provider primary care group with $8 million in revenue, that's $1.8 to $2.5 million in administrative cost annually, a chunk of which is addressable.

Where does the $250 billion go?

Not all administrative spend is created equal. Some is value-generating (the portion that ensures quality, safety, and appropriate reimbursement). Some is pure waste (duplicate work, failed handoffs, workarounds for bad systems). Most sits in the middle.

Here's how the estimated $250 billion of wasteful administrative spend breaks down across the five major cost buckets, drawing from CAQH Index data, AMA survey data, JAMA studies, and MGMA benchmarks.

Totals are approximate and vary across published sources, but the relative scale holds across analyses: billing and claims is the largest bucket, PA and coordination are next, and patient access is the fastest-automating category.

Bucket 1: Billing, claims, and revenue cycle (~$80B)

The largest single administrative cost bucket, and the most heavily studied. The CAQH Index tracks 10-plus transaction types (eligibility verification, claims submission, claim status, remittance advice, prior authorization, and others) and reports that roughly half of these transactions are still partially manual, with per-transaction cost gaps of $10 to $30 between manual and fully electronic processing.

Where the cost concentrates: denial management (workforce to resubmit and appeal denied claims), rework (correcting claims that failed submission the first time), eligibility verification errors (claims denied because coverage was wrong at submission), and coding corrections.

What automation addresses: eligibility verification at appointment scheduling, electronic claim status checking, denial pattern detection and automated appeals, coding consistency checks. What it doesn't address easily: complex clinical appeals, payer-specific manual review processes, and payer-side decisions that require human negotiation.

Bucket 2: Prior authorization and utilization management (~$35B)

The administrative cost most visible to physicians and patients, and the most universally reviled. AMA data puts practices at an average of 43 PA requests per physician per week, with 14 hours of staff and physician time per week spent on PA activities. For a 10-physician practice, that's roughly 7,000 PA transactions per year and one full-time FTE dedicated entirely to PA work.

Where the cost concentrates: portal submissions across dozens of payers with different formats, status tracking on pending submissions, peer-to-peer scheduling for denials, clinical appeal writing.

What automation addresses: form pre-filling and submission, status check polling, peer-to-peer scheduling coordination, clinical documentation extraction from the chart. What it doesn't address: payer-side decision delays, contested clinical cases requiring medical judgment, payer-specific process changes that require continual adaptation. For a deeper take see prior authorization automation with AI.

Bucket 3: Care coordination and referrals (~$30B)

The administrative category that has the most direct impact on patient outcomes because coordination failures translate into missed care, late diagnoses, and avoidable ED visits. MGMA benchmarks put coordinator cost at 25 to 30 minutes per referral processed manually, and practices run between 150 and 500 referrals per week depending on size and specialty mix.

Where the cost concentrates: inbound fax and EHR referral intake, patient outreach for scheduling, eligibility and prior auth coordination on referred services, closed-loop documentation back to referring providers.

What automation addresses: fax-to-data extraction, patient SMS and voice outreach, self-scheduling, appointment confirmation, closed-loop reporting. What it doesn't address: the ~15 percent of cases with genuine coordination complexity (multi-specialist referrals, unusual coverage, social barriers to care) that still require human judgment. For why these breakdowns happen see why referrals get lost between primary care and specialists.

Bucket 4: Scheduling, patient access, and front office (~$25B)

The most automation-ready of the major categories, because the work is high-volume, rule-based, and patient-facing. Practices running no-show rates of 18 to 25 percent, abandoned call rates above 15 percent, and time-to-first-appointment of three-plus weeks see the cost across all of these metrics.

Where the cost concentrates: inbound call handling, appointment booking, reminders, no-show recovery, waitlist management, schedule template maintenance.

What automation addresses: voice AI-based inbound call handling and scheduling, SMS-based self-scheduling, multi-cadence reminders, no-show recovery outreach, intelligent waitlist management. What it doesn't address: complex new-patient intake requiring unusual insurance workflows, scheduling that requires clinical triage judgment.

Bucket 5: Credentialing, compliance, and quality reporting (~$20B)

The least-automated major category, for structural reasons. Credentialing depends on individual payer processes that are slow to change. Quality reporting depends on data that lives across EHR, claims, and registry systems.

Where the cost concentrates: provider enrollment and payer credentialing (6 to 9 months per new provider across all payers), HEDIS and quality measure reporting, audit response, compliance documentation, regulatory filings.

What automation addresses: document management and tracking, reminder workflows for renewal deadlines, quality measure data aggregation across sources, some audit workflow preparation. What it doesn't address: the payer-side credentialing cycle itself, novel compliance questions requiring expertise.

How much of this is realistically automatable in 2026?

The question that matters for operators is not the size of the wasteful spend. It's how much of it is addressable with today's technology at today's economics.

The honest answer across buckets is roughly $100 billion to $130 billion of the $250 billion is addressable through operational automation in the 2026 to 2028 timeframe. The rest is structural, requires payer-side change, or requires regulatory reform.

The addressable portion breaks down roughly as follows: 45 to 60 percent of coordination and referral costs, 65 to 80 percent of scheduling and patient access costs, 60 to 75 percent of PA costs, 35 to 55 percent of revenue cycle costs, and 25 to 40 percent of credentialing and compliance costs. These are within-reach percentages based on what existing AI agent and automation platforms deliver in production today.

For a 10-provider specialty practice, that translates to approximately $400,000 to $700,000 in addressable annual administrative cost, against an automation investment typically running $60,000 to $120,000 per year. For a 100-provider multi-specialty group, it's $4M to $7M against investment of $500K to $900K. For an FQHC running Medicaid and Medicare panels, the addressable cost plus the quality bonus revenue opportunity combines to a multi-million dollar annual impact. For the dollar-level math on referral workflows specifically see the ROI of AI referral automation.

Why has administrative spend grown instead of shrinking?

The uncomfortable counterpoint to the “automation will fix this” narrative is that healthcare administrative spend has grown, not shrunk, over the past decade despite substantial IT investment. EHR adoption is near-universal, patient portals are ubiquitous, and yet per-capita administrative spending has kept rising.

Three structural forces explain the growth.

First, payer proliferation and complexity. The number of distinct payers, plan designs, and benefit structures has grown, and each variation creates incremental administrative burden. Medicare Advantage growth specifically has added complexity without simplifying fee-for-service Medicare equivalents.

Second, quality and value-based reporting requirements have grown. Each new CMS payment model, each new HEDIS measure, each new state quality program adds reporting overhead. These are not necessarily wasteful in aggregate (they're aimed at driving better care), but they sit on top of existing workflows.

Third, PA requirements have expanded. Payers have added prior auth to more service categories over the past decade as a cost containment tool. Each new PA adds administrative burden on the provider side that isn't offset by savings visible to the provider.

Automation can offset growth in administrative burden. It hasn't yet reversed it because the burden keeps expanding. The organizations that will come out ahead in the next five years are the ones that automate the coordination, scheduling, and PA layers fast enough to absorb incoming burden without staff growth. That's especially true for PE-backed multi-location groups running roll-up integration playbooks.

Best fit and less ideal fit for this framing

This breakdown fits best for: CEOs and COOs at multi-site practices and health systems making platform-level technology decisions, CFOs evaluating ROI on operational automation investments, private equity operating partners at healthcare portfolio companies, board members and investors evaluating healthcare operations companies.

Less ideal for: clinical leaders focused on care quality and patient outcomes (the care outcome story is implicit here but not the focus), revenue cycle specialists looking for tactical tooling recommendations (the $80B billing bucket is discussed at bucket level, not at tool selection level).

Frequently asked questions

Is the $250 billion figure still current?

The frequently-cited $265 billion estimate comes from the JAMA “Health Care Administrative Costs in the United States” analysis by Shrank, Rogstad, and Parekh. When inflation-adjusted and extrapolated against current total health spend, the 2026 equivalent is in the $300 billion range. The “$250 billion” framing is a conservative rounding of the original figure and remains directionally accurate.

Why don't EHRs solve this?

EHRs are systems of record. They document the work that happens and route tasks among people. They don't do the work themselves. Administrative waste concentrates in the handoffs between systems, between organizations, and between work steps, which is where automation agents live. EHRs plus automation is a better operating model than EHRs alone.

How much do payers spend on administration?

Payer-side administrative spending runs in the same $400 to $500 billion annual range as provider-side spending, according to Commonwealth Fund and NAIC data. Payer-side automation (claims processing, PA review, member services) has its own investment cycle and its own addressable opportunity. The provider-side conversation addresses approximately half of total administrative cost. Payer-side reform addresses the other half.

What's the fastest-payback administrative automation for a mid-size practice?

Patient outreach and scheduling automation typically has the fastest payback (60 to 90 days) because the no-show reduction and care gap closure show up in revenue quickly. Referral intake automation and PA automation have longer payback windows (90 to 180 days) but larger total impact because the per-transaction labor cost is higher. For a 10 to 50 provider practice, the typical sequence is outreach first, then referral intake, then PA. FQHC operators should also see FQHC care gap closure with AI.

Does administrative automation reduce staff headcount?

Typically it holds headcount flat while absorbing 40 to 80 percent growth in workflow volume. The most common pattern is that practices implementing automation do not lay off staff but do stop backfilling turnover while the system scales. Over 12 to 24 months, effective coordinator ratios improve 2 to 3x. The role itself changes from volume processing to exception handling and patient advocacy.

Pulling it together

The $250 billion figure is an abstraction. What matters for operators is the breakdown. Coordination and scheduling and prior auth are the buckets most addressable in 2026. Billing and revenue cycle is partially addressable. Credentialing and compliance are mostly structural.

Organizations that have started automating the addressable buckets in 2024 and 2025 are already pulling ahead on per-provider administrative cost, quality performance, and staff retention. Organizations still deliberating on whether to automate in 2026 are paying the full freight on a growing administrative burden. That's a decision with compounding consequences.