$1.6 Trillion Healthcare Waste: Hidden Risk in Employer Health Plans

Every year, the United States spends an estimated $5.6 trillion on healthcare, more than any other country on earth. That figure dwarfs peer nations on a per-capita basis. Despite that spending, Americans experience shorter life expectancies, higher rates of preventable disease, and worse outcomes for chronic conditions than citizens in most comparable countries.

The explanation is not a shortage of resources. Research from the Journal of the American Medical Association (JAMA) and the Centers for Medicare and Medicaid Services (CMS) shows up to 30% of all U.S. healthcare spending, approximately $1.6 trillion annually as of 2025, qualifies as waste. Expressed differently, if that waste were a standalone economy, it would rank among the 15 largest in the world, ahead of Spain, Indonesia, and Switzerland.

For the organizations directly funding this care, self-insured employers covering more than 160 million Americans, this is not an abstract policy problem. It is a quantifiable financial exposure. It appears in most claims submitted for payment, shows up on most medical bills, and repeats itself every renewal cycle. For most organizations, it goes unmanaged.

This article examines the sources of that exposure, why conventional oversight structures have failed to contain it, and how artificial intelligence is giving plan sponsors the tools to confront it systematically.

Where $1.6 Trillion Goes Every Year

A landmark 2019 JAMA study led by Dr. William Shrank documented six distinct waste categories embedded in U.S. healthcare, placing total annual waste between $760 billion and $935 billion at the time. Adjusted for current CMS expenditure data, those figures project to approximately $1.6 trillion in 2025.

Billing errors and outright fraud represent the single largest category at roughly 25%, or $400 billion annually. The National Health Care Anti-Fraud Association (NHCAA) estimates between 3% and 10% of all healthcare spending is lost to fraud each year. Industry analysis suggests up to 80% of hospital bills contain errors. The American Medical Association has documented a 20% claims-processing error rate among commercial insurers, representing an estimated $17 billion in misdirected payments.

Administrative complexity accounts for another 22% of waste, approximately $352 billion per year. The JAMA study attributed $265.6 billion to this category alone. The U.S. spends more than $1,000 per capita on healthcare administration, roughly five times the per-capita administrative cost in comparable countries. Unnecessary services, pricing failures, fragmented care coordination, and clinical inefficiencies make up the balance.

Shrank’s core finding was unambiguous: proven interventions could recover $191 billion to $282 billion annually. The savings are there. The question is whether the organizations absorbing these costs are positioned to find them and act on them.

How Fragmented Oversight Became a Permanent Cost

The conventional response to healthcare cost overruns has been periodic manual auditing, where a third-party reviewer examines a random sample of claims (typically once per year) and delivers a summary to the plan sponsor. In a plan processing hundreds of thousands of claims annually, sampling 5% to 10% of transactions leaves the overwhelming majority of spending unexamined and the associated errors unreported.

Structural conflicts of interest reinforce the problem. Many third-party administrators (TPAs) process claims on behalf of self-insured employers while maintaining financial relationships with the same vendors whose charges they are meant to review. Pharmacy benefit managers (PBMs) operating on opaque pricing models earn more when drug prices rise. Under these types of arrangements, waste is no longer a failure the system seeks to correct; it becomes an outcome the system is designed to sustain.

Former Centers for Medicare and Medicaid Services (CMS) Administrator Dr. Donald Berwick placed the opportunity in clear terms, estimating as much as $800 billion in recoverable waste sits untapped across the U.S. healthcare system. That money is not hidden. It is sitting in claims datasets which most plan sponsors never fully interrogate, in vendor contracts which most employers don’t fully understand, and in pharmacy arrangements which most organizations accept without independent review.

AI changes what is operationally possible. Where a human auditor can review thousands of claims in a week, a machine learning system can process millions of claims in hours. Where a statistical sample captures a fraction of errors, a system auditing 100% of claims in near real time catches patterns no sample would surface. The scale advantage alone is decisive.

AI as the Detection Layer

The most immediate and financially measurable application of AI in healthcare cost containment is claims auditing. AI systems can review every line item on every claim, compare billed amounts against plan allowed amounts, regional benchmarks, Medicare rates, and contracted terms, and flag things like duplicate charges, unbundled services, upcoded procedures, and charges for services with no corresponding clinical record, all in near real time.

The results produced by these systems can be significant. Cotiviti, a major healthcare analytics firm serving more than 180 health plans, identified approximately $15 billion in suspect claims across its client base in 2023 alone. Matthew Hawley, executive vice president of payment integrity operations at Cotiviti, stated at the 2024 National Health Care Anti-Fraud Association conference that fraud, waste, and abuse “remain a major challenge to healthcare payment integrity, costing payers and the entire healthcare system billions of dollars each year,” and it is “critical for plans to adopt artificial intelligence and other advanced technologies that can identify and mitigate FWA while preventing further losses.”

The Centers for Medicare and Medicaid Services (CMS) confirmed their own deployment of these tools in a 2023 statement: “We use artificial intelligence and machine learning to find potential fraud that would not be apparent to the human eye. We try to use the latest technology to make potential fraud easier to detect more quickly.” Confirmed fraud schemes are addressed through vulnerability analyses, law enforcement referrals, and regulatory action.

Predictive modeling is a second high-impact application. These systems analyze historical claims data, clinical indicators, demographic profiles, and pharmacy utilization to identify plan members likely to require more help the months ahead. That early signal gives care managers the opportunity to intervene before a chronic condition deteriorates into a $200,000 hospitalization. Prevention at that scale is not achievable through manual data review.

Price transparency data, now available through federally mandated machine-readable payer files, gives employers the raw material to benchmark provider pricing market by market. AI can convert raw data into actionable comparisons, identifying where a plan is overpaying relative to the local market and where members can be directed toward higher-value providers. David Pierre, CEO of a payment integrity entity built from the merger of three specialized analytics firms, put it plainly in a 2024 interview: “Now you have technology, machine learning, AI that can look at things, serve it up to the experts so they can actually make the decision…I think this will really change the trajectory of healthcare spend across the U.S.”

The Legal Exposure Is Real

AI adoption in healthcare is also accelerating because the legal consequences of passive plan management are no longer hypothetical. Under the Employee Retirement Income Security Act (ERISA) and the Consolidated Appropriations Act (CAA) of 2021, self-insured employers have a fiduciary duty to ensure plan assets are spent prudently and solely in the best interest of plan participants.

That obligation is now being tested in the courts. In 2024 and 2025, Johnson & Johnson, Wells Fargo, JPMorgan Chase, and Mayo Clinic each faced lawsuits alleging fiduciary failures tied to prescription drug pricing, excessive administrative fees, and insufficient vendor oversight. A National Alliance for Healthcare Purchaser Coalitions survey found 65% of employers reported growing concern about litigation exposure related to their fiduciary duties.

Plan sponsors should respond by positioning AI-powered analytics as documented evidence of prudent oversight. A sponsor who can demonstrate continuous review of 100% of claims, market-benchmarked provider pricing, and regular pharmacy audits occupies a materially stronger legal position than one relying on annual random-sample reviews conducted by a vendor with competing financial interests.

Elizabeth Mitchell, president and CEO of the Purchaser Business Group on Health (PBGH), outlined the employer obligation without qualification: “Under the Consolidated Appropriations Act of 2021, employers are legally accountable as fiduciaries for their health plans, requiring them to provide employees with the best healthcare benefits for the best price.” Technology has become the most direct path to meeting that standard.

Prevention Is the Next Frontier

The current generation of AI applications in healthcare focuses primarily on detection: billing anomalies, fraud patterns, utilization outliers, and pricing irregularities, largely identified after the fact. The next generation is shifting the intervention point earlier, toward prediction and prevention.

AI systems trained on multi-year claims datasets can model chronic disease trajectories at the individual member level. When a predictive model identifies a member as high-risk for a cardiovascular event six months before it appears in claims data, a care manager can reach out proactively. It’s important to note the model itself does not make clinical decisions; it flags risk signals which enable humans to act before the high-cost event occurs.

Natural language processing is being applied to prior authorization workflows, targeting the 13 hours of weekly staff time physician practices spend on authorization requests. Automated routing of routine requests, combined with AI-driven clinical criteria review, can accelerate approval timelines and reduce the 93% physician-reported rate of care delays tied directly to the prior authorization process.

Pharmacy analytics are growing sharper. AI systems can flag members on branded medications where clinically equivalent generics exist, flag lower-cost therapeutic alternatives, and identify formulary patterns which reflect PBM incentive misalignment rather than clinical best practice. Organizations moving to transparent, pass-through pharmacy pricing models are documenting savings of 15% to 30% on pharmacy spend.

Technology Informs, Humans Decide

AI systems produce findings. People act on them. Technology deployments delivering the most consistent results pair analytical output with human decision-making: claims audit findings leading to vendor contract renegotiation, predictive risk flags triggering care management outreach, and pharmacy inefficiency analytics driving formulary redesign. Technology accelerates the process but the ultimate strategy still requires human judgment.

Bill Fera, principal consultant at Deloitte, described the shift precisely: “We’re taking the mystery away. And there is a fact base. There’s a core piece of information that can be interrogated. It’s just now, it can be interrogated very quickly.” A claims dataset that would occupy a human analyst team for months can now be processed overnight, with findings prioritized by financial impact and ready for decision makes to act upon the following morning.

For self-insured organizations funding health coverage for over 160 million Americans, this is a financial management issue as much as a technological one. A plan sponsor applying AI-powered claims auditing to a $50 million health plan and identifying a 14% payment inaccuracy rate can recover more than $7 million annually which would otherwise flow to vendor margins rather than applied to employee benefits.

The $1.6 trillion in annual healthcare waste is not distributed randomly. It tends to cluster in specific, repeating patterns and spread across duplicate charges, inflated facility fees, upcoded procedures, PBM spread pricing, and avoidable or redundant utilization. AI cannot dismantle the incentive structures that create those patterns. What it can do is make them visible, measurable, and actionable at a scale which changes what plan sponsors can demand from their data and their vendors.

For organizations that have treated healthcare expenditure as an unmanageable fixed cost, that shift in capability can be consequential.