Ask a clinic owner what their biggest expense is and they will say salaries, rent, or equipment. Ask what their administrative work costs and you usually get a shrug. That is the problem in one sentence. Administrative work is one of the largest cost centres in healthcare, but most of it never appears as a line item. It is buried in staff time, in the hours a physician spends after clinic, and in the friction of a payment system with many moving parts.

This piece is the money chapter of our series. The earlier pieces covered where clinical time goes and how the in-basket grew. Here we put numbers on the page, the ones a health system leader actually needs, and we are careful about a trap that catches almost everyone who writes about this.

The 7 percent versus 34 percent trap

If you read that administrative costs are "7 percent of healthcare spending" and then read that they are "34 percent", you have not found a contradiction. You have found two different definitions, and confusing them is the single most common error in this literature.

The narrow definition counts only the cost of running the insurance machinery: the overhead of private insurers and public programmes. By that measure, the US Centers for Medicare and Medicaid Services put the "net cost of health insurance" at about 7.0 percent of national health spending in 2024, down from a pandemic peak of 8.4 percent in 2020. Against a total national health expenditure of roughly 5.3 trillion dollars, that is a large number on its own, but it deliberately excludes everything hospitals and physicians spend on their own billing, prior authorisation, and paperwork.

The broad definition adds all of that provider-side work back in. Using it, a 2020 study in Annals of Internal Medicine estimated total US health administration at 812 billion dollars in 2017, or 34.2 percent of national health expenditure. Per person, that worked out to 2,497 dollars in the US against 551 dollars in Canada, where a single-payer system removes much of the billing complexity. The authors' summary is blunt: the administrative spending gap between the two countries is "large and widening".

Neither number is wrong. When someone quotes one at you, the only useful question is which one, because they describe very different things. For anyone running a provider organisation, the broad figure is the one that lands on your floor.

What it costs to get paid, per patient

The abstraction becomes concrete when you measure the cost of billing a single encounter. A 2018 study in JAMA did exactly that at a large academic health system with a certified electronic health record, tracing the billing and insurance-related (BIR) cost of five encounter types down to the minute.

Encounter	Staff time	BIR cost	Share of professional revenue
Primary care visit	13 min	$20.49	14.5%
Emergency department visit	32 min	$61.54	25.2%
General inpatient stay	73 min	$124.26	8.0%
Ambulatory surgical procedure	75 min	$170.40	13.4%
Inpatient surgical procedure	100 min	$215.10	3.1%

Read the primary care row again. It costs a little over 20 dollars and 13 minutes of combined clinician and staff time just to get paid for a routine visit, and that consumes almost 15 cents of every revenue dollar. Across the five encounter types, BIR work ate between 3 and 25 percent of professional revenue.

The most important detail is easy to miss: this was a well-run academic centre with a modern EHR, and the authors concluded the high costs were not the result of local inefficiency or duplicated effort. This is what billing costs when it is done properly. The complexity is in the system, not in one clinic's bad habits.

The waste lens, and the uncomfortable finding

Zoom out to the whole system and the framing shifts from cost to waste. A widely cited 2019 review in JAMA estimated total waste in US healthcare at 760 to 935 billion dollars a year, roughly a quarter of all spending. Within that, the "administrative complexity" category came to 265.6 billion dollars.

Here is the part that rarely makes the headline. The same review estimated potential savings for most categories of waste, but for administrative complexity it identified no intervention with demonstrated savings. None of that 265.6 billion was counted as recoverable. The reason is structural: administrative complexity is generated by the way payers, coding rules, and coverage policies interact, and no single provider can automate its way out of a problem it did not create alone.

That is a sobering message for anyone selling, or buying, the idea that software will simply make admin disappear. Some of it will not, because it is a feature of the system rather than a bug in your workflow.

So what is actually automatable?

Plenty, as long as you are honest about which slice. The recoverable admin is the transactional layer: the routine, repetitive exchanges between providers and payers such as eligibility checks, claim submissions, claim status enquiries, and prior authorisations. These are high-volume, rule-bound, and increasingly standardised, which is exactly the kind of work software handles well.

The CAQH Index, an annual measurement of these transactions across hundreds of health plans and provider organisations, put it in 2024 at roughly 90 billion dollars in annual cost for the transactions it tracks, with about 20 billion dollars (near 22 percent) recoverable by shifting the remaining manual and partly manual transactions to fully electronic ones. That is a real, bankable opportunity, but notice its size relative to the 812 billion dollar total. The automatable transactional layer is a meaningful minority of the whole, not the whole.

What this means if you run the numbers

Three takeaways for clinic owners and health system leaders:

Separate the automatable from the structural. Prior authorisation, eligibility, and claims processing are genuine targets for automation and yield measurable savings. The broader complexity that comes from a fragmented payment system will not yield to any product, and a vendor who promises otherwise is overselling.
Measure your own BIR cost before you buy anything. The JAMA per-encounter figures are a benchmark, not your number. Knowing what a paid visit actually costs you, in staff minutes and dollars, is the only way to judge whether a tool earns its subscription. Our playbook piece covers how to measure this.
Count clinician time as the real currency. The largest administrative cost in most practices is not a software licence or a billing clerk. It is skilled clinical time diverted into paperwork, which is both expensive and the thing your clinicians most resent losing.

The economics point in a consistent direction. Administrative work is a first-order cost, a chunk of it is genuinely reducible through automation, and a chunk of it is not. The next piece looks hard at what today's automation honestly fixes, and where the savings are quieter than the sales deck suggests.

Browse the full Admin Tax series.

References

Centers for Medicare and Medicaid Services. National Health Expenditure Data, NHE Fact Sheet (2024). https://www.cms.gov/data-research/statistics-trends-and-reports/national-health-expenditure-data/nhe-fact-sheet
Peterson-KFF Health System Tracker. How has US spending on healthcare changed over time? https://www.healthsystemtracker.org/chart-collection/u-s-spending-healthcare-changed-time/
Himmelstein DU, Campbell T, Woolhandler S. Health Care Administrative Costs in the United States and Canada, 2017. Annals of Internal Medicine. 2020;172(2):134-142. https://www.acpjournals.org/doi/10.7326/M19-2818
Tseng P, Kaplan RS, Richman BD, Shah MA, Schulman KA. Administrative Costs Associated With Physician Billing and Insurance-Related Activities at an Academic Health Care System. JAMA. 2018;319(7):691-697. https://jamanetwork.com/journals/jama/fullarticle/2673148
Shrank WH, Rogstad TL, Parekh N. Waste in the US Health Care System: Estimated Costs and Potential for Savings. JAMA. 2019;322(15):1501-1509. https://jamanetwork.com/journals/jama/fullarticle/2752664
CAQH. 2024 CAQH Index Report. https://www.caqh.org/hubfs/Index/2024 Index Report/CAQH_IndexReport_2024_FINAL.pdf