Only in America: Administrative Cost in Perspective

The cost of running a claim through American health care is not a fact of medicine. It is a fact of the system that pays for medicine, and it can be measured against systems that do the same job differently. The peer-reviewed literature has measured it repeatedly, always the same way: take administrative spending — billing, coding, claims adjudication, the insurance-related paperwork around care — and set the United States beside a comparable high-income country. The gap is large, it has been documented for two decades, and it is the subject of this brief.

In 2017, US insurers and providers spent $812 billion on administration, or $2,497 per capita — 34.2% of national health expenditures. Canada, over the same year, spent $551 per capita, or 17.0%. The figures are Himmelstein, Campbell, and Woolhandler's, published in the Annals of Internal Medicine in 2020.¹

The Annals figure is the most recent full accounting, but it is not the first. The same research group measured hospital administration across eight nations for Health Affairs in 2014, using 2010–2011 accounting data. Administrative work accounted for 25.3% of US hospital spending — $667 per capita — against 12% in Canada ($158 per capita). The Netherlands sat at 20%, England at 16%, Scotland level with Canada near the bottom. Reducing US per-capita hospital administration to Canada's level, the authors calculated, would have saved $158 billion in 2011.²

Each panel below is one documented metric from one named study. The point is not any single number but the consistency of the shape across measurements taken years apart, by different methods, at different levels of the system.

The physician panel is the one closest to a small US practice. In the same 2017 accounting, physicians' insurance-related administrative costs ran $465 per capita in the United States against $87 in Canada— the billing, coding, and claims-handling overhead that attaches to a doctor's office rather than a hospital.¹ It is the closest the cross-national record comes to pricing the daily transaction this publication otherwise measures domestically.

The figures are not interchangeable. One counts hospital costs as a share of hospital spending; another counts all administration as a share of all health spending; a third isolates physician overhead. Reading them precisely is the whole discipline of the comparison — so the studies, their exact metrics, and their years are set out plainly.

The cross-national gap describes a level; the domestic record describes what moving off that level is worth. CAQH, which has measured the cost of US health care's administrative transactions annually since 2011, estimates that automating standard transactions avoided $258 billion in administrative cost in 2024, with a further $21 billion still recoverable through full automation of the manual and partially manual work that remains.³ That $21 billion is the same residue this publication has measured in 25-minute phone calls — the manual transactions a US practice still runs by hand.

The cross-national studies and the domestic index measure different things and cannot be added together: one is a level relative to another country, the other is a domestic opportunity relative to full automation. They converge only in direction. Clinicians in the comparator systems do not spend the same hours, or the same dollars per capita, on the insurance-related work around care. The cost gap is documented. The reader draws the conclusion.