Medical Malpractice Payments in the U.S.: What the Federal Data Actually Shows (2026)

By John Mahoney · August 31, 2026 · A MedLegal AI original-data study · Every figure below is cited to a public federal source or peer-reviewed journal. Reuse permitted with attribution and a link to this page.

Most published "medical malpractice statistics" are recycled from other marketing pages, uncited, or simply wrong. This study takes a different approach: every number below comes directly from the federal government's own National Practitioner Data Bank (NPDB) — the confidential repository that federal law has required medical malpractice payers to report into since September 1990^[1] — or from peer-reviewed research in NEJM and JAMA Internal Medicine. We generated the counts ourselves from HRSA's official NPDB Data Analysis Tool on August 31, 2026 (data through June 30, 2026)^[2], and the full methodology, including what this data can and cannot say, is at the bottom of the page.

Three findings journalists can quote

First, what this data is — and what it is not

The NPDB is not a verdict database and not a claims database. A Medical Malpractice Payment Report (MMPR) is filed when a payment is actually made — by an insurer or other payer — for the benefit of an individual licensed practitioner as a result of a settlement or judgment of a written claim or complaint.^[3] That definition has three consequences anyone using these numbers should understand:

• Unpaid claims never appear. Claims that are dropped, dismissed, or defensed at trial generate no report.
• Payments made solely on behalf of institutions never appear. A hospital-only settlement with no individually named practitioner is not reportable to the NPDB, so the true institutional liability footprint is larger than these counts.
• One claim can generate multiple reports. Reports are per practitioner, so a single case naming two covered practitioners can produce two MMPRs.

An NPDB report also is not a finding of malpractice — federal guidance is explicit that a payment "shall not be construed as creating a presumption that medical malpractice has occurred."^[3] For a practitioner-side walkthrough of what happens when a claim resolves and a report is filed, see our companion explainer: What an NPDB Report Means When a Malpractice Claim Resolves.

Why you will not find an "average malpractice payout" here

Nearly every marketing page on this topic quotes an "average malpractice settlement." The federal government publishes no such official statistic. In the NPDB's Public Use Data File, payment amounts are deliberately range-coded — the format specification states that "payment amounts are coded into ranges" (for example, all payments of $100 or less are coded as $50), and warns that "the grouping of payment amounts has the effect of slightly lowering the apparent mean and median payment amounts," with exact statistics available only on request to the Division of Practitioner Data Bank.^[4] Because reports are per practitioner rather than per claim, dividing total dollars by report counts also does not produce a true "average payout per case." We therefore publish the counts, the aggregate dollars, and the official payment-range distribution — and leave manufactured averages to others.

The national trend, 2015–2025

All figures in this table were generated from the NPDB Data Analysis Tool (report counts and dollar amounts in millions, nominal, all practitioner types, all states and territories).^[2]

Three things stand out. First, report volume has been remarkably stable at roughly 11,200–12,000 per year for a decade, apart from a sharp pandemic-era trough in 2020–2021 (9,685 and 9,307 reports) that is consistent with court closures and slowed claim resolution rather than any change in underlying care. Second, volume never fully returned to its 2015 level: 2025's 11,263 reports are 6.1% below 2015. Third, dollars decoupled from volume — total reported payments climbed every year from 2021 through 2025 and now sit 30.2% above 2015 in nominal terms.^[2] For the first half of 2026, the tool shows 4,499 reports, but the NPDB cautions that partial-year data lag due to the 30-day reporting window.^[2]

The payment-size shift: fewer small payments, many more seven-figure ones

The NPDB classifies each payment report into an official payment range. Here is the national distribution for 2021–2025:^[2]

The distribution is migrating upward in real time. Reports of $500,000 or more went from 23.9% of the 2021 total to 33.3% of the 2025 total. The $1 million-plus bands alone grew 76% in count (831 to 1,466) over the same five years, and the $2 million-plus band more than doubled (160 to 346).^[2] This is the payment-size mechanism behind the dollar growth in the first table — and it matters equally to plaintiff case selection and to defense exposure modeling.

State-level variation, 2021–2025

Report counts by the practitioner's state, from the same tool, sorted by five-year total. Share is each state's percentage of the 55,076 U.S. total. Raw counts are presented deliberately: differences between states reflect population and practitioner supply as well as legal climate, so treat rank, not raw magnitude, as the story (methodology note below).^[2]

A few state-level observations worth flagging: Pennsylvania's counts rose sharply within the window (619 reports in 2021 to 993 in 2025 — a 60% increase, the largest among high-volume states), while Texas — the second-largest state by population — produced fewer reports than New Jersey.^[2] Whether a state's tort rules amplify or dampen its raw counts is exactly the comparison our 50-state medical malpractice law hub is built for.

Who gets named: practitioner types

Physicians remain the overwhelming subject of malpractice payment reports: of the 55,076 MMPRs recorded for 2021–2025, 35,953 named an MD and 3,919 named a DO — 72.4% combined. Dentists (5,747), advanced practice nurses (2,425), registered nurses (1,718), physician assistants (1,268), and podiatrists (1,106) account for most of the remainder.^[2]

The growth story inside those numbers tracks the changing shape of the American care team: payment reports naming advanced practice nurses (nurse practitioners, nurse anesthetists, nurse midwives, clinical nurse specialists) rose from 387 in 2021 to 549 in 2025 — up 41.9% — while physician (MD) counts rose from their pandemic trough of 6,007 in 2021 to 7,279 in 2025.^[2] As states expand APRN scope of practice, the liability data is beginning to follow the workload.

What the peer-reviewed literature adds on specialty risk

The NPDB's public data does not break out physician specialty, but the peer-reviewed literature built on malpractice insurer and NPDB data does, and three findings have held up over time:

• Specialty risk varies enormously. The landmark New England Journal of Medicine analysis of 25 specialties found that surgical and procedural specialties (neurosurgery, cardiothoracic surgery, general surgery) face annual claim risks several times those of the lowest-risk specialties, and that a large majority of physicians in high-risk specialties can expect to face at least one malpractice claim over a long career — most of which close without any payment.^[5]
• Paid-claim rates per physician fell for two decades even as the biggest payments grew. A JAMA Internal Medicine study of NPDB data covering 1992–2014 documented a large decline in the rate of paid claims per physician-year across specialties, alongside rising mean payments and a growing share of payments above $1 million — a pattern our 2021–2025 band data shows is still running.^[6]
• Paid claims concentrate in a small subset of physicians. A New England Journal of Medicine analysis of NPDB paid claims found that a small fraction of physicians accounts for a substantially outsized share of all paid claims, with recurrence risk rising steeply with each prior paid claim.^[7]

Methodology

This study uses exclusively public, verifiable federal data and peer-reviewed literature. No proprietary datasets, no estimates, no third-party statistics pages.

• Primary source. All report counts, dollar aggregates, payment-range distributions, state counts, and practitioner-type counts were generated by the author from the HRSA NPDB Data Analysis Tool on August 31, 2026, filtering Report Type to "Medical Malpractice Payment Report."^[2] The tool covers September 1, 1990 through June 30, 2026; we treat 2025 as the most recent complete year. Recommended citation per HRSA: Division of Practitioner Data Bank, Bureau of Health Workforce, Health Resources and Services Administration, U.S. Department of Health and Human Services.
• Internal consistency checks. The 2021–2025 national total of 55,076 reports reconciles independently across all three cuts we extracted: the year-by-year table, the payment-range table (column sums 9,307 / 11,203 / 11,676 / 11,627 / 11,263), and the state table.
• Dollar figures are the tool's nominal "Dollar Amounts ($M)" series (inflation adjustment set to "No"). Percentage changes on dollars are therefore nominal, and we label them as such.
• What we deliberately did not compute. No mean or median payment appears on this page: the NPDB's public file range-codes payment amounts and HRSA warns that computed means/medians from it are approximations,^[4] and per-practitioner reporting means report-level division misstates per-claim values. We also omitted population-adjusted state rates rather than mixing in a non-NPDB denominator; state counts are presented raw with that caveat attached.
• Known limitations. MMPRs exclude unpaid claims and institution-only payments; one claim may generate multiple reports; partial-year 2026 data (4,499 reports through June 30) is incomplete due to the 30-day reporting window; and a small number of 1990–2003 reports are excluded from the tool by HRSA for data-quality reasons.^[2]
• Verification. Every figure can be reproduced in minutes: open the Data Analysis Tool, set Report Type to Medical Malpractice Payment Report, and apply the year/state/practitioner/payment-range views described above. Journalists who want our extracted tables in spreadsheet form can email [email protected].