What the new claims data shows
The Doctors Company, a physician-owned medical malpractice insurer and part of TDC Group, analyzed 735 radiology malpractice claims it handled between 2010 and 2025. The average indemnity payment across those claims was about $456,000, according to the insurer's November 2026 "Malpractice Risk Review," as reported by Radiology Business and AuntMinnie.
"Every malpractice claim magnifies where patient care can be improved," said Julie Ritzman, MBA, CPHRM, senior vice president of patient safety and risk management at The Doctors Company, in a statement accompanying the release. The insurer's point in publishing claims data at this level of detail isn't academic — it's meant to tell practices exactly where their exposure concentrates, in dollar terms.
The one common cause: diagnosis, not technique
About 75% of the claims The Doctors Company analyzed stemmed from diagnosis-related errors. Narrowing further, clinical judgment issues were present in about 80% of radiology claims overall — a category the insurer defines as misinterpretation of diagnostic studies, a narrow diagnostic focus (reading for the indication and stopping, rather than scanning the whole study), and a failure or delay in ordering the tests needed to reach a diagnosis.
That is not an isolated finding. A separate, independent analysis of the National Practitioner Data Bank, published in RSNA's journal Radiology in August 2026, reviewed nearly 400,000 physician malpractice payments from September 1990 through December 2024 and found diagnostic errors accounted for 71.3% of radiology malpractice claims — versus 27.6% of claims across other specialties. Two different data sets, two different time windows, the same conclusion: in radiology, the claims problem is overwhelmingly a diagnosis problem.
Where the cost actually concentrates
The Doctors Company's review also breaks out indemnity by contributing factor, which matters for anyone trying to prioritize risk-reduction spend rather than treat "malpractice risk" as one undifferentiated number:
| Contributing factor | Present in claims | Avg. indemnity when present |
|---|---|---|
| Clinical judgment (misread, anchoring, missed order) | ~80% | ~$487,000 |
| Communication breakdown (clinician-to-clinician) | ~26% | ~$518,000 |
| Behavior-related (patient non-adherence to follow-up) | ~20% | ~$444,000 |
Communication failures carry the highest average payout of the three — a reminder that a correct finding never acted on is nearly as costly as a finding never made. (The Doctors Company's November release includes a case example: a suspected lung mass on a chest X-ray that was never effectively communicated from radiologist to cardiologist, delaying a lung cancer diagnosis until it had metastasized.) But clinical judgment is both the most frequent factor and close behind on cost per claim — which is why it is the single largest line item in radiology's malpractice exposure.
Does workflow design actually move the liability needle?
Claims data shows what drives cost after the fact. A separate, prospective line of research asks a sharper question: does *how* a radiologist uses AI change whether a miss turns into a malpractice finding at all? A 2026 mock-jury study published in Nature Health tested exactly that, using a hypothetical stroke-rule-out CT case in which a radiologist missed a brain bleed that the AI tool under review had flagged correctly.
Across 282 mock-juror participants, the only variable changed was the workflow: in a single-read condition, the radiologist viewed the AI output once; in a double-read condition, the radiologist read the scan once before seeing the AI flag and once after. With identical clinical facts and the identical miss, 74.7% of participants in the single-read condition sided with the plaintiff, versus 52.9% in the double-read condition — roughly a 22-percentage-point swing driven by workflow structure alone, not by the underlying error. The result doesn't prove structured review lowers actual claim rates, but it is a direct, controlled demonstration that how a diagnostic process is structured affects how defensible a miss looks — which is exactly the lever insurers are pointing practices toward.
What The Doctors Company recommends — and where AI-drafted reporting fits
With clinical judgment identified as the primary driver of claims, The Doctors Company's own recommendations to practices are notably concrete: address cognitive biases such as anchoring and satisfaction of search, build awareness of commonly missed diagnoses, use AI to flag abnormalities, and obtain second reads on uncertain cases.
That maps closely onto how structured AI-drafted CT reporting is designed to work. Rather than a single read against a blank template, the AI produces a comprehensive draft report that systematically prompts for findings across the full study — not just the organ system tied to the clinical indication — which directly targets "narrow diagnostic focus," the specific failure mode named in the insurer's own data. xAID's in-house radiologist reviews every preliminary before it reaches the client, and the report is delivered ready-to-sign to the client's reading radiologist. It is a built-in second read on every study, not only the uncertain ones — applied to the exact failure mode the claims data says is most expensive.
Frequently asked questions
What is the average radiology malpractice claim worth?
According to The Doctors Company's analysis of 735 closed radiology malpractice claims handled between 2010 and 2025, the average indemnity payment was about $456,000. Claims where clinical judgment was a key contributing factor averaged about $487,000, and claims involving a communication breakdown averaged about $518,000 — the highest of the categories the insurer tracked.
What is the most common cause of radiology malpractice claims?
Diagnosis-related errors, not technical or procedural mistakes. The Doctors Company found about 75% of radiology malpractice claims stemmed from diagnosis-related errors, and clinical judgment issues — including misinterpretation of diagnostic studies, narrow diagnostic focus, and failure or delay in ordering necessary tests — were present in about 80% of claims. A separate 2026 analysis of the National Practitioner Data Bank, published in RSNA's journal Radiology, independently found diagnostic errors accounted for 71.3% of radiology malpractice claims, versus 27.6% in other specialties.
Do communication breakdowns contribute to radiology malpractice claims?
Yes. The Doctors Company's review found communication breakdowns, most often between clinicians, present in about 26% of radiology claims, and those claims carried the highest average indemnity of the categories studied, at about $518,000. Behavior-related factors, such as patients not following up on recommended imaging or treatment, were present in about 20% of claims.
Can AI-assisted reporting reduce radiology malpractice risk?
Early evidence suggests workflow design matters. A 2026 mock-jury study published in Nature Health found that when a radiologist's miss was reviewed in a single AI-assisted read, 74.7% of participants sided with the plaintiff, versus 52.9% when the same case involved a double read (once before AI input, once after) — a roughly 22-point swing from workflow alone, with the same underlying facts. That does not prove AI-assisted structured reporting lowers claim rates directly, but it is consistent with insurers' own advice to use AI to flag abnormalities and get second reads on uncertain findings.
Source: The Doctors Company, "Radiology Claims: November Malpractice Risk Review," analysis of 735 closed claims (2010–2025), as reported by Radiology Business and AuntMinnie; "Insights into Radiology Malpractice Cases in the United States from the National Practitioner Data Bank," Radiology (2026), doi.org/10.1148/radiol.253352; and the mock-jury workflow study in Nature Health (2026), doi.org/10.1038/s44360-026-00085-2. Figures are rounded as reported.