41% and Climbing

Since 2022, Experian Health has been asking a few hundred of the people who run provider revenue cycles a blunt question: how much of what you bill comes back denied? In the first survey, 30% said at least one in ten claims. By the 2024 wave, 38%. In the third edition — 250 respondents in financial, billing, or claims-management roles, fielded June–July 2025 — the share reached 41%.¹

This is a survey of perceptions, not an audit of claims — the caveats matter and sit in the methodology block below. But the direction is consistent across three waves, and it agrees with everything else the same respondents report about their working conditions: 54% say claim errors are increasing, and 68% say submitting a clean claim is harder than it was a year earlier.¹

The series is short — three waves, with no survey in 2023 — but each wave has moved toward the worse answer. Experian's own cross-wave comparison, published with the 2025 edition, draws the trend:²

The headline number does not climb alone. Nearly every internal trend the survey tracks deteriorates in the same direction between waves — errors, payer policy churn, reimbursement lag, and the respondents' own confidence in their tooling:

Read together, the respondents are describing a squeeze from both ends: the rules move faster — more payer policy changes, more prior-authorization friction, more data requirements — while the tolerance for error shrinks. Missing or inaccurate claim data was the most-cited cause of denials in 2025 (50%, up from 46% in 2024), ahead of authorization issues and incomplete patient registration data.¹

The structural aggravator underneath the trend is staffing. In the 2022 wave, more than half of respondents said staff shortages were already slowing claims submission.⁵ In 2025, 43% report being understaffed — and the same survey finds that 90% of denied claims still pass through human review before resubmission.¹ The denial burden lands, almost entirely, on exactly the people there are fewer of.

The 2025 edition frames AI as the candidate answer, and the gap it documents is stark: 67% of respondents believe AI can improve the claims process, while 14% report actually using it to reduce denials.¹ Among the minority who do, 69% report fewer denials or better resubmission success.¹ Adoption has barely moved across the whole series — in 2022, 11% of respondents said they were using artificial intelligence in the claims process.⁵

A self-reported survey cannot say whether the adopters' results would generalize. What it can document is the posture: 41% of respondents upgraded their claims technology in the past year, and the share who call their current stack sufficient has fallen 21 points since 2022¹; 59% say they plan to invest in claims-management technology within six months.⁶ The people closest to the problem are not describing a workflow they expect to keep working by hand.

What this series cannot tell you is the actual national denial rate — respondents estimate their own organization's experience, and nobody is checking their math against remittance data. What it can tell you is that the people who process claims for a living, asked comparable questions three waves running, keep choosing the worse answer — about denial frequency, about error rates, about payer behavior, and about their own capacity. For a small practice, the practical reading is a ratio: the denial workload is climbing while the staff available to work it is not. Every year that ratio worsens, the rational-but-costly habit of writing denials off gets easier to justify — and more expensive in aggregate.