This is a reference document. It sets out what the Department of Veterans Affairs published about its own claims and appeals outcomes, alongside a count we ran ourselves against the Board of Veterans’ Appeals decision archive.
It does not tell you what to do with your claim. It contains no recommendation, and nothing here is legal or medical advice. It is a set of numbers and an explanation of what they describe, so that a veteran, a representative, or a reporter can see the same picture the agency’s own reporting shows.
Who compiled it. Flat Rate Nexus prepares independent medical opinions for veterans. We are paid to do that, which gives us an interest in how you read the figures below. That is why every number here is linked to its source and every reporting period is stated. Check them.
After a VA decision, three review paths exist. They differ in one respect that governs everything else: what evidence each is allowed to look at.
Higher-Level Review (VA Form 20-0996). A more senior adjudicator at the same office re-examines the existing file. No new evidence may be submitted or considered. The reviewer is looking for an error or a difference of opinion in what is already there.
Supplemental Claim (VA Form 20-0995). The claim is reopened on the basis of new and relevant evidence. Something not previously in the file has to be added for this path to function.
Appeal to the Board of Veterans’ Appeals (VA Form 10182). A Veterans Law Judge, independent of the office that issued the decision, reviews the record.
Those are the rules of the three lanes, not a suggestion about which to use.
Higher-Level Review, fiscal year 2025. VA granted 24.2 percent — 68,853 of 283,975. On a file to which nothing was added.
That rate has risen every reporting period: 14.5 percent in 2019, 18.7 percent in 2022, 24.2 percent in 2025.
Supplemental Claim, fiscal year 2025. VA granted 48.2 percent — 330,209 of 685,669.
Board of Veterans’ Appeals, January 2021 through April 2025. In a classification of the Board’s published decisions, the Board granted at least one previously denied benefit in 33.3 percent of decisions — 96,935 of 290,661 parsed, 95% confidence interval 33.2 to 33.5. The method, the code and the derived data are published so the figure can be recomputed independently.
Set the Higher-Level Review figure beside the Supplemental Claim figure.
Roughly a quarter of decisions changed when a second, more senior person read the same file. Nothing was added to it. Whatever moved those outcomes was already in the record when the first decision was made.
Roughly half changed in the lane where new evidence enters.
The two lanes are not measuring the same population and the comparison is not an experiment — a veteran who has new evidence to file is not the same veteran as one who does not. But the difference between the two published rates is the clearest available indication of how much the contents of the file drive the result.
We maintain a copy of the Board’s published decision archive. Within it, we counted outcomes according to whether a private medical opinion was present in the record.
Across January 2021 to April 2025, among the 103,083 decisions the Board granted or denied outright:
Read that carefully, because it is the number on this page we have the most interest in you misreading. It is an association, not a demonstrated effect. Veterans who obtain a private medical opinion are not a random sample: they differ in resources, in representation, and quite possibly in how strong their case looked before they sought one. This count cannot separate the opinion from the veteran who went and got it. It is also our own classification rather than a VA statistic, and unlike the 33.3 percent figure it is not part of the published replication package.
VBA reports claims processing accuracy in the low-to-mid 90 percent range, drawn from its Systematic Technical Accuracy Review programme. The VA Office of Inspector General has reported that STAR has not adequately identified and corrected the agency’s claims-processing deficiencies.
Most compensation and pension examinations are conducted by contract vendors rather than VA clinicians. VA has spent nearly $6.8 billion on these contracts since fiscal 2017. The contracts set a 92 percent accuracy floor; the Inspector General reported all three vendors below it since at least 2017, during a period when the available contractual disincentives went unused. The contracts have since been redesigned to tie payment to accuracy, and the Government Accountability Office reported that the revised mechanism was itself poorly administered, including calculation errors that overpaid contractors about $2.3 million in fiscal 2024.
These are findings by the VA’s own Inspector General and by GAO. They are reported here as findings, without any claim about why they occurred.
VA Appeals Modernization Act metrics — Higher-Level Review and Supplemental Claim grant rates, federal fiscal year 2025, and the Higher-Level Review trend for 2019, 2022 and 2025. benefits.va.gov/REPORTS/ama/
Board of Veterans’ Appeals decision archive — our classification of 290,661 parsed decisions issued January 2021 through April 2025. April 2025 is where our copy of the archive ends; 2025 is a partial year of four months and is not a full calendar or fiscal year. Method, code and derived data: github.com/gohokies2005/bva-reversal-census
Private-medical-opinion count — same archive and period as above; our own classification, not published as part of the replication package.
VA Office of Inspector General — STAR programme review and contract examination programme.
Government Accountability Office — GAO-25-107483, examination contract accuracy incentives.
How to cite this: Flat Rate Nexus, The 2025 VA Claims and Appeals Report, 17 September 2026. VBA figures are federal fiscal year 2025 from VA’s published AMA metrics. Board figures are our classification of 290,661 decisions issued January 2021 through April 2025, with method and code published for independent replication.