← Back to Resources

AI Nexus Letters and the BVA: What 295,000 Decisions Tell Us

Not sure if your VA claim is worth pursuing?

Run your case against 295,756 actual BVA appeal decisions. 5 minutes. No payment. No obligation.

Free Viability Check →
April 19, 2026 - Full data analysis with charts

We downloaded and analyzed 295,756 Board of Veterans Appeals decisions issued between January 2021 and April 2025. Every decision is a public record published by the VA. We parsed each one for outcome, whether a private medical opinion was cited, what weight the Board gave it, and why.

This is what the data shows.

Private medical opinions matter

BVA Favorable Outcome Rate: With vs. Without Private Medical Opinion
With private IMO (35,887 cases)
64% favorable
Without private IMO (259,869 cases)
48% favorable
With IMO - fully granted
26%
Without IMO - fully granted
15%
Source: 295,756 BVA decisions, 2021-2025. "Favorable" = granted or partially granted.

Veterans who submitted a private medical opinion had a 64% favorable outcome rate at the BVA. Without one: 48%. The fully granted rate nearly doubled (26% vs 15%).

But not all opinions are equal.

Quality determines weight

The BVA doesn't just note whether a private opinion exists. It evaluates the opinion's "probative value" using the framework from Nieves-Rodriguez v. Peake (2008): Is it factually accurate? Is the reasoning fully articulated? Is it based on sound medical principles applied to this specific veteran?

BVA Outcome by IMO Quality Rating
Probative / High Weight (9,172 cases)
79% favorable
Low Weight / Not Probative (2,108 cases)
41% favorable
No private opinion submitted (259,869 cases)
48% favorable
An opinion rated "not probative" performs worse than having no opinion at all.

A "probative" opinion nearly doubles the favorable rate compared to having no opinion (79% vs 48%). An opinion rated "not probative" actually performs worse than submitting nothing, because the Board's explicit rejection of the opinion can strengthen the denial rationale.

This is the core problem with low-quality letters, including AI-generated ones. If the BVA rates your opinion as not probative, you may have been better off without it.

Why the BVA rejects medical opinions

We extracted the specific language the BVA used when rejecting private medical opinions. The top reasons:

Top Reasons the BVA Assigns Low or No Probative Weight
"Speculative" - language like "could have," "may be," "possible"
Most common
Warren v. Brown: "could have been" is not probative. Tirpak v. Derwinski: "may or may not" is not probative.
"Conclusory" - opinion stated without explanation
Very common
A conclusion without reasoning has no probative value regardless of who signed it.
"No rationale" - no medical reasoning provided
Common
The BVA needs to see WHY, not just that someone agrees.
"Did not consider" contrary evidence
Frequent
Failing to address alternative etiologies or contradictory findings.
"Did not review" claims file
Notable
Opinion not based on actual record review.
Extracted from BVA decision text where the Board explicitly stated the reason for reduced probative weight.

Every one of these failure modes is a default behavior of AI text generation. AI hedges (speculative), states conclusions without building reasoning from specific records (conclusory), and has not actually reviewed any claims file.

The C&P exam battlefield

One of the most important functions of a private medical opinion is surviving a disagreement with the C&P examiner. When the examiner opines against the veteran, the BVA weighs the private opinion against the examiner's opinion.

When the C&P Exam is Unfavorable
Unfavorable C&P + private IMO present (4,752 cases)
67% favorable
Unfavorable C&P + no private IMO (16,732 cases)
52% favorable
C&P exam overturned AND private IMO present (1,792 cases)
82% favorable
A well-reasoned private opinion can overcome a negative C&P exam. Quality is the deciding factor.

The trend: IMOs are getting more effective

IMO Impact on Favorable Outcomes by Year
2021
59%
43%
+16
2022
58%
40%
+18
2023
67%
45%
+22
2024
76%
56%
+20
2025
72%
54%
+18
With IMO Without IMO Delta
Private medical opinions are becoming more influential at the BVA, not less.

IMO impact by condition

Condition Decisions With IMO Without Delta
Radiculopathy13,85478%60%+18
Tinnitus30,92676%62%+14
Migraines10,21376%63%+13
Erectile Dysfunction11,67875%60%+15
Hearing Loss41,51474%56%+18
GERD10,15473%57%+16
Sleep Apnea29,44872%53%+19
PTSD53,02972%55%+17
Knee52,75470%53%+17
Shoulder23,69770%54%+16
Hypertension24,98568%54%+14
Low Back20,61466%50%+16

Sleep apnea shows the largest absolute swing (+19 points). Radiculopathy has the highest favorable rate with an IMO (78%). In every condition, across every year, a quality private medical opinion improves outcomes.

What this means for AI-generated letters

An AI-generated nexus letter is text that resembles a medical opinion. The VA's evaluation framework does not reward resemblance. It rewards:

AI can help veterans prepare personal statements, understand medical terminology, and organize their thoughts. Those are legitimate uses. But there is no data in 295,756 BVA decisions that AI-generated letters, submitted without genuine physician review and signature, improve claim outcomes.

The quality bar is going up, not down. As IMOs become more influential at the BVA, the gap between a well-reasoned physician opinion and an AI-generated template will only widen.

Methodology: We downloaded and parsed 295,756 Board of Veterans Appeals decisions issued between January 2021 and April 2025 from the VA's public decision archive (va.gov/vetapp/). Each decision was analyzed for outcome classification, presence of private medical opinions, probative weight assigned, denial reasons, conditions at issue, claim type, and other structured fields. All decisions are public records. No veteran identifying information was used in this analysis. Data updated April 2026.

Start Your Nexus Letter

$50 record review. $500 only if we can support your claim. Every letter physician-reviewed and signed.