Andrews v. McDonough — reversal instead of remand

34 Vet. App. 216 (2021), argued December 9, 2020, decided June 22, 2021. Court of Appeals for Veterans Claims, No. 19-0352, Judge Toth writing, before Judges Allen, Toth and Falvey. Precedential. Board denial of service connection reversed, and remanded only to assign a rating and effective date.

There are two precedential decisions called Andrews v. McDonough from 2021. This is the one at 34 Vet. App. 216, about remedies. The other, at 34 Vet. App. 151, is about the evidentiary record in the Direct Review docket. Different veteran, different question, seven weeks apart.

If you have read Bufkin and come away thinking a court will never simply grant a claim, this is the counterexample — and the reason it is rare.

Why this matters even though it is about hepatitis C

Both sides agreed the Board had relied on an inadequate medical examination. The whole dispute was about what the court should do about it. VA asked for a remand to get yet another opinion. Mr. Andrews asked the court to grant the claim.

That is the position a lot of people end up in: you win the point and VA gets another try, which can end in another denial. This decision is the narrow case where the court says no.

The default really is remand

The court was candid that its reluctance to reverse is a feature of appellate review and not timidity. The standards:

• Remand where the Board "has incorrectly applied the law" or "failed to provide an adequate statement of reasons or bases for its determinations, or where the record is otherwise inadequate" (Tucker v. West).
• Reversal where the court "is left with the definite and firm conviction that," on the entirety of the evidence, the Board made a mistake (Gilbert v. Derwinski, borrowing the standard from U.S. Gypsum).
• But "[w]hen there are facts that remain to be found in the first instance, a remand is the proper course," absent rare circumstances — such as the government conceding the relevant facts (Byron v. Shinseki).

And the court refused the broad rule Mr. Andrews argued for. He wanted reversal whenever the lay and medical evidence together reach "approximate balance," even with substantial uncertainty remaining. The court "part[ed] ways" with him. So this page is not authority for a general right to reversal on close evidence; the opinion is explicit that it is agreeing "for a narrower reason than he advances."

The exception that decided it

From Adams v. Principi, 256 F.3d 1318 (Fed. Cir. 2001) — not the Adams on the obesity question, which is a different case entirely:

"Remand is inappropriate where the predominant purpose is not to allow the Board to make a fully informed decision unencumbered by error but to allow VA to obtain more evidence so that it can properly deny the claim."

And the test the court drew from it, which is the usable sentence on this page:

"For a reviewing court, the key factor is not the weight of the evidence already of record but whether the remand serves a purpose beyond allowing VA to obtain evidence to deny a claim."

Why that fit this record

The reasoning runs through VA's own adjudication manual, the M21-1. Three steps:

One — the manual is not binding, until the Board adopts it. The M21-1 does not bind the Board or the court, and the Board cannot simply cite a provision without analysis. But when the Board cited the hepatitis C risk-factor provision in its own 2017 remand order and developed the claim under it, it made that provision the standard by which the claim would be judged. The Federal Circuit's phrasing in an earlier, nonprecedential case is the one the court found persuasive: "It cannot be that the VA may tell a veteran how to establish a service connection for his [condition] only to move the goalposts once he has done so."

Two — that provision contains a tiebreaker. The manual lists risk factors for hepatitis C — blood transfusions, tattoos and piercings, puncture with non-sterile needles, high-risk sexual activity, and air gun injections, the last with caveats — and then instructs adjudicators to "[r]esolve reasonable doubt … in favor of the Veteran when the evidence favoring risk factor(s) in service is equal to the evidence favoring risk factor(s) before or after service."

Three — on this record, there was nothing on the other side. The Board never disputed that the in-service risk factors occurred; it only discounted how likely they were to have caused the condition, and the opinions it used to do that were conceded inadequate. It never questioned his statement that he had no risk factors outside service, which the court took as an implicit finding that he was credible. So:

"Regardless of the relative probative value of the evidence supporting Mr. Andrews in-service risk factors, this much is clear: it at least matches the zero evidence suggestive of pre- or postservice risk factors."

Nine years of development, two VA examinations, numerous private records — and VA had never identified a single non-service risk factor. Another examiner "would have no warrant to reject the fact that no non-service risk factors exist," so a remand would have served no purpose except to look for one.

The four circumstances the court named, together, are the holding:

"the adoption of the M21-1 standard, the conceded presence of in-service risk factors, the absence of any non-service risk factors, and the lack of any reasonable basis to believe that additional development could alter the approximate balance of evidence."

What to take from it

It is not a hepatitis C rule and it is not a general rule. All four circumstances were present. The most transferable piece is the first one: if the Board told VA to develop your claim under a particular manual provision, that provision is the standard, and the Board cannot later decide your evidence was aimed at the wrong target. Keep the remand order — the same advice Stegall earns for a different reason.

The second transferable piece is about the shape of a record rather than its weight. What made this one reversible was not that the favourable evidence was strong. It was that there was nothing at all on the other side, after years in which VA had every chance to find some.

How it sits with Bufkin

They do not conflict, and reading them together is more useful than reading either alone.

Bufkin (Supreme Court, 2025) holds that VA's decision about whether the evidence was in approximate balance is predominantly factual, so the Veterans Court reviews it only for clear error and will not reweigh the record. Andrews is a case where the court found clear error — and said so expressly — because an adopted standard had a tiebreaker in it and one side of the scale was empty.

The practical reading is the same one on the court page: the place to win on close evidence is before the Board, not after it. Andrews is what it takes for the exception to apply.

A limit worth knowing

A Veterans Court memorandum decision the following year distinguished Andrews on exactly the three things that were missing from its own record: the Board had not mentioned the manual at all, had found the veteran's statements not competent rather than not probative, and there was only one medical opinion with its adequacy still in dispute. That decision is not precedent and is not cited here as authority — see the case law hub — but it shows how narrow the gate is.

Finding the opinion

Published at 34 Vet. App. 216. The official PDF is Andrews v. McDonough, No. 19-0352 — the AndrewsDA in the filename distinguishes it from the other 2021 Andrews.

The holding was read against the full official opinion, and a 2022 Veterans Court decision in the same source batch applies it as live law. No commercial citator was run; see sources and corrections.

Reviewed October 6, 2026