VA Rating Criteria, Explained

It is a regulation. 38 CFR part 4, titled Schedule for Rating Disabilities, is the text VA rating staff apply to decide what percentage a service-connected condition carries. It is public and free to read at ecfr.gov, and the wording of a criterion is the wording of a criterion. Nobody has a better copy.

Congress set the frame. 38 U.S.C. 1155 directs the Secretary to adopt a schedule of ratings of reductions in earning capacity, based 'as far as practicable, upon the average impairments of earning capacity resulting from such injuries in civil occupations.' The same section says the schedule shall provide 'ten grades of disability and no more,' naming them: 10, 20, 30, 40, 50, 60, 70, 80, 90, and total, 100 percent. There is no 35 percent. There never has been.

Part 4 comes in two halves. Subpart A, sections 4.1 through 4.31, is general policy: how evidence is weighed, what happens when a record is ambiguous, how percentages combine. Subpart B holds the schedules themselves, organized by body system, running from the musculoskeletal system at 4.71a through to dental and oral conditions at 4.150. Mental disorders are at 4.130. Neurological conditions and convulsive disorders are at 4.124a. Respiratory is at 4.97.

Subpart A carries more weight than most readers expect. A criterion in Subpart B is a few lines of text. The rules for reading those lines live in Subpart A, and they change the answer.

Severity. This is the thing most often got backwards about the schedule, and getting it backwards makes the rest of it look arbitrary when it isn't.

The diagnosis decides which criteria apply. The record decides which level within those criteria is met. Two separate questions, answered from two different parts of the file.

Migraine is a clean illustration. A migraine diagnosis is what sends a file to diagnostic code 8100 in 4.124a. What happens after that has almost nothing to do with the diagnosis. The 8100 entry offers four levels, 0, 10, 30 and 50, separated by how often attacks occur and how disabling they are when they happen. A veteran with a migraine diagnosis can sit at any of the four, and the diagnosis alone does not indicate which.

4.1 says the same thing in the language of 1976. The percentages 'represent as far as can practicably be determined the average impairment in earning capacity resulting from such diseases and injuries and their residual conditions in civil occupations.' Earning capacity, measured through function. Not a name for a condition.

The schedule is also not scored like a checklist, and 4.21 says so outright: 'In view of the number of atypical instances it is not expected, especially with the more fully described grades of disabilities, that all cases will show all the findings specified.' What 4.21 does expect in every instance is 'coordination of rating with impairment of function.'

4.27 answers this and is blunter than most people expect. The code numbers 'are arbitrary numbers for the purpose of showing the basis of the evaluation assigned and for statistical analysis in the Department of Veterans Affairs,' and they 'extend from 5000 to a possible 9999.' A diagnostic code is a label recording which criteria were used. It carries no percentage of its own.

The hyphen in a code like 5002-5240 has a specific meaning, and 4.27 gives two different uses for it. The first is for conditions with no listing at all. When an unlisted disease, injury or residual condition has to be rated by analogy, the code is 'built-up': the first two digits come from the part of the schedule most closely identifying the body part or system involved, and the last two digits are '99'. So a code ending in 99 is telling you the condition has no entry of its own.

The second use is for a disease rated on what it left behind. 4.27: 'With diseases, preference is to be given to the number assigned to the disease itself; if the rating is determined on the basis of residual conditions, the number appropriate to the residual condition will be added, preceded by a hyphen.' The regulation supplies its own example. Rheumatoid (atrophic) arthritis rated as ankylosis of the lumbar spine is coded 5002-5240. Left of the hyphen, what the condition is. Right of the hyphen, the criteria used to measure it.

One more line in 4.27 is worth knowing about, because it governs what a rating sheet is allowed to say: 'the diagnostic terminology will be that of the medical examiner, with no attempt to translate the terms into schedule nomenclature.'

Average impairment in earning capacity. That phrase from 4.1 is the whole definition, and it explains percentages that otherwise look strange. A percentage is not a pain score and it is not a measure of how bad a particular week was.

4.1 goes on: 'Generally, the degrees of disability specified are considered adequate to compensate for considerable loss of working time from exacerbations or illnesses proportionate to the severity of the several grades of disability.' Flare-ups are treated as already inside the level, rather than as separate events to be counted up.

4.10 names what is being measured. The basis of evaluation is 'the ability of the body as a whole, or of the psyche, or of a system or organ of the body to function under the ordinary conditions of daily life including employment,' and evaluations are 'based upon lack of usefulness, of these parts or systems, especially in self-support.' 4.10 closes with a sentence that anticipates the obvious objection: 'a person may be too disabled to engage in employment although he or she is up and about and fairly comfortable at home or upon limited activity.'

4.1 also asks for history rather than a snapshot. 'It is thus essential, both in the examination and in the evaluation of disability, that each disability be viewed in relation to its history.' A single examination report is one document in the record, not the record.

Separate conditions are each rated on their own, then combined under the table at 4.25. Combining is math done after the criteria have been applied, and it is a different operation from deciding a level.

4.7 is the section, and it is two sentences long: 'Where there is a question as to which of two evaluations shall be applied, the higher evaluation will be assigned if the disability picture more nearly approximates the criteria required for that rating. Otherwise, the lower rating will be assigned.'

Both sentences are operative. 4.7 is not a rule that resolves upward on request. It applies where a genuine question exists between two adjacent levels, and it turns on which set of criteria the documented picture more nearly approximates. Where the picture sits clearly below, the second sentence is the one that governs.

4.3 handles a different kind of uncertainty. 'When after careful consideration of all procurable and assembled data, a reasonable doubt arises regarding the degree of disability such doubt will be resolved in favor of the claimant.' It cross-references 38 CFR 3.102, which is where the same principle sits for the rest of the claim.

Reasonable doubt is about evidence that is roughly in balance. An empty file does not produce reasonable doubt. It produces no evidence, which is a different situation with a different outcome.

The rule is called pyramiding and it lives at 4.14. Its first sentence: 'The evaluation of the same disability under various diagnoses is to be avoided.'

The unit the rule works on is the manifestation, not the diagnosis. 4.14 closes by prohibiting 'the evaluation of the same manifestation under different diagnoses.' Two diagnoses that produce one symptom get one evaluation of that symptom, not two.

4.14 also flags where overlap is most likely: 'Disability from injuries to the muscles, nerves, and joints of an extremity may overlap to a great extent, so that special rules are included in the appropriate bodily system for their evaluation.' The same section warns against the reverse error, using manifestations that do not come from a service-connected disease or injury to build up a service-connected evaluation.

Genuinely separate manifestations can carry separate evaluations, and sometimes the schedule instructs it. Note (1) to the General Rating Formula for Diseases and Injuries of the Spine at 4.71a: 'Evaluate any associated objective neurologic abnormalities, including, but not limited to, bowel or bladder impairment, separately, under an appropriate diagnostic code.' Radiating nerve symptoms from a back condition are not the back condition being rated a second time.

Yes, and it is two findings rather than one. A 0 percent rating says the condition is service connected, and it says the record does not meet the schedule's lowest paying level for that code. Only the second half is about money.

4.31 is short: 'In every instance where the schedule does not provide a zero percent evaluation for a diagnostic code, a zero percent evaluation shall be assigned when the requirements for a compensable evaluation are not met.' Some codes write a 0 level into themselves anyway. Diagnostic code 8100 does, with a level reading in full 'With less frequent attacks.'

What a 0 percent rating does carry: service connection, which is the finding that is usually harder to establish and does not have to be established again. What it does not carry: any contribution to the combined figure computed under 4.25, because combining a 0 changes nothing.

Conditions are also not fixed at the level first assigned. 4.1 anticipates change in both directions: 'Over a period of many years, a veteran's disability claim may require reratings in accordance with changes in laws, medical knowledge and his or her physical or mental condition.'

It sets levels. It does not decide whether a condition is service connected at all. That question is answered under 38 CFR part 3, a different regulation with different rules, and it is settled before part 4 is opened.

The schedule also cannot forecast a result. Two files carrying the same diagnosis and the same diagnostic code can land on different percentages, because the criteria measure what a record documents about function, and records differ. Anyone quoting a percentage from a diagnosis alone is describing something the regulation does not contain.

An accredited Veterans Service Organization representative will sit with a veteran, read a rating decision against the criteria it cites, and file a claim, at no charge. Accreditation is a status VA grants and publishes, so it can be checked before anyone is trusted with a file.