# VA C-File Review Services — DBQ & 38 CFR > C-file review for VA examination providers, accredited agents and VSOs. Service treatment and private records organised against 38 CFR Part 4 and the DBQ. Canonical page: https://medrecords.ai/solutions/va-c-file-review/ --- [Home](https://medrecords.ai/) ›[Solutions](https://medrecords.ai/solutions/) ›[Veteran Disability](https://medrecords.ai/solutions/veteran-disability/) ›VA C-File Review VA C-FILE REVIEW SERVICES ## Fifteen hundred pages of c-file, read before the examiner opens it. VA C-file review services organise a claims file for contracted examination providers, accredited agents and veterans service officers: service treatment records, private treatment, prior rating decisions and presumptive-condition evidence turned into a cited pre-read keyed to the DBQ and the 38 CFR Part 4 criteria that decide the rating. [Test a file ](https://medrecords.ai/test-a-file/?src=va-c-file-review) [Book a demo](https://medrecords.ai/demo/?src=va-c-file-review) Click any row → the source page it cites C-file pre-read · Case #IME-4812 indexed Adams, Timothy · c-file production 342 pp / 2 packets received logged 11 duplicate pages removed free p.140 — wrong patient quarantined Pages 342 Documents 27 Cited 100% ### The criteria are published. So is the scorecard. We do not build for a case type until we can name the standard that defines a correct output and the person who grades the file against it. In VA disability examinations, both are written down. That is what makes the c-file tractable: the pre-read has a fixed target. Every line we surface exists because a diagnostic code, a DBQ field, or an adequacy rule asks for it. The rulebook and the referee The rulebook - 38 CFR Part 4, the Schedule for Rating Disabilities — each diagnostic code carries explicit criteria - 38 CFR §3.303, §3.304, §3.307 and §3.309, which govern service connection and presumptive conditions - 38 CFR §4.2: an inadequate examination must be returned - Roughly 70 DBQ forms, each with its own required fields - Nieves-Rodriguez v. Peake: an opinion without a reasoned record review has zero probative value The referee - The MDEO 10-question Quality Criteria Checklist, sampled quarterly. FY24 Q3 alone produced roughly 1,200 errors, 6% of them requiring a re-do - Contract SLAs, which impose financial penalties below 92% accuracy and pay incentives above 97% - The VBA rater, who returns an exam that will not support a decision - The Board of Veterans' Appeals and the Court of Appeals for Veterans Claims, on remand Contracted examinations average roughly $1,563 each; an exam returned as inadequate costs the provider the full fee plus an SLA penalty Flat 10¢/page here, duplicates free C-file pre-read · Case #IME-4812 cited 100% Production342 pp / 2 packets · 11 deduplicatedlogged Service treatment recordsTyped and handwritten, page-citedcited Prior rating decisionsIndexed by diagnostic codesorted Scan qualityDegraded pages flagged, not skippeddegraded Gaps in the recordListed as gaps, never filled insurfaced 27 documents · cited 100% · no rating calls What the file actually looks like ### Three hundred pages on a good day. Five thousand on a bad one. A c-file in this work runs from roughly 300 to 5,000 pages, with the typical file near 2,650. It rarely arrives as one clean production. Service treatment records come from one source, private treatment from another, prior rating decisions and correspondence from the VBMS export, and supplemental packets keep landing after the exam is scheduled. What makes it hard is not the length. It is decades-old scans photographed at an angle, handwritten entries from a battalion aid station, the same discharge examination appearing in four packets, and a single line buried at page 1,900 that decides whether a presumptive condition is documented at all. 3.2M Contracted examinations in FY2024 (GAO-25-107483), up from 1.8M in FY2022. ~93% Share of all VA disability examinations performed by contractors, at an FY24 cost above $5 billion. What we'd build with you ### One real file first. Then the build. The c-file pre-read is a co-build, not a shelf product. Examination providers, accredited agents and veterans service officers each index a file differently, and the DBQ set in front of the clinician differs by contention. So we start from your file, not from a template we wrote for someone else. The boundary holds throughout. The platform organises, cites and surfaces. The clinician examines the veteran, completes and signs the DBQ, and forms the opinion. The rater decides the percentage. Nothing here scores a claim or predicts an outcome. Neighbouring case types in this lane: [Camp Lejeune claim review](https://medrecords.ai/solutions/camp-lejeune-claim-review/) and [VCF and WTC claim review](https://medrecords.ai/solutions/vcf-wtc-claim-review/) . 01 #### Send one real file A c-file you have already worked, in whatever state it arrived: flat VBMS export, mixed packets, degraded scans. In the demo case that was 342 pages across 2 packets, with a wrong-patient page quarantined before the read began. 02 #### We return the read Duplicates removed, documents sorted, a chronology built, and the record findings each DBQ section asks about surfaced with a page citation. You check it against the file you already know. 03 #### We scope the build Your output template, your field set, your review workflow. The demo report carries 38 template fields; yours carries whatever your DBQ set and quality checklist require, in your own house format. 04 #### You run it Your clinicians and reviewers work in it. Every access is logged, every line traces to a page, and files are deleted 30 days after delivery. Why Medrecords AI ### The rules the platform never breaks. EVERY LINE CITED CASE #IME-4812 · ADAMS, T.342 pp 2/14 — ER visit, right knee p.4 4/18 — arthroscopic surgery p.61 p.140 — wrong patient quarantined Medrecords AI Read every page · cite every line #### HIPAA, under a signed BAA Every file is handled under our Business Associate Agreement, from the first byte. #### Never trains a model Your records are never used to train any AI model — ours or anyone else's. #### Every line cited If we can't cite it, we don't say it. Every sentence links to its source page. #### Deleted after delivery Files are deleted 30 days after delivery, with a full audit log of every access. Powered by the platform ### Four capabilities behind every c-file pre-read. The c-file pre-read is one deliverable of the same platform that reads, sorts, and cites the whole record. [ #### Medical Chronology In-service complaint through post-service treatment, one continuous timeline built from the c-file and synced to every source page. IN ACTION · 7 visits improving since 4/02, each page-cited ](https://medrecords.ai/product/chronology/) [ #### Document Sorting & Packet Builder A flat VBMS export becomes discrete documents: service treatment, private treatment, prior decisions, correspondence, each in its own section. IN ACTION · 342 pp / 2 packets → 27 documents ](https://medrecords.ai/product/document-sorting-packet-builder/) [ #### Missing Records Identification Referenced-but-absent records are listed as gaps before the examination, not discovered when §4.2 sends the exam back. IN ACTION · gaps listed, never filled in ](https://medrecords.ai/product/missing-records-identification/) [ #### Negative Findings Detection A DBQ field often turns on what the record does not contain. Documented normal findings and explicit denials are surfaced and cited, not left implied. IN ACTION · absence recorded as a citation, not a guess ](https://medrecords.ai/product/negative-findings-detection/) FAQ ### C-file review, answered. A cited pre-read of the claims file. Service treatment records, private treatment, prior rating decisions and presumptive-condition evidence are sorted, deduplicated and organised into a chronology, with every line linked back to the page it came from. The clinician opens a file that is already indexed against the DBQ sections and the 38 CFR Part 4 criteria, instead of paging through a raw production for the findings that matter. No. The DBQ is completed and signed by the examining clinician, and the opinion is theirs alone. We organise and cite the underlying record so the reasoned review behind that opinion is visible on the page. Nieves-Rodriguez v. Peake holds that an opinion without a reasoned record review has zero probative value; making that review traceable is our job, forming the opinion is not. The same way we handle any production. C-files in this work run from roughly 300 to 5,000 pages and often arrive as a single flat export with no document boundaries, mixing typed service treatment records with handwritten notes and degraded scans. The platform splits it into documents, removes duplicate pages at no charge, and rebuilds a navigable index. In the demo case, 342 pages across 2 packets produced 11 duplicate pages removed and one page quarantined. No. Each diagnostic code in the Schedule for Rating Disabilities carries explicit criteria, and applying them is the VBA rater's decision. We surface and cite the findings in the record those criteria refer to, including range-of-motion figures, treatment frequency and documented functional limitation. What percentage those findings support is decided by VA, not by us. No. Medrecords AI does not retrieve records from VA, from providers, or from facilities. You bring the c-file and any private treatment you already hold, and the review starts in minutes from upload. Retrieval vendors take days; you can keep yours and still cut the read down to minutes. ### Send one c-file. We'll tell you what we can read. No obligation. If the case type is buildable we'll scope it; if it isn't, we'll say so. [Test a file ](https://medrecords.ai/test-a-file/?src=va-c-file-review) [Book a demo](https://medrecords.ai/demo/?src=va-c-file-review) [All veteran disability work →](https://medrecords.ai/solutions/veteran-disability/) [See every case type we process →](https://medrecords.ai/solutions/case-types/)