How to Build a Medical Chronology From a 400-Page Claim File

A medical chronology turns a 400-page injury file into a date-ordered account of treatment, with every entry traceable to the page it came from. This is the seven-step process, and the evidentiary standard the finished document has to meet.

‍

A chronology is an evidentiary document, not a reading aid

‍

A medical chronology is a date-ordered account of every clinically and legally relevant event in a claimant's treatment history. Each entry names the date, the provider, what happened, and the page it came from.

‍

That last field is the one that matters. A chronology without page citations is a set of notes. A chronology with page citations is a summary a party can stand behind.

‍

Federal practice makes the distinction explicit. Under Federal Rule of Evidence 1006, a court may admit a summary that proves the content of voluminous records. The rule reaches records "that cannot be conveniently examined in court." The proponent must make the underlying originals available to the other side. A 2024 amendment took effect on December 1, 2024. It confirmed that such a summary comes in as substantive evidence. A court may not tell the jury it is not evidence.

‍

The records themselves usually enter through Rule 803(6), the business-records exception. That rule asks three things of a record. It was made at or near the time by someone with knowledge. It was kept in the course of a regularly conducted activity. And making it was a regular practice of that activity.

‍

So the standard for a defensible chronology is set by what the records have to survive, not by what is convenient to write. Build to that standard from the first row.

‍

Before the first page: confirm what you are holding

‍

A 400-page production is not the same thing as the complete record. Two checks come first.

‍

Check the record set for holes

‍

The HIPAA right of access covers protected health information held in a designated record set. A covered entity "must act on a request for access no later than 30 days after receipt of the request." It may take one extension, of no more than 30 days (45 CFR 164.524).

‍

That timeline tells you something practical. If a demand package arrived with records from four providers, and the treatment narrative points to a fifth, the fifth set is obtainable. Note the absence before you start reading, not after.

‍

Build a provider list from three places: the intake or emergency department record, the referral lines inside each note, and the billing. Billing catches providers whose notes were never produced.

‍

Fix the pagination before you fix anything else

‍

Records arrive in reverse chronological order, in provider-by-provider stacks, or in whatever order the copy service scanned them. If the file has Bates numbers, every entry cites the Bates number. If it does not, apply a stable page numbering scheme and never renumber it.

‍

Renumbering mid-build destroys every citation already written. This is the single most common way a chronology becomes unusable.

‍

Step 1: Inventory the file before you read it

‍

Do not start on page one. Spend the first pass building a map.

‍

Log six fields for each document. Start page, end page, provider, facility, document type, and the date range it covers. A 400-page file often resolves to a few dozen documents rather than hundreds.

‍

The inventory gives you three things immediately. You see which providers dominate the file. You see where the dates jump. You see which documents are duplicates. Duplicates are common. The same emergency department record often arrives three times: from the hospital, from plaintiff's counsel, and from the treating orthopedist. Separating documents inside a large PDF and removing duplicates is also the first job of claim document intake software.

‍

Mark the duplicates but do not delete them. A document produced three times with three different page counts is sometimes three different versions.

‍

Step 2: Build one date-ordered spine

‍

Now convert the document map into an event list, sorted by the date the event occurred, not the date the document was created.

‍

One row per encounter

‍

The unit is the encounter, not the page and not the document. A ten-page hospital admission is one row with a page range, unless something inside it matters on its own. A single-page office note is one row.

‍

Splitting by page produces a 400-row table nobody reads. Splitting by encounter produces a table that tracks the actual course of care.

‍

Handle dates that do not cooperate

‍

Three date problems recur:

‍

• Dictated later. A note dated two weeks after the visit describes the visit. Sort by the visit date and record the dictation date in a separate column.
• Date ranges. Physical therapy records often cover six weeks in one document. Enter the first and last session as rows, and treat the block as a single course.
• No date. Undated pages get placed by context and flagged as placed by inference. Never place an undated page silently.

Step 3: Write entries the other side cannot attack

‍

The description column is where chronologies go wrong. Two disciplines keep it clean.

‍

Use the record's language, not your own

‍

Write what the record says. "Patient reports pain 8/10 radiating to left leg" is an entry. "Patient claims severe pain" is an argument, and it invites a fight about the summarizer's neutrality instead of about the treatment.

‍

Plain-language explanation belongs in a separate bracket or column. Translate the terminology, then leave the quotation intact.

‍

Separate what was said from what was found

‍

Every clinical note contains at least two categories of information with very different evidentiary weight.

‍

The history of present illness records what the claimant told the clinician. Rule 803(4) admits a statement made for, and reasonably pertinent to, medical diagnosis or treatment. The statement may describe medical history, symptoms, their inception, or their general cause. The mechanism description sits here, and it is admissible for that reason.

‍

The examination findings, imaging results, and diagnoses are the clinician's own observations. They carry different weight and often conflict with the history.

‍

Keep them in separate fields. Consider two entries: "patient reported the fall caused the knee injury" and "MRI shows chronic tricompartmental degeneration." Merged into one narrative sentence, they erase the most important tension in the file.

‍

Step 4: Code the diagnoses consistently

‍

Diagnosis codes give you a way to compare providers who use different words for the same thing, and to spot the moment a diagnosis changes.

‍

The ICD-10-CM Official Guidelines for Coding and Reporting, FY 2026 govern diagnosis coding from October 1, 2025 through September 30, 2026. CMS and NCHS publish them, and the four Cooperating Parties approve them. Adherence is required under HIPAA in all healthcare settings. So the codes on the claimant's own records were assigned under a published rule set.

‍

Three coding facts change how a chronology reads:

‍

• Injury codes carry a seventh character marking the encounter as initial, subsequent, or sequela. A run of codes that never leaves "initial encounter" tells you about the biller's habits, not about the patient.
• Personal history codes record a condition that is no longer present but that may affect care. Their appearance and disappearance across providers is worth a note.
• Signs and symptoms are coded differently from confirmed diagnoses. A file that codes symptoms for eight months and a definitive diagnosis in month nine has a diagnostic history, not a stable one.

Record the codes as billed. Do not correct them. A coding error is itself a finding.

‍

Step 5: Run the gap pass and the contradiction pass separately

‍

These are two different reads, and combining them means doing neither well.

‍

The gap pass looks only at the date column. Measure the interval from injury to first treatment, then every interval between consecutive encounters. Flag the intervals, and record what the next note says about the interval, because clinicians often explain them.

‍

The contradiction pass looks only across providers on overlapping dates. The two passes differ in every respect:

‍

• What you read. Gap pass: The date column only. Contradiction pass: Descriptions across providers.
• What you compare. Gap pass: Consecutive encounters. Contradiction pass: Overlapping encounters.
• Typical finding. Gap pass: An unexplained interval. Contradiction pass: Two records that disagree.
• What you record. Gap pass: Interval length, plus any explanation in the next note. Contradiction pass: Both page cites, side by side.

The contradictions that recur in an auto liability claim file:

‍

• Mechanism of injury described differently to the emergency department and to the specialist.
• Body parts that appear for the first time weeks into treatment.
• Work or activity restrictions in one provider's note and full activity in another's on the same week.
• Prior injury denied in one intake form and documented in another provider's history.

Both passes produce a flag list, not a conclusion. The chronology records that two documents disagree and cites both pages. Someone else decides what the disagreement means.

‍

Step 6: Put billing next to treatment

‍

Run a parallel column for charges. For each encounter, capture three things. The amount billed. The amount paid or adjusted, where visible. And whether a letter of protection or lien appears anywhere in the production.

‍

The treatment column and the billing column answer different questions. Treatment tells you what care the record supports. Billing tells you what the demand is built on. They diverge more often than the demand narrative admits.

‍

Billing sometimes exists for a provider whose notes are missing. That is a records gap with a dollar figure attached. It goes straight back to step one.

‍

Step 7: Make the finished document checkable

‍

Before the chronology leaves your desk, three things must be true.

‍

Every factual entry cites a page. Anything you inferred is labeled as inference. Every flag names both documents it compares.

‍

A reviewer should be able to pick any row at random, open the cited page, and see the row's source in under a minute. That test is the whole quality standard. If it fails on a row, the row is not finished.

‍

Keep an open-items list beside the chronology: missing providers, illegible pages, undated documents, records requested and not yet received. The list travels with the document. An acknowledged gap is a work product; an unacknowledged gap is an error waiting for the other side to find.

‍

Working a 400-page file when the clock is short

‍

The full build above is the right process. Some files get an hour.

‍

In that case, run steps 1, 2, and 5 only. Inventory the documents, build the encounter spine with page cites, and run the gap pass. That gives you the shape of the treatment course, the date intervals, and a citation trail.

‍

Skip the billing column and the coding pass on the first hour. Neither changes the shape of the file, and both are additive later. What you cannot skip is the page citation, because retrofitting citations means reading the file twice.

‍

Software changes the arithmetic here rather than the method. Some tools read the unstructured documents in an injury file directly. AI-assisted bodily injury claim file review handles the sorting, deduplication and coding passes mechanically. The judgment steps do not move. Someone still has to decide what a contradiction means. Someone still has to open the cited page. In workers' compensation claim handling, consistency across providers carries the causation question. That verification step is the work.

‍

Where chronologies fail

‍

Five failure modes account for most unusable chronologies.

‍

The document was renumbered after citations were written. The description column drifted from quotation into characterization. Gaps were noticed but never written down, so the next reader repeated the work. Duplicates were merged, which hid the fact that two versions of one record exist. And the open-items list was left out, so an incomplete chronology was handed over as a complete one.

‍

Each of those is a process failure, not a reading failure. Fix the process once and the same 400-page file becomes a repeatable job.

‍

Key takeaways

‍

• A chronology is an evidentiary summary, not a reading aid. Under Federal Rule of Evidence 1006, a summary of voluminous records is substantive evidence, and the underlying originals must be available to the other side.
• Fix pagination before writing a single entry. Renumbering mid-build destroys every citation already written, and that is the most common way a chronology becomes unusable.
• Inventory the file before reading it. Log start page, end page, provider, facility, document type and date range for each document, then build the spine one row per encounter, not per page.
• Run the gap pass and the contradiction pass separately. The gap pass reads dates only; the contradiction pass compares providers on overlapping dates. Both produce flags with page cites, not conclusions.
• The quality test is one sentence: a reviewer should be able to pick any row at random, open the cited page, and see the source in under a minute.

‍