A litigation-ready medical chronology maps every treatment event, provider, finding, and gap in chronological order. It forms the spine of the demand letter's injury narrative and is the single most important document, other than the records themselves, for deposition preparation. Building one from scratch, reading through hundreds of pages of records and organizing entries by hand, takes four to eight hours on a typical personal injury file. With a disciplined input protocol, Claude can help you produce a structured, attorney-reviewed chronology in under an hour. This post is that protocol.
Why the Chronology Is the Center of the File
Defense counsel looks for two things first: a weakness in liability and an angle to attack damages. On the damages side, your medical chronology is their road map. It documents the treating providers, the progression of documented injuries, the treatment ordered, and the gaps that a defense expert will exploit if you do not surface and frame them first. The attorney who controls the chronology controls the damages narrative. That fight is won in case preparation, not in front of a jury.
The traditional process is slow and expensive. A paralegal reads each record set, builds a table, flags problems, and sends a draft to the attorney for review. The attorney makes corrections. Someone formats it for the demand. AI does not replace the attorney's judgment in reading the records, but it absorbs the non-billable clerical work of formatting, sorting, and initial pattern-spotting that consumes the bulk of that four-to-eight-hour window.
Step 1: The Input Format: The Treatment Event List
Never paste raw OCR'd medical records into a public AI tool. Under HIPAA, covered entities cannot transmit protected health information (PHI) to a third-party vendor without a signed Business Associate Agreement (BAA). Even with a BAA in place, work-product privilege concerns add a second layer of risk. Following this protocol means the raw records never touch the AI directly.
The protocol requires one manual step first. The attorney or paralegal handling records review reads each document and extracts the relevant clinical entries into a structured text file called the Treatment Event List. This is the only document Claude receives. Each entry contains:
- Date: the date of service or record entry
- Provider: name and specialty
- Facility type: emergency department, outpatient clinic, imaging center, etc.
- Visit type: initial evaluation, follow-up, imaging, surgery, discharge, etc.
- Findings or diagnosis: a quoted or closely paraphrased excerpt from the record
- Treatment ordered: prescriptions, referrals, procedures, restrictions
- Gaps noted: any observations about follow-up not appearing in subsequent records
In most files, this extraction takes 30 to 60 minutes depending on volume. The resulting list is the attorney's work product. What Claude receives is a structured text summary, not the underlying records.
Step 2: The Construction Prompt
Once the Treatment Event List is ready, you prompt Claude to generate the chronology. For most demand letter workflows, start with a table and then request a narrative section separately for the letter's injury paragraph.
PROMPT: You are a legal analyst assisting a personal injury attorney. Below is a Treatment Event List extracted from the client's medical records. Build a litigation-ready medical chronology in table format with the following columns: Date | Provider | Facility Type | Visit Type | Findings/Diagnosis | Treatment Ordered | Notes. Sort entries in strict chronological order. Preserve quoted language from the findings column exactly as written. Flag any entry where the gap before the next treatment event exceeds 30 days by adding [GAP] in the Notes column. [PASTE TREATMENT EVENT LIST HERE]
Claude returns a formatted table. The attorney then checks it against the original Treatment Event List for accuracy before using it in any demand, discovery response, or filing.
Step 3: The Gap Detection Scan
After the chronology is built, run a second prompt for the red-flag scan. This mirrors the review a defense expert will conduct on your records. Running it yourself first lets you address the weaknesses in the demand letter rather than leave them for opposing counsel to highlight in discovery. This defensive analysis is a key part of a sound Claude AI settlement negotiation protocol.
PROMPT: Using the medical chronology above, identify the following: (a) all treatment gaps exceeding 30 days between consecutive entries, including the exact date range and duration; (b) any referrals mentioned in a provider note that do not appear as a subsequent entry in the chronology; (c) any apparent inconsistency between symptoms reported by the patient and findings documented by the provider. Present your findings as a numbered list with the specific entry dates cited for each item.
The output of this scan goes directly into your pre-demand review notes. It is also useful when preparing the client for an independent medical examination.
Step 4: ICD-10 Notation
A useful addition to the chronology for demand letter drafting and preexisting-condition analysis is a column for ICD-10 code ranges. You can prompt Claude to map injury descriptions to approximate code ranges. A qualified coder must verify the codes before they are used for billing, but for litigation prep and demand letter drafting the notation is an effective organizational tool.
PROMPT: Review the findings column of the chronology above. For each distinct injury type documented, identify the applicable ICD-10 code range (e.g., S13 for cervical sprain) and list: the injury description as stated in the chronology, the applicable ICD-10 range and descriptor, and whether any entry in the chronology suggests the condition predated the incident. Format as a table: Injury Type | ICD-10 Range | First Documented Entry | Possible Preexisting Indicator. Note that ICD-10 mapping is for organizational purposes and should be verified by a qualified coder before clinical or billing use.
The preexisting-condition column is the real value here. Defense counsel will argue any prior treatment for a related condition undercuts causation. The chronology should confront that before the demand goes out, not after.
Step 5: Cross-Document Reconciliation
In cases involving multiple treating providers for the same injury, each provider's narrative of the condition can differ. An orthopedic surgeon's account of a lumbar injury may not match a pain management clinic's account, particularly on mechanism of injury or symptom onset. When accounts diverge, the demand letter must reconcile them, not pretend they are not there.
After the full chronology is assembled, prompt Claude to map the multi-provider narrative for any contested condition:
PROMPT: Focus on entries involving [INJURY TYPE, e.g., lumbar disc herniation]. List every provider who documented this condition in the chronology and summarize what each provider recorded about: (1) mechanism of injury or onset; (2) severity and functional limitations; (3) prognosis or treatment plan. Then identify any points where the providers' accounts diverge in material ways. Cite the specific chronology entries for each finding.
The output of this prompt becomes your working notes for the injury narrative section of the demand letter. For how to move from the chronology to the letter itself, see our guides on writing a demand letter with Claude and the personal injury demand letter template.
What This Protocol Does Not Cover
Claude is a text-based AI. It cannot ingest raw PDFs or scanned medical records in bulk without file-by-file upload. It cannot reliably OCR handwritten physician notes. It cannot access a database of comparable claim outcomes to benchmark your damages request against settlement history. Purpose-built tools like EvenUp and CounselorAI are designed for those functions.
This protocol works because it plays to what Claude does well: receiving structured text and producing organized, formatted output under attorney direction. The attorney reads the records. Claude structures the product of that reading. The supervising attorney reviews and verifies before anything goes into a filing or a letter. This workflow avoids the AI citation hallucination risks that arise from unverified AI output in legal documents.
Time Savings in Practice
A straightforward motor vehicle accident file with records from an emergency department, a treating orthopedist, a physical therapy program, and a pain management clinic might involve 150 to 300 pages of records. Manual chronology build: 4 to 6 hours. With this protocol, the Treatment Event List extraction typically takes about 45 minutes, and prompting and reviewing the structured chronology, gap scan, and ICD-10 table takes another 15 minutes. That is a realistic reduction of 3 to 5 hours per file, on work that does not require attorney-level judgment for most of its tasks.
Multiply that across a caseload of 50 active PI files and the math speaks for itself.
If you want to build this capability into your practice workflow, the full protocol is covered in the Leveraged Attorney course, which walks through AI use for demand letters, deposition preparation, and case strategy from intake through resolution.