How to read a sample chronology
Answer in brief: read a sample chronology in 5 passes: the header and the question it serves, the records list, the flags, 3 citations you open yourself, then the columns against the case type. A sample tells you about a format and a set of habits; it does not tell you how a vendor or a reviewer will do on your worst file.
Limits: Every sample here is synthetic: no real patient, invented names and invented page numbers. Rules quoted for QME and Social Security work come from the cited regulation, read 2026-11-12; they are not legal advice, and facts about other sample libraries come from each publisher's page on that date.
What a sample is for, and what it hides
You meet samples in 3 places: a vendor or service sends one before you buy, a colleague hands you one as a model, or you open one on a site like this to set your own format. In all 3 cases the sample was picked. A service shows its cleanest case. A colleague shows the one they are proudest of. Our samples are synthetic and short on purpose.
So read a sample for 3 things only: the columns, the way rows are written, and the habits it shows on hard records (conflicts, gaps, late entries, duplicates). Do not read it as evidence of how fast or how well the work gets done on a real file. For that, the only test is your own file.
Read in this order: 5 passes
- Header and question. Does the sample say what it was built for? A chronology for a referral question reads differently from one built for a demand letter. If the header has no case type, no date range and no referral question, the rows below were chosen by someone guessing.
- Records list. Look for a list of every source with its page range and date range. Then look for what was set aside: duplicates, non-medical pages, records from another patient. A sample with no set-aside list either had a perfect file or did not look.
- Flags only. Skim the flag or note column and nothing else. Count the conflicts, gaps, late entries and missing records. A sample with 0 flags on a file of any size is a warning sign, not a good sign.
- Open 3 citations. Pick 3 rows, ideally the ones that matter most to the question, and check that the page holds what the row says. On a vendor sample, ask for the source pages. If they cannot send them, the citations cannot be tested.
- Columns against the case type. Compare the columns with the sample for the same case type on this site. A workers comp chronology with no work status, or a med mal chronology with dates but no times, is missing what its reader needs.
The note key used on this site
The samples use a short, fixed set of tags in the note column. Each tag has 1 meaning, so you can count them. If you build your own format, a fixed list beats free text: "see note" tells a reader nothing.
| Tag | Meaning in the samples | What the reader does with it |
|---|---|---|
| Prior | The event predates the injury or claim and involves the same body part or condition | Weighs it for causation, aggravation or apportionment |
| Conflict | 2 records disagree on a fact; both rows stay, side by side | Decides which to credit, and says why in the report |
| Gap | No care for a stated number of days between 2 dated rows | Looks for the explanation in the next note |
| Late entry | The note was signed or entered well after the event it describes; both times shown | Weighs reliability, may request the EHR audit trail |
| Verbatim | Quote marks: the words are the record's own | Can read the line back exactly, at deposition if needed |
| Series | A run of similar visits grouped into 1 row with first and last values | Opens the range only if the trend matters |
| Not produced | A record is referenced but not in the file | Requests it, or notes the limit in the report |
| Opinion | The row is a clinician's stated opinion, not a finding | Weighs it against the findings that support it |
| Set aside | A duplicate or non-medical page left out of the rows and logged | Checks the log if a count or a page is disputed |
Score a sample: 12 checks
Use this on any sample a vendor, a service or a colleague sends you. Tick what the sample shows. It runs in your browser; nothing is sent anywhere.
The score bands are our judgment, not a standard. The last check carries the most weight: a sample whose citations you cannot test is a picture of a chronology, not evidence of one.
Reading a sample against your own case type
Open the sample on this site that matches your work and hold it next to the one you were sent.
- IME: does each row map to a referral question?
- QME: does it show who excerpted the records, and does the records list end in a page count?
- Disability: can you see function across at least 12 months, and is each fact marked as exam, test or report?
- Workers comp: is work status a column, and is every authorization request matched to a decision?
- Med mal: are there clock times, roles and signing times?
- Life care plan: does every future care item point to a recommending note?
Questions about reading samples
Should I trust a vendor's sample chronology?
Trust it as a picture of the format. Do not treat it as evidence of quality on your files. Run a trial on 1 of your own cases and open the citations yourself.
What is the fastest check on a sample?
Open 3 cited pages and confirm each row. It takes a few minutes and tells you more than reading the whole table.
Why does a sample with no flags worry you?
Real files almost always contain a conflict, a gap or a duplicate. A sample with none either came from an unusually clean file or from a process that does not look for them.