Prompt library
The medical record review prompt library
The medical record review prompt library is 71 free prompts for reading medical records with AI. Twenty-two shared extraction prompts cover the chronology, the gaps and the causation statements. Seven role packs then reframe them for plaintiff firms, defense firms, legal nurse consultants, IME physicians, experts, life care planners and adjusters.
How it is organised
Seven professions read the same medical records for different reasons. The extraction work underneath is largely shared; the framing, the priorities and the professional exposure are not. So this library has one spine of 22 role-neutral prompts and seven lane packs that select from it, reframe toward the document that lane produces, and add prompts that exist only for that role. 71 prompts in total.
Read the method once. Run the spine prompts your lane marks as priority. Then run your lane pack. If you work both sides of a file, or move between roles, read the other lanes too: knowing what the opposing analyst is asking the record is worth more than any single prompt here.
Before you run anything
Two rules carry most of the value, and both are in the method. First, force every answer into one of three states: record-supported with a citation, record-silent, or inferred. Most weak output collapses all three into one confident voice. Second, absence is a finding. The gap, the unfilled referral, the surgery never scheduled and the maximum medical improvement no provider ever documented are usually worth more than anything the extraction returns, and none of them produce text to extract.
Every prompt below is written to be pasted and run. Placeholders in angle brackets need filling in. Nothing here depends on a particular tool.
Run these on a file you already have
Send one real record set. You get back a cited chronology, and you review it, you revise it, you sign it.