The standard-of-care record review process, step by step
Answer in brief: a standard-of-care review runs in 9 steps: scope and conflicts, written retention terms, a full record inventory, a page-cited chronology, the standard as it stood on the date of care, a decision-by-decision comparison, a separate causation test, a hindsight check, and a write-up that cites the page behind every fact and names the records you did not see.
Limits: Prices, turnarounds and claims come from each provider's own pages or the cited survey and statute, read 2026-11-14. We bought no review from anyone listed and cannot vouch for the quality of any reviewer, and nothing here is legal advice on who may testify in your state.
This page is for the physician doing the review. Hospital teams commissioning one will find the same steps useful as a checklist for what to ask of an outside reviewer. Each step lists what to do, why, and the most common way it goes wrong.
Intake: steps 1 to 3
1. Scope the questions and check for conflicts
Write the questions down in the requester's words before you open the file: which providers, which dates, standard of care only or causation too, and what form the answer takes (a phone screen, a letter, an affidavit, a full report). Then check fit. The AMA's Opinion 9.7.1 asks physician experts to testify only in areas where they have appropriate training and recent, substantive experience. Many state statutes go further and make it a legal requirement, covered on the who reviews page.
Where it goes wrong: a reviewer accepts a case at the edge of their practice because the specialty label matches. A hospitalist reviewing an ICU decision, or a general surgeon reviewing a bariatric revision, invites a qualification challenge later.
2. Agree retention terms in writing
Rate, retainer, what time is billable (reading, research, calls, writing, travel), payment terms, cancellation, and who holds the work product. In the SEAK 2024 survey of 1,633 experts, 64% required retaining counsel to sign a written retention agreement, and 29% reported at least 1 instance where a fee owed to them was never collected. The AMA opinion rules out fees that depend on the outcome.
Where it goes wrong: starting work on a verbal agreement, then finding that the requester expected a flat fee or a lien.
3. Collect and inventory the record
List every source (each facility, each office, imaging, pharmacy, ambulance), the date range each covers, and the page count. Then list what the record refers to but the file lacks. Florida's expert statute, section 766.102(5), requires an expert on the standard of care to conduct "a complete review of the pertinent medical records". Records that commonly go missing in a malpractice file:
- Medication administration records and nursing flowsheets, which hold the times that decide many cases.
- Fetal monitoring strips, telemetry strips and anesthesia records.
- The imaging itself, as opposed to the radiology report.
- Orders with timestamps, including cancelled and verbal orders.
- The EHR audit trail, which shows when each note was written and changed.
- Institutional policies in force on the date of care.
Where it goes wrong: reviewing the curated set the requester sent. Ask for the full record and write down what you asked for and when.
Analysis: steps 4 to 6
4. Build a page-cited chronology
Every order, result, note, vital sign, medication and communication that bears on the questions, in time order, each with the page it came from. Times matter more than dates in most acute care cases; record whether a time is when the event happened or when someone charted it. Mark late entries and notes that were copied forward. The chronology is the factual base of the opinion, so every line has to survive a check against the page.
Where it goes wrong: summarizing instead of citing. A chronology without pages cannot be checked by you a month later or by opposing counsel at deposition.
5. Fix the standard as it stood on the date of care
Identify the decision points, then for each one find what a reasonable physician in that specialty and setting would have done at that time: specialty society guidelines and their publication dates, drug labels, the hospital's own policies, and what was taught and practiced then. The AMA opinion asks experts to consider "standards that prevailed at the time the event under review occurred". A guideline updated after the event describes today's standard, not the one the defendant worked under.
Where it goes wrong: quoting the current version of a guideline. Check the edition and date of every source you cite.
6. Compare care to the standard, decision by decision
For each decision point, write 4 things: what the physician knew then (with pages), what the standard required, what was done (with pages), and whether that was a departure. Allow a third answer, "can't tell from the record", and say what record would settle it. The worksheet lays this out as a table.
Where it goes wrong: judging the whole course of care at once. A global impression is where outcome bias does its work.
Opinion: steps 7 to 9
7. Test causation separately
A departure from the standard and harm caused by that departure are 2 findings. For each departure, ask what would most likely have happened with care that met the standard, and whether the record supports that to the degree of certainty your jurisdiction requires. Some departures change nothing; some outcomes would have happened anyway. Say so.
8. Run a hindsight check
The evidence on this is clear. In Caplan and colleagues' 1991 study, practicing anesthesiologists judged identical care less appropriate when the outcome was permanent injury than when it was temporary, by 31 percentage points on average. A practical check: for each departure you found, re-read your reasoning as if the patient had recovered fully. If the finding only stands because of the outcome, it is not a standard-of-care finding.
When the stakes justify it, a second reviewer helps. Goldman's 1992 review found agreement between peer reviewers mostly poor and listed multiple reviewers and more structured assessment among the fixes. Hayward and Hofer measured a reliability of 0.34 for 2 reviewers even with a structured instrument, so a second read is a check, not a guarantee.
9. Write it up with page citations and limits
The report names the questions, the records reviewed with date ranges, the records requested but not received, each opinion with the standard applied and its source, the pages behind every fact, and what new information would change the opinion. Keep opinions inside your specialty. If the requester asked for causation and you can only speak to the standard, say so in the report.
What the deliverable looks like at each stage
| Deliverable | When it is asked for | What it must show | Steps it needs |
|---|---|---|---|
| Verbal or email screen | Before a case is taken on | Whether a departure is plausible and which records would decide it | 1, 3, a quick 4 and 6 |
| Affidavit or certificate of merit | At filing, or within a set period after it; New Jersey allows 60 days after the answer, plus 1 extension of up to 60 | A reasonable probability that care fell outside acceptable standards, signed by a qualified person | 1 to 6 |
| Florida presuit expert opinion | With the notice of intent to sue | A verified written opinion corroborating reasonable grounds for the claim | 1 to 7 |
| Full expert report | Expert disclosure deadlines | Every opinion, its basis, the records reviewed, with pages | All 9 |
| Peer review opinion | Medical staff review, OPPE or FPPE, privilege action | Whether care met the standard, written for the committee, kept within the peer review privilege | All 9, on the hospital's form |
Sources: N.J. Rev. Stat. 2A:53A-27 and Fla. Stat. 766.203, read 2026-11-14. Rules differ by state; check yours with counsel.
How long each step takes
No published study times a standard-of-care review step by step, so we won't invent one. What drives the total is clear: the page count, how much of the record is handwritten or scanned, the number of providers and decision points, and how many records have to be chased. Steps 3 and 4 scale with pages; steps 5 to 9 scale with the number of decision points. Use the calculator with your own pace to estimate hours before you quote.