Medical Record Review Services: Three Different Things With One Name

What are medical record review services?

The term covers three distinct services. Clinical review means a nurse or physician offering a medical opinion. AI review means software summarizing records you've already collected. Paralegal-level review means organizing and indexing records, building the chronology, flagging gaps, and identifying what's missing so it can be requested. Most firms need more than one.

The three get conflated because they all involve reading medical records. They differ in who does it, what it costs, and what question it answers.

What's the difference between clinical review and paralegal review?

Clinical review answers medical questions and requires medical credentials. Paralegal review answers organizational questions and requires legal training and persistence. A paralegal can tell you the treatment history has a four-month gap. Only a clinician can tell you whether that gap matters.

The distinction is worth being precise about, because it determines what you're buying and what it should cost.

Clinical review is a nurse consultant, a physician reviewer, or a retained expert reading records and forming an opinion — on causation, standard of care, the reasonableness of treatment, or the extent of injury. It's priced like expert time because it is expert time.

Paralegal-level review is receiving the production, checking it for completeness, indexing by provider and date, building the chronology, flagging what's clinically interesting for someone else to assess, and identifying what hasn't arrived.

Firms lose money in both directions. Sending everything to a clinical reviewer means paying expert rates for indexing. Expecting a paralegal to opine on causation means asking for something they shouldn't give and you shouldn't rely on.

The division that works: paralegal builds the chronology and flags the questions, attorney decides which questions need a clinical answer, clinician answers those.

Can AI do medical chronologies now?

Increasingly yes, for summarization — and that is a real capability worth using. What AI does not do is get the records in the first place, notice that a production is incomplete, or chase a provider who has ignored four requests. The tools are good at the part that was never the bottleneck.

This deserves an honest answer rather than a defensive one.

A number of well-funded companies now produce chronologies from uploaded records, and the output has improved substantially. If you have a complete production and want it summarized, that is a job software does quickly and will keep doing better.

The constraint was never summarization. Ask any personal injury paralegal where the hours go and the answer is retrieval and follow-up — requesting, waiting, requesting again, discovering the production is partial, and starting over with a records department that has outsourced its function to a vendor who doesn't answer email.

A tool summarizes what you feed it. It cannot know that the client mentioned a physical therapist in the intake call whose records nobody requested. It cannot tell that a fourteen-page production from a facility the client attended for six months is obviously incomplete. It cannot call.

The sensible position is both. Use the tools for what they're good at. Staff the part they can't touch.

What does paralegal-level medical record review include?

Records and billing requests, follow-up and escalation, completeness checking, indexing by provider and date, chronology construction, flagging treatment gaps and notes bearing on causation, and identifying missing records for further request. Retrieval and follow-up consume more hours than everything else combined.

Requesting. Preparing and submitting requests with the right authorizations to every provider, including the ones the client mentioned once.

Following up. The largest single time cost. Providers are slow, outsource their records functions, lose requests, and produce partially.

Completeness checking. Noticing that a production skips a date range, that pages are missing, that a referenced imaging study isn't included. This is the step most often skipped and the most expensive to skip.

Indexing. Organizing by provider and date so the file is navigable.

Chronology construction. Turning records into a narrative — treatment dates, findings, gaps, referrals, and what the notes say. Done well it becomes the spine of the demand.

Flagging. Marking what an attorney or clinician should look at. Not assessing it.

What can't a paralegal do with medical records?

Offer any clinical opinion — on causation, standard of care, reasonableness of treatment, prognosis, or whether the records support a claimed injury. Those are medical judgments requiring medical credentials, and no amount of experience reading records changes that. A paralegal who has seen four hundred files still cannot supply one.

The line is clear in principle and gets tested in practice, usually by a well-meaning question.

An attorney asking "does this gap hurt us?" is asking a clinical question. A paralegal who has read four hundred files may have an instinct about it, and that instinct is worth flagging — but flagging is not opining, and the distinction matters if the file is ever examined.

A paralegal can say: there is a four-month gap between the emergency visit and the first orthopedic appointment, the client's intake notes mention a prior injury to the same shoulder, and the imaging report references a comparison study we don't have.

A paralegal cannot say: whether any of that damages the case.

How do you know which service you need?

Start by asking whether your problem is understanding the records or getting them in the first place. If files are stalling before anyone can analyze anything, the constraint is retrieval — and neither clinical review nor AI summarization does anything about that particular bottleneck.

Two questions that usually settle it:

How many of your open files are waiting on records right now? If it's a meaningful share of the caseload, you have a retrieval problem, and buying analysis capacity solves nothing.

When a production arrives, does anyone check it for completeness before it's filed away? If the answer is "not really," you'll discover the gaps later — usually when someone is building the demand and the timeline doesn't work.

What do medical record review services cost?

Amata publishes $90 an hour for paralegal support and $70 an hour for administrative support, billed in sixth-of-an-hour increments, or $2,520 a month for forty paralegal hours. Clinical review and AI tools are priced separately and differently, because they're different services.

Comparing the three on price is misleading, but the ranges are worth knowing.

Clinical review is priced like expert time and varies widely by credential and scope.

AI summarization is typically per-page or subscription, and it is inexpensive relative to human time — which is exactly why it should be used for what it does well.

Paralegal-level work is hourly, and the honest way to evaluate it is against what stalled files cost you. In contingency practice that's settlement timing, lien interest accruing, and capital tied up — the arithmetic is in the cost breakdown.

One thing worth noting: the cheapest possible hour is a poor way to buy records work. Completeness checking is the step that prevents an expensive surprise, and it's the first thing a rushed or under-trained person skips.

Frequently asked questions

The part that has no category

What strikes me about this market is that two of the three services have a name and a marketing category, and the third doesn't.

Clinical review is a recognized profession with credentials. AI summarization is a funded software category with a dozen companies competing. The work of getting complete records — calling, waiting, calling again, noticing what didn't arrive — is nobody's product. It shows up as an assumption in everyone else's pitch: upload your records and we'll handle the rest.

Somebody has to get them first.

That work is unglamorous enough that firms tend to leave it to whoever has an hour, which means it happens inconsistently, which is why files sit. It's also the step where a mistake compounds — a production nobody checked becomes a demand built on an incomplete record.

The firms that fix it don't usually do anything clever. They just make it somebody's actual job.

Questions about any of this? 312.924.0200 or [[email protected]](mailto:[email protected]).