Skip to main content
The Practice
Clinical operationsOctober 4, 2026

Six Note Mistakes That Trigger Denials — and the Corrected Lines

The six documentation mistakes payer reviewers deny first — missing minutes, unsigned notes, cloned text, unskilled language, floating data, lapsed certifications — each paired with the line that fixes it.

Callie Editorial 15 min read
The documentation issue
Denial-proof
Session note
S

Relevant report

Caregiver reports carryover at home

O

Observable change

78% accuracy · minimal verbal cue

A

Clinical meaning

Self-monitoring is emerging

P

Next decision

Progress to conversational retell

At a glance

What you’ll leave with

  • Most denial-linked note mistakes are mechanical — missing minutes, missing signatures, lapsed certifications — and a reviewer checks them before reading a word of clinical reasoning.
  • Every mistake here has a one-line fix. The corrected line is rarely longer than the broken one; it is just specific: the minutes, the baseline, the cue, the decision.
  • Copy-forward is the expensive habit. Medicare contractors treat cloned notes as misrepresentation of medical necessity, which risks denial and recoupment across every visit the pattern touched.

When a therapy claim is denied, the instinct is to re-argue the clinical case: the patient needed the care, the treatment was appropriate, the progress was real. All of that can be true and the denial can still be correct — because the reviewer never saw the session. They saw a note with no treatment minutes, or no signature, or a narrative identical to the previous four visits. The most common denial-linked documentation mistakes are not failures of clinical judgment. They are mechanical habits, each one visible in a single line of the note, and each one fixable in a single line too.

This article walks through six of those habits. For each one: what the broken line looks like, what a reviewer concludes from it, and the corrected line that closes the gap. Then one full daily note gets rewritten end to end, because the mistakes rarely travel alone. If you want the deeper treatment of any single thread, the skilled-care language, audit review order, and timed-code minutes each have a dedicated article; this one is the field guide to the mistakes themselves.

The pattern

The denial usually happens at the desk, not in the session

Payer reviewers work in a predictable order, and it is the opposite of the order therapists care about. Clinical reasoning is read last, if at all. First come the checks that take seconds and require no clinical training: is the note signed and dated, do the documented minutes support the units billed, is there a certified plan of care covering this date of service, does this note say anything the last one did not. A chart can lose on every one of those before anyone evaluates whether the care was good. That is bad news for a clinician who writes thoughtful assessments above an unsigned, minute-less note — and good news operationally, because mechanical mistakes have mechanical fixes. You do not need to write longer notes. You need specific lines present and specific habits gone.

Mistake one

The minutes are on the claim but not in the note

The claim says three units of a timed code. The note lists the interventions, describes the session, and never states a minute count. To the therapist, the minutes feel implied — the session was clearly a full session. To a reviewer, unsupported units are the fastest denial available, because Medicare’s rule is explicit: the treatment note must document the total timed-code minutes and the total treatment time for the visit (Medicare Benefit Policy Manual, Pub. 100-02, Ch. 15, §220.3). No minutes means the units cannot be verified, and a unit that cannot be verified is not paid. The fix costs one line: “Total timed code minutes: 38. Total treatment time: 45 minutes.” Note that these are two different numbers with two different jobs — total treatment time also covers untimed-code work — and the note needs both. You do not need start and stop times per intervention; you need the two totals, every visit.

Mistake two

The note nobody signed — or signed as “J.D.”

Unsigned notes cluster in two places: drafts that were never finalized in the EHR, and entries authenticated with bare initials or an illegible scrawl that never resolves to a credentialed person. Medicare reviewers are instructed to look for signed and dated documentation that identifies the author, and claims whose records fail the signature requirement can be denied on that basis alone (CMS, “Complying with Medicare Signature Requirements,” MLN905364). Worse, this one cannot be fixed after the fact: adding a late signature once a reviewer asks is specifically not permitted — the remedies are a signature log or an attestation statement, both of which are paperwork you produce under audit pressure rather than a workflow you want. The durable fix is procedural: the note is signed, dated, and carries your professional identification (“Jane Doe, PT, DPT”) the day of the visit, and someone owns a weekly sweep of the EHR’s unsigned-note queue so drafts cannot age silently.

Mistake three

The note that matches last Tuesday’s

Copy-forward feels like efficiency: the plan has not changed, the activities are similar, so last visit’s note becomes this visit’s note with the date updated. Reviewers have a word for the result — cloned documentation — and Medicare contractors treat it severely. CGS, a Medicare Administrative Contractor, puts it flatly: cloned documentation is considered misrepresentation of the medical necessity requirement, and identifying it “will lead to denial of services for lack of medical necessity and recoupment of all overpayments made.” That last clause is what makes this the expensive mistake on the list: a cloned-note pattern does not risk one visit, it risks every visit the pattern touched. The fix is not to ban templates — consistent structure is fine and good. It is to require that three things change in every note because they genuinely change in every session: the data you measured, the patient’s response, and the decision you made about next visit. If those three lines are honest, the note cannot clone itself.

Mistake four

The exercise log wearing a note’s clothes

“Pt completed 3x10 sit-to-stand with SBA. Tolerated well. Continue POC.” Every word is true, and that is the problem: it is also everything a family member supervising home practice could truthfully write. Medicare covers outpatient therapy only when the service requires the skills of a therapist (Pub. 100-02, Ch. 15, §220.2), and the note is the only evidence that it did. A note that lists activities and tolerance without a single decision — which cue you chose, what you modified when the first approach failed, what the response tells you about next visit — reads as unskilled supervision, and unskilled supervision is not covered. The corrected line adds the decision, not the word count: “Progressed sit-to-stand to lower surface height after patient achieved 3x10 without upper-extremity push-off; verbal cue for anterior weight shift reduced from every repetition to first repetition only.” The full before-and-after treatment of this mistake, with discipline-specific examples, is in the medical necessity article.

Mistake five

Data with no baseline to stand on

“80% accuracy with minimal cues” sounds objective, but a number with no referent is decoration, not data. Eighty percent of what task, at what level of the cueing hierarchy, compared to what? If the evaluation said 60% and today says 80%, the note demonstrates progress. If no note in the chart ever anchors the number, a reviewer reading the progress report cannot tell improvement from noise — and Medicare’s progress-report requirement is precisely an assessment of improvement toward the goals in the plan of care (Pub. 100-02, Ch. 15, §220.3). The fix is a referent on every data point that matters: “Word retrieval in structured conversation: 80% with first-sound cues, up from 60% independent-cue baseline at evaluation; long-term goal 90%.” One clause ties today’s number to where the patient started and where the plan says they are going. Data points that never move between visits belong in mistake three.

Mistake six

The plan of care that quietly expired

This mistake lives outside the daily note, which is why it survives in practices full of excellent writers: nobody’s note is wrong, and the chart still cannot support the claim. Medicare requires the physician or non-physician practitioner to certify the initial plan of care within 30 calendar days of the first treatment day, and to recertify at least every 90 calendar days — or sooner when the plan changes significantly (Pub. 100-02, Ch. 15, §220.1.3). Visits delivered under a lapsed certification are deniable no matter how skilled the documentation of the visits themselves. There is a delayed-certification process, but it is a salvage path with its own documentation burden, not a workflow. The fix is operational: a tracked recertification date for every active plan of care, reviewed weekly, with the request going out before day 75 — because the slow part is rarely your paperwork, it is the referring provider’s signature queue.

30 days

Initial certification window

The physician or NPP certifies the initial plan of care within 30 calendar days of the first treatment day (Pub. 100-02, Ch. 15, §220.1.3).

90 days

Maximum recertification interval

Recertification is due at least every 90 calendar days, or sooner when the plan of care changes significantly (§220.1.3).

2 totals

Minutes required in every treatment note

Total timed-code minutes and total treatment time, documented per visit (§220.3). Per-intervention start and stop times are not required.

The centerpiece

One note, four mistakes, one rewrite

The mistakes rarely travel alone, because they share a cause: the note was written to close the chart, not to answer a reviewer. Here is a daily note that commits four of the six at once — then the same visit, rewritten. The rewrite is barely longer. It is just specific.

Worked example · fictional case

Physical therapy: gait training three weeks after total knee replacement

An adult three weeks post total knee arthroplasty, treated in outpatient PT for gait training and knee range of motion, with two units of therapeutic exercise and one unit of gait training on the claim.

The note as written

“Pt seen for therex and gait training. Completed quad sets, SLR 3x10, heel slides. Ambulated in clinic with rolling walker. Tolerated treatment well, no adverse reactions. Will continue POC.”

What a reviewer sees

No minutes anywhere, so none of the three billed units can be verified against §220.3. The text is word-for-word identical to the two previous visits except the date — a cloned-documentation flag. Every sentence describes activity and tolerance; nothing required a therapist. And there is no data at all: no distance, no assist level, no range of motion, nothing to compare to baseline. Four mistakes, one short paragraph, zero payable units.

The note rewritten

“Knee flexion 92° today (85° last visit; goal 110°). Progressed heel slides to seated active flexion with contract-relax at end range after passive stretch plateaued — pt initially guarded, achieved +4° within session after diaphragmatic-breathing cue. Gait: 150 ft with rolling walker, supervision; corrected persistent knee-extension avoidance at initial contact with verbal cue “land soft, then straighten,” needed on 3 of 10 passes (was 8 of 10 on Tuesday). Next visit: trial single-point cane in parallel bars if cue frequency holds below 3/10. Total timed code minutes: 38. Total treatment time: 44 minutes.”

What changed

Two time totals appeared, so the units are now verifiable. Every number has a referent — last visit, the goal, or Tuesday’s cue count — so progress is visible. The skilled decisions are on the page: the technique change when the stretch plateaued, the specific cue and its tapering frequency, the conditional plan for next visit. And because the data, response, and decision are visit-specific, this note cannot clone itself forward. It is roughly four sentences longer than the original, and every added sentence is one a reviewer pays for.

Reference

All six mistakes at a glance

What the reviewer concludes, and the fix

The mistakeWhat a reviewer concludesThe fix
No minutes in the noteThe billed units cannot be verifiedTwo totals every visit: total timed-code minutes and total treatment time
Unsigned, undated, or initials-onlyThe record has no accountable authorSign with date and credential the day of the visit; sweep the unsigned queue weekly
Cloned text from prior visitsMisrepresentation of medical necessityMake the data, the response, and the next decision change in every note
Activities and tolerance onlyNo skilled service was requiredDocument the cue, the modification, and the decision it produced
Numbers with no referentProgress cannot be distinguished from noiseAnchor each data point to baseline, last visit, or the plan-of-care goal
Lapsed plan-of-care certificationVisits after the lapse are deniable regardless of note qualityTrack recertification dates; request signatures before day 75

The habit

The 30-second check before you sign

None of these fixes require writing more at the end of a long day; they require checking six things before the signature goes on. Run this against today’s last note. If a line fails, the repair takes seconds now and a denial cycle later.

Field checklist

06 items

Six lines, before you sign

  • Both time totals are in the note — total timed-code minutes and total treatment time — and they support the units on the claim.
  • The note will carry your full name, credential, and today’s date the moment you sign — and no note from earlier this week is still sitting in draft.
  • At least three things differ from this patient’s last note: the data, the response, and the next-visit decision.
  • The note records at least one decision only a therapist could have made — a cue selected, a task modified, a progression chosen, and why.
  • Every number has a referent: baseline, last visit, or the plan-of-care goal.
  • The plan of care covering today’s date is certified, and its recertification date is more than two weeks away — if not, the request is already out.
Do these documentation rules apply to commercial insurance and Medicaid, or just Medicare?

The specific citations here — the two time totals, the 30- and 90-day certification windows — are Medicare’s, from the Benefit Policy Manual. Commercial payers and state Medicaid programs write their own rules, but most borrow heavily from Medicare’s framework, and the underlying failure modes (unverifiable units, unsigned notes, cloned text, unskilled language) are denial risks under essentially every payer’s review standards. Check the payer’s provider manual for the exact requirements; treat Medicare’s as the floor.

Can I fix a note after a claim is denied?

Only within narrow limits, which is why the pre-signature check matters. A missing signature specifically cannot be added late — Medicare’s remedy is a signature attestation or signature log, not a backdated signature. A genuine omission can be handled with an amendment that is clearly labeled as such, separately signed, and dated the day it was written; what you must never do is alter the original entry to look like it always said something it did not. An amendment that appears after a denial also carries less weight with a reviewer than the same fact recorded on the day of the visit.

Is “tolerated treatment well” ever worth writing?

As a standalone sentence, no — it is the signature line of the activity-log note, and it tells a reviewer nothing a layperson could not observe. Tolerance becomes clinically meaningful when it feeds a decision: “no increase in reported pain at the higher surface height, so progression holds” earns its place because the tolerance observation justified a skilled judgment. If a sentence about tolerance does not connect to a decision, it can usually be deleted without the note losing anything.

Will using EHR templates get my notes flagged as cloned?

Templates and consistent structure are not the problem — reviewers read structured notes all day, and CMS guidance on EHR documentation targets content, not format. The flag is narrative content that repeats verbatim across visits or across patients: the same objective data, the same assessment sentences, the same plan, with only the date changed. A template is safe exactly to the degree that its fields force visit-specific entries — the data you measured today, today’s response, and what you decided for next visit.

How long does a defensible daily note need to be?

Length is not the standard; specificity is. The rewritten note in the worked example above is about four sentences longer than the deniable version, and each added sentence does a job: verifying units, anchoring a number, or recording a decision. A short note that states both time totals, visit-specific data with referents, and one skilled decision will outperform a long note that paraphrases the plan of care. If you are adding sentences to make a note look thorough, you are adding risk, not protection.

Primary sources

Bibliography / 7
  1. 01Medicare Benefit Policy Manual, Pub. 100-02, Chapter 15, §220–230: conditions of coverage, plan-of-care certification (§220.1.3), and documentation requirements (§220.3) for outpatient therapyCenters for Medicare & Medicaid Services
  2. 02Complying with Outpatient Rehabilitation Therapy Documentation Requirements (MLN905365)Centers for Medicare & Medicaid Services
  3. 03Complying with Medicare Signature Requirements (MLN905364)Centers for Medicare & Medicaid Services
  4. 04Medical Record Cloning (Part B news, August 2012): cloned documentation as misrepresentation of medical necessityCGS Administrators (Medicare Administrative Contractor)
  5. 05Electronic Health Records fact sheet, Documentation Matters Toolkit: appropriate use of copy-paste and documentation integrityCenters for Medicare & Medicaid Services
  6. 06Medicare documentation requirements for SLPsAmerican Speech-Language-Hearing Association
  7. 07Documentation: Medicare Part B requirementsAmerican Physical Therapy Association

Written by Callie Editorial

Published October 4, 2026

Educational content, not legal, billing, or patient-specific clinical advice.