Appendix H — Worked Coding Scenarios
Sixteen encounters, worked end to end: documentation in, codes out, with the reasoning shown.
Every other appendix in this book is a reference. This one is practice. It exists because the skill this book is actually trying to build is not a list of codes — it is a path: main term, subterm, verify in the Tabular, read the conventions, check the guidelines, check the edits, and be able to explain two years later, to a stranger, why you assigned what you assigned. A code with no path is a memorized fact, and memorized facts in this field expire on a published schedule.
How to use these. Read the documentation. Read the task. Then stop and work it yourself — on paper, with your code books open, the way you would at a desk. Only then read the solution. A worked solution you read straight through teaches you that the answer was obvious, which is the opposite of what happened. The value is in the gap between what you assigned and what the path produces, and you cannot see that gap if you never assign anything.
Each scenario has the same six parts:
- The documentation — a constructed chart excerpt, in the form you would actually receive it.
- The task — exactly what you are being asked to produce.
- The path — the reasoning, step by step.
- The wrong answer, named — what a reasonable person codes instead, and precisely why it fails. The near-miss is where learning happens; a solution that only shows the right answer teaches half the lesson.
- The transferable point — one or two sentences you can carry to a chart that looks nothing like this one.
- Where the book teaches it — the chapter and section that owns the reasoning, so you can go back for the full argument rather than trusting a summary.
⚠️ Read this before you use anything below.
Every chart, note, operative report, remittance, and claim in this appendix is constructed. None of them is a real patient's record, none describes a real person, and no provider, hospital, or health plan named here exists. The documents are written to be realistic enough to work from, which is exactly why they must be labeled.
Every dollar figure, allowed amount, relative weight, and payment rate is illustrative. They are built so the arithmetic is checkable, not because they are anyone's current rate. Wherever a number appears, treat it as
[constructed — verify current values at the source].And the codes move. ICD-10-CM changes every October 1. CPT changes every January 1. HCPCS Level II changes quarterly. NCCI edits and medically unlikely edit values change quarterly. Medicare's fee schedule, the inpatient prospective payment system's relative weights, and the outpatient rates change annually by rule. Local coverage determinations change continuously. Code from the current year's book or encoder and from the payer's current published policy — never from a textbook, including this one. Descriptors here are paraphrased to teach; CPT is a copyrighted work of the American Medical Association and the descriptor that governs a claim is the one printed in the current book.
Where a code cannot be assigned with certainty from the documentation, this appendix says so rather than supplying a number. That is not evasion. It is the answer.
One more thing before the first chart, and it is the discipline the whole appendix is built on. Several of these scenarios end with an answer nobody wants: query the provider, remove the charge, this cannot be coded from this documentation. Those are correct answers, they are common answers, and a book that only showed the satisfying ones would be teaching you to expect a satisfaction the work does not provide. Recognizing that a note does not support a code is the skill. What to add to the note is never the coder's decision — that is the provider's, and Chapter 4 §4.9 governs how you ask.
H.1 An established office visit, leveled on medical decision making
What this tests: selecting an office visit level from the three medical decision making elements, without counting a single history or examination bullet.
CONSTRUCTED TEACHING EXAMPLE — not a real patient record
─────────────────────────────────────────────────────────────────────────
OFFICE NOTE — family medicine, established patient, age 61
DATE OF SERVICE: (constructed) Signed electronically, same day
CHIEF COMPLAINT
Routine follow-up, chronic conditions. Reports fatigue for three weeks.
HISTORY OF PRESENT ILLNESS
Fatigue for approximately three weeks, gradual onset, no fever, no
weight loss, no chest pain, no dyspnea on exertion. Sleeping poorly.
Takes all medications as prescribed. No new medications, no over-the-
counter supplements. No recent illness or travel.
MEDICATIONS (reviewed)
metformin 1000 mg twice daily
lisinopril 20 mg daily
atorvastatin 40 mg daily
EXAMINATION
BP 138/86. HR 72. Afebrile. Weight stable.
Cardiac: regular rate and rhythm, no murmur. Lungs clear.
No thyromegaly. No pallor. Extremities without edema.
ASSESSMENT AND PLAN
1. Type 2 diabetes mellitus — A1c 6.9 three months ago, at goal.
Continue metformin 1000 mg twice daily. Recheck A1c today.
2. Essential hypertension — 138/86, slightly above goal but improving
on current therapy. Continue lisinopril 20 mg daily; recheck in
three months. No change today.
3. Hyperlipidemia — continue atorvastatin 40 mg daily. Lipid panel
ordered today.
4. Fatigue, three weeks — etiology not established. Could reflect
poor sleep, could reflect an endocrine or hematologic cause.
Ordering TSH and CBC with differential today. Will reassess when
results return; consider sleep evaluation if labs unremarkable.
Time was not used for level selection on this encounter.
─────────────────────────────────────────────────────────────────────────
THE TASK. Assign the evaluation and management code and the diagnosis codes, and show how each of the three medical decision making elements scored.
The path
Start with the element that decides most visits: problems addressed. Items 1, 2, and 3 are chronic illnesses, each assessed with a stated status and a stated plan — three stable chronic illnesses, addressed. "Two or more stable chronic illnesses" is the moderate category by itself (Chapter 15 §15.5).
Two cautions live inside that sentence and both are examined on certification exams. "Stable" means at treatment goal, not unchanged. The diabetes is at goal and is stable. The hypertension is documented as "slightly above goal but improving" — that is arguably not stable, which would push it into the exacerbation-or-progression language and, if anything, raise the element rather than lower it. Either reading lands the element at moderate, so the visit does not turn on it; but notice that the intuitive direction is backward. The poorly controlled patient generates the higher problem score, because managing a patient who is not responding is more work.
Item 4 is the one to look at hardest. Fatigue with no established etiology, a differential stated, and a workup ordered is one undiagnosed new problem with uncertain prognosis — which is moderate on this element by itself, with no chronic conditions required. Coders underuse that category badly. It is not "a symptom, therefore low."
Problems addressed: MODERATE.
Now data. Count unique tests ordered (Chapter 15 §15.6). A hemoglobin A1c, a lipid panel, a thyroid stimulating hormone, and a complete blood count with differential — four unique tests ordered, four Category 1 items. Three items in Category 1 is the moderate threshold; four exceeds it. There is no independent interpretation of anyone else's test and no discussion with an external professional, so Categories 2 and 3 are empty and the element cannot reach extensive on two of three categories.
Data: MODERATE, on Category 1 volume alone.
And risk. Three prescriptions are reviewed and each is continued with a documented decision to continue. That is prescription drug management, which is squarely moderate (Chapter 15 §15.7). Note carefully what makes it management rather than a list: the drugs appear under Medications as a list and under Assessment and Plan as decisions. The Assessment and Plan entry is what does the work. Nothing in the note reaches the high tier — no intensive monitoring for toxicity, no hospitalization decision, no major surgery.
Risk: MODERATE.
Three elements at moderate. The rule needs two.
ELEMENT THIS VISIT
─────────────────────────────────────────────────
Problems addressed MODERATE (2+ stable chronic;
AND 1 new undiagnosed
problem, uncertain
prognosis)
Data MODERATE (4 unique tests ordered)
Risk MODERATE (prescription drug
management)
─────────────────────────────────────────────────
Two of three at moderate ► MODERATE MDM
► 99214, established patient
The code: 99214.
Diagnoses. The first-listed diagnosis is the condition chiefly responsible for the services at this encounter (Chapter 9 §9.3). The visit is a chronic-conditions follow-up at which a new complaint was worked up; the reason for the encounter is the follow-up, and the fatigue is a legitimate additional diagnosis, not the reason the patient was on the schedule. Report the chronic conditions that were addressed — E11.9, I10, E78.5 — and the symptom.
The fatigue is coded as a symptom, because no definitive diagnosis was established (Chapter 9 §9.6). That is the correct answer, explicitly endorsed by Section IV of the Official Guidelines, and it is not a failure. Look it up: main term in the Alphabetic Index, then verify in the Tabular List before you assign — the symptom chapter has its own conventions and the Index is a pointer, not an answer (Chapter 8 §8.1).
The wrong answer, named
99215. A reasonable coder reads four problems, four lab orders, and three medications and concludes that this is a busy visit. It is. Busy is not high. High requires either a chronic illness with severe exacerbation or a problem posing a threat to life or bodily function, plus a second element at high — and neither the data element (which needs two of three categories, and has one) nor the risk element (which needs intensive toxicity monitoring or a hospitalization-grade decision) reaches it. Volume of documentation is not complexity of decision making, and this is the single most common upcoding pattern in primary care.
99213. The mirror error, and it is not the safe choice. A coder who sees "stable, continue, no change" three times reads the visit as maintenance and drops a level. That undercounts a moderate problem element, ignores four unique test orders, and ignores prescription drug management. It underpays the practice, it distorts the practice's own quality and risk data, and it is exactly as inaccurate as 99215 (Chapter 5 §5.8). Downcoding is not conservatism.
Leveling by time. The note states that time was not used. A coder who reconstructs a time-based level from the visit's apparent length has invented a fact (Chapter 15 §15.8).
The transferable point
The three elements are scored independently and only two have to agree. Score each one on its own evidence before you look at the code, and never let the length of a note substitute for the difficulty of a decision.
Where the book teaches it: Chapter 15 §15.4 (the two-of-three rule), §15.5 (problems addressed), §15.6 (data), §15.7 (risk), §15.10 (the same note leveled four ways) · Chapter 9 §9.3, §9.6 · Chapter 8 §8.1.
H.2 A minor procedure, and the E/M that should not be on the claim
What this tests: modifier 25 — the most useful modifier in outpatient coding and the most audited, for the same reason.
CONSTRUCTED TEACHING EXAMPLE — not a real patient record
─────────────────────────────────────────────────────────────────────────
OFFICE NOTE — orthopedics, established patient, age 57
DATE OF SERVICE: (constructed)
CHIEF COMPLAINT
Presents for scheduled left knee injection.
HISTORY
Seen three weeks ago for left knee osteoarthritis. At that visit,
conservative therapy and injection were discussed and the patient
elected to proceed. Returns today for the injection as planned.
No new complaints. Knee unchanged since the last visit.
EXAMINATION
Left knee: no effusion, no erythema. Range of motion unchanged.
PROCEDURE
Left knee intra-articular injection. Verbal consent obtained after
discussion of risks and benefits. Lateral approach, 22-gauge needle,
methylprednisolone acetate 40 mg with 3 mL of 1% lidocaine. No
aspirate obtained. No imaging guidance used. Dressing applied.
Patient tolerated the procedure well.
ASSESSMENT AND PLAN
Left knee osteoarthritis — injection performed today as planned.
Return in six weeks or sooner if symptoms worsen.
─────────────────────────────────────────────────────────────────────────
THE TASK. Build the claim lines. Decide whether an office visit is separately reportable and say what decides it.
The path
Two codes are obviously present: the injection and the drug. The injection of a major joint without imaging guidance is 20610, with LT for the left side (Chapter 14 §14.8). The drug is J1030 — methylprednisolone acetate, 40 mg per unit — at one unit for a 40 mg dose (Chapter 20 §20.3).
The lidocaine does not go on the claim, and there are three independent reasons. It is inside the surgical package as local anesthesia (Chapter 17 §17.1); it is a supply usually included in the procedure (Chapter 20 §20.10); and reporting it separately runs into the standards of medical and surgical practice that the National Correct Coding Initiative is built on (Chapter 21 §21.6). Any one of the three would be enough. A claim with a separate line for the local anesthetic is telling a payer something about the practice, and none of it is good.
Now the question the scenario exists for: is there an E/M?
Start from the package, not from the modifier. CPT's surgical package includes the evaluation and management service on the date of the procedure after the decision for surgery has been made — including the history, the examination, and the process of obtaining consent (Chapter 17 §17.1). Read the note against that sentence. The decision was made three weeks ago. Today's history is "no new complaints, knee unchanged." Today's examination is limited to the joint being injected. Today's consent discussion is explicitly part of the procedure note.
Every element of today's evaluation is inside the package. There is nothing above and beyond it.
The claim:
LINE CODE MOD UNITS WHAT IT ASSERTS
─────────────────────────────────────────────────────────────
1 20610 LT 1 A major joint injection was
performed on the left side,
without imaging guidance
2 J1030 — 1 40 mg of methylprednisolone
acetate was administered
─────────────────────────────────────────────────────────────
TWO LINES. NO OFFICE VISIT.
Diagnosis: a code from the M17.- category for osteoarthritis of the knee, at the laterality and specificity the record supports — here, unilateral primary osteoarthritis of the left knee. Look it up and verify it in the Tabular; do not assign a laterality character from memory (Chapter 7 §7.7).
And the honest note about the drug line: this claim does not say what happened to the rest of the vial. If the methylprednisolone came from a single-dose container, the claim owes either a JZ attesting that nothing was discarded or a JW line reporting what was (Chapter 20 §20.4). A claim with neither is incomplete. Scenario H.7 works that arithmetic in full.
The wrong answer, named
99213-25 on line 1. This is the most common modifier 25 error in the country and it is entirely sympathetic: the physician saw the patient, took a history, examined the knee, and discussed the procedure. All of that happened. All of it is inside the package.
Modifier 25 asserts a significant, separately identifiable evaluation and management service above and beyond the usual preoperative and postoperative care associated with the procedure. Ask the question Chapter 14 §14.4 asks: where in the documentation is the thing this modifier says? Point at the sentence. In this note there is no sentence to point at — and a modifier whose supporting sentence you cannot find is a factual assertion you cannot defend, made on a document that carries a certification (Chapter 5 §5.1).
The second wrong answer is subtler and it is worth more money over a year: 20611. The descriptor for the ultrasound-guided version of this injection requires imaging guidance with permanent recording and report. The note says "no imaging guidance used" — which is a documented negative placed there deliberately, and it forecloses 20611 entirely. Coders reach for 20611 because it pays more; the note is the reason they cannot.
And a third: appending modifier 51. There is one surgical procedure on this claim. The multiple-procedure reduction needs two (Chapter 18 §18.8).
The transferable point
Modifier 25 is a question, not a habit. Before you append it, find the sentence in the record that describes work the procedure's package does not already include — and if the decision to perform the procedure was made at an earlier visit, the bar is higher, not lower, because the evaluation that justified it has already been paid for.
Where the book teaches it: Chapter 14 §14.4 · Chapter 17 §17.1, §17.2, §17.7 · Chapter 20 §20.3, §20.4, §20.10 · Chapter 21 §21.6 · Appendix B §B.3.
H.3 The knee with no side: when the answer is "query the provider"
What this tests: recognizing that an available code is not the same as a correct code, and knowing which gap is a query and which is not.
CONSTRUCTED TEACHING EXAMPLE — not a real patient record
─────────────────────────────────────────────────────────────────────────
OFFICE NOTE — family medicine, established patient, age 44
DATE OF SERVICE: (constructed)
CHIEF COMPLAINT
Knee pain.
HISTORY OF PRESENT ILLNESS
Four weeks of knee pain, worse with stairs and after prolonged
sitting. No injury recalled. No locking, no giving way, no swelling
noted by the patient. Has taken ibuprofen intermittently with
partial relief.
EXAMINATION
Knee: no effusion, no erythema or warmth. Medial joint line
tenderness. Crepitus with passive range of motion. Ligamentously
stable. McMurray equivocal.
ASSESSMENT AND PLAN
Knee pain, likely degenerative. No definitive diagnosis today.
Continue NSAIDs as tolerated. Home exercise program discussed.
Return in four weeks; consider imaging if not improving.
─────────────────────────────────────────────────────────────────────────
THE TASK. Assign the diagnosis code. If you cannot, say exactly what you would do instead.
The path
Work the Index the long way (Chapter 8 §8.2). Main term Pain, subterm joint, subterm knee — which lands you in the M25.56- subcategory. Verify in the Tabular (Chapter 8 §8.4), and the Tabular tells you something the Index alone did not: this subcategory requires a sixth character for laterality — right, left, or unspecified.
Now read the note again for a side. There isn't one. Not in the chief complaint, not in the history, not in the examination, not in the assessment. The knee is examined and described in clinical detail, and the record never says which knee it is.
A code exists for exactly this situation: M25.569, pain in unspecified knee. It is a real code, it is in the Tabular, it will pass every format check, and a claim carrying it will very likely be paid.
Assigning it is still the wrong answer.
Here is the reasoning, and it is the reasoning that separates a coder from a data-entry function. "Unspecified" in ICD-10-CM has two entirely different uses. It correctly describes a condition whose specifics are genuinely not knowable — an unspecified organism when no culture was drawn, an unspecified site when the record honestly does not localize it. It is being misused when it describes a fact the provider plainly knows and simply did not write down. The physician examined one knee. They know which one. A body part with two sides and no laterality in the record is not a clinical unknown; it is a documentation gap (Chapter 7 §7.9).
And this is squarely inside the query criteria. Chapter 4 §4.9 lists, among the situations that warrant a query, specificity the code set requires is absent — laterality, stage, type, acuity, linkage. That is this note, exactly.
The query, in the shape the rules require (Chapter 4 §4.9): it presents the clinical evidence from the record — four weeks of knee pain, medial joint line tenderness and crepitus documented on examination. It asks an open question: which knee was examined and is being treated at this encounter? It offers clinically reasonable options including the honest one — right, left, both, or undetermined from this encounter. It does not name a code, does not state a financial consequence, and does not indicate a preferred answer. It is documented and retained, and the response becomes part of the record.
And what if the query is not answered? Then you assign what the documentation supports, which is M25.569, and you have an accurate code and an inaccurate description of the patient. That is a real outcome and you should not pretend otherwise. It is also a measurable one: unspecified-laterality rates by provider are a standard clinical documentation integrity report, and the fix is upstream, in Chapter 38's territory, not on this claim.
What is not a query here. The note says "likely degenerative" and "no definitive diagnosis today." Do not query to convert that into osteoarthritis. The provider has stated a clinical impression and explicitly declined to establish a diagnosis; in the outpatient setting a probable or suspected condition is not coded as though it existed (Chapter 9 §9.5), and querying to obtain a more specific — and better-paying — diagnosis that the provider has deliberately not made is the query the rules forbid. The laterality is a gap. The diagnosis is a decision.
The wrong answer, named
Assigning M25.569 and moving on. It is the fastest answer, it is defensible in the narrow sense that the code matches the record, and it is how unspecified laterality reaches thirty percent of a practice's musculoskeletal claims without anyone deciding it should. The claim is not wrong. The record is — and the coder is the only person in the building positioned to notice.
Assigning a laterality from context. A coder who sees "medial joint line tenderness" and a prior note about a right knee and assigns M25.561 has decided what happened rather than read what happened. That is inference, and Chapter 4 §4.7 forecloses it: you may ask what happened; you may not decide what happened.
Querying for the diagnosis instead of the laterality. Chapter 4 §4.9's "when not to query" list names this directly — do not query to obtain a higher-paying code, and do not query when the honest resolution is the less specific code.
The transferable point
The existence of a code is not permission to use it. Ask whether "unspecified" describes something genuinely unknown or something merely unwritten — and where it is unwritten, the mechanism is a non-leading query, not a better guess.
Where the book teaches it: Chapter 4 §4.7, §4.9 · Chapter 7 §7.7, §7.9 · Chapter 8 §8.2, §8.4 · Chapter 9 §9.5 · Chapter 38.
H.4 A symptom, a suspicion, and the sequencing between them
What this tests: the outpatient uncertain-diagnosis rule, which runs opposite to the inpatient one, and when a symptom code is the right answer rather than a fallback.
CONSTRUCTED TEACHING EXAMPLE — not a real patient record
─────────────────────────────────────────────────────────────────────────
OFFICE NOTE — internal medicine, established patient, age 66
DATE OF SERVICE: (constructed)
CHIEF COMPLAINT
Shortness of breath with exertion, two months.
HISTORY OF PRESENT ILLNESS
Progressive dyspnea on exertion over roughly two months. Now
symptomatic climbing one flight. No orthopnea, no paroxysmal
nocturnal dyspnea, no chest pain. No cough, no fever. Former
smoker, quit eleven years ago. No known cardiac history.
EXAMINATION
BP 128/78. HR 84. Oxygen saturation 95% on room air.
Lungs: mildly diminished breath sounds at the bases, no wheeze.
Cardiac: regular, no murmur, no gallop. No peripheral edema.
No jugular venous distension.
ASSESSMENT AND PLAN
Dyspnea on exertion, etiology not established. Differential
includes cardiac and pulmonary causes; rule out heart failure
versus obstructive lung disease.
Ordering: chest radiograph, two views. ECG with interpretation
performed in office today. Pulmonary function testing scheduled.
Return when results are available.
─────────────────────────────────────────────────────────────────────────
THE TASK. Assign the diagnosis code or codes for this encounter and state the sequence. Say what supports the electrocardiogram.
The path
The whole scenario turns on one sentence in the note: "etiology not established … rule out heart failure versus obstructive lung disease."
In the outpatient setting, you do not code a rule-out. Section IV of the Official Guidelines is explicit and it is the mirror image of the inpatient rule: a condition documented as probable, suspected, questionable, or "rule out" is not coded as though it existed. Code the condition to its highest degree of certainty for that encounter — which here means the symptom (Chapter 9 §9.5).
And coding the symptom is the correct answer, not a consolation prize. Chapter 9 §9.6 states the rule in two halves and most coders remember only the first. The half that applies here: when no definitive diagnosis has been established, code the symptom. It is endorsed, it is accurate, and it is what the record supports.
Look it up. Main term Dyspnea in the Alphabetic Index, then verify in the Tabular. The shortness-of-breath code is R06.02, in the symptoms chapter.
One code. R06.02, first-listed. Nothing else on this encounter meets Section IV's test for an additional diagnosis: the former smoking status is history rather than a condition affecting this encounter's care, and neither heart failure nor obstructive lung disease exists as a diagnosis yet.
What supports the electrocardiogram is the same code. The professional and technical components of an in-office ECG with interpretation and report are reported together as 93000, and the diagnosis pointer on that line points at R06.02 (Chapter 25 §25.5). This is the point at which a lot of billing offices get nervous, because a symptom code feels like a weak justification for a diagnostic test. It is not. A diagnostic test is ordered precisely because the diagnosis is not known, and the symptom is the medical necessity. What would be improper is pointing the line at a diagnosis the patient has not been given, in order to match a coverage policy — which is coding to the policy rather than to the record (Chapter 22 §22.6).
The chest radiograph is ordered, not performed here, so it is not on this claim; the imaging facility bills it. And the pulmonary function testing has not happened yet. Order and perform are different verbs and only one of them produces a claim line (Chapter 19 §19.5).
The wrong answer, named
Coding heart failure, or a code from the COPD family, because the differential names them. This is the outpatient uncertain-diagnosis rule violated in the most sympathetic possible way: the physician said the words, and the coder reported them. The physician said them as a question. Reporting a diagnosis the patient has not been given puts a condition on the patient's permanent claims history, follows them to the next payer and the next risk-adjustment cycle, and is a false statement on a document that carries a certification.
Coding both the symptom and a suspected diagnosis "to be safe." This is worse, not safer. It asserts the diagnosis and adds noise.
Reporting the smoking history as though it explained the encounter. A status or history code describes the patient; it does not answer why is this person here today (Chapter 12 §12.9).
And the inpatient reflex. A coder who moves between settings will eventually apply the inpatient uncertain-diagnosis rule here, because inpatient coding does the opposite — a condition documented at discharge as probable or suspected is coded as if established. Same words, opposite instruction, and the only thing that decides is the setting (Chapter 9 §9.5, Chapter 33 §33.2).
The transferable point
Code to the highest degree of certainty the record actually reaches, and know which setting you are in before you decide what "certainty" means. A symptom code on a diagnostic workup is not a weakness in the claim; it is an accurate description of what the physician knows on the day they wrote it.
Where the book teaches it: Chapter 9 §9.3, §9.5, §9.6 · Chapter 12 §12.8, §12.9 · Chapter 19 §19.5 · Chapter 22 §22.6 · Chapter 25 §25.5 · Chapter 33 §33.2.
H.5 Three encounters, one fracture: the seventh character
What this tests: the most misapplied convention in ICD-10-CM, and the fact that the seventh character describes a kind of care rather than a position in a sequence.
CONSTRUCTED TEACHING EXAMPLE — not a real patient record
─────────────────────────────────────────────────────────────────────────
THREE ENCOUNTERS, ONE PATIENT, ONE INJURY
Patient age 68. Constructed teaching file.
ENCOUNTER A — urgent care, day 0
Fell on an outstretched hand at home. Radiographs of the right
wrist demonstrate a fracture of the distal radius. The record does
not state whether the fracture is displaced. No wound; skin intact.
Closed reduction not required. Short arm splint applied. Referred
to orthopedics.
ENCOUNTER B — orthopedics, day 6
First visit to this practice. Fracture confirmed on repeat films.
Definitive cast applied today. Fracture care assumed by this
practice.
ENCOUNTER C — orthopedics, day 34
Cast removed. Radiographs show healing in good alignment. No
further immobilization required. Home exercise program given.
ENCOUNTER D — orthopedics, one year later
Returns with persistent stiffness and reduced supination of the
right wrist, attributed by the physician to the prior fracture.
The fracture is healed. Treatment today is directed at the
stiffness.
─────────────────────────────────────────────────────────────────────────
THE TASK. Assign the seventh character for each of the four encounters, and state the sequencing rule that applies at encounter D.
The path
Ask the right question, and it is never "which visit is this?" (Chapter 12 §12.3). The question is: is the patient receiving active treatment for this condition?
Encounter A — active treatment. Radiographs, diagnosis, splinting. Seventh character A. With the site and the closed status, this is the initial encounter for a closed fracture of the lower end of the right radius — S52.501A if the fracture type remains unspecified.
Two defaults apply and both are guideline instructions rather than judgment calls, and both are examined. A fracture not documented as open or closed defaults to closed — which is what makes the closed seventh character correct here rather than a guess. And a fracture not documented as displaced or non-displaced defaults to displaced, which runs in the direction most people do not expect (Chapter 12 §12.4).
Read the second default carefully, because it bites only where the classification offers the distinction. The displaced and non-displaced options live in the subcategories that name a specific fracture pattern. This record names no pattern at all — it says "a fracture of the distal radius" — which lands you in the unspecified-type subcategory, where that axis does not appear. The default did not change your code; the missing pattern did. If the radiology report names a pattern that the physician's note does not, that discrepancy is a query, not a coder's choice (Chapter 4 §4.9).
Encounter B is the one that catches everyone. It is the patient's first visit to this practice, and it takes A, not D. Definitive casting is active treatment. "Initial" does not mean the first visit and "subsequent" does not mean the second one. A patient seen four times during active treatment has four encounters coded A.
Encounter C — routine healing. Active treatment is complete; the visit is cast removal and a healing check. Seventh character D — S52.501D.
And note the trap sitting next to it. Routine fracture aftercare is not reported with an aftercare Z-code. It is reported with the acute fracture code plus the appropriate subsequent-encounter seventh character. This is a specific, frequently missed rule and it is the one place where the aftercare category that seems obviously right is specifically wrong (Chapter 12 §12.4).
Encounter D — sequela. The acute phase is over and a residual effect remains. Seventh character S — S52.501S.
But S never travels alone, and it is never first. The sequencing rule is fixed: code the residual condition first, then the injury code with S (Chapter 12 §12.3). So encounter D is a code for the joint stiffness of the wrist — look it up and verify the site and laterality in the Tabular — followed by S52.501S. The S code says what caused the residual; it does not say what is being treated today.
THE FOUR ENCOUNTERS
─────────────────────────────────────────────────────────────
A day 0 splint, diagnosis ACTIVE ► ...A
B day 6 definitive cast ACTIVE ► ...A
C day 34 cast off, healing check ROUTINE ► ...D
D 1 year stiffness from the RESIDUAL ► residual
healed fracture FIRST,
then ...S
─────────────────────────────────────────────────────────────
NOT the ordinal position of the visit. The KIND OF CARE.
The wrong answer, named
Coding encounter B with D because it is the second encounter, or because it is a different practice. This is the error the convention's name invites, and it is the single most common seventh-character mistake. The character has nothing to do with counting.
Coding encounter C with an aftercare Z-code. Reasonable, tidy, and specifically forbidden for fracture aftercare.
Reporting S52.501S alone at encounter D, or sequencing it first. A sequela code alone tells the payer that a healed fracture is being treated, which is not what happened. The residual is the reason for the encounter and it goes first.
And assigning a displaced or non-displaced character at encounter A because the radiology report implies one. If the record does not state it, the default applies; if you believe the provider knows and did not write it, that is a query (Chapter 4 §4.9), not an inference.
The transferable point
A, D, and S describe what kind of care is being delivered, not where the visit falls in a sequence. Ask whether treatment is active, whether healing is routine, or whether the acute phase is over and something is left behind — and when something is left behind, the leftover is coded first.
Where the book teaches it: Chapter 7 §7.6 · Chapter 12 §12.2, §12.3, §12.4 · Chapter 17 §17.7 (the fracture-care decision on the procedure side) · Chapter 35 §35.3 · Appendix A §A.6.
H.6 One patient, two organs, two different rules
What this tests: the "with" convention, which presumes some causal relationships and refuses to presume others — in the same note, for the same patient.
CONSTRUCTED TEACHING EXAMPLE — not a real patient record
─────────────────────────────────────────────────────────────────────────
TWO VERSIONS OF THE SAME ASSESSMENT
Cardiology office note, established patient, age 73.
The clinical facts are identical. Only the wording differs.
VERSION 1
ASSESSMENT AND PLAN
1. Essential hypertension — 136/82 on current therapy. Continue
lisinopril 20 mg daily.
2. Chronic kidney disease, stage 3a — stable, estimated GFR 52.
Avoid nephrotoxic agents. Nephrology following. [see note]
3. Chronic diastolic heart failure — compensated. Continue
furosemide 20 mg daily. Weight stable, no edema.
VERSION 2
ASSESSMENT AND PLAN
1. Hypertensive heart and kidney disease.
2. Chronic diastolic heart failure due to hypertension —
compensated. Continue furosemide 20 mg daily.
3. Chronic kidney disease, stage 3a — stable, estimated GFR 52.
[see note] In version 1, item 2 also records: "Reviewed with the
patient; medication list adjusted to avoid nephrotoxic agents."
The kidney disease is addressed in both versions.
─────────────────────────────────────────────────────────────────────────
THE TASK. Code the diagnoses for each version. Explain why the two versions do not produce the same codes, and identify which condition is treated differently from the others.
The path
The convention first. The word "with" — and "in" — appearing in a code title, in the Alphabetic Index, or in an instructional note in the Tabular is interpreted as "associated with" or "due to," and the classification presumes a causal relationship between the two conditions it links. Those conditions are coded as related even without provider documentation stating the link, unless the record affirmatively says they are unrelated (Chapter 9 §9.7).
Read that carefully, because it is a genuine exception to everything Chapter 4 taught. The coder is not inferring anything. The classification made the causal judgment, as a matter of classification design, and the coder is following an instruction.
Now the asymmetry, which is the most tested thing in circulatory coding (Chapter 11 §11.1).
Hypertension and chronic kidney disease: the relationship is PRESUMED. Where both are documented, the classification joins them into hypertensive chronic kidney disease — a code from the I12.- category — with no stated linkage required. And a code for the CKD stage is reported additionally; the Tabular's "use additional code" instruction says so (Chapter 8 §8.6). Here that is N18.31, stage 3a.
Hypertension and a heart condition: the relationship must be STATED or IMPLIED. "Due to hypertension," or the word "hypertensive," or an equivalent. Without it, the two are coded separately. The epidemiological reason is worth remembering because it makes the rule stick: many heart conditions have causes other than hypertension, and the classification declines to guess.
Version 1, coded. Hypertension and CKD are both documented, so the presumption fires: a code from the I12.- category, plus N18.31. The heart failure is documented with no stated relationship to the hypertension, so it stands on its own: I50.32, chronic diastolic congestive heart failure. And I10 is not additionally reported — the I12.- code already contains the hypertension, and reporting both would report it twice.
Version 2, coded. Now the record states the linkage for the heart as well — "hypertensive heart and kidney disease," and "heart failure due to hypertension." Both organs are linked, and the classification has a combination category for exactly that situation. I13.0 — hypertensive heart and chronic kidney disease with heart failure and stage 1 through 4 chronic kidney disease — plus I50.32 to identify the type of heart failure and N18.31 to identify the CKD stage, both under "use additional code" instructions.
SAME PATIENT. SAME NOTE. DIFFERENT RULE PER ORGAN.
HYPERTENSION + CKD
relationship PRESUMED by the classification
► combination code, no linkage statement needed
► PLUS the CKD stage code
HYPERTENSION + HEART CONDITION
relationship must be STATED or IMPLIED
► without it: TWO SEPARATE CODES
► with it: the combination category
HYPERTENSION + BOTH, both linked
► a further combination category
One boundary that matters more than it looks. The presumption applies where the Index or a Tabular instruction links the terms — not everywhere the word "with" appears in a clinical note. A provider writing "hypertension with a knee injury" has triggered nothing. The convention is about the classification's use of the word, not the physician's (Chapter 9 §9.7).
And a prior question governs all of this. Before you ask whether two conditions are linked, ask whether the condition was addressed at this encounter — evaluated, treated, or requiring management (Chapter 9 §9.3). The linkage question is second. Both versions of this note address the kidney disease, which is why it is reportable here; a chronic condition sitting on a problem list, mentioned nowhere in the assessment, does not become reportable just because the classification would happily link it to something.
The wrong answer, named
Version 1 coded as I10 + N18.31 + I50.32 — three separate codes. This is the mirror image of the exam question and it is a real error, not a trick: a coder who has correctly learned "never infer a causal relationship" applies that rule to the kidney and misses that the classification has already made the inference for them. Chapter 4's rule and the "with" convention are not in conflict; the convention is the classification speaking, not the coder.
Version 1 coded with a hypertensive-heart combination code. The opposite error, and the more dangerous one, because it asserts a causal relationship the physician did not document.
Reporting I10 alongside the I12.- code. Double-reporting the hypertension. The Tabular's Excludes notes are where this gets caught, and reading them is not optional (Chapter 8 §8.5).
And generalizing the presumption. "The classification presumes hypertension causes kidney disease, so it probably presumes this other pair too." It does not. The presumption extends only where the classification has linked the terms, and where chapter-specific guidance requires an explicitly documented relationship, that guidance governs (Chapter 9 §9.7).
The transferable point
"Never infer" and "the classification presumes" are both true, and the Index and Tabular tell you which one you are in. Read the instructional notes before you decide whether a combination code applies — the answer changes by organ, in the same note, for the same patient.
Where the book teaches it: Chapter 7 §7.8 · Chapter 8 §8.5, §8.6 · Chapter 9 §9.3, §9.7 · Chapter 10 §10.6 · Chapter 11 §11.1, §11.8.
H.7 A vial, a dose, and the units nobody checks
What this tests: the single most consequential arithmetic in HCPCS Level II, and the two attestations that must accompany it.
CONSTRUCTED TEACHING EXAMPLE — not a real patient record
─────────────────────────────────────────────────────────────────────────
PROCEDURE NOTE AND MEDICATION ADMINISTRATION RECORD
Rheumatology office, established patient. Constructed teaching file.
PROCEDURE
Right shoulder subacromial injection. Consent obtained. Sterile
prep. Triamcinolone acetonide 60 mg with 4 mL of 1% lidocaine
instilled. No aspirate. No imaging guidance used. Dressing applied.
MEDICATION ADMINISTRATION RECORD
Drug ................. triamcinolone acetonide
Container ............ 80 mg single-dose vial, one vial opened
Administered ......... 60 mg
Remaining in vial .... 20 mg, discarded
Witnessed and documented by nursing at the time of administration
RELEVANT CODE DESCRIPTOR (paraphrased)
J3301 — injection, triamcinolone acetonide, not otherwise
specified, 10 mg
─────────────────────────────────────────────────────────────────────────
THE TASK. Build the drug lines. State the units on each and prove the arithmetic. Say what a claim missing one of these lines tells the payer.
The path
One sentence governs the whole family: the dosage is in the descriptor, and the units are the arithmetic (Chapter 20 §20.3). A J-code does not mean "this drug." It means a specific quantity of this drug. The descriptor here says 10 mg. Ten milligrams is the unit — not the vial, not the injection, not the patient.
The administered line.
ADMINISTERED 60 mg ÷ 10 mg per unit = 6 units
J3301 × 6.
The discarded line. The vial was a single-dose container, 80 mg was drawn, 60 mg went into the patient, and 20 mg could not be used for anyone else. That is waste, and it is reported on a separate line with modifier JW — drug amount discarded and not administered to any patient (Chapter 20 §20.4).
DISCARDED 20 mg ÷ 10 mg per unit = 2 units
J3301-JW × 2.
Now the check that makes the claim defensible, and it takes four seconds:
THE UNITS MUST ACCOUNT FOR THE WHOLE CONTAINER
administered 6 units × 10 mg = 60 mg
discarded 2 units × 10 mg = 20 mg
──────
container 80 mg ✓ one vial
A claim where those do not add to the container is arithmetically wrong on its face, and it is wrong in a way an edit can see without opening a chart.
Modifier JZ does not belong on this claim, and knowing why is the point of the pair. JZ attests that zero drug amount was discarded. Something was discarded here, so JZ would be a false statement. JW and JZ are mutually exclusive on the same administration, and each is an attestation rather than a formatting convention (Chapter 20 §20.4).
Why JZ exists at all is worth thirty seconds, because it explains a whole species of claim field. JW alone created an ambiguity: a claim with no JW line could mean nothing was wasted or nobody bothered to report the waste, and those are very different facts that looked identical. JZ resolves it by requiring the negative to be stated. A claim now says one of two things — or it is incomplete:
J-code line + JW line ....... "we discarded this much"
J-code line + JZ ............ "we discarded nothing"
J-code line + neither ....... AN INCOMPLETE CLAIM
And the waste must be documented in the record, not merely asserted on the claim. The medication administration record above is the evidence; the claim is the assertion (Chapter 5 §5.1). Notice that the note documents the container size, the administered dose, and the discarded amount at the time of administration — which is the only time that documentation can honestly be created.
Two more lines belong on this encounter and are outside the drug question. The injection itself is coded from the arthrocentesis family by joint size — the subacromial bursa is a major joint or bursa, which is 20610, without imaging guidance because the note documents that none was used, with RT. And the lidocaine is inside the package and does not appear on the claim (Chapter 17 §17.1, Chapter 20 §20.10, Chapter 21 §21.6).
The wrong answer, named
J3301 × 1, reading the code as "the drug" or "the injection." At a 10 mg descriptor and a 60 mg dose, this is a six-fold underpayment. It will be paid, it will produce no denial, no rejection, and no exception report, and it will repeat on every claim until somebody measures allowed amounts against expected allowed amounts (Chapter 28 §28.8). The silent error is the expensive one.
J3301 × 8, reporting the vial rather than the dose. A substantial overpayment, and one that medically unlikely edits are specifically built to catch — which means it will be caught, and the only question is when and how far back (Chapter 21 §21.4).
J3301 × 6 with no JW and no JZ. Incomplete. It does not overstate or understate the dose; it declines to answer a question the payer requires an answer to.
J3301 × 8 on one line, waste included. This reports 80 mg as administered. It is the same money as the correct two-line claim on many fee schedules, which is exactly what makes it tempting — and it is a false statement about what went into the patient.
And JW on a multi-dose vial. Multi-dose containers hold doses for more than one patient by definition, so the unused portion is not waste (Chapter 20 §20.4).
The transferable point
Divide the dose by the descriptor, then prove the units add up to the container. Every J-code line is two numbers that must reconcile to a third, and no other line on a claim gives you a free arithmetic check that catches both directions of error.
Where the book teaches it: Chapter 20 §20.3, §20.4, §20.8, §20.10 · Chapter 21 §21.4 · Chapter 17 §17.7 · Chapter 28 §28.8 · Appendix A §A.17.
H.8 An excision, a repair, and four measurements
What this tests: the integumentary measurement rule, and the fact that two numbers from the same operation are measured from different things.
CONSTRUCTED TEACHING EXAMPLE — not a real patient record
─────────────────────────────────────────────────────────────────────────
OPERATIVE REPORT — outpatient surgery, constructed teaching file
PREOPERATIVE DIAGNOSIS Lesion, right upper back.
POSTOPERATIVE DIAGNOSIS Same, pending pathology.
PROCEDURE PERFORMED Excision of lesion, back.
BODY
After informed consent and sterile prep, the lesion was identified
and measured. It measured 1.2 cm in greatest dimension. An elliptical
excision was designed with 0.3 cm margins circumferentially and the
lesion was excised full thickness. The specimen was oriented and
submitted to pathology.
The resulting defect measured 3.4 cm in greatest length. Hemostasis
was obtained with electrocautery. The deep dermal layer was
approximated with interrupted 4-0 absorbable suture and the skin was
closed with a running 4-0 subcuticular stitch. Sterile dressing
applied. The patient tolerated the procedure well.
PATHOLOGY REPORT (returned four days later)
Specimen: skin, right upper back, measuring 2.9 cm in greatest
dimension. Diagnosis: benign intradermal nevus, margins free.
─────────────────────────────────────────────────────────────────────────
THE TASK. Assign the procedure codes and the size used for each. State when this claim may be submitted.
The path
Read the body, not the heading. The "Procedure Performed" heading is a summary and is not authoritative; where it disagrees with the body, the body wins (Chapter 17 §17.3). The heading names one procedure. The body describes two.
The measurement rule, which is the reason this chapter exists (Chapter 17 §17.4):
The excision size is the LESION'S GREATEST DIAMETER PLUS THE NARROWEST MARGINS REQUIRED, measured BEFORE excision. Not the pathology specimen. Not the defect.
0.3 + 1.2 + 0.3 = 1.8 cm excised diameter
─── ─── ───
margin lesion margin
1.8 cm. With a benign pathology result and a trunk site, that lands in the 1.1 to 2.0 cm tier — 11402.
Now the second procedure. "The deep dermal layer was approximated … and the skin was closed" is a layered closure, which is an intermediate repair, and intermediate and complex repairs are separately reportable with an excision (Chapter 17 §17.5). Simple repair would not be — it is included in the excision code — and that distinction is the whole difference between a two-line claim and a one-line claim.
And the repair is measured from the DEFECT, not the excision. The defect is 3.4 cm. Trunk, layered closure, 2.6 to 7.5 cm — 12032.
Four numbers appear in this operative report and only two of them are used:
| Number | What it is | Use it for |
|---|---|---|
| 1.2 cm | the lesion alone | nothing — it is not the excision size |
| 1.8 cm | lesion + margins | the excision code |
| 3.4 cm | the defect | the repair code |
| 2.9 cm | the formalin-shrunken specimen | nothing |
When may the claim go out? Not on the day of surgery. The excision code family is divided by benign versus malignant, and the pathology report determines which — which means the code frequently cannot be assigned until pathology returns. A practice that codes excisions before pathology is guessing, and it will guess wrong in the expensive direction some of the time (Chapter 17 §17.4). Here pathology returned benign on day four, and 11402 is the benign trunk code. Hold the claim; the four-day delay is correct and it is not an accounts receivable problem.
Sequencing on the claim. The excision is the higher-valued procedure and goes on line 1; the repair follows (Chapter 18 §18.8). Whether a multiple-procedure reduction applies to line 2, and whether the payer wants modifier 51, is a payer-convention question, and the Medicare Physician Fee Schedule relative value file carries the indicator that answers it (Chapter 17 §17.2, Chapter 18 §18.8).
The wrong answer, named
Coding 1.2 cm. The lesion without its margins. Systematically undersizes every excision in the practice, systematically undercodes and underpays, on every claim, forever, with no signal that anything is wrong.
Coding 2.9 cm from the pathology report. Tissue shrinks in formalin. The specimen measured in the laboratory is smaller than the tissue that left the patient, and the pathology report is the wrong source for a size, even though it is the right source for benign versus malignant. Coding from it is the same error wearing a more authoritative costume.
Coding 3.4 cm for the excision. Using the defect for the excision code. The defect belongs to the repair.
Reporting a simple repair, or reporting the closure at all as simple. A simple repair is included in the excision. Reporting it separately is unbundling.
Reporting only the heading's procedure. The most expensive error on this list, because it is silent: the repair is separately reportable, it was performed, it was documented in the body, and nobody billed it. This is the money left on the table that Chapter 14 keeps naming.
And submitting on the day of surgery. Which forces a guess on benign versus malignant.
The transferable point
One operation can produce several measurements, and each code has its own source for the number. Excision takes lesion plus margins before cutting; repair takes the defect; pathology's measurement is for pathology's purposes and belongs in no code at all.
Where the book teaches it: Chapter 17 §17.3, §17.4, §17.5, §17.10 · Chapter 18 §18.8 · Chapter 13 §13.7.
H.9 Inside a ninety-day global: three visits, three answers
What this tests: the surgical package's duration, and the modifiers that identify the exceptions to it.
CONSTRUCTED TEACHING EXAMPLE — not a real patient record
─────────────────────────────────────────────────────────────────────────
ORTHOPEDIC PRACTICE — one surgeon, one patient, four contacts
Constructed teaching file. Patient age 69.
DAY 0 Total knee arthroplasty, RIGHT knee. Performed by Dr. A of
this practice. (CPT 27447; global period per the fee
schedule file: 090.)
DAY 12 Office. Staples removed, wound inspected and clean, range of
motion reviewed, therapy plan reinforced. No new problem.
Seen by Dr. A.
DAY 30 Office, Dr. A. Patient reports six weeks of LEFT knee pain,
separate from the surgical knee, worsening. History taken,
left knee examined: medial joint line tenderness, crepitus,
no effusion, ligamentously stable. Radiographs of the left
knee reviewed from an outside facility. Options discussed;
patient elects injection today. LEFT knee intra-articular
injection performed: methylprednisolone acetate 40 mg with
lidocaine, lateral approach, no aspirate, no imaging
guidance used.
DAY 62 Returned to the operating room by Dr. A for evacuation of a
hematoma at the RIGHT knee surgical site. Unplanned;
a complication of the index procedure.
DAY 71 Office, Dr. A. Right knee wound check following the day-62
procedure. No new problem.
─────────────────────────────────────────────────────────────────────────
THE TASK. For each contact after day 0, state whether it is separately reportable and which modifier, if any, applies.
The path
First, look the global period up rather than deciding it. 27447 carries a 090 indicator — a major procedure: one preoperative day, the day of surgery, and ninety postoperative days. The Medicare Physician Fee Schedule relative value file carries the indicator in a column, free, for every code, and the lookup takes about eleven seconds (Chapter 17 §17.2). Minor versus major is a lookup, not a judgment about how big the operation felt — and it is the same lookup that decides modifier 25 versus modifier 57.
Day 12 — not separately reportable. Staple removal, a wound check, range-of-motion review, and reassurance are typical postoperative follow-up care, which is inside the package (Chapter 17 §17.1). No code, no modifier, no charge. This is the single most common surprise for practices new to surgical billing, and it is also the most common source of a denial that is entirely correct.
Day 30 — this is the interesting one, and it produces two lines.
The left knee is a different problem, in a different joint, on the other side, with its own history, its own examination, and its own management decision. It is unrelated to the right knee arthroplasty. Two exceptions to the package therefore both apply at once.
The evaluation and management service takes modifier 24 — an unrelated evaluation and management service by the same physician during a postoperative period (Chapter 14 §14.9). Without it, the payer's system sees an office visit inside a 090 global and denies it, correctly, on the information it has.
The injection takes modifier 79 — an unrelated procedure by the same physician during the postoperative period. Modifier 79 tells the payer this is not related to the index surgery, and it does not restart or extend the original global period.
And here is the wrinkle that makes this scenario worth working: the E/M also needs modifier 25. 20610 carries a 000 global of its own, so an evaluation and management service on the same date as the injection is a modifier 25 question in its own right (Chapter 14 §14.4) — and this note answers it cleanly. The history, the examination, the review of outside radiographs, and the decision to inject are all documented, and none of it is inside 20610's package.
So the office visit line carries both 24 and 25. Which goes first is a sequencing question, and Chapter 14 §14.3's rule is that payment modifiers precede informational ones — but both of these are payment modifiers, and payer conventions differ on the order. Check the payer's companion guide (Chapter 25 §25.9). This is not a cop-out: it is the honest answer, and the companion guide is a published document the practice already has a right to.
DAY 30 — THE CLAIM
─────────────────────────────────────────────────────────────
1 99213 24, 25 LEFT knee: new problem, unrelated to the
RIGHT knee's 090 global; and significant
and separately identifiable from 20610
2 20610 79, LT unrelated procedure during another
procedure's postoperative period, left
3 J1030 — 40 mg, one unit
─────────────────────────────────────────────────────────────
The E/M level comes from the documentation, not from this table.
Day 62 — reportable, with modifier 78. A complication requiring a return to the operating room is excluded from the package (Chapter 17 §17.1). Modifier 78 identifies an unplanned return to the operating or procedure room by the same physician for a related procedure during the postoperative period. Two consequences the book froze and that people get backward: 78 does NOT start a new global period, and it typically pays at a reduced rate — the intraoperative portion only, because the preoperative and postoperative work was already paid in the index procedure.
Day 71 — not separately reportable. A wound check following the day-62 procedure, inside the original global, related to the surgical knee. Same answer as day 12, for the same reason.
The wrong answer, named
Modifier 58 on day 62. This is the most common global-period modifier error and the two look adjacent on a chart. 58 is planned or staged; 78 is not. A hematoma evacuation is an unplanned return for a complication. The distinction is worth money in both directions: 58 generally pays in full and starts a new global period; 78 pays reduced and does not (Chapter 14 §14.9). Choosing 58 here overstates payment and misstates what happened.
Modifier 79 on the day-30 office visit. 79 is for procedures; 24 is the E/M counterpart. They are not interchangeable and a claim using the wrong one will be denied for a modifier inconsistency.
Billing the day-12 visit with modifier 24. The visit is related and routine. Modifier 24 asserts it is unrelated, which would be a false statement, and appending a modifier to defeat an edit is the exact behavior Chapter 21 §21.8 is written against.
Writing off day 30 entirely because "it's in the global." The opposite error, and it is a silent loss rather than a visible one — a practice can do this for years and see nothing but a clean denial report (Chapter 14's silent underpayment, in the highest-dollar setting in the book).
And one that is not the practice's error at all: if a different physician, in a different group or a different specialty, had seen this patient on day 30, no global-period modifier would be needed. The global period binds the operating surgeon and the physicians in the same group and specialty, not everyone who touches the patient.
The transferable point
Look the global period up, then ask two questions of every postoperative contact: is it related, and did it go back to the operating room. Those two answers select among 24, 58, 78, and 79 — and "related and routine" is not a modifier at all, it is a service already paid for.
Where the book teaches it: Chapter 14 §14.4, §14.9 · Chapter 17 §17.1, §17.2 · Chapter 18 §18.8 · Chapter 25 §25.9 · Appendix B §B.6.
H.10 An edit, an indicator, and a modifier you should not append
What this tests: what a National Correct Coding Initiative modifier indicator actually permits, and the difference between a system accepting a modifier and a record supporting one.
CONSTRUCTED TEACHING EXAMPLE — not a real patient record
─────────────────────────────────────────────────────────────────────────
OPERATIVE REPORT EXCERPT — shoulder arthroscopy
Constructed teaching file. This is NOT Account 31-2245; it is a
separate constructed case built to be read against it.
PROCEDURES PERFORMED
1. Arthroscopic rotator cuff repair, right shoulder.
2. Arthroscopic subacromial decompression, right shoulder.
3. Arthroscopic limited debridement, right shoulder.
BODY (excerpt)
... The subacromial space was entered. Bursal tissue and frayed
soft tissue overlying the cuff were debrided to expose the tear
and to allow visualization of the repair site. A subacromial
decompression with partial acromioplasty was performed. The
rotator cuff tear was then repaired with two suture anchors,
with restoration of the footprint ...
WHAT THE BILLING SYSTEM DID
The practice management system applied modifier 59 to line 3
automatically, by rule, because an edit was present and the
edit's modifier indicator was 1. No coder read the operative
report before submission.
THE EDIT FILE ROW (as looked up)
Column One Column Two Effective Deleted Modifier
29827 29822 (date) ---- 1
─────────────────────────────────────────────────────────────────────────
THE TASK. Decide what the claim should contain. If a line comes off, say what happens to the charge. If the claim is denied, say whether it is appealable.
The path
Read the edit row in words first (Chapter 21 §21.2). "29822 is not separately payable when reported with 29827 for the same patient on the same date by the same provider. This edit is active. A modifier MAY override it if the criteria are met."
Now read what the row does not say. It does not say a modifier is appropriate here. The indicator is a permission, not an authorization (Chapter 21 §21.3), and the entire compliance content of the edit system lives in the gap between those two sentences.
| Indicator | What it means |
|---|---|
| 0 | A modifier will not override this edit, under any circumstance |
| 1 | A modifier may override this edit — if the clinical circumstances justify it and the documentation supports it |
| 9 | The edit does not apply; it has been deleted |
So the indicator tells you the system will accept a modifier. Whether the modifier is true is a question for the operative note. Chapter 14 §14.5's discipline is one sentence long: before you append modifier 59, find the sentence in the operative note.
Go find it. The body says the debridement was performed "to expose the tear and to allow visualization of the repair site." That is not a distinct procedural service. That is access. The debridement is in the same anatomic region as the repair, on the same structures, in service of the repair — which is precisely the situation the edit exists to describe, and it falls under the standards of medical and surgical practice that the edit system is built on (Chapter 21 §21.6).
Line 3 comes off the claim. Not appealed later — off, now, before submission.
What about line 2? 29826 is an add-on code with a ZZZ global — it lives inside the primary procedure's global period, it cannot be billed alone, and modifier 51 does not belong on it, because add-on codes are exempt from the multiple-procedure reduction (Chapter 13 §13.7, Chapter 17 §17.2). Reported with 29827, it is correct.
THE CLAIM AS IT SHOULD GO OUT
─────────────────────────────────────────────────────────────
1 29827 RT rotator cuff repair (090 global)
2 +29826 RT subacromial decompression (ZZZ, add-on)
─────────────────────────────────────────────────────────────
Line 3 removed. Not appealed — removed.
And the disposition question, if it went out wrong and denied. Run Chapter 21 §21.11's six-way triage:
NCCI, indicator 0 ........... DO NOT APPEAL. Remove the charge.
NCCI, indicator 1,
documentation supports ..... APPEAL, quoting the Policy Manual
section and the documented
circumstance
NCCI, indicator 1,
documentation does NOT ..... DO NOT APPEAL. Correct the claim.
MUE, MAI 2 .................. DO NOT APPEAL. Find the units error.
MUE, MAI 1 or 3 ............. workable — split the lines, or
appeal with documentation
PROPRIETARY edit ............ GET THE POLICY FIRST, then decide
This claim is the third branch: indicator 1, documentation does not support. Correct the claim. Three of the six branches say do not appeal, and a billing office that internalizes that redirects a great deal of effort from letter-writing to charge correction, which is faster, cheaper, and correct.
One thing to check before you decide any of this: which edit file did you open? The practitioner procedure-to-procedure file and the outpatient hospital file are different files and they differ. Using the wrong one gives you a wrong answer confidently (Chapter 21 §21.2).
The wrong answer, named
Appending modifier 59 because the indicator is 1. This is the error, it is the one the scenario is built on, and it is nearly always made by a system rather than a person. A macro cannot read an operative report. It can only read the indicator, and the indicator is the same on every claim regardless of what the surgeon did.
Substituting an X modifier and calling it more precise. XE, XS, XP, and XU are a more specific subset of 59, and specificity does not manufacture a distinct service (Chapter 14 §14.5). XS — a separate structure — would be a false statement here, because the structures debrided are the structures repaired. Choosing a narrower modifier to describe something that is not true makes the claim worse, not better, because it asserts more.
Appealing the denial. The documentation does not support the override. The appeal will deny, and the second denial is not a failure of the appeal — there is nothing to appeal to.
And the pattern-level error, which is the one that produces an audit rather than a denial. An indicator-1 edit overridden at a rate approaching one hundred percent is a self-report. Indicator 1 exists because a minority of cases genuinely separate; a practice overriding it on essentially every occurrence has asserted that the minority case is its normal case — a claim, visible in claims data without anyone opening a chart, that the practice made without knowing it made it (Chapter 21 §21.3, Chapter 17 §17.9). The measurement is a single query: for each pair you override, what percentage of the time do you override it?
The transferable point
An edit's modifier indicator tells you what the system will accept; the operative note tells you what is true. Before appending 59 or any of its X-subset, point at the sentence — and if you cannot find it, the correct action is to remove the line, not to write a letter about it later.
Where the book teaches it: Chapter 13 §13.7 · Chapter 14 §14.5, §14.6 · Chapter 17 §17.2, §17.9 · Chapter 21 §21.2, §21.3, §21.5, §21.6, §21.8, §21.9, §21.11 · Appendix B §B.4 · Appendix E.
H.11 The screening that became diagnostic
What this tests: intent versus outcome in Z-code sequencing, and the two different modifiers two different kinds of payer look for.
CONSTRUCTED TEACHING EXAMPLE — not a real patient record
─────────────────────────────────────────────────────────────────────────
ENDOSCOPY REPORT AND SCHEDULING RECORD
Hospital outpatient department. Constructed teaching file.
Patient age 64, Medicare beneficiary, average risk, no personal or
family history of colorectal neoplasia, no symptoms.
SCHEDULING RECORD
Reason for visit: screening colonoscopy, average risk.
Last colonoscopy: none on file.
PROCEDURE REPORT
Indication: colorectal cancer screening.
The colonoscope was advanced to the cecum, identified by the
appendiceal orifice and ileocecal valve. Preparation adequate.
A single 8 mm sessile polyp was identified in the sigmoid
colon and removed by snare technique. No other lesions. The
scope was withdrawn. No immediate complications.
Specimen submitted to pathology.
PATHOLOGY (returned three days later)
Colon, sigmoid, polypectomy: tubular adenoma. Benign.
─────────────────────────────────────────────────────────────────────────
THE TASK. Assign the diagnosis codes in order, the procedure code, and the modifier. Then say what changes if the patient is covered by a commercial plan instead.
The path
Start with why the patient came, because that is what the first-listed diagnosis reports. The scheduling record and the indication both say screening. A screening is an intent, not an outcome (Chapter 12 §12.9), and finding something does not retroactively convert the reason the patient walked in the door.
First-listed: Z12.11, encounter for screening for malignant neoplasm of colon. The screening intent survives the polyp.
Then the finding. At the time of service the record documents a polyp of the colon: K63.5. When pathology returns benign three days later, the final coding uses D12.5 — benign neoplasm of the sigmoid colon. Look the site up rather than reaching for the code you used last time: the benign neoplasm category is divided by colonic segment, and a polyp in the descending or transverse colon takes a different code than this one (Chapter 8 §8.4, Chapter 10 §10.4). This is one of the places where waiting for pathology changes a code, exactly as it did in scenario H.8.
Now the procedure, and this is where the money is. The screening became diagnostic the moment therapeutic work was performed. The procedure billed is the snare polypectomy — 45385 — not the screening code. G0121 describes a screening colonoscopy for a beneficiary not meeting high-risk criteria; that is not what happened once a polyp came out (Chapter 18 §18.2, Chapter 20 §20.6).
And the diagnostic colonoscopy is not separately reported. 45378 is the family's base code, and a diagnostic endoscopy is never reported alongside a surgical endoscopy of the same family (Chapter 18 §18.1). One code: 45385.
The modifier is the part that decides what the patient pays. For Medicare, a colorectal cancer screening test converted to a diagnostic test is reported with modifier PT on the diagnostic procedure code. PT is what keeps the screening benefit's cost-sharing protections attached — the deductible is waived, and the coinsurance follows the statutory phase-down (Chapter 34 §34.11).
Without PT, the claim is an ordinary diagnostic colonoscopy with ordinary cost-sharing, and the patient pays for a coding omission. That sentence is the entire reason this scenario is in the book.
For a commercial plan, the vocabulary changes and the logic does not. Under the preventive-services coverage rules, plans generally look for modifier 33 — preventive service — to keep a converted screening inside the zero-cost-sharing preventive benefit. Payer-specific, as always, and worth verifying per plan (Chapter 22 §22.5).
The pathology is its own claim line and often its own claim — 88305, level IV surgical pathology — billed by whoever performed the examination, with its own professional and technical components (Chapter 19 §19.1).
THE ENCOUNTER, ASSEMBLED
─────────────────────────────────────────────────────────────
DIAGNOSES Z12.11 first-listed — why she came
K63.5 the finding at the time of service
D12.5 or the site-appropriate benign neoplasm
code, once pathology returns
PROCEDURE 45385 snare polypectomy
+ PT Medicare: screening converted to
diagnostic
+ 33 commercial: preventive service
PATHOLOGY 88305 separately reported
─────────────────────────────────────────────────────────────
NOT G0121. NOT 45378 in addition. Intent first, finding second.
And a note on the money, because a patient is going to ask. Under the statutory phase-down for a
screening colonoscopy that becomes diagnostic, the beneficiary's coinsurance is being reduced in steps
toward zero by calendar year. On the book's constructed facility example — an illustrative payment rate
of \$1,020.00 for a level 4 endoscopy [constructed — verify current rates at the source] — twenty
percent would be \$204.00 and an illustrative fifteen percent step is \$153.00, a difference of \$51.00
to the patient. Verify the current step before you quote anyone a number. A percentage from last
year is an estimate discipline failing in a new costume (Chapter 32 §32.2, Chapter 34 §34.11).
The wrong answer, named
Reporting G0121. The most common error in this family, and it is sympathetic: the visit was scheduled as a screening, so the coder reports the screening code. The procedure code reports what was done. A polyp was snared; that is a therapeutic colonoscopy.
Dropping Z12.11 and leading with the polyp. This is the mirror error and it is the more expensive one for the patient, because the first-listed diagnosis is where a payer looks to decide whether the preventive benefit applies. Lead with the finding and the encounter stops looking like a screening, the benefit does not attach, and the patient receives a bill for something they were told was free.
Reporting 45378 in addition to 45385. The base code is inside the surgical code. This is the endoscopic base-code family rule and it is not optional (Chapter 18 §18.1).
Reporting four codes for four polyps removed by snare. Not this case, but the single most common colonoscopy error and it runs in the expensive direction: same technique, multiple lesions, one code (Chapter 18 §18.2). Different technique on a different lesion is a different question, subject to the edits.
And omitting PT because the claim will pay without it. It will. The patient pays the difference, and nothing on the remittance will say that anything went wrong.
The transferable point
Diagnosis codes report why the patient came; procedure codes report what was done; and when a screening becomes diagnostic, the claim has to carry both facts at once. The modifier is the only field that says the encounter began as prevention.
Where the book teaches it: Chapter 12 §12.9 · Chapter 18 §18.1, §18.2 · Chapter 19 §19.1 · Chapter 20 §20.6 · Chapter 22 §22.5 · Chapter 32 §32.2 · Chapter 34 §34.3, §34.5, §34.11 · Appendix B §B.12.
H.12 One visit, two claims: the facility and the professional split
What this tests: the distinction this book has kept alive since Chapter 1 — one encounter, two claims — at the point where it confuses patients and billers most.
CONSTRUCTED TEACHING EXAMPLE — not a real patient record
─────────────────────────────────────────────────────────────────────────
HOSPITAL OUTPATIENT CLINIC ENCOUNTER
Ridgeview Regional Medical Center (constructed), on-campus
provider-based internal medicine clinic. Medicare beneficiary,
established patient. Constructed teaching file; illustrative
figures throughout.
WHAT HAPPENED
A thirty-minute office-type visit for management of two chronic
conditions and one new complaint. The physician is employed by
the health system. The room, the nurse, the vital signs, the
supplies, and the medical record system belong to the hospital.
No procedure was performed.
WHAT THE DOCUMENTATION SUPPORTS
Medical decision making at the LOW level: two problems, limited
data, low risk. (Leveled by the rules in Chapter 15.)
WHAT WENT OUT
Two claims, to the same payer, for the same thirty minutes.
─────────────────────────────────────────────────────────────────────────
THE TASK. State what goes on each claim — form, bill type or place of service, code, revenue code where applicable. Then explain to the patient why two bills arrived for one visit.
The path
Two claims, because two entities furnished something (Chapter 1 §1.5, Chapter 16 §16.9).
The professional claim asserts the clinician's work. It goes on a CMS-1500, with place of service 22 for an on-campus outpatient hospital (Chapter 25 §25.8), and it carries 99213 — the established-patient office or other outpatient visit at low medical decision making, selected exactly as it would be anywhere else. What changes is not the code; it is the price. Because the building is not the physician's cost, the practice expense component of the relative value calculation is paid at the facility rate rather than the non-facility rate — the site-of-service differential (Chapter 23 §23.5, Chapter 34 §34.8).
The facility claim asserts the institution's resources, and here Medicare does something the rest of this book has not prepared you for. The hospital does not report an E/M level at all. It reports G0463 — a single HCPCS Level II code for a hospital outpatient clinic visit of any level — on a UB-04, type of bill 131, with revenue code 0510 for the clinic (Chapter 26 §26.3, §26.4, §26.5, Chapter 34 §34.8).
One visit, one leveled code and one flat one. The flat one exists because CMS concluded that facility clinic-visit resources did not vary enough by E/M level to price five ways — the same G-code logic Chapter 20 §20.6 taught, with Medicare inventing vocabulary where CPT's did not fit its payment policy.
ONE ENCOUNTER, TWO CLAIMS
─────────────────────────────────────────────────────────────────
PROFESSIONAL FACILITY
Form CMS-1500 UB-04
Identifier POS 22 TOB 131
Code 99213 G0463
Revenue code — 0510 (clinic)
What it asserts the clinician's the institution's
decision making rooms, staff,
equipment, supplies
Priced through RVUs at the FACILITY the outpatient
practice-expense rate prospective payment
system
─────────────────────────────────────────────────────────────────
Neither is a duplicate of the other.
Now the money [constructed — verify current values at the source]:
| Line | Charge | Allowed | Contractual adj. | Patient | Plan |
|---|---|---|---|---|---|
| Professional — 99213 | 120.00 | 62.00 | 58.00 | 12.40 | 49.60 |
| Facility — G0463 | 310.00 | 145.00 | 165.00 | 29.00 | 116.00 |
| Total | 430.00 | 207.00 | 223.00 | 41.40 | 165.60 |
Checks: 430.00 − 207.00 = 223.00 ✓ · 207.00 − 41.40 = 165.60 ✓ · 12.40 + 29.00 = 41.40 ✓ · 49.60 + 116.00 = 165.60 ✓ (coinsurance shown at twenty percent of the allowed amount on each claim, deductible assumed met).
And the sentence a biller owes the patient, because this is the call that comes in three weeks later: "You had one visit. The doctor's bill covers the doctor's work. The hospital's bill covers the room, the nurse, and the equipment, because this clinic is part of the hospital. They are not the same charge and neither one is a duplicate." Two claims means two adjudications, two coinsurance streams, and two statements arriving weeks apart for one afternoon. The biller who can identify which claim a balance came from, in one sentence, prevents most of the phone call (Chapter 34 §34.8, Chapter 32 §32.10).
The wrong answer, named
Reporting 99213 on the facility claim. The most direct error and it will be rejected or denied by the outpatient code editor. The hospital's claim reports the hospital's resources, and for Medicare clinic visits the vocabulary for that is G0463.
Appending modifier 26 or TC to sort out who is billing what. 26 and TC split a single code's professional and technical components on professional claims (Chapter 19 §19.1). The facility's claim is the institutional side; it does not need a modifier to say so (Chapter 34 §34.8).
Pricing the professional claim at the non-facility rate. A practice that bills the same charge and expects the same allowed amount it gets in a freestanding office will see a variance it cannot explain, and it is not an underpayment — it is the site-of-service differential doing exactly what it is designed to do (Chapter 23 §23.5).
Treating one of the two claims as a duplicate and voiding it. This happens, it happens most often when a patient calls to complain, and it writes off money the organization earned.
And a subtler one worth knowing about. If a procedure had been performed and both claims carried an E/M with modifier 25, the two claims would be leveled by two different rulebooks — the professional side by the national E/M guidelines, the facility side by the hospital's own internal, consistently applied criteria, because CMS has never published national facility E/M leveling criteria. The two levels can legitimately differ, and a payer analyst comparing them for "mismatch" is asking the wrong question. The right question, and the one an auditor asks, is to see the facility's written criteria and evidence that they are applied consistently (Chapter 34 §34.8, Chapter 16 §16.9).
The transferable point
One encounter can generate two claims against one account, and they are not versions of each other — they report different work by different entities under different payment systems. Learn to say which is which in one sentence, because the patient cannot see any of it.
Where the book teaches it: Chapter 1 §1.5 · Chapter 16 §16.9 · Chapter 19 §19.1 · Chapter 20 §20.6 · Chapter 23 §23.5 · Chapter 25 §25.8 · Chapter 26 §26.3, §26.4, §26.5, §26.9 · Chapter 32 §32.10 · Chapter 34 §34.8 · Appendix D.
H.13 An admission: the principal diagnosis, the tier, and the day-three ulcer
What this tests: "after study," the severity tier, and what the present-on-admission indicator does to a code that would otherwise be worth money.
CONSTRUCTED TEACHING EXAMPLE — not a real patient record
─────────────────────────────────────────────────────────────────────────
INPATIENT RECORD SUMMARY
Ridgeview Regional Medical Center (constructed), 310-bed community
hospital. Medicare beneficiary, age 74. Length of stay: five days.
Constructed teaching file.
EMERGENCY DEPARTMENT, DAY 0
Presented with shortness of breath and cough for three days.
Oxygen saturation 84% on room air. Admitting diagnosis recorded
as shortness of breath. Admitted to a medical bed.
HISTORY AND PHYSICAL, DAY 0
Impression: hypoxic respiratory failure. Community-acquired
pneumonia versus heart failure exacerbation. Chest radiograph
shows a right lower lobe infiltrate. Started on supplemental
oxygen and empiric antibiotics.
PROGRESS NOTES
Day 1 Acute respiratory failure with hypoxia — continues on
supplemental oxygen. Community-acquired pneumonia,
organism not identified; cultures negative to date.
Continue antibiotics.
Day 3 Nursing documents a stage 3 pressure ulcer of the sacral
region, newly identified. Skin assessment on admission
documented intact skin. Wound care consulted.
Day 4 Oxygen weaned. Pneumonia improving.
DISCHARGE SUMMARY, DAY 5
Final diagnoses:
Community-acquired pneumonia, organism not identified
Acute respiratory failure with hypoxia, present on admission
Essential hypertension
Hyperlipidemia
Stage 3 sacral pressure ulcer, hospital-acquired
Discharged home with home health.
─────────────────────────────────────────────────────────────────────────
THE TASK. Select the principal diagnosis and justify it. Identify the secondary diagnosis that sets the severity tier. State the present-on-admission indicator for the pressure ulcer and what it does.
The path
The definition, word by word (Chapter 33 §33.2):
The principal diagnosis is the condition established after study to be chiefly responsible for occasioning the admission of the patient to the hospital for care.
"After study" — not the presenting symptom, not the admitting impression. The emergency department recorded shortness of breath; the workup found a pneumonia. The UB-04 holds both honestly: the admitting diagnosis lives in one form locator and the principal in another, and the disagreement between them is the story of the admission, not an error (Chapter 26 §26.2).
So R06.02 is not the principal diagnosis, even though it is what brought the patient in and even though it is a perfectly correct code for the encounter that generated it.
"Chiefly responsible" and "occasioning the admission." The question is why this patient needed to be in a hospital bed. Both the pneumonia and the acute respiratory failure were present on admission, both were worked up, and both were treated — which puts this squarely in the situation Chapter 33 §33.2 warns is not self-executing: two or more conditions that each meet the definition. Where the Guidelines' sequencing instructions, the Tabular's conventions, and the circumstances of admission do not establish one over the other, either may be sequenced first.
That discretion is real and it is narrow. Read this record for what it actually says. The respiratory failure is the manifestation of the pneumonia; the antibiotics are directed at the pneumonia; the oxygen supports the patient while the antibiotics work; and the discharge summary lists the pneumonia first. The record supports the pneumonia as the condition chiefly responsible.
Principal diagnosis: J18.9 — pneumonia, unspecified organism. The organism was not identified and cultures were negative; that is a genuine clinical unknown and the unspecified code is correct here, which is a different situation from scenario H.3's missing laterality.
Now the tier. A CC is a secondary diagnosis CMS has designated as meaningfully increasing the resources a stay consumes. An MCC is the higher tier — conditions of the gravity of acute respiratory failure, acute renal failure, sepsis. Every secondary diagnosis is an MCC, a CC, or neither, per the current year's list, and the lists are revised annually in the inpatient prospective payment system final rule (Chapter 33 §33.3). Codes move on, off, and between tiers every year, so verify against the current Definitions Manual and never against a textbook.
J96.01, acute respiratory failure with hypoxia, documented as present on admission, is the MCC, and the highest-severity secondary on the record decides the tier. One MCC puts the stay in the "with MCC" tier: MS-DRG 193, simple pneumonia and pleurisy with MCC, rather than 194 (with CC) or 195 (without CC/MCC).
And now the pressure ulcer, which is the point of the scenario. L89.153 — pressure ulcer of the sacral region, stage 3 — is documented on day 3, and the admission skin assessment documented intact skin. Present on admission indicator: N. Not present on admission.
A condition that would otherwise raise the severity tier does not function that way when it was acquired in the hospital and sits on the hospital-acquired condition list (Chapter 33 §33.7). Severity the hospital caused is not severity the hospital is paid for. The code still goes on the claim — it is a true fact about the stay and it feeds quality measurement and the case mix index — but it does not move the payment.
Two more rules that decide what else belongs on this record.
Severity does not accumulate. A record with one MCC and a record with one MCC plus six CCs group to the same DRG and pay the same. Once the tier is set, additional CCs change nothing about this claim's payment — though they still belong on the claim where Section III supports them, because case mix index, quality measurement, and risk adjustment all read the full record (Chapter 33 §33.3, §33.6).
And the principal diagnosis cannot be its own CC. The tiers are built from secondary diagnoses, and the grouper applies exclusion logic that keeps a secondary too close to the principal from counting as additional burden.
What the tier is worth. Do not compute it from memory. The structure is relative weight × base
rate, both of which change annually (Chapter 33 §33.5). The book's worked example is Account 22-8891,
where a chronic obstructive pulmonary disease admission's three tiers carry constructed relative weights
of 1.1015, 0.8003, and 0.6555 against a constructed base rate of \$6,200.00
[constructed — verify current weights in the annual IPPS final rule and the hospital's own base rate]:
190 COPD with MCC 1.1015 x 6,200.00 = 6,829.30
191 COPD with CC 0.8003 x 6,200.00 = 4,961.86
192 COPD without CC/MCC 0.6555 x 6,200.00 = 4,064.10
190 - 191 = 1,867.44 the value of one documented phrase
191 - 192 = 897.76
That first difference is the value of writing "acute respiratory failure with hypoxia" instead of "hypoxia," and it is the book's central clinical documentation example (Chapter 33 §33.10).
The wrong answer, named
R06.02 as the principal diagnosis. The admitting diagnosis reported as the principal. It is the error the phrase "after study" exists to prevent, and it is common in facilities where the coder works from a face sheet rather than the record.
Sequencing J96.01 first because it is the MCC. This is the error with a financial motive attached and it is worth being blunt about it: sequencing to reach a payment rather than to describe the admission is not an application of the "either may be sequenced" rule, it is an abandonment of it. The rule permits a choice the record supports; it does not permit a choice the record does not. Chapter 33 §33.10's compliance discussion is about exactly this, and a facility whose sequencing choices in discretionary cases run uniformly toward the higher-weighted option has made a visible pattern claim.
Treating the pressure ulcer as an MCC. It was acquired in the hospital, the present-on-admission indicator says so, and the payment consequence is removed. A coder who assigns present on admission = Y because the discharge summary lists it among the final diagnoses has not read the day-3 note or the admission skin assessment.
Omitting the pressure ulcer entirely because it does not pay. It is a true fact about the stay, it belongs on the claim, and omitting it corrupts the hospital's own quality data.
And assuming that adding more secondary diagnoses raises the payment. Once an MCC is on the record, it does not. Coders who have learned outpatient habits — where each additional diagnosis can support an additional service — routinely expect otherwise.
The transferable point
Inpatient payment turns on two questions that sound similar and are not: what occasioned the admission, and what else was wrong with the patient when they arrived. The present-on-admission indicator is the field that separates the second question from a third one nobody wants to answer — what happened to the patient here.
Where the book teaches it: Chapter 9 §9.4, §9.5 · Chapter 11 §11.5 · Chapter 26 §26.2, §26.3 · Chapter 33 §33.2, §33.3, §33.4, §33.5, §33.6, §33.7, §33.10.
H.14 A remittance to post and five lines to dispose of
What this tests: reading an 835 line by line, and knowing that four denials on one claim can have four different correct next actions, only one of which is an appeal.
CONSTRUCTED TEACHING EXAMPLE — not a real patient record
─────────────────────────────────────────────────────────────────────────
THE CLAIM AS SUBMITTED
Specialty office, commercial PPO, in network. New patient.
Constructed teaching file; illustrative figures throughout.
LINE CODE MOD UNITS CHARGE
1 99204 25 1 295.00
2 20605 LT 1 165.00
3 80053 — 1 68.00
4 82947 — 1 22.00
5 J1030 — 2 18.00
──────
TOTAL 568.00
The record documents a 40 mg dose of methylprednisolone acetate.
The comprehensive metabolic panel and the glucose were drawn from
the same specimen at the same encounter.
─────────────────────────────────────────────────────────────────────────
THE REMITTANCE (835), AS RECEIVED
LINE 1 99204-25 charge 295.00
CO-45 113.00 (charge exceeds the fee schedule)
CO-97 182.00 RARC N19
PAID 0.00
LINE 2 20605-LT charge 165.00
CO-45 77.00
PR-2 17.60
PAID 70.40
LINE 3 80053 charge 68.00
CO-45 53.80
PR-2 2.84
PAID 11.36
LINE 4 82947 charge 22.00
CO-97 22.00 RARC N19
PAID 0.00
LINE 5 J1030 charge 18.00 units billed 2
CO-151 18.00
PAID 0.00
ELECTRONIC FUNDS TRANSFER THIS CLAIM ......... 81.76
PATIENT RESPONSIBILITY CREATED ............... 20.44
─────────────────────────────────────────────────────────────────────────
THE TASK. Post the remittance. Then state, for each denied line, the disposition and the reason — appeal, corrected claim, or remove the charge.
The path
Read every adjustment as three parts: a group code that says who bears it, a reason code that says why, and an amount (Chapter 28 §28.3). The group code is the consequential one, because it decides whether an amount moves to the patient or is absorbed by the practice.
Line 1 first, because it is the largest and the most interesting. It carries two CO adjustments, and reading them in order tells you what the payer did.
CO-45 \$113.00** prices the service: charge \$295.00 less \$113.00 leaves an allowed amount of \$182.00. **CO-97 \$182.00 then removes the whole allowed amount, with RARC N19 — procedure code incidental to the primary procedure.
Read that sequence carefully, because it decides the appeal. The payer priced the service and then declined to pay for it. The claim was adjudicated; the E/M has a contracted value; the payer applied a bundling edit on top of it. Therefore the appealable amount is \$182.00, not \$295.00 — you never had a claim on the \$113.00, which is the contractual spread you agreed to years ago.
And note what the denial does not say (Chapter 28 §28.4). CARC 97 says the benefit is included in another service's allowance. That is a bundling decision, not a statement that the modifier was wrong. Nothing in an 835 tells you what a human looked at.
Disposition of line 1: APPEAL, if — and only if — the office note supports the assertion modifier 25 makes. Go read it before you write anything. If three chronic conditions were separately assessed with plans, if medications were managed, if tests were ordered with stated clinical reasons, and if the new problem carried its own history, examination, and management decision, the documentation demonstrates a significant, separately identifiable service and the appeal has evidence rather than an argument (Chapter 30 §30.2, §30.4). If it does not, do not write the letter.
Lines 2 and 3: post them, and then do the thing almost nobody does. They paid. Check the allowed amount against what the contract says it should be (Chapter 28 §28.8). An underpayment does not deny, does not reject, appears on no exception report, and arrives as a payment — and because the denominator shrinks with the numerator, a silent underpayment raises the net collection rate. The only way to see one is to compare expected allowed with actual allowed, line by line.
Line 4: 82947, denied CO-97 with N19. Do not appeal this one. A quantitative blood glucose is a component of a comprehensive metabolic panel; the panel rule says you report the panel, not the panel plus its parts (Chapter 19 §19.4). Both were drawn from the same specimen at the same encounter. Remove the charge.
And classify the root cause correctly, because the category matters more than the fix (Chapter 29 §29.4). This is not a coding error in the sense of a coder choosing the wrong code. The order entry or the charge capture put both the panel and its component on the encounter, and no amount of coder training fixes an order set. A denial names the line; it does not name the cause.
Line 5: J1030 at two units, denied CARC 151 — the payer deems the information does not support this many services. Do the arithmetic (Chapter 20 §20.3). The descriptor is 40 mg per unit; the record documents a 40 mg dose; 40 ÷ 40 = 1 unit. Two units was wrong.
Disposition: a corrected claim, not an appeal. There is nothing to argue about; the units were overstated and the fix is on the claim, not in a letter (Chapter 29 §29.6). And the corrected claim still owes an answer to a question the original never addressed — whether anything was discarded from a single-dose container, which is JW or JZ (Chapter 20 §20.4).
Check the units before you assume the disposition, though, because CARC 151 has two branches that run in opposite directions. Either you billed the wrong units — this case — or you billed the right units and the payer's limit is wrong, which is a genuinely appealable denial supported by the documentation and, if it recurs, by the medically unlikely edit's adjudication indicator (Chapter 20 §20.3, Chapter 21 §21.4, Appendix E). The arithmetic decides which branch you are on, and it takes ten seconds.
Now the posting itself, and the trap in the third column. Here is what the claim should have earned,
had lines 1, 4, and 5 been right [constructed — verify current values at the source]:
| Line | Charge | Allowed | Contractual adj. | Patient | Plan |
|---|---|---|---|---|---|
| 99204-25 | 295.00 | 182.00 | 113.00 | 36.40 | 145.60 |
| 20605-LT | 165.00 | 88.00 | 77.00 | 17.60 | 70.40 |
| 80053 | 68.00 | 14.20 | 53.80 | 2.84 | 11.36 |
| J1030 (1 unit) | 18.00 | 7.10 | 10.90 | 1.42 | 5.68 |
| Total | 546.00 | 291.30 | 254.70 | 58.26 | 233.04 |
Checks: 546.00 − 291.30 = 254.70 ✓ · 291.30 − 58.26 = 233.04 ✓ · 36.40 + 17.60 + 2.84 + 1.42 = 58.26 ✓ · 145.60 + 70.40 + 11.36 + 5.68 = 233.04 ✓ (coinsurance at twenty percent of the allowed amount; deductible assumed met; 82947 does not appear, because it should never have been on the claim).
Against that, the remittance paid \$81.76 and created \$20.44 of patient responsibility. Check: 70.40 + 11.36 = 81.76 ✓ and 17.60 + 2.84 = 20.44 ✓.
⚠️ The most damaging thing you can do with this remittance is post all three CO amounts as contractual adjustments.
A contractual adjustment and a write-off are two different decisions (Chapter 28 §28.6). The CO-45 amounts are genuine contractual adjustments — the spread you agreed to. The CO-97 on line 1, the CO-97 on line 4, and the CO-151 on line 5 are not. They are open questions: one to appeal, one to remove, one to correct. Posting them as adjustments closes all three silently, and a contractual adjustment is exactly what a contractual adjustment looks like on every report the practice runs.
A rule that says any CO becomes a contractual adjustment can be correct when it is written and stop being correct when a payer's edits change, with no event inside the practice to trigger a review.
And one patient-facing detail. The \$20.44 becomes a patient balance. If anything was collected at check-in, the statement must show it as a credit, or the practice will ask a patient for money they already handed over (Chapter 28 §28.11, Chapter 32 §32.6).
The wrong answer, named
Appealing all four denials. The instinct is to fight everything, and it converts a billing office into a letter-writing office. Only one of these four is appealable. Chapter 21 §21.11's triage has three branches out of six that say do not appeal, and this remittance is a demonstration of why.
Appealing \$295.00 on line 1. The payer priced the service. You are owed the allowed amount, not the charge, and an appeal demanding the charge tells the reviewer you have not read your own contract.
Correcting the claim on line 1. There is nothing to correct. The coding is right; the payer applied a policy. A denial reason code tells you what the payer did, not what you did wrong — and correcting a claim that is already correct converts an appealable denial into a resubmission that will deny again.
Rebilling line 4 with modifier 91. Modifier 91 identifies a repeat clinical diagnostic laboratory test — a second, separate test on a separate specimen to obtain subsequent results (Chapter 19 §19.7). This was one specimen and one draw. Appending 91 to defeat a panel edit is a false statement, and it is the exact behavior Chapter 21 §21.8 forbids.
Writing off line 5 as a contractual adjustment. Eighteen dollars is below most practices' review thresholds, which is precisely why this error survives: it is individually too small to look at and collectively large enough to matter.
The transferable point
Four denials on one claim can require four different actions, and identifying which is which takes about ninety seconds per line. Read the group code, then the reason code, then the record — and never let a posting rule make the decision for you.
Where the book teaches it: Chapter 19 §19.4, §19.7 · Chapter 20 §20.3, §20.4 · Chapter 21 §21.8, §21.11 · Chapter 28 §28.3, §28.4, §28.5, §28.6, §28.8, §28.11 · Chapter 29 §29.1, §29.4, §29.5, §29.6 · Chapter 30 §30.1, §30.2, §30.4 · Chapter 32 §32.6 · Appendix E.
H.15 Six conditions on a problem list, three on the claim
What this tests: whether a condition was addressed at this encounter — the question risk adjustment turns on, and the one most charts answer badly.
CONSTRUCTED TEACHING EXAMPLE — not a real patient record
─────────────────────────────────────────────────────────────────────────
OFFICE NOTE — primary care, established patient, age 70
Medicare Advantage member. Constructed teaching file.
PROBLEM LIST (carried forward, reviewed)
Type 2 diabetes mellitus
Chronic kidney disease, stage 3a
Chronic diastolic heart failure
Hyperlipidemia
Osteoarthritis, right knee
History of cholecystectomy, remote
HISTORY OF PRESENT ILLNESS
Here for chronic disease follow-up. Reports mild ankle swelling in
the evenings, improved from last visit. Knee is doing well since
the injection three months ago; no pain today. Taking insulin as
prescribed.
MEDICATIONS
insulin glargine 24 units nightly
furosemide 20 mg daily
atorvastatin 40 mg daily
ASSESSMENT AND PLAN
1. Type 2 diabetes mellitus with diabetic chronic kidney disease,
stage 3a — A1c 7.4, estimated GFR 51 and stable. Continue
insulin glargine 24 units nightly. Avoid nephrotoxic agents.
Repeat metabolic panel in three months.
2. Chronic diastolic heart failure — compensated. Mild evening
edema, improved. Continue furosemide 20 mg daily. Daily weights
reviewed with the patient.
3. Hyperlipidemia — continue atorvastatin.
Nothing further documented regarding the knee or the surgical
history at this encounter.
─────────────────────────────────────────────────────────────────────────
THE TASK. Decide which conditions are reportable from this encounter and code them. Say what happens to the ones that are not.
The path
Ask one question of every condition on the record: did the clinician do work about it, at this encounter, and does the note show it? The industry mnemonic for the evidence that answers yes is MEAT — Monitor, Evaluate, Assess or address, Treat — and one of the four is enough (Chapter 36 §36.7).
And know what MEAT is. It is not a federal regulation and there is no CMS rule that says the word. It is a teaching device the risk-adjustment community built to operationalize requirements that are binding: Section IV's instruction to report conditions that coexist and require or affect patient care, treatment, or management; the general documentation standards; and the payment rules' requirement of a face-to-face encounter with an assessment. Some organizations use a different mnemonic entirely. Use whichever yours uses, and know that the authority is the Guidelines and the payment rules, not the acronym — a coder who cites MEAT to a physician as though it were law has invited a challenge they cannot answer.
Now walk the list.
Diabetes with chronic kidney disease — reportable, and this is the one to get right. The assessment monitors it (A1c 7.4, estimated GFR 51), assesses it (states a status), and treats it (insulin continued). That is three of the four. And the provider has documented the linkage explicitly — "type 2 diabetes mellitus with diabetic chronic kidney disease" — which means the combination code is supported without needing the "with" presumption at all.
E11.22 — type 2 diabetes mellitus with diabetic chronic kidney disease — plus N18.31 for the stage, under the "use additional code" instruction (Chapter 8 §8.6, Chapter 11 §11.8).
And the insulin. The record documents long-term insulin use, which is reported additionally: Z79.4 (Chapter 12 §12.9).
Chronic diastolic heart failure — reportable. Monitored (edema, daily weights), assessed (compensated, improved), treated (furosemide continued). I50.32.
Hyperlipidemia — reportable. Treated: atorvastatin continued with a documented decision to continue. E78.5. Thin, but it is one of the four and one is enough.
Osteoarthritis of the knee — NOT reportable from this encounter. The history of present illness mentions it — "knee is doing well since the injection three months ago; no pain today" — and the assessment says nothing about it. This is the most common near-miss in real charts and it is worth saying slowly: the history is not the assessment. A condition narrated in the history of present illness, the past medical history, or the review of systems has been described. MEAT lives where the clinician states what they concluded and what they are doing. The information exists and it is in the wrong section of the note.
History of cholecystectomy — not a current condition. A status or history code may describe the patient, but nothing at this encounter was affected by it, and Section IV directs you not to report conditions previously treated that no longer exist or that have no bearing on the current encounter (Chapter 9 §9.3).
THE ENCOUNTER'S DIAGNOSES
─────────────────────────────────────────────────────────────
E11.22 diabetes with diabetic CKD M, A, T documented
N18.31 CKD stage 3a "use additional code"
I50.32 chronic diastolic HF M, A, T documented
Z79.4 long-term insulin use documented
E78.5 hyperlipidemia T documented
─────────────────────────────────────────────────────────────
NOT reported this encounter:
osteoarthritis of the knee ► HPI only, not assessed
history of cholecystectomy ► no bearing on this encounter
Why this matters beyond the claim, stated carefully. In a risk-adjusted payment arrangement, the conditions reported from face-to-face encounters during the year describe the patient the plan is being paid to care for. Chronic conditions do not carry forward on their own — the model resets annually, and a condition that is real and permanent still has to be documented and reported again, from an encounter, in the year it is being counted (Chapter 36 §36.6). A knee that was assessed at the last visit and merely mentioned at this one contributes nothing from this one.
And the version-proof way to say what E11.22 does that E11.9 would not. Diabetes without complication and diabetes with chronic complications go to different categories in every version of these models, and the with-complications category carries the larger coefficient. It is a smaller coefficient versus a larger one — not nothing versus something, which is a claim a student could disprove from the published model tables. Never state a specific category number, coefficient, or mapping as current; the model version in force differs by contract year, and even whether a given stage of kidney disease is itself payment-eligible has differed between versions (Chapter 36 §36.3, §36.4).
The wrong answer, named
Reporting everything on the problem list. Six conditions in, six codes out. The problem list is not documentation — it is a list, and a condition on a list has been recorded rather than addressed (Chapter 36 §36.7). This is the single most common risk-adjustment error, it is what retrospective chart reviews are looking for, and it is a false statement about what happened at the encounter.
Reporting the osteoarthritis because it is obviously still true. It is still true. It was not addressed here. "I know it is true" is not a documentation standard, and the fix is upstream — a conversation about where information belongs in a note, which is Chapter 38's territory, not a coding decision.
Reporting E11.9 instead of E11.22. Not wrong in the sense of being false, but incomplete as a description of the patient, and the difference between those two statements is worth money. The assessment states the linkage; the specific code is available and supported; there is no reason to use the less specific one (Chapter 36 §36.11).
Querying to have the knee added to the assessment. Do not. A query asking a physician to document work they did not do at an encounter is what Chapter 4 §4.9 forbids, and the fact that a payment model would reward it makes it worse rather than better.
And the opposite error, worth naming because it is the overcorrection: reporting nothing beyond the visit's chief reason, on the theory that a coder cannot infer anything. Three conditions here have explicit monitoring, assessment, and treatment documented in the assessment and plan. Under-reporting what the note supports is as inaccurate as over-reporting what it does not, and it distorts the same data in the opposite direction.
The transferable point
Ask of every condition: what did the clinician do about this, today, in writing? One documented instance of monitoring, evaluating, assessing, or treating is enough — and a condition that appears only in the history of present illness has been mentioned, not managed.
Where the book teaches it: Chapter 4 §4.9 · Chapter 8 §8.6 · Chapter 9 §9.3, §9.7 · Chapter 11 §11.8 · Chapter 12 §12.9 · Chapter 36 §36.3, §36.4, §36.6, §36.7, §36.11 · Chapter 38.
H.16 The note that cannot be coded
What this tests: the hardest professional judgment in the book — recognizing that the correct answer is not a code at all.
CONSTRUCTED TEACHING EXAMPLE — not a real patient record
─────────────────────────────────────────────────────────────────────────
FOUR CONSECUTIVE OFFICE NOTES, ONE PATIENT, ONE PROVIDER
Constructed teaching file. Patient age 52. Visits four weeks apart.
The coder receives all four in one work queue on the same day.
WHAT IS IDENTICAL ACROSS ALL FOUR NOTES
Chief complaint ......... identical text
History of present
illness ............... identical text, word for word,
including "symptoms began three
weeks ago"
Review of systems ....... identical, 10 systems, all negative
Physical examination .... identical, including the same vital
signs to the decimal: BP 128/76,
HR 72, weight 184.6 lb
Assessment and plan ..... identical, including "return in four
weeks"
WHAT DIFFERS ACROSS THE FOUR NOTES
The date of service.
The electronic signature date.
ALSO IN THE RECORD
Visit 3's nursing intake, entered separately, records weight
191.2 lb and a blood pressure of 148/92, and a note that the
patient "reports the new medication is making her dizzy."
None of that appears in the physician note for visit 3.
WHAT THE BILLING SYSTEM SHOWS
All four visits were charged as 99214 by a charge-capture
default. No coder has reviewed any of them.
─────────────────────────────────────────────────────────────────────────
THE TASK. Code these four encounters. If you cannot, say what you do instead, and say it precisely enough that someone could act on it tomorrow morning.
The path
Start where the book has started forty times: if it isn't documented, it didn't happen — and its corollary, which this scenario is about. A coder does not know what the provider did; a coder knows what the provider wrote. Here the coder knows what the provider wrote four times, and the four are the same document.
Work the level anyway, for one moment, to see what happens. Problems addressed: whatever the identical assessment says. Data: whatever the identical plan orders. Risk: whatever the identical medication list supports. You will get an answer, and the answer will be the same for all four visits, including the visit at which the patient's weight was up seven pounds, her blood pressure was 148/92, and she reported a side effect.
That is the proof. A leveling method that produces the identical result for a visit where something clearly changed is not measuring the visit. It is measuring the template.
Name what this is. Documentation carried forward from a prior encounter without being updated to describe the current one — cloned documentation, and it is one of the named errors (Chapter 4 §4.6, Chapter 5 §5.8). The identical vital signs to the decimal are the tell. Two visits four weeks apart do not produce identical weights and blood pressures, and a review of systems that is negative in ten systems on four consecutive occasions with no variation is a copied paragraph rather than an examination.
Now the internal contradiction, which is the more serious half. Visit 3's nursing intake records a different weight, a different blood pressure, and a reported medication side effect. The physician note for that visit contains none of it and contradicts two of it. The record now says two incompatible things about the same afternoon.
So what do you do?
1. You do not code these from this documentation, and you say so in writing. Not "these are level 3 instead of level 4." The level is not the problem. There is no reliable way to know what happened at visits 2, 3, and 4, and a code is an assertion about what happened. Assigning any level would be asserting a fact the record does not support.
2. You do not hold them silently. A held charge is invisible; a documented hold is a control. Write down what you found, on which dates, and what you need — the same discipline the book recommends when a correct analysis is declined (Chapter 29 §29.9). The dates matter, because timely filing is running (Chapter 27 §27.7).
3. You raise it, and you raise it to the right person. This is not a query. A query resolves a gap in a single encounter's documentation (Chapter 4 §4.9); this is a pattern across four encounters with an internal contradiction inside one of them, and the questions it raises — whether the note describes care that was delivered, whether the template is doing the documenting, whether other patients' notes look like this — are compliance and clinical documentation questions, not coding questions. Compliance officer and clinical documentation integrity, today, and your organization's policy tells you the route (Chapter 5 §5.9).
4. You flag the charge-capture default separately, because it is its own finding. Four visits were charged 99214 by a system rule, with no coder review. That is a configuration making an assertion nobody chose, and this book has shown a dozen versions of it. The claims have not gone out yet, which is the only piece of good news here.
5. And you say the thing that is uncomfortable and true: the fix is not a coder's fix. What belongs in a note is the provider's decision and the organization's policy. Recognizing that a note does not support a code is instruction. Telling a provider what to add to a note so that it will is not, and the line between those two sentences is the one that keeps people employable.
What can be salvaged? Possibly visit 1 — it is the origin document, it has no earlier note to have been copied from, and it may be an accurate description of a real encounter. Code it on its own merits, from its own content, and treat the other three as unresolved. That is a partial answer and it is an honest one.
The wrong answer, named
Coding all four as 99214 because the documentation supports 99214. It does, on its face, four times. That is exactly the problem — and it is the answer a coder gives who has learned to score a note without asking whether the note describes an encounter. Every element is present, and the presence of every element in four identical copies is evidence against the note, not for it.
Downcoding all four to 99213 to be safe. The single most tempting wrong answer in this appendix, and it is wrong on three counts. It is inaccurate, because nobody knows what these visits were. It underpays the practice for whatever real work was done. And it converts a documentation problem into a revenue problem and hides it, which means nothing gets fixed and the notes keep coming (Chapter 5 §5.8). Downcoding is not a defense and it is not conservatism.
Querying the provider to update the notes. Amendments and late entries have their own rules and their own audit trail (Chapter 4 §4.5), and a coder asking a physician to rewrite four notes to support charges is asking for something the book forbids in every chapter that touches it.
Coding visit 3 from the nursing intake. The nursing documentation is real and it is part of the record, but it is not the physician's evaluation and management service and it cannot substitute for one.
And doing nothing. Sending the charges back to a work queue with no note, no date, and no owner is the failure mode this book has named in a dozen case studies: a report is not a control; a person who reads a report is a control. The four charges will age, someone will eventually release them to keep the accounts receivable clean, and the record of the concern will not exist.
The transferable point
Sometimes the correct output of a coding review is not a code. A note that could have been written without the patient in the room does not describe an encounter, and no amount of scoring skill converts it into one — the answer is to stop, document what you found, and route it to the people whose decision it actually is.
Where the book teaches it: Chapter 4 §4.4, §4.5, §4.6, §4.7, §4.9 · Chapter 5 §5.1, §5.8, §5.9 · Chapter 15 §15.12 · Chapter 27 §27.7 · Chapter 29 §29.9 · Chapter 37 §37.3 · Chapter 38.
H.17 Where to go from here
These sixteen are a beginning, not a curriculum. Sixteen charts will not make anyone a coder, and they are not meant to. What they are meant to do is show you the shape of the reasoning often enough that you start producing it yourself — main term, subterm, verify in the Tabular, read the conventions, check the guidelines, check the edits, and then say out loud what a reasonable person would code instead and why it fails.
The book's own practice material is much larger than this appendix, and it is organized by topic rather than by scenario:
- Every chapter's
exercises.mdcarries twenty-five to forty items, graduated and mixed — recall, applied reasoning, a "code this chart" task from constructed documentation, a "complete the claim" field-by-field drill, a "work this denial" task, an "audit this claim" diagnostic, a written query or appeal letter, an ethics dilemma, and certification-style questions. Items marked with a dagger get worked solutions in the answers appendix. - Every chapter's
quiz.mdis twenty to thirty self-check questions with an answer key. - Every chapter's
🔢 Code Itcallouts are shorter versions of what you have just read, embedded at the point where the rule is taught. They are the best single index to the reasoning in this book, and working through them chapter by chapter is a better use of a week than re-reading anything. - Appendix C is the progressive project's workbook, blank, so you can code Account 10-4471 yourself rather than watch it being coded.
And the reference appendices are meant to be open while you work these. Appendix A for the code-set structures and the conventions, Appendix B for what a modifier asserts, Appendix D for the claim form fields, Appendix E for what to do about a denial code, Appendix G for documentation requirements and query templates, and Appendix I if a certification exam is the next thing on your calendar.
One last time, because it is the habit that outlasts every code in this appendix. ICD-10-CM changes October 1. CPT changes January 1. HCPCS Level II and the edit files change quarterly. The fee schedules and payment rates change annually, by rule. Local coverage policies change continuously. Every specific code, descriptor, edit, indicator, weight, rate, and percentage above will eventually be wrong. The path will not be. That is why these scenarios show it, and it is why the answer to "which code?" is always, in the end, "here is how I found out."