Chapter 35: Specialty Coding: Cardiology, Orthopedics, OB/GYN, Pediatrics, Emergency Medicine, and Anesthesia

"Nobody hires you because you know cardiology. They hire you because you know how to find out what cardiology does differently, and you can say what you do not know yet without flinching." — constructed

Overview

Start with a Monday. You have thirty-four chapters of this book behind you, a certification course half finished, and a job offer from a nine-physician cardiology group. On your first morning somebody puts a cardiac catheterization report in front of you — six pages, three vessels, contrast volumes, a diagram of the coronary tree — and you realize that the code sets you have learned describe this document perfectly and you cannot read it.

That is what specialty coding actually feels like, and the honest news is better than it looks. There is no separate code set for cardiology. There is no orthopedic ICD-10-CM, no pediatric CPT®, no anesthesia HCPCS Level II. The three books are the same books, the Official Guidelines are the same Guidelines, the National Correct Coding Initiative (NCCI) edit file is the same file, and every modifier rule from Chapter 14 still holds. What changes in a specialty is narrower and more specific than the anxiety suggests: a small slice of the code set that carries almost all the volume, one or two payment conventions peculiar to the work, a source document that is not an office note, a concentrated body of payer policy, and a vocabulary of documented words that decide codes.

This chapter takes six specialties plus behavioral health and works each one to that level — not to make you a specialist, which takes years, but to show you what the pattern looks like enough times that you can find it yourself in the seventh. Cardiology's bundled catheterization families and its professional/technical split (§35.2). Orthopedics' fracture-care fork and its life inside global periods (§35.3). The obstetric package that spans nine months and the services that fall outside it (§35.4). Pediatrics' preventive visits and the two-code vaccine habit (§35.5). The emergency department from both sides of Chapter 34 §34.8's split, including the facility acuity leveling that has no national rulebook (§35.6). Anesthesia's units and its clock (§35.7). The time-based code families and the documentation time requires (§35.8). Telehealth, which is not a specialty but crosses all of them and changes more often than anything else in this book (§35.9). And then §35.10, which is the chapter's real deliverable: a two-week method for entering a specialty you have never coded.

In this chapter, you will learn to:

• Separate what actually differs by specialty from what only feels different
• Navigate a catheterization family and decide the component question
• Make and document the orthopedic fracture-care decision
• Report an obstetric episode whole, or in the pieces actually furnished
• Keep the vaccine product and its administration together, every time
• Level an emergency department facility claim against written criteria
• Build an anesthesia claim from the anesthesia record, minute by minute
• Match a time-based code to the right time convention
• Report telehealth structurally and verify every coverage rule at its source
• Enter a new specialty in two weeks, and know what that does not buy

35.1 What "specialty coding" actually means

Start with the decision on the desk. You have been hired into a specialty. You have one week before you are expected to produce, and a shelf of general knowledge that took you eight months to build. The question is not "what do I not know?" — the answer to that is everything — but the far more useful one: which of the things I already know still work here, and which of them will get me in trouble?

What transfers, unchanged. All of it, structurally. The Alphabetic Index and the Tabular List still govern diagnosis lookup (Chapter 8 §8.1). The Official Guidelines still decide sequencing (Chapter 9). CPT's section guidelines and parenthetical notes still carry the instructions (Chapter 13 §13.4, §13.6). Modifier 25 still means significant and separately identifiable (Chapter 14 §14.4), and 59 still means distinct (§14.5). The NCCI procedure-to-procedure edits and medically unlikely edits (MUEs) still run (Chapter 21). Medical necessity is still a coverage word (Chapter 22 §22.1). The claim is still a CMS-1500 or a UB-04, and the money still comes from a contract. A coder who is good at the general work is most of the way to being good at the specialty work, and the largest single mistake new specialty coders make is assuming otherwise and abandoning their own discipline at exactly the moment it is most valuable.

What is genuinely specialty-specific is five things, and the whole chapter is an inventory of them:

WHAT ACTUALLY DIFFERS IN A SPECIALTY

  [1] THE VOLUME SLICE       Twenty codes are most of the work. In a general
                             practice the codes are broad and shallow; in a
                             specialty they are narrow and deep. You will code
                             the same eight procedures a hundred times a month
                             and meet the tail twice a year.

  [2] THE PAYMENT CONVENTION What is the unit being purchased?
                               a service ............ most of this book
                               a session ............ therapy, behavioral health
                               an EPISODE ........... obstetrics, fracture care
                               a UNIT OF TIME ....... anesthesia, timed therapy
                               an ACUITY LEVEL ...... ED facility claims
                               a CALENDAR WINDOW .... transitional care
                             Get this wrong and every claim is wrong the same way.

  [3] THE SOURCE DOCUMENT    Not an office note. A catheterization report. An
                             operative report. An anesthesia record. A therapy
                             flowsheet. An immunization record with a funding
                             source in it. Learn its layout before its content.

  [4] THE POLICY LANDSCAPE   A specialty's coverage exposure concentrates in a
                             handful of services. Three or four local coverage
                             determinations and two commercial medical policies
                             govern most of the money.

  [5] THE DOCUMENTED WORDS   "Complete." "With manipulation." "Permanent
                             recording and report." "Counseling provided by the
                             physician." Words in the record that decide codes —
                             and whose absence is not a small problem.
                                                       [schematic — not to scale]

Read [2] again, because it is the one that produces catastrophic rather than merely annoying errors. Every other item on that list produces a wrong code on one claim. The payment convention produces a wrong shape on every claim: an obstetric practice that bills each prenatal visit as an office visit is not making a coding error, it is running the wrong business model through the claim system, and it will do so for months before anything denies loudly enough to be noticed.

The five-question intake

Here is the intake, and it is the spine of §35.10's two-week method. Before you touch a claim in a specialty you have not worked in, answer these in writing:

1. What is the source document, and who signs it?
2. What is bundled into what? Which family rule governs — an endoscopic base code (Chapter 18 §18.1), a surgical package (Chapter 17 §17.1), a catheterization's included components (§35.2), an episode package (§35.4)?
3. What is the unit of payment? Service, session, episode, unit of time, acuity level, calendar window.
4. Which policy files govern the top services? The NCCI Policy Manual chapter for that CPT section (Chapter 21 §21.5), the Medicare Administrative Contractor's (MAC's) local coverage determinations (LCDs) and billing-and-coding articles (Chapter 22 §22.4), and the two or three largest commercial plans' medical policies (Chapter 22 §22.10).
5. Which twenty codes are the volume, and which twenty are the dollars? They are different lists.

Two more facts belong in this section because they are what a specialty does to your calendar.

The update cycle hits specialties unevenly. ICD-10-CM changes every October 1. CPT changes every January 1. HCPCS Level II changes quarterly, and so do the NCCI edits and the MUE values. That is the general rhythm this book has repeated since Chapter 6 §6.7 — but in a specialty it arrives in lumps. Most years a given CPT section barely moves; in the year its section is restructured, half of what you know about that family expires at once, and every reference published before the change becomes actively wrong rather than merely stale. Chapter 18 §18.4 named that exact event for cardiac catheterization and Chapter 18 §18.11 named it again for moderate sedation. Code from the current book or encoder, never from a textbook — including this one.

And the specialty's incentive structure is not neutral. A specialty practice's revenue is concentrated in a few families, which means a single convention applied one notch too generously, across a year, is a large number — and a single convention applied one notch too conservatively is an equally large number in the other direction. Chapter 5 §5.8's rule holds with no softening: upcoding and downcoding are both errors, and in a specialty they are both errors at volume.

🎓 Exam Watch

The certification exams do not test specialty expertise. They test whether you know the conventions that make a specialty different — which is precisely what this chapter is.

The Certified Professional Coder (CPC) exam samples across CPT's sections, and the specialty questions are conventions questions: the catheterization scenario that lists the components separately (Chapter 18 §18.4), the fracture scenario that turns on who is providing definitive care, the obstetric scenario where the patient transferred care at 28 weeks, the vaccine scenario missing its administration line, the anesthesia scenario that gives you room time and anesthesia time and hopes you use the wrong one. The Certified Coding Specialist (CCS) and the Certified Outpatient Coder (COC) push the facility side, including emergency department acuity. AAPC also maintains a family of specialty credentials — cardiology, orthopaedic surgery, obstetrics and gynecology, pediatrics, emergency department, and anesthesia and pain management among them — taken after a core credential rather than instead of one; Chapter 39 §39.3 covers the landscape.

The trap in the stem is almost always an included component listed as though it were a separate service. When a specialty scenario reads like a list, ask what the primary code already contains before you report anything from the list.

35.2 Cardiology: catheterization families and the component question

Cardiology is where two of this book's general rules do their hardest work at once: a procedural family that bundles far more than it looks like it bundles, and a professional/technical split that decides who may bill what. Get both right and most of the specialty's claims are ordinary. Get either wrong and you are either unbundling or billing somebody else's service.

The component question, first, because it decides more claims

Chapter 19 §19.1 built the four arrangements — a global service, a professional component, a technical component, and the two billed by different entities. Cardiology is where you will use them every day, because its highest-volume diagnostic services are exactly the ones that split.

The electrocardiogram (ECG) is the cleanest example in the book.

• 93000 — electrocardiogram, routine, with interpretation and report. The whole service: somebody's machine produced the tracing and somebody wrote and signed an interpretation, and both somebodies work for the entity billing the code.
• 93010 — interpretation and report only. The cardiologist read a tracing that was produced on equipment they do not own, in a place they do not bill for.
• And a third code in the family reports the tracing alone, without the interpretation — the technical half, billed by whoever owns the machine.

Notice what decides among the three: not what happened to the patient. The same twelve leads, the same twelve seconds, the same paper. What decides is who owned the equipment and who wrote the report — a fact about business arrangements, not about medicine. This is Chapter 19 §19.1's point with a specialty's volume behind it, and it is why a cardiology group that reads hospital ECGs and a cardiology group that runs its own ECG room bill the same clinical service with different codes forever.

And "interpretation and report" is a documentation requirement, not a courtesy. It means a separate, retrievable, signed written interpretation — the findings, the conclusion, the reader, the date. A number scrawled on the tracing is not a report. Chapter 17 §17.7 made the identical point about 20611's "permanent recording and report," and Chapter 14 §14.7 made it about the professional component generally: no report, no professional service. A cardiology practice that reads three hundred tracings a month and produces no retrievable interpretations has three hundred claims it cannot defend.

Do not assume a code splits. Some do, some do not, and the authority is the Medicare Physician Fee Schedule's own indicator for each code rather than intuition — Chapter 19 §19.1 taught how to read it, Chapter 23 §23.6 taught what it does to the price. Look it up per code. In this specialty, per code, every time.

The transthoracic echocardiogram carries the same split and adds a word. 93306 describes a complete transthoracic echocardiogram with spectral Doppler and color flow Doppler. The load-bearing word is complete — the descriptor defines a set of elements the study must include and the report must document. A limited or follow-up study is a different code, and the difference is not the machine's setting; it is what the report says was examined. The code is chosen from the report's content, and the report's content is not a coder's judgment. Where the study is genuinely limited, report the limited code; where the study is complete and the report is thin, that is a documentation problem, and Chapter 38 §38.3 owns the query that addresses it compliantly.

The catheterization families

Chapter 18 §18.4 froze the rule and this section applies it. Cardiac catheterization codes were restructured so that the codes now include the catheter placement, the injection procedures, and the imaging supervision and interpretation that older references report separately. If a reference tells you to bill those components separately, that reference predates the restructuring.

What you actually need in front of a catheterization report is a two-axis question:

   THE CATHETERIZATION QUESTION — two axes, one code

     AXIS 1 — WHAT WAS CATHETERIZED?
        left heart · right heart · both

     AXIS 2 — WHAT WAS IMAGED, AND WAS IT INCLUDED?
        coronary arteries          ─┐
        bypass grafts               ├─ combinations are BUILT INTO
        left ventriculography      ─┘   the code, not added to it

     93458 = LEFT HEART CATHETERIZATION with CORONARY ANGIOGRAPHY,
             including intraprocedural injection(s) for coronary
             angiography and imaging supervision and interpretation.

     The words "including ..." in the descriptor are the whole lesson.
     Everything after them is already paid for.
                                                 [schematic — not to scale]

93458 is the family's workhorse and the one to learn first: a left heart catheterization with coronary angiography, with the injections and the imaging supervision and interpretation inside it. The neighboring codes in the family answer the same two questions differently — a right heart catheterization instead, both sides, a study that included bypass grafts, a study that did or did not include left ventriculography. You do not memorize the family. You read the report, answer the two questions, and take the descriptor that matches both.

Two more habits carry the specialty:

Interventions are frequently coded per vessel or per territory, with add-on codes for additional vessels (Chapter 13 §13.7's add-on rules govern them). Which means the operative report must name the vessel — Chapter 17 §17.9's compartment convention, in a different anatomy. A report that says "stent placed" without naming the vessel has not documented a billable specificity, and the query is about anatomy, not about payment.

Device families follow one pattern — insertion, removal, replacement, repositioning, and programming or evaluation — and the code depends on which components were involved: the generator, the leads, or both. Ask which components, always (Chapter 18 §18.4).

📋 Read the Chart

text FIGURE 35.1 — "Six pages, one code" [constructed teaching example] THE DOCUMENT Cardiac catheterization report, hospital catheterization laboratory, signed by the performing cardiologist the same day. Six pages: indications, access, hemodynamics, angiographic findings by vessel, a coronary diagram, conclusions and plan. THE CONTEXT A nine-physician cardiology group. The physicians perform in the hospital's laboratory; the group owns no equipment there and employs none of the staff. WHAT IT SHOWS Femoral arterial access. A catheter advanced to the left ventricle; left ventricular pressures recorded. Selective injections of the left main, left anterior descending, circumflex, and right coronary arteries, with cine acquisition. Findings described vessel by vessel. A signed interpretation and conclusion. No intervention performed; the plan is medical therapy and a return visit. WHAT IT DOESN'T It does not tell you who owns the room, the equipment, or the contrast — and that is the fact the component question turns on. It does not state a ventriculogram was performed (pressures are not an angiogram). And it does not describe any intervention, so nothing in the intervention families is in play. THE DECISION One code from the catheterization family, selected on the two axes: LEFT heart, CORONARY angiography — 93458. Do NOT separately report the catheter placement, the injections, or the imaging supervision and interpretation; the descriptor already contains them. The hospital bills the facility resources on its own claim (Chapter 34 §34.8). THE LESSON A long document does not mean a long claim. In this family the descriptor absorbs the components, and the length of the report is a measure of the work, not of the number of codes.

🔢 Code It

The documentation (constructed teaching example): a cardiologist in the group's office reviews a twelve-lead tracing performed that morning in the hospital's emergency department, on the hospital's equipment, at the emergency physician's request. The cardiologist dictates a signed interpretation — rate, rhythm, axis, intervals, ST-segment findings, comparison with a prior tracing, and a conclusion — which is filed in the hospital record and in the group's chart.

The path: ask the two component questions before you look at the code book. Who produced the tracing? The hospital. Who wrote the interpretation and report? The cardiologist. That is the professional component alone → 93010, interpretation and report only.

The plausible wrong answer: 93000. It is the code everyone reaches for, because it is the one that "means ECG." It asserts that the billing entity produced the tracing and the report. The group did not own the machine, was not in the room, and paid for none of it. Reporting 93000 bills the hospital's technical resource on the physician's claim — and the hospital, billing its own claim for the same tracing, produces the duplicate that eventually surfaces this (Chapter 27 §27.7's distinction decides whether what comes back is a rejection or a denial).

The second wrong answer is subtler: reporting nothing. A practice that reads tracings all day and has never billed 93010 because "the hospital bills the ECG" is leaving a professional service unreported — Chapter 15 §15.9a's principle in a new specialty: a code you never report can never be denied, and it also never pays.

35.3 Orthopedics: fracture care, global periods, and the surgical decision

Chapter 17 §17.7 gave you the fracture-care fork and Chapter 17 §17.2 gave you the global periods. This section is what an orthopedic practice does with them on a Tuesday, which is a different skill from knowing the rule.

The decision, and the sentence that decides it

The fork, in one line: a physician treating a fracture may report global fracture care — one code, a 090-day global period, including the initial treatment, the first cast or splint, and normal follow-up through healing — or an evaluation and management (E/M) service plus a casting or splinting procedure, with no global period and every subsequent visit billable. Chapter 17 §17.7 has the full comparison and the reasons neither is a default.

What §35.3 owns is the specialty decision: how the documentation decides it, and what the practice does about it.

The decision turns on who is providing definitive care through healing — and that is a clinical plan, stated in the record, not a billing preference discovered later. The sentence that decides it is short and it belongs in the plan:

• "Will follow in this office until union; return in two weeks for repeat films." → the physician is undertaking fracture management. Global fracture care.
• "Splinted; referred to orthopedics for definitive management." → the physician stabilized and handed off. E/M plus the splint application, with no global period, and reporting fracture care here would assert ninety days of management that will not happen.

Notice that the coder did not make this decision. The physician made it, in the plan, and the coder read it. When the plan is silent — "fracture care rendered" and nothing else — the coder has a genuine ambiguity that costs the practice ninety days of billing rights in one direction and a denial in the other, and the correct response is a query about the treatment plan (Chapter 4 §4.9; Chapter 38 §38.3 owns the compliant form of it), not a guess.

The practice-level answer is a written rule, and Chapter 17 §17.7 already named why: without one, a macro makes the decision. In this organization, in this setting, fracture care is reported when these conditions are met. One meeting, written down, applied consistently — and reviewed when the practice adds a setting, because an urgent care and an operating room do not make the same decision.

📋 Read the Chart

text FIGURE 35.2 — "Two plans, two claims" [constructed teaching example] THE DOCUMENT Two office notes for two patients, same orthopedic practice, same afternoon, both distal radius fractures seen after an urgent care referral. THE CONTEXT Both patients arrive splinted. Both notes document the history, the examination, the films reviewed, and a plan. WHAT IT SHOWS NOTE A, plan: "Nondisplaced fracture, acceptable alignment. Short arm cast applied today. Will follow in this office weekly x2 then biweekly until union; repeat films at 3 weeks. Anticipate 6 weeks of immobilization." NOTE B, plan: "Displaced, intra-articular extension. Patient wishes to be treated near home. Splint reapplied for comfort. Records and images released to the patient; referred to the orthopedic group in her county for definitive management. No follow-up scheduled here." WHAT IT DOESN'T Neither note says the words "global fracture care" or "E/M plus casting," and neither should. Neither note states a treatment TYPE in the code set's vocabulary — closed, open, percutaneous, with or without manipulation (Chapter 17 §17.7) — and note A will need that before a fracture care code can be selected. THE DECISION Note A: the practice is providing definitive care — global fracture care, once the treatment type is documented, with the cast application INCLUDED and the casting SUPPLY separately reportable (Chapter 17 §17.8). Note B: an E/M service at the level the documented decision making supports, plus the splint application and its supply. No global period. Query note A for the treatment type; do not query note B for anything. THE LESSON The fracture-care decision is documented in the PLAN, in ordinary clinical English, by a physician who is not thinking about global periods. Read the plan first. It tells you which claim you are building before you open the code book.

Living inside a global period

An orthopedic practice spends its year inside 090-day globals, and that has operational consequences no other specialty in this chapter feels as constantly.

The postoperative visits are not billable, and everybody in the building has to know it. A front desk that collects a copay for a routine two-week post-op check has collected money for a service inside the package — which is a refund, a phone call, and a small avoidable indignity for a patient who is already inconvenienced.

But the global period covers the related care, not the patient. The modifier set that says so was frozen in Chapter 14 §14.9 and its mechanics belong to Chapter 17 §17.2; the specialty's job is to apply them accurately:

58 is planned and 78 is not, and only 58 restarts the clock. That distinction was frozen in Chapter 14 and it is the one an orthopedic biller uses most, because a practice that reports 78 where 58 was correct has given away a fresh global period, and a practice that reports 58 where 78 was correct has asserted that a complication was planned.

And the specialty's ordinary disciplines still apply, at volume. Laterality on every line — RT, LT, or the bilateral convention (Chapter 14 §14.8) — because orthopedics is the specialty where "the knee" is always two knees until the claim says otherwise. The knee compartment convention and the arthroscopy rules from Chapter 17 §17.9. The casting supply, separately reportable even when the application is included (Chapter 17 §17.8), which is money practices routinely leave behind. And code from the operative report, not from the scheduled procedure — Chapter 17 §17.3's discipline, in the specialty where the schedule is most often optimistic.

⚠️ Where Claims Die

The global period kills claims in both directions, and only one of them is visible.

The visible direction: a postoperative visit billed without modifier 24 during a 090-day global denies, plainly, with a reason code that names the global period. Somebody works it, appends 24 if the visit was genuinely unrelated, and the claim pays. Annoying, recoverable, and it teaches the practice something.

The invisible direction is worse: modifier 24 appended by rule. A system configured to append 24 to every E/M that falls inside a global period will pay every one of those claims — and each one asserts, in the practice's own words, that the visit was unrelated to the surgery. The routine two-week check is related to the surgery. Nothing denies. Nothing surfaces. The practice discovers it when a reviewer pulls twenty postoperative visits and finds that a documented "doing well, wound healing, incision clean" note is attached to a claim asserting the visit had nothing to do with the operation.

This is the thread this book has been counting since Chapter 14 §14.1: a configuration making an assertion nobody chose to make — the auto-appended modifier, the prefilled time, the fracture-care macro of Chapter 17 §17.7, the modifier-59 macro that produced Account 31-2245's audit. The disciplined answer is the same every time: the modifier is a statement about a record, and a statement about a record has to be made by somebody who read it. A weekly report of every E/M coded inside a global period, reviewed by a human for a few minutes, is the whole control.

35.4 Obstetrics and gynecology: the global package and what falls outside it

Chapter 18 §18.6 defined the maternity global package: antepartum care, delivery, and postpartum care, reported with a single code covering the whole episode. It also gave the decision tree for when the episode breaks and the components must be reported separately. This section starts where that one stopped.

The global obstetric codes you will see most:

• 59400 — routine obstetric care including antepartum care, vaginal delivery, and postpartum care.
• 59510 — routine obstetric care including antepartum care, cesarean delivery, and postpartum care.

And the family continues past both: delivery-only codes, delivery-and-postpartum codes, postpartum-care-only codes, and antepartum-care codes divided by the number of visits furnished. The global code is not the normal case that everything else deviates from — it is one of several correct answers, and the correct one depends on what this practice actually did.

What falls outside the package

Chapter 18 §18.6 listed the exclusions. Here they are as the working inventory an obstetric coder keeps, because the whole specialty's separately reportable revenue is on this list — and so is its whole unbundling exposure:

Two cautions on that table, both large.

"Complication" is a provider's word. A coder may not promote a visit into a complication because extra work happened at it. The record has to characterize the condition, and the ordinary documentation discipline of Chapter 4 §4.7 applies with full force in the specialty where the temptation is strongest.

And this is the most payer-policy-dependent list in the chapter. Plans differ on what they treat as inside the package, on how many antepartum visits the package assumes, and on how they want a complication-related visit reported. Chapter 22 §22.10's discipline is not optional here: read the policy of the plans that carry your practice's volume, and write down what each one says.

The date-of-service problem, and the log that solves it

Chapter 18 §18.6 flagged the structural oddity: the global package is billed at the end of an episode that began nine months earlier, and that single scheduling fact produces a category of error that exists nowhere else in this book — eligibility that changed mid-episode, visits furnished and never counted, and a genuine question about which date of service the claim carries. Payer policies address the date; read them before you need them.

The operational answer is an artifact, and this is the section's contribution: the obstetric episode log. One row per patient, opened at the first prenatal visit and closed when the claim goes out:

   THE OBSTETRIC EPISODE LOG — one row per patient, opened at visit one

   PATIENT ACCT | EDD | PLAN + ID | ELIG CHECKED | ANTEPARTUM VISITS | OUTSIDE SERVICES
   ─────────────┼─────┼───────────┼──────────────┼───────────────────┼──────────────────
   (account)    | ... | ...       | dates: __ __ | tally: 1 2 3 ...  | US 20wk, NST x2,
                |     |           | __ __ __     |                   | GDM visits x3
                |     |           |              |                   |
   RULES
     · The visit tally is incremented AT THE VISIT, by the person who
       rooms the patient. Reconstructing it from the chart in month nine
       takes an hour per patient and is never as accurate.
     · Eligibility is re-verified on a schedule, not once (Chapter 24
       §24.3). Coverage at the first visit is not coverage at delivery.
     · An "outside service" is flagged THE DAY IT HAPPENS. A non-stress
       test identified eight months later is a charge nobody will find.
     · The delivery type closes the row and selects the code: vaginal
       (59400) or cesarean (59510) — or, if the episode broke, the
       components actually furnished (Chapter 18 §18.6's decision tree).
                                                 [schematic — not to scale]

The log is the specialty's single highest-value process control, and it is a spreadsheet. Chapter 24 §24.3's eligibility discipline, applied on a nine-month cadence instead of a same-day one, plus Chapter 23 §23.9's charge capture, applied to services that will not be billed for months.

The gynecology side

The other half of the specialty is ordinary outpatient coding with two recurring questions.

Preventive versus problem. A well-woman visit is a preventive medicine service (§35.5's family); a visit for a specific complaint is a problem-oriented E/M. When both happen — the annual exam at which a new problem is evaluated and managed — the preventive service and a problem-oriented E/M with modifier 25 may both be reportable, on the same documentation standard Chapter 14 §14.4 set. It is the same structure as §35.5's well-child visit and it produces the same patient-facing surprise, for the same reason.

Screening versus diagnostic. The distinction Chapter 12 §12.9 built and Chapter 22 §22.7's frequency limitations enforce governs cervical cancer screening and screening mammography (77067 is the roster's example) exactly as it governed Account 22-9107's colonoscopy: the intent codes first, and modifier 33 identifies a service that began as preventive. Chapter 34 §34.11 worked the full financial resolution of that pattern; the obstetric-gynecologic version differs only in the anatomy.

And the diagnosis side has its own conventions, owned by Chapter 11 §11.9: trimester, the sequencing rules for pregnancy-related conditions, and the roster's O09.512 as the shape of a supervision code. Cite them; do not rebuild them.

📞 On the Phone

The obstetric coordinator, calling the practice's biller about a patient who transferred in at 30 weeks.

"She's ours from here. Her old practice in another state saw her from eight weeks to twenty-nine. We're delivering. What do I bill?"

The biller who has read Chapter 18 §18.6: "Not the global — we didn't provide antepartum care from the beginning, so no single code describes what we did. We report what we furnished: the antepartum visits we provide from here to delivery, by count, plus the delivery, plus the postpartum care. That means I need the count, exactly, and I need it as we go — not in March."

"Her old practice will bill their part?"

"They'll bill antepartum care by their count, and that's their claim, not ours. What I need from you is three things: her visit count starting today, a note in the record that care was transferred at 30 weeks and why, and her eligibility re-verified now — because her coverage may have moved with her."

The failure modes, all three of them common. Billing the global anyway, because the practice delivered and the global code is the one everybody knows — which claims antepartum care the practice did not furnish, and which the other practice's claim will eventually contradict. Billing individual office visits for the antepartum care instead of the antepartum-care codes, because office visits are familiar. And reconstructing the count from the chart in month nine, which is where the log in this section comes from.

And the sentence that ends the call well: "Write down today's date and today's count on the log, and we'll both be right in March."

35.5 Pediatrics: preventive visits, vaccines, and the two-code habit

Here is what actually unifies this section, and it is worth stating before the codes: in primary care and pediatrics, more of the codes depend on facts that are not in the note you are reading than in any other specialty in this chapter. The patient's age, which lives in the registration record. The vaccine's funding source, which lives in an inventory system. The date the patient left the hospital and the date somebody called her, which live in a log. "If it isn't documented, it didn't happen" acquires a corollary here: some of it is documented somewhere else, and the coder has to go and get it.

The preventive medicine services

CPT carries a family of preventive medicine services — comprehensive, periodic evaluations for patients without a presenting complaint — and it is the second E/M family in this book that is not leveled by medical decision making (MDM) or by time. (Chapter 19 §19.8's eye codes were the first.)

Preventive medicine service codes are selected by two facts only: whether the patient is new or established, and the patient's age. 99381–99387 for new patients and 99391–99397 for established patients, each range divided into age bands from infancy through 65 and over.

Three consequences follow immediately.

The age band is the patient's age at the visit, which is a registration fact, and a birthday between the scheduling and the visit changes the code. The new-versus-established rule is Chapter 15 §15.2's three-year rule, unchanged, and it is the same trap here: a patient seen by any physician of the same specialty in the same group within three years is established. And the service is defined by content, not by level — an age- and gender-appropriate history, examination, counseling, anticipatory guidance, risk-factor reduction, and the ordering of appropriate laboratory or diagnostic procedures. A note that documents a problem-oriented visit does not support a preventive code merely because the appointment was labeled a physical.

And here is the place in this book where CPT's code and the payer's benefit most visibly fail to line up. A commercial plan generally pays a preventive medicine service under the preventive benefit. Medicare does not cover a routine physical examination at all. What it covers instead is its own benefit — the annual wellness visit, reported with HCPCS Level II G-codes: one for the first such visit and G0439 for each subsequent one. Chapter 20 §20.6 explained exactly why that code exists — CPT has no code for the service, because the benefit exists before the description does — and Chapter 3 explained why Medicare's benefit structure is built that way. The clinical work overlaps substantially; the codes do not overlap at all, and a practice that bills a preventive medicine service to Medicare has billed a statutorily excluded service (Chapter 22 §22.9's GY territory).

The two-code habit

Chapter 19 §19.10 stated the rule and pediatrics is where it is exercised twenty times a day:

An immunization requires TWO codes: the vaccine product and the administration. Both. Every time.

90686 is a vaccine product code — influenza vaccine, quadrivalent, preservative free, for intramuscular use. 90471 is the administration of the first vaccine at the encounter; 90472 is each additional administration. Two vaccines at one visit are 90471 once and 90472 once; three are 90471 once and 90472 twice. The counting is per administration, not per component and not per visit.

The failure mode is a pure underpayment and it is silent. Report the product without the administration and the claim pays — for the vaccine. The administration is simply absent. Nothing denies, nothing appears on an exception report, and nobody in the building sees an absence. This is Chapter 28 §28.8's silent-underpayment shape in a different medium: the money that does not arrive leaves no evidence of itself.

And the reverse case is frequently correct. A practice that received doses through a public supply program did not purchase them and may not bill for the product — but it performed the administration and may bill that. A claim with an administration code and no product code is not automatically an error, and a well-meaning biller who "fixes" it has just billed for a vaccine somebody else paid for. Know which supply each dose came from. That is an inventory fact, it lives outside the medical record entirely, and it is the clearest example in this book of a code that cannot be selected from the chart alone.

One more family belongs here and pediatrics is its home. A separate set of administration codes applies to patients through 18 years of age when the physician or other qualified health care professional provides counseling to the patient and family. Chapter 19 §19.10 named the two rules: the counseling must be provided by the physician or qualified health care professional — not by clinical staff — and it must be documented. A note stating that a vaccine was given does not support them, and the counting convention in that family differs from the ordinary one. Read the descriptors; do not carry a habit from one family into the other.

The two other codes primary care lives on: transitional care management

The same "facts outside the note" principle produces one more pair worth knowing, because it is the clearest example in the book of a service defined by a calendar window rather than by an encounter:

Transitional care management (TCM) covers the 30-day period beginning on the date a patient is discharged from an inpatient, observation, or skilled nursing setting back to a community setting. 99495 is the moderate-complexity service; 99496 is the high-complexity service.

Three requirements decide whether the service exists at all, and only one of them happens in an exam room:

1. Interactive contact with the patient or caregiver — direct, by telephone, or electronic — within two business days of discharge.
2. A face-to-face visit within a defined window: 14 calendar days of discharge for 99495, 7 calendar days for 99496.
3. Medication reconciliation, no later than the date of that face-to-face visit.

The face-to-face visit is part of the TCM service and is not separately reported, which surprises practices that have been billing it as an office visit. And the complexity that separates the two codes is the medical decision making during the service period — Chapter 15 §15.4's grid, applied across a month instead of a day. Verify the current requirements before you build a workflow on them; the elements have been refined more than once, and CPT changes every January 1.

Why this matters operationally more than technically: the two-business-day contact is the requirement most often missed, and it is missed because nobody knows the patient was discharged. The practice's part of the work is a discharge notification process — Chapter 24's front-end discipline pointed at a hospital's census rather than at a schedule — and the documentation is a log of the contact attempt, its date, its time, its outcome, and who made it. A phone call nobody wrote down is a phone call that did not happen.

🔢 Code It

The documentation (constructed teaching example): a 9-year-old established patient presents for an annual well-child visit. The note documents a comprehensive age-appropriate history and examination, growth chart review, developmental and behavioral screening, anticipatory guidance on nutrition, screen time, and helmet use, and the immunization status reviewed. A separate, clearly headed paragraph reads: "Mother additionally reports two weeks of intermittent right ear pain. On examination, right tympanic membrane erythematous and bulging with loss of landmarks. Assessment: acute otitis media, right. Plan: amoxicillin, dosing reviewed, return if no improvement in 48–72 hours." The nurse administered an influenza vaccine, quadrivalent, preservative free, intramuscular, from practice-purchased stock. The note documents no counseling by the physician regarding the vaccine.

The path:

• The preventive service: established patient, age 9 → the 5–11 age band of the established preventive medicine range → 99393.
• The problem: a separate, significant, separately identifiable problem with its own history, examination, assessment, and plan. Level it on MDM (Chapter 15 §15.4): one acute uncomplicated illness (low problems), no data reviewed, prescription drug management (moderate risk) — two of three at low → 99213, appended with modifier 25 (Chapter 14 §14.4).
• The vaccine: product 90686 plus administration 90471. The practice bought the dose, so the product is billable. One vaccine, so no 90472.

Four plausible wrong answers, each named and rejected:

"You can't bill both the well visit and a sick visit." You can, when the record supports it, and Chapter 14 §14.4 says exactly when. Refusing to is downcoding, which Chapter 5 §5.8 refuses to call the safe option.

The counseling-based administration code. Tempting — the patient is 9 — but the note documents no counseling by the physician or qualified health care professional. The documentation does not support it (Chapter 19 §19.10), and the ordinary administration code is correct.

Omitting 90471. The silent underpayment above.

Reporting 90472. One vaccine was given. 90472 is each additional, and there is no additional.

And the honest patient-facing footnote: the preventive service will likely process at 100% under the preventive benefit and the 99213-25 will likely process against the deductible or coinsurance. The family was told the visit was free and will receive a bill. That is not a coding error — it is the benefit design working as written — and the practice that explains it at check-out rather than by letter three weeks later saves a phone call and a great deal of goodwill (Chapter 32 §32.10).

🧮 Run the Numbers

What the missing administration line is worth in one influenza season [constructed teaching figures — verify current allowed amounts in your own contracts].

A pediatric practice administers 1,200 influenza doses in a season. The administration code's allowed amount under its largest contract is \$26.40. A charge-capture audit finds that on 9% of those encounters the product line was billed and the administration line was not.

```text Doses administered ............................ 1,200 Encounters missing the administration line 1,200 × 0.09 ................................ 108

Value of one administration ................... $26.40

108 × $26.40 = $2,851.20 ```

Checks: 1,200 × 0.09 = 108 ✓ · 108 × 26.40 = 2,851.20 ✓

Interpretation. Two thousand eight hundred fifty-one dollars and twenty cents, from one code, in one season, in one practice — and not one claim denied. There is nothing in a work queue to find, nothing on an aging report, nothing on a denial log. The vaccine paid; the administration was never asked for.

The only way to see it is to compare two counts you already have: doses out of inventory, and administration lines out of the billing system. They should match. When they do not, the difference is either a documentation gap or a charge-capture gap, and both are fixable in an afternoon — Chapter 23 §23.9's charge capture and Chapter 28 §28.8's method, pointed at a specialty's highest-volume line.

35.6 Emergency medicine: acuity, the facility side, and the professional side

Chapter 34 §34.8 built facility-versus-professional for the same service in general terms. The emergency department (ED) is where that split is sharpest, because the two claims describe the same hour in the same room, both may carry the same five-digit code, and they were produced by two different rulebooks — one national, one written by the hospital itself.

The professional side

Chapter 16 §16.6 owns it and two structural facts define the family: 99281–99285 carry no new-versus-established distinction and no time-based selection option. Everything is leveled on medical decision making — Chapter 15's grid, all three elements, two of three. Critical care (99291 for the first 30–74 minutes and 99292 for each additional 30) is Chapter 16 §16.7's.

What §35.6 adds is the ED professional coder's actual working conditions, because they shape the errors:

The volume is high and the charts are short. An ED coder may work three to five times as many encounters per day as an office coder, on documentation that was produced under interruption. The discipline that survives that pace is a fixed reading order — disposition first, then the medical decision making paragraph, then procedures, then the diagnosis — rather than reading each chart from the top as though it were new.

The presenting problem is not the diagnosis, and the diagnosis is not the level. Chapter 16 §16.6 built this on a chest-pain chart that ended in reflux: the MDM reflects the work the presentation required, not what it turned out to be. The ED is the setting where the final diagnosis most routinely understates the service, and it is why ED documentation should state what was being ruled out. In diagnosis coding, Chapter 12 §12.3's rule holds — code to the highest degree of certainty established — which is why R07.9 and its relatives are correct outcomes rather than placeholders.

And procedures hide in the narrative. A laceration repair, a splint, a foreign body removal, a reduction — these are documented inside the note rather than in a separate operative report, and an ED coder who reads only for the E/M level leaves them behind. The ED anchor's professional claim (Account 10-7789, constructed) is exactly this shape: 99284-25 and 12002 — the visit and the repair, with modifier 25 doing Chapter 14 §14.4's work.

The facility side: acuity leveling

This is the part with no national rulebook, and it is the specialty discipline this section owns.

Chapter 16 §16.9 stated the fact and Chapter 34 §34.8 restated it in payment terms: CMS has never published national facility E/M leveling criteria for the emergency department. Each hospital develops its own written criteria, applies them consistently, and must be able to show that the resulting levels reasonably relate to the resources the department actually consumed.

What the criteria measure is not what the physician decided. The facility's level reports the department's resource intensity — nursing interventions, monitoring, medications administered, procedures supported, time in the department, ancillary staff involvement. The physician's level reports cognitive work. They are different quantities measured on different scales, which is why the same visit can legitimately produce a level 4 facility claim and a level 3 professional claim, and why a payer analyst comparing them for "mismatch" is asking the wrong question (Chapter 34 §34.8).

In practice, two families of criteria are used, and both are defensible:

• A point or intervention model. Each nursing intervention carries a weight; the weights total to a score; score bands map to the five levels. Reproducible, auditable, and easy to apply at volume.
• A critical-intervention or decision-tree model. The presence of a defined intervention — cardiac monitoring, procedural support, multiple medication administrations — places the encounter at a level directly. Fewer moving parts; more judgment at the boundaries.

What makes either one defensible is not its design. It is five properties, and Chapter 16 §16.9 named them: written down, related to resources actually consumed, producing a distribution that reflects the department's case mix, reproducible by two independent readers, and auditable by someone outside the department. The one that fails in real hospitals is the last. Criteria that live in a senior coder's head, or in a document last revised nine years ago, cannot demonstrate any of the others.

And there is a sixth test the specialty applies to itself: the distribution test.

Plot the department's facility levels as a distribution and compare it with the department's own case mix and with peer departments. A distribution weighted heavily toward the top levels without a case mix that explains it is visible from outside the organization without anyone reading a single chart — which is Chapter 21 §21.10's pattern detection and Chapter 26's lesson that a facility can be detectable from outside before it is detectable from inside. ED facility level distributions have been a recurring subject of federal and payer attention for years, and the reason is arithmetic rather than suspicion: the data required to notice is already in the claims.

Two honest cautions about that test. A distribution is a question, not a finding — a department serving a high-acuity population should level higher, and the burden is to explain it, not to apologize for it. And the fix for a skewed distribution is never "level lower." It is to re-derive the criteria from resource consumption and apply them; if the levels then come out high, they come out high with a defensible reason attached.

📋 Read the Chart

```text FIGURE 35.3 — "The criteria, and one chart against them" [constructed teaching example] THE DOCUMENT Two things side by side: an excerpt from a hospital's written ED facility leveling criteria (revised this year, approved by the revenue integrity committee), and one ED nursing record scored against it. THE CONTEXT Ridgeview Regional Medical Center (constructed). The criteria are the hospital's own; there is no national set.

THE CRITERIA (excerpt, intervention-point model) Triage and vital signs, initial ......................... 1 Repeat vital signs, each set beyond the first ........... 1 Continuous cardiac / pulse oximetry monitoring .......... 3 Medication administered, oral ........................... 1 Medication administered, IM / SC / IV ................... 3 Intravenous line established and maintained ............. 3 Specimen collection and handling ........................ 1 Assistance with a physician procedure ................... 3 Splint / immobilization applied by nursing .............. 2 Discharge teaching and follow-up arrangement ............ 1 ───────────────────────────────────────────────────────── LEVEL BANDS 1-3 → 99281 4-7 → 99282 8-12 → 99283 13-18 → 99284 19+ → 99285

THE CHART Triage and vitals (1). Two repeat vital sign sets (2). Continuous monitoring for 90 minutes (3). IV established (3). One IV medication (3). Blood drawn and sent (1). Nursing assisted the physician's procedure (3). Discharge teaching with a follow-up appointment made (1). TOTAL = 1+2+3+3+3+1+3+1 = 17 → band 13-18 → 99284

WHAT IT SHOWS A level produced by a written rule from documented nursing work — reproducible by anyone with the record and the criteria sheet. WHAT IT DOESN'T It says NOTHING about the physician's level. The professional coder scores the same encounter on medical decision making and may land on 99283 or 99285. Neither number is evidence about the other. THE DECISION Report 99284 on the facility claim, with revenue code 0450 (Chapter 26 §26.4). Keep the scored sheet — or the system's audit trail of the score — because the reproducibility test is only passable with it. THE LESSON Facility acuity leveling is not a softer version of E/M leveling. It is a different measurement of a different quantity, and its entire defense is a written rule applied the same way to everyone. ```

One encounter, two claims, one patient

The ED anchor makes the split concrete with figures this book froze in Chapter 1. Account 22-7788 is the facility claim (UB-04, type of bill 131); Account 10-7789 is the emergency physician's professional claim. Both carry 99284, determined by two different rulebooks, and both are correct.

Checks: 3,842.00 − 1,196.40 = 2,645.60 ✓ · 1,196.40 − 439.28 = 757.12 ✓ · 680.00 − 318.60 = 361.40 ✓ · 318.60 − 63.72 = 254.88 ✓ · 4,522.00 − 1,515.00 = 3,007.00 ✓ · 1,515.00 − 503.00 = 1,012.00 ✓ · 439.28 + 63.72 = 503.00 ✓ · 757.12 + 254.88 = 1,012.00 ✓

The patient received two bills, weeks apart, from two organizations, for one Sunday evening — and in many hospitals the physician group is a separate company with its own billing office and its own phone number. She will call one of them about the other. The biller who can say, in one sentence, which claim a balance came from and why there are two prevents most of that call; Chapter 32 §32.10 owns the conversation and Chapter 1 decoded this exact bill.

🔍 Check Your Understanding

1. An ED encounter produces a facility claim at level 4 and a professional claim at level 3. A payer analyst flags the "mismatch." What is the correct response, and what evidence supports it?
2. A hospital's ED facility levels are 55% level 5. Name the two things that could produce that, the one document that decides which, and the response that is not appropriate.
3. Why does the ED professional family have no time-based selection option, and what does that remove from the coder's toolkit that exists in every office visit?

Answers: (1) They measure different quantities on different scales — cognitive work versus departmental resource intensity — and are expected to correlate, not to match. The evidence is the hospital's written criteria plus the scored record (Figure 35.3), which shows the facility level was produced by a rule rather than copied. Chapter 16 §16.9 adds the sharper point: a facility level that mirrors the physician's on every claim suggests the facility is not applying its own criteria at all. (2) A genuinely high-acuity case mix, or criteria that are wrong or wrongly applied. The written criteria plus a sample of scored records decide which. Lowering the levels to improve the distribution is not appropriate — it substitutes one unsupported number for another. (3) Because ED work is interleaved by design; a physician manages several patients at once and total time on the date is not a meaningful measure of one patient's care (Chapter 16 §16.6). It removes the time-based alternative that Chapter 15 §15.8 provides everywhere else, which means the MDM documentation carries the entire level.

35.7 Anesthesia: base units, time units, and modifiers

Chapter 18 §18.11 owns the formula — base units plus time units plus modifying units, multiplied by an anesthesia conversion factor — and it defines the physical status modifiers P1 through P6 and the anesthesia time boundaries. This section owns the workflow: what an anesthesia coder actually does with a record, in what order, and where the specialty's money and exposure live.

Three sentences of recap so the section stands on its own, and then no re-teaching:

Base units are published per anesthesia code and reflect the complexity of the procedure being anesthetized. Time units are computed from anesthesia time, converted by a payer-specific increment. Modifying units come from physical status (P1–P6, describing the patient's condition from a normal healthy patient to a declared brain-dead organ donor) and from qualifying circumstances add-on codes. Chapter 18 §18.11 has all three in full.

The workflow, in order

1. The source document is the anesthesia record. Not the operative report. Times, agents, monitoring, the physical status assignment, the provider's presence and its transitions, the airway management — none of that is in the surgeon's note, and the surgeon's note will contain a time that is not the time you need. An anesthesia claim built from an operative report is built from the wrong document.

2. Select the anesthesia code from the anesthesia section, not from the surgeon's CPT code. The anesthesia section has its own codes, organized anatomically, and the mapping from a surgical procedure to its anesthesia code is not one to one. When two or more procedures are performed under a single anesthetic, one anesthesia service is reported — the code with the highest base unit value. One anesthetic, one anesthesia code, however many things the surgeon did.

3. Compute anesthesia time from the record's own boundaries. Anesthesia time begins when the anesthesia professional begins preparing the patient for induction and ends when the patient may be safely placed under postoperative supervision (Chapter 18 §18.11). It is neither room time nor surgical time, and the difference is where the specialty's most common error lives — see the arithmetic below.

4. Assign physical status from the record, never from the chart's severity. P1 through P6 describe the patient's condition and the assignment is made by the anesthesia professional and documented. A coder may not upgrade it because the patient's problem list looks impressive — Chapter 4 §4.7's line, in a specialty where the temptation has a per-unit price attached. And payer recognition varies: not every plan adds units for P3 through P5, and some recognize only P4 and P5.

5. Report the provider-arrangement modifier accurately, because it is the audited field. Chapter 18 §18.11 named the four arrangements — personally performed, medical direction of concurrent procedures, medical supervision of more than four, and a certified registered nurse anesthetist with or without medical direction. The concurrency count is a fact about the schedule, not about the case, which means the modifier cannot be selected from the anesthesia record alone; it requires the day's assignment board. And medical direction has a documented checklist of required steps — the preanesthetic examination, prescribing the plan, presence at induction and emergence, remaining available, monitoring — and it is audited as a checklist.

6. Then, and only then, the units. Increment and rounding rule are payer policy, published, and different. Chapter 18 §18.11's 📞 On the Phone is the call you will make about it.

🧮 Run the Numbers

One anesthesia record, three clocks, two answers [constructed teaching figures — base units, increments, rounding rules, and conversion factors are payer-specific and published; verify all four]. The conversion factor is Chapter 18 §18.11's constructed \$22.00.

The record:

text Patient into the operating room ................ 07:42 Anesthesia start (preparation for induction) ... 07:46 Surgical incision .............................. 08:05 Surgery end .................................... 09:18 Anesthesia end (safe for postop supervision) ... 09:24 Patient out of the room ........................ 09:28

The procedure's anesthesia code carries 8 base units. The patient is P2 — mild systemic disease — which adds no units under this payer's methodology. The payer's increment is 15 minutes and its rule is to truncate.

```text ANESTHESIA TIME 07:46 → 09:24 = 98 minutes time units 98 ÷ 15 = 6.53 → truncated → 6

  Base units .................................  8
  Time units .................................  6
  Modifying units (P2) .......................  0
                                              ───
  Total units ................................ 14

  14  ×  $22.00  =  $308.00

THE ERROR — using ROOM TIME instead 07:42 → 09:28 = 106 minutes 106 ÷ 15 = 7.07 → truncated → 7 time units 8 + 7 + 0 = 15 units × $22.00 = $330.00 ```

Checks: 8 + 6 + 0 = 14 ✓ · 14 × 22.00 = 308.00 ✓ · 8 + 7 + 0 = 15 ✓ · 15 × 22.00 = 330.00 ✓ · 330.00 − 308.00 = 22.00 ✓

Interpretation, and the distinction that matters most in this section. Eight minutes of room time that are not anesthesia time produce one extra unit and \$22.00 on this case — every case, every day, across a group's entire volume. That is an error, and it is an error that overstates a claim.

Now change nothing but the payer. Under a plan that rounds to the nearest unit rather than truncating, the same 98 minutes is 6.53 → 7 time units, 15 units, \$330.00 — the identical figure, reached legitimately. The same record produces \$308.00 or \$330.00 depending on a published policy, and produces \$330.00 wrongly if you read the wrong clock. The two look alike on a remittance and are not alike at all: one difference is policy and the other is a false statement. Know which one you are looking at, and know the increment and rounding rule of every payer that carries your volume — in writing, in a file, because the reason anybody makes that phone call is that nobody wrote it down last time.

⚖️ Compliance Check

Anesthesia is one of the most reliably examined billing areas in American health care, and the reason is structural: two of its three payment inputs are assertions about things that happened in a room, minute by minute, and only one document records them.

Time reporting. A claim reporting anesthesia time the record does not support is a false statement about a service, and the False Claims Act theory (Chapter 5 §5.3) does not require anyone to have intended fraud — a pattern of reported times that consistently exceeds documented times is exactly the pattern a data analysis finds without reading a chart (Chapter 21 §21.10). Federal and commercial enforcement actions involving anesthesia time reporting and medical-direction requirements are part of the documented public record; this book asserts no figures and you should distrust any source that does without a citation to the settlement itself.

Medical direction. The required steps are a checklist and are audited as one. An organization that bills medical direction should be able to produce, for any case, the documentation of each step — and the concurrency count that made the arrangement permissible in the first place.

Physical status. P3 through P5 carry units under many methodologies. That makes the assignment a payment input, and a payment input assigned by anyone other than the anesthesia professional who examined the patient is an assertion nobody was entitled to make.

The standing caveats apply and are not boilerplate: these requirements change, state law and payer policy vary, and the arrangement rules in particular have been revisited more than once. Verify with your compliance officer and the primary source — the Medicare Claims Processing Manual, your MAC's articles, and each plan's published anesthesia policy — before you rely on any of it.

35.8 Behavioral health and the time-based codes

Two things before the codes, and neither is optional.

Patients are not a coding category. Behavioral health records describe people in the parts of their lives they are least able to choose to disclose. The minimum-necessary discipline (Chapter 5 §5.7) is not more relaxed here because the records are routine to you — it is tighter, because the consequences of disclosure are larger. An appeal packet includes the records at issue and no others. A query quotes the sentence it needs and stops. And substance use disorder treatment records from federally assisted programs carry their own federal confidentiality regime (42 CFR Part 2) in addition to HIPAA, with consent requirements that differ from the ones you learned in Chapter 5. If your work touches those records, learn that rule specifically and do not reason by analogy from HIPAA.

And a practical fairness point. Behavioral health is among the most prior-authorized, most frequency-limited, most narrow-network areas of American coverage — which means its patients absorb more administrative burden than most, at moments when they have the least capacity for it. The Mental Health Parity and Addiction Equity Act exists because that pattern was documented. A revenue cycle professional cannot fix coverage policy; they can decline to treat the resulting phone call as an inconvenience.

The families, structurally

This book names no behavioral health CPT codes, deliberately — the families are learnable from the descriptors and naming individual codes from memory is exactly the risk Chapter 6 §6.10's habit exists to prevent. What you need is the shape:

• The psychiatric diagnostic evaluation — the initial assessment, in two versions: with medical services and without. Which one applies depends on who performed it and what medical services were furnished, not on how long it took.
• Psychotherapy, in timed increments — codes describing sessions of differing typical durations, selected by the time actually spent in the therapeutic encounter, with a published floor below which the service is not reportable at all.
• Psychotherapy furnished with an E/M service — reported as add-on codes to the E/M (Chapter 13 §13.7 governs add-ons). ⚠️ When psychotherapy is reported with an E/M service, the E/M is selected on medical decision making only — not on time — because the time is being used to select the psychotherapy code and cannot be counted twice. Verify the current rule; this is the most-missed convention in the specialty.
• Crisis psychotherapy — an initial period plus add-on codes for additional increments.
• Family and group psychotherapy, each with its own descriptor conventions about who is present and whether the patient is.

Four time conventions, one book

Here is the section's real content, and it applies far beyond behavioral health. This book has now taught four different ways to convert time into a code, and each of them is plausible in the others' territory.

   FOUR TIME CONVENTIONS — and the documentation each requires

   [1] E/M TOTAL TIME ON THE DATE                     Ch. 15 §15.8
       All qualifying time by the physician/QHP on the date, face-to-face
       and not. Thresholds per level. → Document TOTAL TIME and what was
       done in it.

   [2] TIMED THERAPY UNITS (the "8-minute rule")      Ch. 19 §19.11
       TOTAL the timed minutes for the date FIRST, then convert: 8-22 = 1
       unit, 23-37 = 2, and so on. Constant attendance required.
       → Document MINUTES PER SERVICE and the total.

   [3] PSYCHOTHERAPY TIME                             §35.8
       The code describes a typical duration; select by time actually
       spent, with a floor below which nothing is reportable.
       → Document START AND STOP, or total duration, plus content.

   [4] ANESTHESIA TIME                                Ch. 18 §18.11, §35.7
       Continuous, from a defined start to a defined end, converted by the
       PAYER's increment and rounding rule.
       → Document the CLOCK TIMES on the anesthesia record.

   THE TRAP: each convention is a reasonable-sounding answer in the other
   three's territory. Totaling before converting is right in [2] and wrong
   in [4]. A published range is right in [3] and meaningless in [1].
                                                 [schematic — not to scale]

97110 — therapeutic exercises, each 15 minutes — is the book's worked example of convention [2], and Chapter 19 §19.11's check-your-understanding shows exactly how the shortcut of converting each service separately produces a different, usually higher, answer than the correct method. Two services of 8 minutes each are 1 unit, not 2.

What time documentation actually has to say, in every convention: the duration, and what happened in it. A duration alone documents attendance. Content alone documents a service of unknown extent. The code requires both, and the second is what an auditor reads first.

⚠️ Where Claims Die

The template that prints the time.

A behavioral health practice's note template includes a session-duration field with a default. Over a year, the overwhelming majority of that practice's notes read "session duration: 50 minutes." The claims are consistent with the notes. Nothing denies.

Then a payer reviews forty charts and observes that the appointment schedule shows 45-minute slots with 15 minutes between, that several documented 50-minute sessions overlap in the clinician's day, and that the duration field never varies — including on two dates where the same clinician documented eleven 50-minute sessions.

Every one of those numbers was produced by a default and signed by a clinician who did not change it. This is Chapter 15's prefilled-time problem in a different specialty, and it is another entry in the thread this book has counted since Chapter 14 §14.1: a configuration making an assertion nobody chose to make. The clinician did not decide that the session was 50 minutes; the template did, and the signature made it an attestation.

It is also worse than the ordinary version, for the same reason Chapter 29's Case Study 2 was worse: the assertion is in the clinical record, the clinician's signature is on it, and a note can be amended but never un-asserted.

The disciplined alternatives are unglamorous. Remove the default. Require start and stop times rather than a duration. Reconcile documented session time against the schedule once a quarter — a report the practice can run against its own two systems, which is the only kind of control that catches a problem no claim will ever surface.

35.9 Telehealth across specialties

Telehealth is not a specialty. It is a modality, and it crosses every specialty in this chapter — which is why it gets its own section rather than a paragraph in each.

It is also, by a wide margin, the least stable body of rules in this book. Coverage has been expanded, contracted, extended with end dates, and revised again; lists of covered services have been republished; place-of-service and modifier conventions have changed; audio-only coverage has moved more than once; and state law adds its own layer on top of federal and commercial policy. This section teaches the structure and refuses to state a current coverage rule, because any such statement in a printed book has a short and unannounced expiration. Chapter 23 §23.5 and Chapter 25 §25.8 both said to verify rather than to carry a habit. Here is the structure that makes verification quick.

The four questions a telehealth claim answers

1. Is this service payable by this payer, furnished this way, right now? Medicare maintains a list of services payable when furnished via telehealth; commercial plans maintain their own, and they are not the same list. The list is the first lookup, and the date on the copy you are reading is part of the answer. Chapter 6 §6.6 taught how to find free authoritative sources; Chapter 22 §22.5 taught how to find the policy that governs a specific claim. Both apply.

2. Where was the patient, and where was the practitioner? The vocabulary is originating site (where the patient is) and distant site (where the practitioner is). On the professional claim, the place of service field (Chapter 25 §25.8) carries it:

• 02 — telehealth provided other than in the patient's home
• 10 — telehealth provided in the patient's home

Why two codes for what feels like one situation? Because the originating site can generate its own charge. When the patient is at a clinic or hospital rather than at home, a facility may bill an originating-site facility fee, and HCPCS Level II carries a code for it. When the patient is at her kitchen table, there is no originating site to pay. The distinction is about who else is billing, not about the quality of the video.

3. What technology was used? Synchronous audio-and-video is the baseline. Audio-only is a different question with a different answer per payer and per service, and there are separate code families for asynchronous store-and-forward, remote physiologic monitoring, and brief virtual check-ins. A visit conducted by telephone is not automatically a telehealth visit — it may be a different service entirely, or none.

4. What does the modifier assert? Modifier 95 identifies a synchronous telemedicine service rendered via a real-time interactive audio and video telecommunications system. Chapter 14 §14.1's principle governs: a modifier is an assertion, made to a stranger, about a fact. Modifier 95 asserts a technology. If the encounter was audio-only, 95 is a false statement regardless of how good the encounter was.

And here is the operational trap that catches practices every year: the place of service and the modifier are not two ways of saying the same thing, and payers combine them differently. Some want POS 02 or 10 with no modifier. Some want the in-person place of service plus modifier 95. Some want both. The combination affects payment, because place of service drives facility versus non-facility pricing (Chapter 23 §23.5) — which is exactly why this cannot be a habit. Look it up per payer, write it down, and put a review date on the note.

What each specialty does with it

• Behavioral health has been the most durably covered telehealth service across payers, and it is where the modality has changed practice most.
• Cardiology carries remote physiologic and cardiac monitoring families with their own frequency rules and their own component questions — a different animal from a video visit, and one where Chapter 22 §22.7's frequency limitations do real work.
• Orthopedics uses it for postoperative checks — which, inside a 090-day global (§35.3), are usually not separately reportable no matter how they are conducted. The modality does not change the package.
• Pediatrics meets the boundary honestly: a vaccine cannot be administered over video. Some of a well-child visit's content transfers and some does not, and the preventive medicine service's content definition (§35.5) is the test.
• Emergency medicine raises the originating-site question in its clearest form: a specialist consulting into an ED by video is at the distant site, while the hospital is the originating site and bills accordingly.

⚖️ Compliance Check

Two rules, and one of them is about this book.

The documentation rule. A telehealth encounter's note should state the modality (audio and video, or audio only), the location of the patient and of the practitioner, the participants, and consent where the payer or state law requires it. A note that reads exactly like an in-person visit, for an encounter conducted by video, is a record that does not describe what happened — and the claim's modifier and place of service will be asserting facts the record does not support. That is the same failure as the anesthesia record and the time template, in a third medium.

The verification rule. Telehealth coverage is set by a mix of federal statute with expiration dates, annual rulemaking, state law, and individual payer policy, and it has changed repeatedly and will change again. Do not rely on this section, your notes from last year, or a colleague's recollection. Check the payer's current published list and policy before you bill a telehealth service you have not billed recently, and check with your compliance officer about state requirements — including licensure across state lines, which is a legal question rather than a coding one and is outside this book entirely.

And the meta-rule, which is why this section is short on specifics: in a field where a rule has an expiration date, the durable skill is knowing where the rule lives, not knowing what it currently says.

35.10 How to learn a new specialty in two weeks

The premise, stated honestly first. Two weeks does not make you a specialty coder. It makes you competent at the twenty codes that are most of the volume, aware of the specialty's characteristic failure modes, and safe at the tail — which means you know when you are outside what you know and you look it up rather than guessing. That is a genuinely valuable state and it is reachable in ten working days. Mastery is a different project measured in years, and Chapter 39 covers the credential that formalizes it.

Here is the method. It has been used, it produces artifacts rather than feelings, and every step answers one of §35.1's five intake questions.

Before day one: obtain four things

1. The current code book, opened to the specialty's section guidelines. Not a study guide. The book. Read the section guidelines cover to cover before you code anything — they are three to eight pages, they are where the family rules live, and Chapter 13 §13.4 already told you they are the part everyone skips. Two hours here saves two months. Then read every parenthetical note in the specialty's subsections (Chapter 13 §13.6); parentheticals are instructions wearing punctuation.

2. The NCCI Policy Manual chapter for that CPT section. It is free, it is organized by section, and it states in prose what the edit file states in code pairs — including the rationale, which is what lets you predict edits you have not met yet. Chapter 21 §21.5 taught you to read it before you override anything; here you read it before you code anything.

3. The coverage policies that govern the specialty's top services. Your MAC's local coverage determinations and billing-and-coding articles (Chapter 22 §22.4, §22.5), plus the published medical policies of the two or three commercial plans that carry the practice's volume (Chapter 22 §22.10). You are not reading them all. You are reading the ones attached to the services on the list you build on day one, which is usually four to eight documents.

4. The specialty society's coding resources. Every major specialty has one — the American College of Cardiology, the American Academy of Orthopaedic Surgeons, the American College of Obstetricians and Gynecologists, the American Academy of Pediatrics, the American College of Emergency Physicians, the American Society of Anesthesiologists — and each publishes coding guidance, frequently-asked-question files, and specialty coding manuals. They are the fastest correct answer to a specialty question and they are written by people who do the work.

⚠️ With one caveat you must hold onto. A society's guidance is an authoritative reading of the code set, and it is also advocacy for its members' interests. Where it conflicts with a payer's published policy, the payer's policy governs the claim — and the society's position is what you cite in an appeal, not what you rely on when you submit. Chapter 22 §22.10's discipline exactly. Society guidance is Tier 2 in this book's terms: attributed, useful, and not the rule.

Week one — build the map

Days 1–2: build the top-twenty list from the practice's own data. Run a frequency report of the practice's CPT and HCPCS codes for the last twelve months, sorted two ways: by volume and by charge dollars. The lists differ, and both matter — the volume list is what you will code every day; the dollar list is what you cannot afford to get wrong. Twenty of each, typically thirty distinct codes between them.

This is the step people skip, and skipping it is why textbook specialty knowledge does not transfer. No textbook knows that this cardiology group does almost no interventional work, or that this orthopedic practice is 60% spine. The practice's own claims history is the only accurate description of what you are about to do all day.

Day 2, second half: annotate each code. One grid, one row per code, six columns:

   THE SPECIALTY GRID — one row per code on the top-twenty lists

   CODE | WHAT THE DESCRIPTOR ACTUALLY SAYS | WHAT'S BUNDLED IN | GLOBAL
        | SPLITS INTO COMPONENTS? | MUE / EDIT NOTES | POLICY THAT GOVERNS

   Rules
     · "What the descriptor actually says" is written in your own words,
       from the current book. If you cannot paraphrase it, you do not know
       the code yet.
     · "What's bundled in" comes from the section guidelines, the
       parentheticals, and the NCCI Policy Manual chapter.
     · "Global" is 000 / 010 / 090 / XXX / YYY / ZZZ (Chapter 17 §17.2).
     · "Splits into components" is a fee schedule lookup, not a guess
       (Chapter 19 §19.1).
     · "Policy that governs" is a document name and a date, not a memory.
                                                 [schematic — not to scale]

Day 3: learn the source document before its content. Sit with ten real examples of whatever document the codes come from — catheterization reports, operative reports, anesthesia records, therapy flowsheets, immunization records. Learn the layout first: where the findings are, where the times are, where laterality is stated, where the signature and date live, what the standard headings are and which one the practice's dictation system tends to leave empty. You are learning to find things, not yet to judge them.

Days 4–5: shadow, in four half-days. Half a day with a clinician. Half a day with whoever currently does this coding. Half a day at the front desk — this is where the specialty's authorization and eligibility burden lives and it is invisible from the coding chair (Chapter 24 §24.5). Half a day with the biller who works the specialty's denials.

Ask one question in each seat and write down the answer: "What do you wish the coder knew?"

Week two — code, compare, and keep a log

Days 6–8: code in parallel. Take a real day's charts and code them independently, then compare with the person who codes them now. Code first, compare after. Checking your work against someone else's as you go teaches you their answers; coding blind and then comparing teaches you your own errors, which is the only thing worth learning. Every disagreement gets written down, with which of you was right and why — including the ones where neither of you was sure, because those are the specialty's genuine ambiguities and they will recur.

Day 9: read the denials. Pull the last ninety days of the specialty's denials and sort them by claim adjustment reason code (CARC) and root cause (Chapter 28 §28.4, Chapter 29 §29.4). This is the single highest-yield hour of the two weeks. The specialty's characteristic failure modes are in that list, ranked by frequency, already paid for by somebody else's mistakes. You are being handed a map of what you are about to get wrong.

Day 10: start the error log, and keep it for a month. One line per error or uncertainty: the date, the code, what you did, what was right, why, and where the answer came from. Review it weekly. The log is the deliverable — not the memory of having learned something. By day 30 it will have a shape: three or four recurring themes, which are your actual gaps, and which are then fixable in an afternoon each rather than by "being more careful."

What two weeks does not buy

It does not buy the tail. You will be competent on the top twenty and genuinely uncertain on the unusual case — which is fine, because the unusual case is exactly where the correct professional behavior is to look it up and to ask. Chapter 6 §6.10 built that habit; a new specialty is where it pays.

It does not buy the clinical fluency that lets you read a report quickly. That comes at about month three and there is no shortcut.

It does not buy authority. In the first month you should be checking your unusual answers with somebody, and saying so out loud. A coder who says "I'm not sure — let me look at the section guideline and the policy and come back to you in an hour" is doing the job correctly. A coder who guesses confidently in a specialty they learned two weeks ago is a compliance problem wearing a willingness to help.

And it does not buy a credential. Chapter 39 §39.3 covers the specialty credentials and what each one signals.

📞 On the Phone

Day 4 of the two weeks, sitting with a physician between cases. One question, asked plainly.

"I'm the new coder. I've read the section guidelines and I've built a list of the twenty codes you generate most. Before I start coding your notes — what do you wish the coder knew?"

What comes back is almost never what you expect, and in this book's experience it is one of three things:

"Nobody ever tells me what they need." — The most common answer, and the most useful. It means the documentation gaps you are about to find have never been described to the person who could close them, and a one-page list of the five phrases that decide codes in this specialty will be received as help rather than as criticism.

"The last coder kept changing my codes and never said why." — This is a relationship to repair rather than a fact to record. The answer is a standing practice: when you change a code, you say which sentence in the note drove it. Every time, in one line.

"Why does the plan keep denying [X]?" — Write the question down verbatim. It is usually the specialty's biggest coverage problem, stated by the person who feels it, and you will find it again on day 9 in the denial report.

The failure modes. Asking for a tutorial on the medicine — the physician's time is expensive and the anatomy is in a book. Arriving with a list of the physician's documentation errors during the first week, which converts a colleague into a defendant. And asking nothing, which is the most common of all, and which costs you the one conversation nobody else in the building can have for you.

🗂️ The Encounter — the same knee, in orthopedics

Account 10-4471 was managed at Northgate Family Medicine and closed at a zero balance on day 100. Nothing below changes any of that. This checkpoint is a lens: the same 58-year-old, the same right knee, the same six weeks of pain — referred instead to an orthopedic practice, and coded by a specialty coder. (Constructed counterfactual; the real file is unchanged, and this checkpoint prices nothing.)

What this chapter contributes: five things about the claim change, and none of them is the knee.

The E/M category changes. She has never been seen by the orthopedic group, so she is a new patient — the 99202–99205 range under Chapter 15 §15.2's three-year rule — not the established-patient 99214 the real file carries. Same patient, same day, different practice, different family.

The elements supporting the E/M change completely, and this is the checkpoint's real finding. The real claim's modifier 25 argument rests on four documented elements frozen in Chapter 4's Figure 4.2, and three of them have nothing to do with the knee: three chronic conditions each separately assessed with a plan, prescription drug management across three medications, and two laboratory tests ordered with stated reasons. The orthopedist addresses none of those. Diabetes, hypertension, and hyperlipidemia are on her problem list and are not this specialist's work — exactly the finding Chapter 16 §16.10 made when the same knee went to the emergency department, and Chapter 15 §15.5's "addressed" rule doing the same job a third time.

So in the orthopedic version, the entire separately-identifiable argument must be built from the knee alone: a new-patient evaluation, a differential, imaging ordered and interpreted, and a management decision. Same modifier, same patient, same joint — and a materially weaker and more contestable argument, because every element in it is arguably the evaluation that precedes the procedure. Chapter 14 §14.4 closed Q1 on the real file with four elements; the counterfactual would have to win with one, and specialists lose that argument to payers routinely. The lesson transfers: modifier 25's strength is not in the modifier. It is in how much of the documented work has nothing to do with the procedure.

The procedure code may change. An orthopedic practice with ultrasound in the room is materially more likely to inject under guidance — which is 20611 rather than 20610, and which makes "permanent recording and report" a documentation requirement rather than a nicety. Chapter 17 §17.7 froze that contrast on this exact file: the real note's clause "no imaging guidance used" is what settles it there. In the counterfactual, the note has to say the opposite thing, and say it with a recording and a report behind it, or 20610 is still the correct code.

The diagnosis may firm up sooner. The orthopedist obtains films at the visit — a knee radiograph, three views (73562), the same code Chapter 16 §16.10's counterfactual used — and a report describing degenerative change would support M17.11 on that date. On the real March 14 encounter it does not, because Figure 4.2's frozen clause states plainly that no definitive diagnosis was established and no prior imaging of this knee was available. M25.561 was correct for March 14, and Chapter 22 §22.11 closed Q5 on exactly that reasoning. The counterfactual changes the imaging, not the history.

And one line simply disappears. The orthopedist does not order the hemoglobin A1c or the lipid panel, so the venipuncture line — 36415, the one line on the real claim that points at diagnosis B rather than at the knee — does not exist, and neither does the \$3.00 Chapter 19 §19.12 traced to it. The real claim's four lines become three — the visit, the injection, and the drug — or two, if the E/M does not survive the separately identifiable test above. And every remaining line points at one diagnosis, because there is only one.

One thing that does not change: the global period. 20610 and 20611 both carry a 000-day global (frozen at Chapter 14 §14.4 and Chapter 17 §17.2). An orthopedic practice's 090-day machinery — the modifier 24/58/78/79 discipline of §35.3 — is not engaged by this encounter at all. Specialty coding is not "the specialty's rules apply to everything the specialist does"; it is "the code's rules apply, and the specialist happens to generate different codes."

What the lens settles. That "the same patient" is not the same claim. The specialty changed the E/M category, the raw material of the modifier 25 argument, the likely procedure code, the certainty of the diagnosis, and the number of lines — without changing one fact about the knee.

What it does not settle. The money. This checkpoint deliberately prices nothing: the real file is frozen at \$367.00 in charges, \$216.28 allowed, \$47.58 to the patient, \$168.70 to the plan, and no counterfactual in this chapter constructs a competing set. A new-patient E/M's allowed amount and 20611's are contract terms, and Chapter 23 §23.6 already showed that Northfield's contract is priced per line rather than by a formula anyone can apply to a code it has not priced.

The open questions. Unchanged. Only Q4 — could the denial have been prevented? — remains open, and it belongs to Chapter 40. This checkpoint adds none.

Summary

There is no separate code set for a specialty. The books, the Guidelines, the edits, and the modifiers are the ones you already know. What differs is five things: the volume slice (twenty codes are most of the work), the payment convention (service, session, episode, unit of time, acuity level, calendar window), the source document, the concentrated policy landscape, and the documented words that decide codes. The payment convention is the one that produces catastrophic rather than annoying errors, because it makes every claim wrong the same way.

Cardiology is a component question and a bundling question. 93000 is the whole electrocardiogram service; 93010 is the interpretation and report only; a third code reports the tracing alone — and what decides among them is who owned the machine and who wrote the report, not what happened to the patient. "Interpretation and report" means a separate, retrievable, signed document. 93306 turns on the word complete. 93458 — left heart catheterization with coronary angiography — already includes the catheter placement, the injections, and the imaging supervision and interpretation (Chapter 18 §18.4); reporting them separately is the exams' favorite trap and a real unbundling risk.

Orthopedics is a decision documented in the plan, and a life inside global periods. Global fracture care (090 days, follow-up included) versus an E/M plus casting is decided by who is providing definitive care through healing, stated in the record — and the practice-level answer is a written rule, because otherwise a macro decides. Inside a global, modifier 24 (unrelated E/M), 58 (planned — restarts the global), 78 (unplanned, related — does not restart), and 79 (unrelated procedure) carry the assertions; 24 appended by rule is the invisible failure.

The obstetric global package spans months, and the errors come from the calendar. 59400 (vaginal) and 59510 (cesarean) cover antepartum care, delivery, and postpartum care in one code; when the episode breaks, report the components actually furnished (Chapter 18 §18.6). Outside the package: unrelated problem visits, documented complications, laboratory beyond routine urinalysis, ultrasounds and non-stress tests, and visits beyond the routine schedule. The operational control is an episode log — visit count kept at the visit, eligibility re-verified on a schedule, outside services flagged the day they happen.

Pediatrics and primary care depend on facts outside the note. Preventive medicine services (99381–99387 new, 99391–99397 established) are selected by age and new-versus-established, not by decision making — and Medicare does not cover a routine physical, paying instead for the annual wellness visit (G0439 for subsequent ones). Every immunization is two codes: product plus administration — 90686 + 90471, and 90472 for each additional. Omitting the administration is a silent underpayment; adding a product the practice did not purchase is billing for somebody else's vaccine. 99495 and 99496 are transitional care management: a two-business-day contact, a face-to-face visit within 14 or 7 days, and medication reconciliation — a service defined by a calendar window, not by an encounter.

The emergency department is two claims and two rulebooks. The professional family (99281–99285) has no new/established distinction and no time option; the facility level reports the department's resource intensity against criteria the hospital writes itself — which must be written, resource- based, reproducible, auditable, and productive of an explainable distribution. A facility level that mirrors the physician's on every claim is itself a finding, and a skewed distribution is visible from outside the organization without a single chart. On the ED anchor, both claims carry 99284, and together they allow \$1,515.00 of \$4,522.00 in charges, \$503.00 of it the patient's.

Anesthesia is built from the anesthesia record, in order. One anesthetic, one anesthesia code (the highest base unit value). Anesthesia time is neither room time nor surgical time. Physical status P1–P6 is assigned by the anesthesia professional, not upgraded by a coder. The provider-arrangement modifier depends on the day's schedule, and medical direction is a documented checklist audited as one. On the worked record, eight minutes of room time that were not anesthesia time were worth one unit — and the same 98 minutes legitimately produces two different unit counts under two published rounding rules. One of those differences is policy; the other is a false statement.

Time has four conventions in this book and each is plausible in the others' territory: E/M total time on the date, the eight-minute rule for timed therapy (97110), psychotherapy's typical durations, and anesthesia's continuous clock. Document the duration and what happened in it — and never let a template supply the number.

Telehealth is a modality, not a specialty, and it is the least stable material here. Place of service 02 (not in the patient's home) and 10 (in the patient's home); modifier 95 asserts synchronous audio and video. Payers combine the place of service and the modifier differently, and the combination moves money. Teach yourself where the list lives rather than what it currently says, and verify before every unfamiliar telehealth claim.

And a specialty is learnable in two weeks to a defined level. Read the section guidelines and the NCCI Policy Manual chapter before coding anything; pull the policies that govern the top services; build the top-twenty lists from the practice's own claims history, twice, by volume and by dollars; annotate each code on a six-column grid; learn the source document's layout; shadow four half-days and ask "what do you wish the coder knew?"; code in parallel and compare after, never during; read ninety days of the specialty's denials; and keep an error log for a month. What that buys is competence at the top twenty and safety at the tail. It does not buy the tail, the clinical fluency, or the credential, and saying so out loud is part of doing the job correctly.

Key Terms

Specialty coding — coding the same three code sets under a specialty's own volume slice, payment convention, source document, policy landscape, and documentation vocabulary; not a separate code set and not a separate discipline. (Ch.35)

The specialty intake — the five questions answered in writing before coding in an unfamiliar specialty: what is the source document, what is bundled into what, what is the unit of payment, which policies govern the top services, and which twenty codes are the volume and which twenty are the dollars. (Ch.35)

Cardiac catheterization family — the group of catheterization codes selected on two axes (what was catheterized; what was imaged), whose descriptors already include the catheter placement, the injection procedures, and the imaging supervision and interpretation. (Ch.35; the bundling rule is Ch. 18 §18.4)

The component question — the per-code determination of whether a service is being reported globally, as a professional component, or as a technical component, decided by who owned the equipment and who produced the signed interpretation rather than by what happened clinically. (Ch.35; the arrangements are Ch. 19 §19.1)

Global orthopedic fracture care — reporting fracture treatment with a code carrying a 090-day global period that includes the initial treatment, the first cast or splint, and normal follow-up through healing; correct only where the reporting physician is providing definitive care, as documented in the plan. (Ch.35; the fork itself is Ch. 17 §17.7)

Obstetric global package — antepartum care, delivery, and postpartum care reported with one code (59400 vaginal, 59510 cesarean); operationally, the episode a practice must be able to prove it furnished in full before reporting it that way. (Ch.35; the package's definition and exclusions are Ch. 18 §18.6)

The obstetric episode log — the per-patient record opened at the first prenatal visit that carries the antepartum visit count kept contemporaneously, the eligibility re-verification dates, and every service furnished outside the package. (Ch.35)

Preventive medicine services — the comprehensive periodic evaluation family (99381–99387 new, 99391–99397 established) selected by patient age and new-versus-established status rather than by medical decision making or time, and defined by required content. (Ch.35)

Vaccine administration pairing — the discipline of reporting the vaccine product and its administration together (90686 with 90471, plus 90472 for each additional administration), including the correct exception where the product was supplied by a public program and only the administration may be billed. (Ch.35; the two-code rule is Ch. 19 §19.10)

Transitional care management (TCM) — 99495 and 99496, covering the 30-day period beginning on the date of discharge to a community setting: interactive contact within two business days, a face-to-face visit within 14 or 7 calendar days, and medication reconciliation; the face-to-face visit is part of the service. (Ch.35)

ED facility acuity leveling — assignment of an emergency department facility visit level from the hospital's own written criteria measuring resource intensity, in the absence of any national criteria; distinct in both scale and subject from the physician's medical-decision-making level. (Ch.35; the professional/facility split is Ch. 16 §16.9)

Acuity criteria set — the hospital's written, resource-based, reproducible, and auditable rules for assigning facility ED levels, together with the record of how each encounter was scored. (Ch.35)

The distribution test — comparing a department's facility level distribution against its own case mix and its peers; a question rather than a finding, answerable only from the written criteria and scored records, and never answered by lowering levels. (Ch.35)

Anesthesia physical status modifier — P1 through P6, appended to an anesthesia code to report the patient's condition; assigned and documented by the anesthesia professional, recognized for additional units by some payers and not others, and never adjusted by a coder. (Ch.35; defined in Ch. 18 §18.11)

Time-based units — units of service derived from documented time, under one of four distinct conventions (E/M total time on the date, the eight-minute rule for timed therapy, psychotherapy's typical durations, and anesthesia's continuous clock), each requiring both a duration and a record of what occurred in it. (Ch.35)

Telehealth place of service — place of service 02 (telehealth other than in the patient's home) and 10 (telehealth in the patient's home) on the professional claim, reported in combination with modifier 95 or not according to each payer's current published policy. (Ch.35)

Spaced Review

1. A cardiology group reads electrocardiogram tracings produced on a hospital's equipment and dictates signed interpretations. Which code applies, which one would be wrong, and what would have to change about the arrangement — not about the patient — for the other code to become correct? Then state what "interpretation and report" requires as a document.

2. (Chapter 34) The same emergency department visit produces a facility claim and a professional claim, both carrying 99284. State what each 99284 measures, who wrote the rulebook for each, and why an auditor asking to see the hospital's written criteria is asking a better question than a payer analyst flagging the two claims as a mismatch.

3. (Chapter 17) A patient with a distal radius fracture is seen by an orthopedist who documents: "Cast applied. Will follow weekly until union." Which side of the fracture-care fork does that plan put the encounter on, what is the global period, which service is included and which supply is still separately reportable, and what one fact must still be documented before a fracture care code can be selected?

4. (Chapter 18) A practice provides antepartum care from 8 weeks, the patient moves at 30 weeks and delivers elsewhere. State what this practice reports and what it does not, name the single piece of information the claim depends on, and explain why that information is nearly impossible to reconstruct in month nine.

5. Four time conventions appear in this book. For each of the following, name the convention and the documentation it requires: an anesthesia case running 98 minutes; 20 minutes of therapeutic exercise plus 10 minutes of manual therapy on one date; a 40-minute psychotherapy session; and an office visit selected on 45 minutes of total time. Which two of the four would give a wrong answer if you applied the eight-minute rule to them, and why?

Next: Chapter 36. The last chapter of Part VII, and the one that reopens the book's oldest question. Under risk adjustment, a diagnosis code stops being a justification for a service and becomes a description of a population's expected cost for the coming year — which is why Account 10-4471's diabetes line, coded E11.9 on day 1 and correct ever since, is about to be revisited.