Chapter 8: Finding the Code: The Alphabetic Index, the Tabular List, and Why You Never Code from Memory

"Every coder who has ever been fired for a coding error was, at the moment they made it, entirely certain." — constructed

Overview

This is the chapter where you learn to do the job.

Chapter 7 was the map. This is navigation, and it consists of one rule that sounds trivial and is not:

Alphabetic Index first. Tabular List second. Always. For your entire career.

Both steps. Every time. Including for codes you have used four hundred times.

Every experienced coder in America has heard that rule and most of them break it daily, because after the four hundredth time you do know that knee pain is M25.561 and looking it up feels like theater. The rule survives anyway, and this chapter is the argument for it — which is not "you might be wrong about the code." It is that the index and the tabular contain different information, and skipping either one means you did not see something.

The index tells you where to look. It does not tell you the rules. The instructions that govern a code — the seventh character requirement, the Excludes note, the "code first," the laterality options — live in the Tabular, at the category level, printed once, several lines above the code you landed on. A coder who codes from the index has not read them and does not know they exist.

That is the whole of it. Everything else in this chapter is the detail.

In this chapter, you will learn to:

• State the two-step rule and say what each step protects against
• Find a main term and follow subterms to a code
• Distinguish essential from nonessential modifiers
• Read the Tabular's instructional notes and know which are mandatory
• Apply Excludes1 and Excludes2, including the exception
• Distinguish "code first," "use additional code," and "code also"
• Distinguish NEC from NOS
• Use the three specialty tables

Learning Paths

🎓 Certification — This is the chapter. The exams are open book and timed, which means they test navigation. §8.5 (Excludes1 versus Excludes2) is examined on every credential without exception, and §8.7 (NEC versus NOS) nearly as often.

💼 New Coder — All of it, and then §8.10 again with your own book open. Do the five lookups yourself before reading the answers.

💵 Biller / AR — §8.5 and §8.6 explain a category of denial you will otherwise find inexplicable: claims denied because two codes cannot appear together, or because a required second code is missing.

🏥 Practice Manager — §8.1's argument is the one to internalize, because it is the answer to "why is coding slower than I expected?"

8.1 The two-step rule, and the career it protects

Step one: the Alphabetic Index. Look up the condition by its name. The index gives you a code, or part of one.

Step two: the Tabular List. Go to that code. Read it. Read the category above it. Read every instructional note attached to either.

Then assign.

Why step one is not optional

You cannot navigate the Tabular directly for most conditions, because the Tabular is organized by classification logic and you are searching by clinical language. A physician writes "cellulitis"; the Tabular files it under a category whose heading you would not have guessed.

More importantly: the index knows synonyms and you do not. It carries the alternate names, the eponyms, the abbreviations, and the lay terms, and it maps all of them to the same place. A coder browsing the Tabular for a term the classification files under a different name will not find it and will conclude, wrongly, that no code exists.

Why step two is not optional

Because everything that governs the code is there and none of it is in the index.

Read that list again and notice what it means: a coder who codes from the index alone can produce a code that is invalid, incomplete, mutually exclusive with another code on the same claim, or missing a required companion — and none of those failures is visible from the index.

⚠️ Where Claims Die

"I know this one."

The most dangerous four words in coding, and the reason the rule is stated as an absolute rather than as advice.

The failure is not that experienced coders misremember codes. It is that codes change and instructions change, and a coder who stopped looking things up in 2019 is coding from a 2019 book they are carrying in their head. ICD-10-CM changes every October 1, and the change that matters most is not a new code — it is a new instruction attached to a code that still exists (Chapter 6 §6.7), which produces no rejection at all.

The coder who looks it up encounters the new instruction. The coder who knows it does not, and will not, until an audit.

And the professional version of the argument: two years from now, someone who was not there will ask you why you assigned this code. "I knew it" is not an answer. "Main term pain, subterm joint, subterm knee, verified in the Tabular, no Excludes applied, no seventh character required" is an answer. You are not looking it up for the code. You are looking it up for the path.

📞 On the Phone

This conversation happens in every business office and it is worth having an answer ready.

Manager: "You've coded this exact visit two hundred times. Why does it take four minutes?"

What not to say: "Because we're supposed to." True, and it concedes the premise that the rule is ceremonial.

What works: "Most of them don't take four minutes — most are under two. The four-minute ones are the ones where I open the book and find something. Last month that was the category that added a laterality character in October; we'd have been billing an invalid code on every one of those. The lookup is cheap and it's the only thing that catches a change nobody told us about."

Then, if you can, bring an example. One real instance of a lookup that caught something is worth more than any argument from principle, and if you keep the question file from Chapter 6 §6.10, you have one.

8.2 The Alphabetic Index: main terms and subterms

The index is organized by main terms, in bold, in alphabetical order, with subterms indented beneath them.

And here is the rule that beginners get wrong more than any other:

Main terms are conditions, not body parts.

You look up what is wrong, not where it is wrong.

• "Knee pain" → look up Pain, not Knee
• "Fractured radius" → look up Fracture, not Radius
• "Kidney infection" → look up Infection or Pyelonephritis, not Kidney
• "Diabetic retinopathy" → look up Diabetes, not Retinopathy or Eye

The index knows you will try the body part anyway, and it has a standard answer:

   Knee — see condition

That is the index telling you that you have looked up the wrong kind of word. It is not a dead end; it is an instruction. Go back and look up the condition.

Main terms can also be

• The condition itself — Pneumonia, Fracture, Cellulitis
• A noun describing the encounter — Encounter, Examination, Screening, Aftercare, History
• An eponym — a condition named for a person
• An abbreviation — many are indexed

Following subterms

Subterms are indented, and indentation level is meaning. Each level narrows the one above it.

   Pain(s)  R52                         ◄── main term, with its own default code
     abdominal  R10.9
     joint  M25.50                      ◄── 1st-level subterm
       ankle  M25.57-
       elbow  M25.52-
       hand  M25.54-
       hip  M25.55-
       knee  M25.56-                    ◄── 2nd-level subterm. The dash means
       shoulder  M25.51-                    MORE CHARACTERS ARE REQUIRED.

Read the indentation carefully. In dense index pages the levels are easy to misread, and following a subterm that belongs to a different parent produces a code from an entirely different category.

The dash

A code ending in a dash — M25.56- — means the index cannot finish the job. More characters are required, and only the Tabular can tell you what they are.

The dash is the index explicitly telling you to go to step two. A coder who reports M25.56 has reported an invalid code, and the index warned them.

🔢 Code It

Four conditions, four main terms. Name the main term you would look up.

The documentation says	Main term
"Right knee pain"	Pain
"Left femoral neck fracture"	Fracture
"Screening colonoscopy"	Screening
"COPD with acute exacerbation"	Disease (chronic obstructive pulmonary) — or COPD as an abbreviation entry

The plausible wrong answers, named:

• Knee, Femur, Colon, Lung. Body parts. The index will say "see condition."
• Colonoscopy for the third one. That is a procedure, and this is the diagnosis index — the reason for the encounter is screening, and the main term is Screening.
• Exacerbation for the fourth. It is an adjective describing a state, not the condition. The condition is the COPD.

And a real subtlety: many conditions have more than one legitimate entry point. Pneumonia can be reached through Pneumonia directly; a diabetic complication through Diabetes. When two paths exist they should converge on the same code — and if they do not, that is a signal to slow down, because one of them is probably a different condition than you think.

8.3 Essential and nonessential modifiers

Two kinds of qualifying words appear in the index, they look similar, and they behave in opposite ways.

Nonessential modifiers — in parentheses

Words in parentheses immediately following a main term or subterm. They are alternate wording that does not change the code, and they do not have to be present in the documentation.

   Hypertension (accelerated) (benign) (essential) (malignant)
   (primary) (systemic)  I10

Every one of those parenthetical words is optional. Documentation saying "hypertension," "essential hypertension," or "benign hypertension" all reach I10. The parentheses are the index telling you: these words do not matter here.

Essential modifiers — indented subterms

Words that appear as indented subterms are essential. They do change the code, and they must be supported by the documentation.

   Pain(s)  R52
     joint  M25.50            ◄── ESSENTIAL. "Joint" must be documented
       knee  M25.56-          ◄── ESSENTIAL. "Knee" must be documented

You may not follow a subterm the record does not support. If the note says "pain" with no site, you do not descend to joint because you suspect a joint. You stay where the documentation is.

   THE DIFFERENCE, IN ONE PICTURE

   Hypertension (accelerated) (benign) (essential) ...  I10
                └──────────── PARENTHESES ───────────┘
                   NONESSENTIAL — do not need to be documented,
                   do not change the code

   Pain(s)  R52
     joint  M25.50
       knee  M25.56-
     └─┬──┘
       INDENTED SUBTERMS
       ESSENTIAL — must be documented, DO change the code

🎓 Exam Watch

This distinction is examined, and the question is usually built to punish the intuition that "more words in the note means a more specific code."

A scenario documents "benign essential hypertension." Four options are offered, three of which are more specific-looking codes. The answer is I10 — because "benign" and "essential" are nonessential modifiers in parentheses and change nothing.

The reverse trap: a scenario documents "pain" with no site, and the options include a joint-specific code. You may not descend to a subterm the documentation does not support, and the answer is the less specific code.

The general rule to hold: parentheses are permission to ignore; indentation is a requirement to document.

8.4 The Tabular List and the instructions that live there

Step two. Here is everything you are reading for.

Inclusion terms

Listed beneath a code, these are examples of conditions the code covers. They are not exhaustive — other conditions may be classified here too, and the index is the arbiter.

"Includes" notes

Appearing at the category level, these further define or give examples of the category's content. They apply to everything beneath.

"Excludes1" and "Excludes2"

§8.5. The two most consequential notes in the book.

"Code first," "use additional code," "code also"

§8.6.

"See" and "see also"

• "See" is mandatory. It means: the code is not here, go where I am pointing you. Knee — see condition.
• "See also" is advisory. It means: there is another entry that may be more appropriate, and you should look at it — but if the entry you are on has the code you need, you may use it.

Where the instructions live

This is the structural point that makes step two necessary, and it is worth a diagram.

INSTRUCTIONS CASCADE DOWNWARD

  CHAPTER          ┌──────────────────────────────────────────┐
                   │  Chapter-level notes and Excludes.       │
                   │  Govern EVERY code in the chapter.       │
                   └──────────────────┬───────────────────────┘
                                      ▼
  BLOCK            ┌──────────────────────────────────────────┐
                   │  Block-level notes and Excludes.         │
                   │  Govern every category in the block.     │
                   └──────────────────┬───────────────────────┘
                                      ▼
  CATEGORY         ┌──────────────────────────────────────────┐
                   │  Includes, Excludes1, Excludes2,         │
                   │  code first / use additional code,       │
                   │  SEVENTH CHARACTER REQUIREMENT.          │
                   │  Govern every code in the category.      │
                   └──────────────────┬───────────────────────┘
                                      ▼
  SUBCATEGORY      ┌──────────────────────────────────────────┐
                   │  Further notes, further Excludes.        │
                   └──────────────────┬───────────────────────┘
                                      ▼
  CODE             ┌──────────────────────────────────────────┐
                   │  The descriptor. Inclusion terms.        │
                   │  THE PART THE INDEX POINTED YOU AT.      │
                   └──────────────────────────────────────────┘

  THE INDEX DROPS YOU AT THE BOTTOM.
  EVERYTHING ABOVE IT STILL APPLIES, AND YOU HAVE NOT READ ANY OF IT.

That diagram is the entire argument of this chapter. The index is a search function. It finds a string. It does not carry the four levels of instruction sitting above what it found, and those instructions are binding.

When the index and the Tabular seem to disagree

It happens, and new coders find it genuinely disorienting. Four situations, and only one of them is a real conflict.

1. The index gives a code the Tabular says is invalid. Almost always a dash you did not notice, or an index entry that has not been read down to its last subterm. Go back to the index and read the entry again, all of it, including the indentation level you were on.

2. The index gives a code and the Tabular attaches an instruction the index never mentioned. Not a conflict — this is the system working exactly as designed. The Tabular's instruction governs. Follow it.

3. The index sends you somewhere the Tabular excludes. Read the Excludes note carefully; it frequently excludes a different condition than the one you have, and the resemblance is in the words rather than in the clinical facts. If it genuinely applies, you are in the wrong category and the index entry you followed was probably not the right main term.

4. A genuine conflict. Rare, and the rule is:

The Tabular governs. The index is a locator; the Tabular is the classification.

But before you conclude that, exhaust the first three — because in practice, three times out of four, what looks like a conflict is a subterm not fully read.

And when a genuine ambiguity survives all of that, it is a question with an answer somewhere: the Official Guidelines, Coding Clinic, or your MAC. Chapter 6 §6.6's routing table exists for exactly this moment, and "I picked one" is not a path you can document.

📞 On the Phone

The conversation with a more experienced colleague, done well.

What not to say: "What's the code for a knee injection follow-up?" You will get a code, you will not get a path, and you will be back next week with the same class of question.

What works: "I've got the index taking me to M25.56-, and the Tabular's giving me laterality in the sixth character, but the note only says 'knee' in the assessment. The exam says right. Do you read that as documented, or do I query?"

That question does three things: it shows the path so a colleague can spot a wrong turn, it names the actual decision point, and it asks about judgment rather than about a code. You will get a better answer, and — the part that matters over a career — you will be building the habit of thinking in paths rather than in answers.

8.5 Excludes1 and Excludes2: the most expensive two words in the book

Two notes, one letter apart, meaning opposite things.

Excludes1 — "NOT CODED HERE"

A pure exclusion. The two conditions cannot occur together and must never be reported together. The classification has decided they are alternatives — a congenital form and an acquired form of the same condition, for instance.

If you land on a code with an Excludes1 note naming a condition your patient has, you have the wrong code.

Excludes2 — "NOT INCLUDED HERE"

The excluded condition is not part of the code you are looking at, but a patient may have both conditions at the same time, and both may be reported together when both are documented.

                    EXCLUDES1                      EXCLUDES2
                    ─────────                      ─────────
   MEANS            "NOT CODED HERE"               "NOT INCLUDED HERE"

   THE TWO          cannot occur together          CAN occur together
   CONDITIONS

   MAY YOU          NO — never both                YES — both, when both
   REPORT BOTH?                                    are documented

   IF YOUR          you have the WRONG code        you may need BOTH codes
   PATIENT HAS
   THE EXCLUDED
   CONDITION

   MEMORY AID       ONE means ONE of them          TWO means you may need TWO

That memory aid is worth keeping: Excludes1 — only one of these. Excludes2 — you might need two.

The exception to Excludes1

Because nothing here is quite that simple.

The Official Guidelines address the situation where an Excludes1 note pairs two conditions that turn out to be unrelated in a particular patient. Where the two conditions are genuinely unrelated to each other, both may be reported despite the Excludes1 note — and where there is doubt about whether they are related, the guidance directs you to query the provider.

This exception exists because the classification's authors, writing at scale, occasionally paired conditions that can genuinely coexist independently. It is narrow, it is not a general escape hatch, and the safe habit is: an Excludes1 note means stop, and the exception is something you establish affirmatively rather than assume.

⚠️ Where Claims Die

The Excludes1 pair that both got reported.

This is one of the most common edit failures in outpatient claims, and it is invisible from the index. Two codes are assigned by two different lookups, both look correct, and neither coder ever saw the Excludes1 note because both stopped at step one.

The claim denies, or — in a facility setting — the grouper produces a different result than expected, and the diagnosis is a lookup nobody performed.

The habit: when you assign a second code to the same claim, check whether either one excludes the other. Thirty seconds in the Tabular, and it catches a category of error that is otherwise found only by an edit.

🎓 Exam Watch

Excludes1 versus Excludes2 is guaranteed on every diagnosis credential. Usually in one of two forms:

Form 1: definitional. "An Excludes2 note indicates that…" Answer: the excluded condition is not part of the code, but both may be reported together if both are documented.

Form 2: applied. A scenario documents two conditions; the Tabular excerpt shows one of the two notes; you decide whether to report one code or two.

The trap is the intuition that "excludes" means "do not report" in both cases. Excludes2 is permission, not prohibition — it is telling you the code you are on does not cover the other thing, so if the patient has the other thing, you need another code.

8.6 "Code first," "use additional code," and "code also"

Three instructions about companion codes, and they differ in what they require and in what they say about sequencing.

"Code first" and "use additional code" are two ends of the same relationship, seen from opposite sides. They frequently appear as a matched pair: the etiology code says "use additional code" and the manifestation code says "code first."

"Code also" is the one people misread. It does not mean "you may also want to consider." It means two codes are needed — and it deliberately declines to tell you which goes first, because that depends on which condition was the reason for the encounter. Chapter 9 §9.9 covers sequencing.

THE MATCHED PAIR, IN PRACTICE

  ┌─────────────────────────────────────────────┐
  │  THE ETIOLOGY CODE                          │
  │  "Use additional code to identify ..."      │
  │                                             │
  │      ────────────────────────────────►      │  sequenced FIRST
  └─────────────────────────────────────────────┘
                      │
                      │  the same relationship,
                      │  documented from both ends
                      ▼
  ┌─────────────────────────────────────────────┐
  │  THE MANIFESTATION CODE                     │
  │  "Code first the underlying disease"        │
  │                                             │
  │      ◄────────────────────────────────      │  sequenced SECOND
  └─────────────────────────────────────────────┘

  IN THE INDEX, the manifestation code appears in BRACKETS after
  the etiology code — which is the index's way of showing you the
  pair and the order at once.

When a code carries "code first" and you cannot identify the underlying condition, you have a documentation problem rather than a coding problem, and §4.9's query is the tool.

8.7 NEC and NOS

Two abbreviations that look interchangeable and tell you opposite things about the documentation.

NOS — "not otherwise specified"

Equivalent to "unspecified." It means the documentation did not provide enough detail to assign a more specific code.

The limitation is the record.

NEC — "not elsewhere classifiable" (or "not elsewhere classified")

The documentation is specific — and the classification has no code for it. The provider told you exactly what the condition is, and ICD-10-CM does not have a category for that particular thing, so it goes in the residual "other specified" code.

The limitation is the code set.

   NOS  —  the DOCUMENTATION isn't specific
           "knee pain"  →  which knee?  →  unspecified

   NEC  —  the DOCUMENTATION is specific, the CODE SET isn't
           a precisely named condition with no category of its own
           →  "other specified ..."

   These point in OPPOSITE directions, and only one of them
   can be fixed by a query.

Only NOS can be fixed by a query. If the documentation is insufficient, ask. If the classification has no code, asking the provider a more precise question will not produce one, and a coder who queries repeatedly hoping for a codeable answer has misdiagnosed the problem.

A practical tell: codes whose descriptors contain "other" are usually the NEC codes; codes whose descriptors contain "unspecified" are usually the NOS codes. It is not a perfect mapping and it is a reliable first guess.

🎓 Exam Watch

The reliable question: "A condition is documented specifically, but ICD-10-CM contains no code for it. The appropriate code is one designated…" Answer: NEC.

And its mirror: "The documentation does not specify the site. The appropriate code is designated…" Answer: NOS / unspecified.

The mnemonic that works: NEC blames the book. NOS blames the note.

8.8 Brackets, parentheses, colons, and braces

The punctuation is not decoration and each mark has one job.

Brackets are the one to be careful with, because they mean two different things in the two books: in the Tabular, explanatory text you can mostly ignore; in the Index, a manifestation code that is part of a mandatory pair and must be sequenced second.

Two words that are conventions, not English

"And" means "and/or." A category titled "Tuberculosis of bones and joints" covers tuberculosis of bones, of joints, or of both. It does not require both to be present. This trips people up constantly.

"With" means "associated with" or "due to" — and, critically, it can create an assumed causal relationship between two conditions when they appear together in the Alphabetic Index or in a Tabular instructional note.

That second one is a large enough idea that it gets its own section in the next chapter. Chapter 9 §9.7 is where the "with" convention is resolved, including the diabetes-and-kidney-disease question that Chapters 4 and 7 both deferred to it.

"See" is mandatory; "see also" is advisory. §8.4.

8.9 The Table of Neoplasms, the Table of Drugs and Chemicals, and the External Cause Index

Three specialty tables, each organized differently from the main index, and each with its own trap.

The Table of Neoplasms

Organized by anatomical site down the left, with six columns across:

                    MALIGNANT                                  UNCERTAIN  UNSPECIFIED
   SITE          Primary  Secondary  CaInSitu    BENIGN         BEHAVIOR   BEHAVIOR
   ─────────────────────────────────────────────────────────────────────────────────
   Colon
     sigmoid      C18.7     C78.5      D01.0      D12.5          D37.4       D49.0

The rule: you must know the behavior before you can use the table. The site tells you the row; the pathology report or the provider's documentation tells you the column, and a coder who guesses the column has guessed the diagnosis.

The trap: a lesion documented as "colon polyp" is not automatically benign. The behavior must be documented. Where it is not, the unspecified-behavior column exists — and Chapter 10 §10.4 walks the whole table properly, including what happens when the pathology arrives after the encounter.

And a second trap: the table is a starting point. You still verify in the Tabular, because neoplasm codes carry instructional notes like everything else.

The Table of Drugs and Chemicals

Organized by substance down the left, with six columns:

                       POISONING                                    ADVERSE   UNDER-
   SUBSTANCE     Accidental  Self-harm  Assault  Undetermined       EFFECT    DOSING
   ────────────────────────────────────────────────────────────────────────────────────
   4-aminophenol
   derivatives    T39.1X1    T39.1X2    T39.1X3    T39.1X4          T39.1X5   T39.1X6

The columns are the whole content of the table, and choosing between them is a clinical determination the documentation must support:

• Poisoning — the drug was taken incorrectly: wrong substance, wrong dose, wrong person, taken with alcohol or another drug against instructions
• Adverse effect — the drug was taken correctly as prescribed and produced a harmful effect
• Underdosing — less was taken than prescribed or instructed

Poisoning and adverse effect are different clinical events with different codes and different sequencing, and the distinction is not a coder's judgment call. Chapter 12 §12.6 works it in full.

Note that the codes above end at six characters and every one of them needs a seventh character for episode of care — which the table does not show. Step two.

The External Cause Index

A separate index, organized by the event or the circumstance rather than by the injury: Fall, Collision, Struck by, Exposure. It leads to chapter 20 codes (V00–Y99) describing how an injury happened.

External cause codes are never first-listed and never reported alone. Chapter 12 §12.7.

🔢 Code It

Three tables, three first questions. Before you can use any of them, you need one fact.

Table	The question you must answer first	Where the answer comes from
Neoplasms	What is the behavior?	The pathology report, or the provider's documentation
Drugs and Chemicals	Was this poisoning, adverse effect, or underdosing?	The provider's documentation of the circumstance
External Causes	What was the event?	The record's account of how it happened

In all three cases the answer is a clinical fact the coder may not supply. That is the pattern: a specialty table asks you a question, and the answer must be in the record. When it is not, the options are the unspecified column, a query, or — for external causes — omission.

8.10 Five lookups, start to finish, with the wrong turns shown

The wrong turns are the lesson. Do these with your own book open.

Lookup 1 — "Right knee pain"

Wrong turn 1. Look up Knee. The index answers: Knee — see condition. This is the single most common beginner error, and the index is built to catch it.

Wrong turn 2. Look up Pain and stop at the main term, which carries R52. R52 is pain, unspecified — generalized pain — and a coder who takes it has coded something the patient does not have.

The path. Main term Pain(s) → subterm joint → subterm knee → M25.56-. The dash sends you to the Tabular. There: M25.561 right, M25.562 left, M25.569 unspecified. No seventh character required. M25.561.

Lookup 2 — "COPD with acute exacerbation"

Wrong turn. Coding COPD and the exacerbation separately. A combination code exists (Chapter 7 §7.8), and where one exists and fully describes the condition, you use it.

The path. Main term Disease → pulmonary → chronic obstructive → with exacerbation (acute) → J44.1. Verify in the Tabular, where the category carries instructional notes about coding any associated infection and about tobacco exposure. J44.1, plus whatever those notes require.

The lesson: the Tabular note added a code the index never mentioned.

Lookup 3 — "Screening colonoscopy, average risk"

Wrong turn. Looking up Colonoscopy. That is a procedure; this is the diagnosis index. There is no diagnosis here at all in the ordinary sense — the reason for the encounter is screening.

The path. Main term Screening → malignant neoplasm → colon → Z12.11. Chapter 21 territory, and Chapter 12 §12.9 covers why a person with no complaint has a diagnosis code.

The lesson: when there is no disease, the encounter itself is the reason, and it has a code.

Lookup 4 — "Fracture of the lower end of the right radius, patient here for cast removal, healing normally"

Wrong turn 1. Assigning the initial-encounter seventh character because this is the first time you have seen the patient. §7.6 — the character describes the kind of care, and this is routine healing.

Wrong turn 2. Stopping at the index, which will give a truncated code and will not tell you the seventh-character options, which are printed at the category level.

The path. Main term Fracture, traumatic → radius → lower end → and the index narrows by type. Then the Tabular, for the complete code and the seventh character. S52.501D — subsequent encounter, routine healing — where the documentation supports the unspecified fracture type.

The lesson: the seventh character is a documentation question, and the index will not ask it.

Lookup 5 — "Rash after taking amoxicillin as prescribed"

Wrong turn 1. The Table of Drugs and Chemicals, poisoning column. The drug was taken as prescribed, so this is an adverse effect, not a poisoning.

Wrong turn 2. Coding only the drug code. An adverse effect is coded with the manifestation first — the rash — and the drug code second. Chapter 12 §12.6.

The path. Code the rash from the main index. Then the Table of Drugs and Chemicals → amoxicillin → adverse effect column → the T-code, with its seventh character.

The lesson: the table gives you one of two codes, in second position, and the index gives you the other.

🔍 Check Your Understanding

1. The index gives you M25.56-. What does the dash mean and what do you do next?
2. A code carries an Excludes2 note naming a condition your patient also has. Report one code or two?
3. Documentation reads "benign essential hypertension." Does "benign" change the code? Why?
4. NEC or NOS: the provider documented the condition precisely and ICD-10-CM has no category for it.

(Answers: 1. More characters are required and only the Tabular can supply them — go to step two. 2. Two. Excludes2 means the excluded condition is not part of this code and both may be reported when both are documented. 3. No. "Benign" is a nonessential modifier in parentheses, and hypertension is I10. 4. NEC — the limitation is the code set, not the note.)

🗂️ The Encounter

🗂️ The Encounter

What this chapter contributes: the path.

Chapter 7 showed where M25.561 lives. Here is how you get there, and the two ways a reasonable person does not.

```text ACCOUNT 10-4471 — finding diagnosis A The note says: "Right knee pain - new complaint this visit."

┌─ WRONG TURN 1 ──────────────────────────────────────────────────┐ │ Look up KNEE. │ │ │ │ Knee - see condition │ │ │ │ The index will not help you find a body part. It indexes │ │ CONDITIONS. Go back and look up what is WRONG, not WHERE. │ └─────────────────────────────────────────────────────────────────┘

┌─ WRONG TURN 2 ──────────────────────────────────────────────────┐ │ Look up PAIN, take the main-term code, stop. │ │ │ │ Pain(s) R52 ◄── this is PAIN, UNSPECIFIED │ │ │ │ R52 is generalized pain. The patient has knee pain. Taking a │ │ main term's own code without reading the subterms is the │ │ second most common index error, and it produces a code that │ │ is valid, payable, and wrong. │ └─────────────────────────────────────────────────────────────────┘

┌─ THE PATH ──────────────────────────────────────────────────────┐ │ │ │ STEP ONE — ALPHABETIC INDEX │ │ │ │ Pain(s) R52 │ │ joint M25.50 ◄── essential modifier │ │ knee M25.56- ◄── essential modifier │ │ └── THE DASH. Index cannot finish. │ │ │ │ STEP TWO — TABULAR LIST │ │ │ │ Chapter 13, block M20-M25, category M25 │ │ ▸ read the CATEGORY notes ......... no 7th character │ │ required │ │ ▸ read the EXCLUDES notes ......... nothing our patient │ │ has │ │ ▸ find the laterality options ..... M25.561 right │ │ M25.562 left │ │ M25.569 unspecified │ │ ▸ check the documentation ......... RIGHT, three places │ │ │ │ ASSIGN: M25.561 │ └─────────────────────────────────────────────────────────────────┘ ```

What step two actually produced here. Nothing changed the code — and that is the honest and important result. The Tabular confirmed no seventh character, confirmed no Excludes note applied, and supplied the laterality character the index had replaced with a dash.

Three of those four findings were negative. That is what step two usually looks like, and it is why coders stop doing it. The fourth one — the laterality character — was not available from the index at all, and without the Tabular there is no complete code.

What this settles. The path, which is the thing you will have to reproduce two years from now for someone who was not there. Not "I knew it was M25.561." "Main term pain, subterm joint, subterm knee, dash sent me to the Tabular, no seventh character, no Excludes, laterality in the sixth character, documentation supports right in three places."

What it does not settle. Whether any guideline changes this. The Tabular's instructions are not the only rules — the Official Guidelines sit above all of it, and Chapter 9 reads them. In particular: is a symptom code appropriate at all when the provider suspects a degenerative process? Chapter 9 §9.6 answers it.

Open questions: Q1, Q5, and Q6 remain open. Diagnosis A is now found as well as located.

Conclusion

Alphabetic Index first. Tabular List second. Always.

What was decided in this chapter. That the two steps protect against different things — the index knows synonyms you do not, and the Tabular carries every instruction that governs the code, none of which is visible from the index. That main terms are conditions, not body parts, and that see condition is the index catching you. That indentation is meaning and a dash means the index cannot finish. That parentheses are permission to ignore and indentation is a requirement to document. That instructions cascade from chapter to block to category to code, and the index drops you at the bottom of that cascade having read none of it. That Excludes1 means only one of them and Excludes2 means you might need two, with a narrow unrelated-conditions exception you establish rather than assume. That "code first" and "use additional code" are the same relationship from opposite ends, and "code also" requires two codes without fixing their order. That NEC blames the book and NOS blames the note, and only one of them is fixable by a query. That brackets mean different things in the two volumes. That "and" means and/or, and that "with" is large enough to need its own section in Chapter 9. And that each specialty table asks you a clinical question first — behavior, circumstance, event — which the record must answer.

What remains open. The rules above the rules. Everything in this chapter is a convention printed in the book; the Official Guidelines govern all of it and resolve the conflicts.

The bridge to Chapter 9. The Guidelines are free, they are reissued every October 1, they are binding for the code set, and most working coders have never read them front to back. Chapter 9 reads the parts that decide real claims: first-listed versus principal diagnosis, the uncertain-diagnosis rule that the two settings treat in exactly opposite ways, when a symptom is integral and when it is coded, the "with" convention and the linkage it assumes, and what to do when two conventions appear to conflict. It also answers a question this chapter deliberately deferred twice.

Key Terms

Alphabetic Index — the volume organized by condition name, used to locate a code. Step one, always. (Ch.8)

Tabular List — the volume organized by classification structure, containing the complete codes and every instruction that governs them. Step two, always. (Ch.8)

Main term — a bold entry in the Alphabetic Index. A condition, not a body part. (Ch.8)

Subterm — an indented entry beneath a main term. Indentation level is meaning, and each level narrows the one above. (Ch.8)

Essential modifier — an indented subterm. It changes the code and must be supported by the documentation. (Ch.8)

Nonessential modifier — a word in parentheses following a main term or subterm. Alternate wording that does not change the code and need not be documented. (Ch.8)

Dash — a trailing hyphen on an index entry, meaning more characters are required and only the Tabular can supply them. (Ch.8)

Excludes1 — "not coded here." The two conditions cannot occur together and must never be reported together, subject to a narrow exception for genuinely unrelated conditions. (Ch.8)

Excludes2 — "not included here." The excluded condition is not part of this code, but a patient may have both and both may be reported. (Ch.8)

Code first — an instruction on a manifestation code requiring the underlying condition to be coded and sequenced first. (Ch.8)

Use additional code — an instruction on an etiology code requiring an additional code, sequenced second. (Ch.8)

Code also — an instruction that two codes are needed, without dictating their order. (Ch.8)

NEC (not elsewhere classifiable) — the documentation is specific and the classification has no code for it. The limitation is the code set. (Ch.8)

NOS (not otherwise specified) — equivalent to unspecified; the documentation lacks the detail needed for a more specific code. The limitation is the record. (Ch.8)

Table of Neoplasms — a site-by-behavior grid with six behavior columns; the behavior must be documented before the table can be used. (Ch.8)

Table of Drugs and Chemicals — a substance-by-circumstance grid with six columns covering poisoning by intent, adverse effect, and underdosing. (Ch.8)

External Cause Index — a separate index organized by event or circumstance, leading to chapter 20 codes describing how an injury occurred. (Ch.8)

Spaced Review

1. State the two-step rule and name three things that live in the Tabular and not in the index.

2. (Chapter 7) The index gives you a code ending in a dash. Explain what that means in terms of code validity, and what happens if you report the code as the index printed it.

3. A patient has condition A and condition B. The Tabular entry for A carries an Excludes1 note naming B. What do you do, and what is the narrow exception?

4. (Chapter 4) A code carries "code first" and the record does not document an underlying condition. Is this a coding problem or a documentation problem, and what is the tool?

5. Documentation reads "pain." The record contains no site anywhere. Name the main term, the code you would reach, and the reason you may not descend to a subterm.