How to Look Up an ICD-10-CM Code (Step by Step)

Ask any experienced medical coder how they find a diagnosis code and they will not say "I type it into a search box and pick the first hit." Accurate coding follows a specific, guideline-mandated workflow: you start in the Alphabetic Index, follow its cross-references to a candidate code, and then verify that code in the Tabular List before you ever put it on a claim. Skipping the verification step is one of the most common reasons a code that "looks right" turns out to be invalid, non-billable, or simply wrong. This guide walks through the correct lookup process end to end, with two worked examples, and shows how our ICD-10 code browser compresses the whole workflow into a few keystrokes while keeping every safeguard in place.

If you are new to the code set itself — what the characters mean or how a code becomes billable — read Understanding ICD-10 Codes and ICD-10-CM Code Structure Explained first. This article is the practical companion: not what a code is, but exactly how to find and confirm the right one.

Why "Index First, Tabular Second" Is the Rule

ICD-10-CM is published as two coordinated parts, and the official ICD-10-CM Official Guidelines for Coding and Reporting are explicit that you must use both together. The Alphabetic Index is an alphabetical listing of diagnostic terms — conditions, diseases, injuries, symptoms, and reasons for a visit — each pointing to one or more candidate codes. The Tabular List is the numeric-alphanumeric listing organized by chapter, block, category, and code, and it is the only place where the full context of a code lives: its instructional notes, required additional characters, and exclusions.

The Index narrows the field, but it never gives you the final answer. The Guidelines state plainly that a code cannot be assigned from the Index alone — it must always be verified in the Tabular List. That is because the Index cannot show you the seventh-character requirements, the Excludes1 conflicts, or the "Code first" instructions that may change or invalidate your selection. Treat the Index as the on-ramp and the Tabular List as the destination.

The Step-by-Step Lookup Workflow

Step 1 — Identify the main term in the documentation

Start from what the clinician actually wrote. Read the note and pull out the main term — usually the condition, disease, or injury, not the anatomical site. For "acute bronchitis" the main term is Bronchitis; for "fracture of the left radius" the main term is Fracture; for "type 2 diabetes with neuropathy" the main term is Diabetes. Anatomical sites, adjectives, and severity are usually subterms nested underneath, not the entry point. Choosing the wrong main term is the single most common way a lookup goes off the rails.

Step 2 — Look up the main term in the Alphabetic Index

Find the main term alphabetically, then follow the indented subterms to add the detail from the documentation — type, site, laterality, acuity. Each level of indentation narrows the code. The Index also uses connecting words and punctuation with specific meaning: a term in parentheses is a nonessential modifier that does not change code selection, while indented subterms are essential modifiers that do.

Step 3 — Follow cross-references (see / see also)

The Index frequently redirects you. A "see" instruction is mandatory: it tells you the term you looked up is not where the code lives and points you to the correct main term instead. A "see also" instruction is advisory: your current entry may be valid, but another term could carry a more appropriate code, so check it when the first path does not fully match the documentation. Follow every "see" before you settle on a candidate. There are also default codes — the code listed immediately beside a main term — which represent the condition most commonly associated with that term when the documentation gives no further detail.

Step 4 — Verify the candidate in the Tabular List

Take the code the Index handed you and locate it in the Tabular List. This is the step you must never skip. In the Tabular List you read the code in the full context of its category and confirm three things: that the code exists as written, that you have satisfied every character the category requires, and that no instructional note blocks or redirects your choice.

Step 5 — Read the instructional notes

At the code and at every level above it (subcategory, category, block, and chapter), check the notes that govern selection:

• Includes and inclusion terms confirm the conditions the code covers.
• Excludes1 means "not coded here" — the two conditions can never be reported together. An Excludes1 pointing at another code that also applies signals a conflict you must resolve.
• Excludes2 means "not included here" — the excluded condition is separate, and a patient may legitimately have both, so you may report both when documented.
• Code first tells you to sequence an underlying cause before this code.
• Use additional code tells you to add a secondary code to capture an associated condition.

These notes are inherited down the hierarchy, so a rule stated at the category level applies to every code beneath it even if it is not repeated on the specific code.

Step 6 — Code to the highest specificity

Assign the most complete code the documentation supports. A code is billable (valid for submission) only when it is a complete leaf with every required character present — up to seven characters. Three-character categories and intermediate subcategories are frequently headers that cannot be billed. Where a seventh character is required, it must sit in the seventh position; if the code is shorter than six characters you insert the placeholder X to hold the seventh character in place. And where the classification distinguishes laterality, you must pick left, right, or (where offered) bilateral or unspecified according to the documentation.

Step 7 — Confirm it is a valid current-year code

ICD-10-CM is refreshed every federal fiscal year, effective October 1, with occasional April 1 additions. A code that was valid last year may have been deleted, split, or expanded. Before a code reaches a claim, confirm it is valid for the edition that matches the date of service — and, for a service today, that it is a valid 2026 code.

Worked Example 1 — "Type 2 diabetes mellitus"

The clinician documents "type 2 diabetes mellitus" with no complications noted.

1. Main term: Diabetes (not "type 2," and not "mellitus").
2. Index: Under Diabetes, diabetic, follow the subterm for type 2. With no complication documented, the default path leads to the "without complications" entry.
3. Candidate: the Index points to E11.9.
4. Verify in the Tabular List: locate category E11, "Type 2 diabetes mellitus." Note the category-level Use additional code instructions (for example, to identify long-term insulin use) and the Excludes notes separating type 2 from other diabetes types. Confirm E11.9 reads "Type 2 diabetes mellitus without complications."
5. Specificity and validity: E11 alone is a non-billable header; E11.9 is a complete, billable, valid 2026 code. Assign E11.9.

Had the note said "type 2 diabetes with diabetic neuropathy," the same main term would have led down a different subterm branch to a combination code in the E11.4- range, verified the same way — a reminder that the documentation, not the shortest path, drives the code.

Worked Example 2 — "Closed fracture of the left radius, initial encounter"

The clinician documents a closed fracture of the shaft of the left radius, first visit for active treatment.

1. Main term: Fracture (the condition), then radius, then shaft as subterms.
2. Index: Under Fracture, traumatic → radius → shaft, the Index points into the S52.3- family for the radial shaft.
3. Verify in the Tabular List: locate S52, "Fracture of forearm." Here the Tabular List does the work the Index cannot: it shows that these codes require a seventh character for the episode of care, that laterality must be specified, and that unspecified fractures are treated as closed unless documented otherwise.
4. Laterality: choose the left-radius subcategory. For an unspecified shaft fracture of the left radius that resolves to S52.302.
5. Seventh character and placeholder: the initial encounter for a closed fracture takes seventh character A. Because the base is only six characters, no additional placeholder X is needed here — the seventh character attaches directly, giving S52.302A.
6. Specificity and validity: confirm S52.302A is a complete seven-character billable code valid for the 2026 edition. If the note specifies the exact shaft-fracture type (for example a greenstick or transverse fracture), a more specific code in the same family applies — always code what is documented.

This example shows why verification is non-negotiable: nothing in the Index would have told you a seventh character was mandatory or that a closed fracture is the default. Only the Tabular List carries those rules.

Doing the Whole Workflow in the Browser

The manual process is correct, but flipping between a printed Index and Tabular List is slow. Our ICD-10 code browser is a free, up-to-date ICD-10 search tool that preserves the workflow while removing the friction. Search by clinical term to mimic the Alphabetic Index, then click a result to land in its Tabular context — parent-code chain, billable badge, and the relevant Excludes and instructional notes shown right on the card. You can also type a three-character category (like E11) to see the whole family and drill to the specific child, or browse the chapter tree when you are learning an unfamiliar area. Because it loads the full CDC/NCHS ICD-10-CM dataset once and runs every search locally in your browser, results are instant and completely private — nothing you type ever leaves your machine.

Used well, the browser does not replace the Index-then-Tabular discipline; it accelerates it. You still identify the main term, still read the notes, still code to the highest specificity, and still confirm the code is valid for the date of service. The tool simply puts the Index lookup and the Tabular verification on the same screen.

Common Lookup Mistakes to Avoid

• Coding straight from the Index. The Index gives a candidate, not an answer. Always verify in the Tabular List.
• Picking the wrong main term. Start from the condition, not the site or the modifier.
• Ignoring a "see" cross-reference. "See" is mandatory; skipping it leads you to the wrong family entirely.
• Billing a header. A three-character category or intermediate subcategory that expects more characters is not valid for submission.
• Forgetting the seventh character or placeholder X. Injury and external-cause codes often require both.
• Missing laterality. When the classification offers left/right, "unspecified" is a last resort, not a default.
• Using a retired code. Confirm the code is valid for the edition matching the date of service.

Putting It Into Practice

The correct ICD-10 code lookup is a short, repeatable loop: read the documentation, identify the main term, work the Alphabetic Index and its cross-references to a candidate, verify that candidate in the Tabular List, read every instructional note, code to the highest specificity with the right seventh character and laterality, and confirm the code is valid for the date of service. Do that every time and your codes will hold up to a claim scrubber and an auditor alike. To see how the chapters are organized and speed up the browse step, then open the ICD-10 code browser and try the two worked examples above — you will have both codes verified in under a minute.

Sources

• CMS — ICD-10-CM Official Guidelines for Coding and Reporting, FY 2026 — the official rules that mandate the Alphabetic Index → Tabular List verification workflow, coding to the highest specificity, and the reading of instructional notes.
• CDC/NCHS — ICD-10-CM — the National Center for Health Statistics overview of the US diagnosis classification and its annual, fiscal-year maintenance cycle.
• CDC/NCHS — ICD-10-CM Files — the official downloadable code files for each edition, including the Alphabetic Index and Tabular List.