Abstracting in medical coding means pulling the non-code facts of an encounter out of the medical record: dates, admission type and source, discharge disposition, the attending physician and procedure dates. The result is a coding abstract. Assigning codes is the separate step of translating documented diagnoses and procedures into ICD-10-CM, ICD-10-PCS or CPT codes.
What does abstracting mean in medical coding?
In medical coding, abstracting means reading a patient's record and copying a defined set of facts about the encounter into fixed fields. A hospital coder typically abstracts who the patient is, when they arrived and left, how they arrived, where they went, which physicians were responsible and when procedures happened. The fields together form the coding abstract.
An abstract, in the everyday sense, is a summary. A coding abstract is a short, structured summary of one encounter that sits beside the codes. It can be a paper form or a screen in coding software; either way, each fact has its own field and the codes go in their own section.
The abstract is built from the whole record, not from one page. The FY 2027 ICD-10-CM Official Guidelines for Coding and Reporting (updated October 1, 2026) state that "the entire record should be reviewed to determine the specific reason for the encounter and the conditions treated." In practice the coder moves between the face sheet, the history and physical, the operative report, the progress notes and the discharge summary to fill the abstract. The history and physical extraction and operative report extraction posts show which fields those two documents yield.
Abstracting versus assigning codes in one medical coding workflow
Abstracting and assigning codes are two tasks in the same medical coding workflow, and they differ in output and skill. Abstracting produces facts copied from the record, such as an admission date or a physician's name. Code assignment produces classification codes that a coder chooses by applying coding conventions and official guidelines to the documented diagnoses and procedures.
How the two tasks differ
| Abstracting | Assigning codes | |
|---|---|---|
| Output | Dates, times, categories, names and identifiers | ICD-10-CM, ICD-10-PCS or CPT codes |
| Source | Whatever page states the fact | The provider's documentation of diagnoses and procedures |
| Judgment | Finding the right value and copying it accurately | Selecting and sequencing codes under official rules |
| Typical errors | Wrong date, wrong physician, wrong disposition | Wrong code, wrong sequence, missed condition |
Code assignment has rules the abstract does not. The CMS guidelines say that code assignment "is based on the documentation by the patient's provider," with a few listed exceptions. They also describe a "joint effort between the healthcare provider and the coder" to achieve complete documentation and coding. Abstracted fields carry no such rule: an admission date is the admission date wherever the chart prints it.
Timing
Both tasks usually happen in one sitting after discharge, once the record is complete. Some fields can be abstracted earlier, because the admission date and admission source are known on day one. Codes for the final diagnoses wait for the provider's documentation. Which cases are complex enough to need a certified coder's judgment is covered in which coding cases need a certified coder.
Which data elements does a coding abstract hold besides the codes?
A coding abstract holds the encounter's dates, its admission type and source, the discharge disposition, the attending physician and the procedure dates. Many of these fields map directly to boxes on the UB-04 institutional claim, so an abstracting error can travel straight onto the bill.
The CMS Medicare Claims Processing Manual, Chapter 25 (Rev. 12423, 12-20-23) describes those boxes, called form locators (FL). FL 12 is the admission date, FL 14 the priority (type) of admission or visit, FL 15 the point of origin for admission or visit, and FL 17 the patient discharge status. FL 69 is the admitting diagnosis, FL 74 the principal procedure code and date, and FL 76 the attending provider, defined as "the individual who has overall responsibility for the patient's medical care and treatment." The manual points to the NUBC data specifications for the code values in FL 14, 15 and 17.
These definitions come from a longer tradition. The CMS coding guidelines note that the Uniform Hospital Discharge Data Set (UHDDS) definitions "are used by hospitals to report inpatient data elements in a standardized manner," and quote the UHDDS definition of principal diagnosis as "that condition established after study to be chiefly responsible for occasioning the admission of the patient to the hospital for care."
A fictional inpatient stay, side by side
The stay below is fictional. The patient, Oskar Brandt, the account number and every date and name are made up for this example. The codes are shown only to illustrate the second column; titles were checked against the NLM Clinical Tables ICD-10-CM lookup.
| Abstracted data elements | Assigned codes |
|---|---|
| Account number: 7731-0042 (example) | Principal diagnosis: J18.9, Pneumonia, unspecified organism |
| Admission date: 2026-09-03 | Secondary diagnosis: E11.9, Type 2 diabetes mellitus without complications |
| Admission type: as recorded on the face sheet | Secondary diagnosis: I10, Essential (primary) hypertension |
| Point of origin: as recorded on the face sheet | Principal procedure: none |
| Admitting diagnosis text: "shortness of breath, fever" | |
| Attending physician: Dr. A. Kowalczyk (example) | |
| Discharge date: 2026-09-07 | |
| Discharge disposition: "Home with self care" as written |
Every value in the left column is copied from a page and can be checked against it. Every value in the right column is a decision made under the official guidelines. Discharge summaries are among the document types Schema Care reads: it pulls their diagnoses, procedures, medications and follow-up instructions, scoring each field and linking it to the spot on the page it came from.
When coding abstracting gets confused with registry or quality abstraction
Coding abstracting gets confused with registry or quality abstraction because all three are called abstraction and all three read the same charts. Coding abstraction fills the fields that travel with the codes. Registry abstraction fills a registry's data dictionary, such as a cancer registry's. Quality abstraction fills measure forms. The fields, rules and audiences differ.
Job listings and course titles often use "abstractor" without saying which kind. A coding abstractor works alongside the codes and the claim. A quality or registry abstractor follows a measure specification or a registry manual, and may never assign a code. Reading the duties listed, not the title, tells you which job is meant.
The broader meaning of the word, outside coding, is covered in what medical record abstraction is, and the full picture of chart abstraction lives in medical record abstraction. What an abstractor does across a working day, in any of these roles, is described in what a medical record abstractor does each day.
The practical test is the output. If the work ends in codes and a claim, it is coding abstraction. If it ends in a registry submission or a quality measure, it is the other kind, even when the same person does both.