Coding Audit Checklist: Guidelines, Metrics, and QA
Most coding auditors know what to look for in a chart. The harder part is making sure every auditor checks the same things, scores them the same way, and points to the same rules behind each finding. A coding audit checklist helps only when every auditor uses the same version and the answers are recorded as audit data. This guide covers what to check, the guidelines and scoring rules behind it, and the rebuttal and QA steps that keep results consistent.
What Makes a Coding Audit Defensible?
A defensible coding audit follows written audit guidelines, applies the same scoring rules to every date of service (DOS), and links each finding to official coding or coverage language. When an auditee, payer, or executive questions a result, the auditor can show which rule applied and where it comes from.
Problems start when the rules live in one auditor’s spreadsheet or in each auditor’s habits. As the team grows, the same documentation gets scored differently, findings rest on personal judgment, and auditees stop trusting the results. The fix is to set the rules before the review, store them where every auditor works, and check that everyone follows them.
What Makes an Audit Checklist Reliable?
Whether a checklist lives on paper, in a spreadsheet, or in audit software, it needs to meet these requirements to produce consistent results:
- One current version. Every auditor works from the same version, dated and tied to the code set in effect for each DOS. Old copies are retired when CPT, ICD-10-CM, NCCI edits, or LCDs change.
- Matched to the service. The billed code or service type decides which checklist applies, so two auditors never choose different checklists for the same DOS.
- Clear, answerable items. Each item checks one requirement, is answered Yes, No, or N/A, and defines when N/A applies.
- A citation for every item. Each item points to the CPT, ICD-10-CM, NCCI, or coverage source it enforces.
- Answers recorded as data. Responses are captured so they can be trended by auditee, checklist item, and auditor, not left as checkmarks on a form.
- Linked to the finding. Each answer stays with the finding it supports, so the evidence trail holds up during a rebuttal.
- A named owner. Someone is responsible for updating checklists when official sources change and for revising items that are frequently overturned.
A small team can meet most of these requirements with shared documents and spreadsheets. As the number of auditors, auditees, and specialties grows, version control and data capture are usually the first to break down.
How to Conduct a Coding Audit, Step by Step
A coding audit runs in three stages: set the rules, review each DOS against them, then report and resolve. Your auditors set the standard, and whatever tool you use should apply it the same way every time.
Before the audit: define the scope and parameters
- Define the scope: claim type (professional, outpatient facility, inpatient facility, or risk adjustment), the auditees under review, the date range, and the code set version for each DOS.
- Choose the sample. Use a statistically valid random sample if you plan to extrapolate results; OIG’s free RAT-STATS software can size it. Use a targeted sample for auditees or services your compliance risk assessment flags.
- Fixed baseline per auditee: the same number of DOS for every auditee.
- Small random sample within categories: a few DOS drawn at random from each service or code group.
- Risk-based targeting: the auditees or services your risk assessment flags.
- Probe and expand: a small sample first, expanded if errors exceed your threshold.
- Denial and recoupment driven: services with recent payer denials or recoupments.
- Triggered by an event: a new auditee, a new service line, or a complaint.
- Focused follow-up: a re-audit of the areas that failed a prior audit.
- Rotating coverage: a different group of auditees or services each cycle.
- List the records in scope, such as progress notes, orders, diagnostic reports, and the claim. Audit the claim version submitted to the payer or, for a prebill review, the version that will be submitted. Check each amendment or late entry against the applicable AMA and CMS guidelines and your medical staff bylaws. At a minimum, CMS Program Integrity Manual, Chapter 3, Section 3.3.2.5, requires it to be clearly identified as an amendment, show its date and author, and leave the original entry intact. If a claim is corrected and resubmitted, re-audit it or run a separate audit to validate claim corrections.
- Set the scoring method (data-element level, such as procedure or diagnosis codes; chart level; DOS level; or a combination), the accuracy thresholds for Pass, Conditional Pass, and Fail, and your error categories.
- Define the data elements to be audited. Create or confirm which checklists should be applied to each data element.
- Set the rebuttal response window and escalation path.
During the review: score each DOS against the guidelines
- Pull the full record for each DOS before scoring it.
- Record results in one standardized tool. Every auditor should capture the same fields in the same format: audit ID, DOS, auditee, billed and audited codes for each data element, checklist answers, error category, citation, financial impact, and comments. Options include:
- A locked spreadsheet template with dropdown lists for answers and error categories, protected accuracy formulas, and a single master file per audit.
- A form-based tool, such as an internal database or online form, that writes each finding to a shared table.
- Audit management software that keeps guidelines, checklists, findings, and reports in one system.
- Enter the billed codes and the audited codes the documentation supports.
- Answer each checklist item that applies to the code: Yes, No, or N/A.
- Classify each variance using your error categories.
- Cite the guideline, NCCI edit, or coverage policy behind each finding.
- Send complex or high-dollar findings to second-level review.
After the audit: report, resolve, and educate
- Calculate accuracy using the method set in the guideline.
- Release the findings report and Key Findings/Recommendations to each auditee, which opens the rebuttal window.
- Resolve disputes and record each outcome with its reason.
- Refer identified overpayments to your compliance team for refund.
- Deliver education tied to the guidelines you cited.
- Schedule focused re-audits for auditees below your accuracy threshold.
- Update guidelines and custom checklists based on what the audit found.
Which Coding Audit Metrics Should You Track?
Your coding audit metrics should show what was measured, how it was calculated, and which errors matter most.
Code-level versus DOS-level accuracy
Code-level accuracy counts each code separately. DOS-level accuracy counts a date of service as correct only when every code on it is correct. If an auditor reviews 10 DOS with 40 codes and finds 3 errors on 2 DOS, code-level accuracy is 92.5% (37 of 40) and DOS-level accuracy is 80% (8 of 10). Code-level results help with coder education; DOS-level results are closer to how a payer reviews a claim.
Many organizations target a 95% coding accuracy rate. That figure is an industry benchmark, and no CMS regulation requires it. Some teams also use a middle band, such as a Conditional Pass result, so an auditee just below target gets education rather than an automatic re-audit.
Error categories
A fixed set of categories lets every auditor classify findings the same way. Four basic categories work for most data elements:
| Error category | What the auditor found | Example | Possible impact |
|---|---|---|---|
| Incorrect | The wrong code or value was selected, but an appropriate code exists | Right knee diagnosis billed; note documents the left knee | Over- or underpayment and an education gap |
| Unsupported | The selected code is not supported by the medical record | Diagnosis with no mention in the note | Denial, recoupment, or a deleted risk adjustment diagnosis |
| Missing | A documented, reportable code was not billed | Documented procedure not billed | Lost revenue and incomplete diagnosis data |
| Invalid | The code was not effective for the date of service | Deleted CPT code billed after its termination date | Claim rejection or denial |
Some data elements produce better audit results when you add error categories unique to that element. These might include:
- Overcoded: a higher E/M level or more units than the documentation supports.
- Undercoded: a lower E/M level or fewer units than the documentation supports.
- Missing parent code: an add-on code billed without its required primary procedure.
- Incorrect sequencing: diagnoses reported in the wrong order, such as the wrong principal or first-listed diagnosis.
Add element-specific categories only where they change the education or follow-up, so the basic four stay comparable across auditors and audits.
Some teams also weight errors by financial or compliance impact and flag borderline items as educational only, so auditees get feedback without a lower score. Define both rules in the guideline.
What Should a Coding Audit Checklist Check?
A coding audit checklist turns written guidelines into specific questions an auditor answers for each code reviewed. Each item checks one requirement and cites its source, such as:
- AMA CPT code descriptors and E/M guidelines.
- ICD-10-CM Official Guidelines for Coding and Reporting, updated each October 1.
- CMS National Correct Coding Initiative (NCCI) edits, updated every quarter.
- National and Local Coverage Determinations (NCDs and LCDs).
- Payer guidelines and payment rules.
Good checklist items are written as questions with three answers: Yes, No, or N/A. “Does the note state the status of each problem addressed?” produces more consistent answers than “Rate the quality of the assessment.” When writing each question:
- Ask one thing. Split “Is the note signed and dated?” into two questions so each answer is clear.
- Keep Yes meaning compliant. Phrase every question so Yes means the requirement is met; mixed wording causes scoring errors.
- Ask what is documented. Auditors can agree on whether something is in the record but rarely agree on whether it is “good.”
- Avoid undefined words. Replace terms like “adequate” or “sufficient” with the specific requirement, such as “Does the note document the data reviewed for each test ordered?”
- Use the source’s language and cite it. Wording that mirrors the guideline makes findings easier to defend and to verify.
- Keep it neutral. Avoid leading questions such as “Isn’t the exam documented?”
- Test it before release. Have two auditors answer a new question on the same charts, and rewrite it if they disagree.
Example: what a DRG audit checklist checks
The items below show the kinds of questions a checklist covers for an inpatient MS-DRG audit. Answer each against the medical record and the MS-DRG grouper version in effect for the
| Checklist item | Answer | Citation |
|---|---|---|
| Is a signed inpatient admission order present in the record? | Yes / No / N/A | 42 CFR 412.3 |
| Is the principal diagnosis the condition established after study to be chiefly responsible for the admission? | Yes / No / N/A | ICD-10-CM Official Guidelines, Section II |
| Is the principal diagnosis supported by provider documentation? | Yes / No / N/A | ICD-10-CM Official Guidelines, Section I.A.19 |
| Is each reported CC or MCC documented by a provider? | Yes / No / N/A | ICD-10-CM Official Guidelines, Section III |
| Does each reported CC or MCC meet the definition of a reportable secondary diagnosis? | Yes / No / N/A | ICD-10-CM Official Guidelines, Section III |
| Is each present on admission (POA) indicator assigned correctly? | Yes / No / N/A | ICD-10-CM Official Guidelines, Appendix I |
| Is each ICD-10-PCS procedure code supported by the procedure documentation? | Yes / No / N/A | ICD-10-PCS Official Guidelines |
| Is the discharge disposition code supported by the record? | Yes / No / N/A | 42 CFR 412.4, post-acute transfer policy |
| Does each physician query follow compliant query practice? | Yes / No / N/A | AHIMA/ACDIS Guidelines for Achieving a Compliant Query Practice |
| Does the billed MS-DRG match the MS-DRG supported by the audited codes? | Yes / No / N/A | CMS MS-DRG Definitions Manual |
How Do You Produce an Audit Report and Handle Auditee Rebuttals?
A finding holds up when it cites the rule, its source, and the documentation the auditor relied on. Name the CPT guideline, ICD-10-CM guideline, NCCI edit, or LCD requirement for each one, so the auditee can check the finding against the same rule and a second reviewer can confirm it later.
What makes a good audit report
A good audit report lets each reader find what they need quickly: executives see the overall result, auditees see their own findings, and compliance sees the follow-up required. It should include:
- Scope and methodology. Claim types, auditees, date range, sample, scoring method, and guideline version, so results can be compared across audits.
- Summary results. Overall accuracy, Pass, Conditional Pass, or Fail status, and results by auditee and data element.
- Finding-level detail. Audit ID and DOS, billed and audited codes, error category, citation, and the auditor’s rationale, with no patient identifiers.
- Financial impact. Potential overpayments to review for refund and underpayments that can still be rebilled, kept separate from educational-only findings.
- Key findings and recommendations. The patterns behind individual errors, with specific education or process changes.
- Trends. Comparison with prior audits to show whether education is working.
- Next steps. The rebuttal deadline, a contact, and the date of any focused re-audit.
Release reports promptly and in the same format every time. A consistent layout lets auditees compare results across audits and helps leaders spot trends across departments.
Set the rebuttal process before results go out
- A response window, set by your organization.
- A response tied to a specific finding or to the Key Findings/Recommendations, with supporting documentation.
- An escalation path to a named reviewer, usually a senior auditor who did not perform the original review.
- A recorded outcome (upheld, overturned, or modified) with the reason.
Keep each rebuttal and its outcome with the original finding, so the full history is available if the finding is questioned later. If the same checklist item is overturned repeatedly, revise its wording or the guideline behind it.
If your team tracks audit findings across spreadsheets and email, request a demo of Audit Manager+ to see them in one record.
How Do You Keep Coding Auditors Consistent?
Auditor quality assurance checks whether every auditor applies your guidelines the same way. A standard checklist means every auditor works from the same questions, but auditors can still answer them differently. These steps measure and close that gap:
- Train and test. Train auditors on your guidelines, checklists, and scoring rules when they join, when guidelines change, and after a knowledge gap is found. Then test them on sample charts with known answers, and require a passing score before they return to live audits.
- Run blind dual reviews. On a set schedule, such as quarterly, have two auditors score the same sample of DOS without seeing each other’s work.
- Measure agreement. Compare results item by item. Divide the number of checklist items where both auditors gave the same answer by the total number of items compared, and set an internal target for that rate.
- Find the cause of each discrepancy. A senior auditor decides whether it came from unclear item wording, a knowledge gap, or a situation the guidelines don’t address.
- Fix the cause. Rewrite unclear custom items, add guidance for gaps, and give targeted education to any auditor who didn’t follow the guidelines.
- Monitor over time. Track agreement by auditor and by checklist item, and increase dual reviews for auditors below target.
Also decide which findings need a second auditor’s sign-off before release, such as findings above a set dollar amount or on high-risk services. If you’re comparing tools that support this kind of review, the medical coding audit software buyer’s guide lists questions to ask vendors.
How Audit Management Software Supports Consistent Coding Audits
The steps in this guide depend on every auditor using the same guidelines, checklists, and scoring rules while code sets and NCCI edits change during the year. That is harder to manage in spreadsheets as the audit team grows.
Audit Manager+ is a web-based platform for provider, coding, and documentation audits, used by more than 150 organizations. It gives audit teams a framework for putting each recommendation in this guide into practice:
- Written standards in one place. Scope, claim type, scoring method, Pass, Conditional Pass, and Fail thresholds, and error categories are saved once in the Audit Guideline, and each project defines which data elements are audited.
- Checklists assigned by code. More than 65 pre-built checklists cover areas such as E/M leveling, modifier use, and telehealth, and teams can add their own. Checklists load automatically by code or chart rule, and required checklists must be completed before the audit can close.
- One standard record of results. Billed and audited codes, checklist answers, error categories, citations, and financial impact are captured the same way at the data-element level.
- Consistent reports. Accuracy is calculated automatically at the code and DOS levels, and results roll up from the line item to the DOS, auditee, department, and organization in reports that use the same format every time.
- Built-in auditor QA. QA routing sends findings that meet your criteria to second-level review, and blind dual reviews can be set up as test audits. AAPC uses Audit Manager+ to audit its own auditors.
If you run a few audits a year with one auditor and your spreadsheet already produces consistent, cited findings, you may not need dedicated audit software.
Hold Every Coding Audit to the Same Written Standard
A defensible audit program depends on every auditor working from the same guidelines, checklists, and scoring rules, and on every finding tracing back to its source. When an auditee disputes a result or a payer asks how a score was reached, that record answers the question. Audit Manager+ brings those standards, along with findings, rebuttals, and reports, into one system. Your team sets the standard, and the platform applies it the same way on every review. To see how your current guidelines and checklists would work in Audit Manager+, request a demo.









