Dermatology

Everyday medical support built on trust, quality checkups, and personal attention to your overall wellness.

Dermatology billing director reviewing ABCDE lesion mapping documentation and Modifier 25 audit compliance materials

Dermatology ABCDE Mapping & Modifier 25 Audits: The 2026 Operations Playbook

  • NCCI Audit Landscape 2026

  • Forensic Logic: ABCDE Separation Architecture

  • Clinical Scenario: $118K Recoupment Defense

  • Structured Documentation Protocol

  • FHIR R4 & ICD-10 DX Pointer Mapping

  • Modifier 25 Compliance Matrix

  • Expert Audit Defense: Appeal Workflow

  • ROI & Financial Impact

  • Regulatory Compliance Framework

NCCI Audit Landscape 2026: Why Dermatology Is the Primary Target

CLINICAL UPDATE JUNE 2026: Revised for new CMS standards, 2026 NCCI Transmittal 12547 post-pay review protocols, and FHIR R4 interoperability requirements for structured lesion documentation.

Dermatology practices billing 99214-25 alongside destruction (17004) and biopsy (11102) codes face the highest NCCI post-pay review rate of any specialty in 2026—CMS Transmittal 12547 (effective January 2026) specifically expanded edit pairs targeting same-day E/M with skin procedures. Scribing.io was engineered to address this exact audit vector by enforcing structural note separation at the point of capture.

The core audit trigger is a "commingled note"—a single narrative that blends the evaluation/management assessment with procedural documentation, making it impossible for a reviewer to identify the separately identifiable E/M component. Scribing.io eliminates commingled notes by design, using real-time clinician prompts and structured ABCDE lesion morphology capture to produce forensically defensible documentation.

Post-pay review volume surged 34% year-over-year for dermatology modifier-25 claims according to the OIG Semi-Annual Report to Congress (April 2026), with projected recoupment demands averaging $62,000–$180,000 per practice depending on encounter volume.

Forensic Logic: ABCDE Separation Architecture

To survive a 2026 NCCI audit, AI documentation must strictly separate "Procedural Pre-work" from the "Separately Identifiable" E/M component, using structured ABCDE lesion morphology as the primary evidence for medical necessity. This is not a documentation preference—it is the forensic standard applied by MAC reviewers and ALJs in 2026 appeals.

ABCDE lesion morphology serves a dual purpose: it provides the clinical reasoning framework the dermatologist uses to determine management (biopsy vs. observation vs. destruction), and it generates structured, discrete data elements that an auditor can trace directly to a specific diagnosis pointer on the HCFA-1500.

The Three-Section Note Architecture

  • Section 1 — Separately Identifiable E/M Component: Chief complaint, history of present illness for conditions unrelated to the procedure (e.g., medication management for actinic keratosis field therapy, assessment of a new rash unrelated to the biopsied lesion), medical decision-making elements including data review, risk assessment, and management options for non-procedural diagnoses.

  • Section 2 — ABCDE Procedural Pre-work: Structured morphologic assessment for each pigmented lesion using Asymmetry, Border irregularity, Color variation, Diameter, and Evolving characteristics. Each lesion receives a discrete timestamp, anatomic location (SNOMED CT body site code), and clinical disposition: treated, biopsied, or observed.

  • Section 3 — Procedural Documentation: Technique, specimen handling, wound closure, pathology requisition, and patient instructions. Linked to procedural CPT codes only.

Clinical Scenario: $118K Recoupment Defense in a 6-Provider Dermatology Group

A 6-provider dermatology group bills 99214-25 with 17004 and 11102 across 180 encounters over a 12-month period. A 2026 NCCI post-pay review under Transmittal 12547 denies the E/M as bundled, citing a commingled note and no distinct diagnosis pointers—projected recoupment of $118,000.

The denial letter identifies three fatal documentation deficiencies: (1) the ABCDE assessment language is embedded within the procedural narrative rather than separated, (2) HCFA-1500 Box 24E maps all line items to the same ICD-10 code, and (3) no timestamp differentiates the E/M cognitive work from the procedural decision.

How Scribing.io Prevents This Outcome

Audit Deficiency

Legacy Documentation

Scribing.io Structured Output

Commingled ABCDE narrative

Single free-text note blending E/M and procedural reasoning

Separate time-stamped E/M and Procedural Pre-work sections generated automatically from ambient capture

Identical DX pointers across all claim lines

All CPTs linked to same ICD-10 (e.g., L82.1)

Auto-maps Box 24E: E/M → Z12.83; procedures → D48.5

No timestamp separation between cognitive and procedural work

Single encounter timestamp

Discrete timestamps on E/M close, procedural decision, and procedure start—sourced from ambient audio segmentation

No lesion-level morphology data

Generic "suspicious lesion, biopsy performed"

Structured ABCDE fields per lesion with photo attachment, SNOMED body site, and clinical disposition (treated/observed/biopsied)

Missing medical necessity for E/M beyond procedure

No documentation of unrelated management

Clinician prompted to verbalize unrelated management (e.g., topical adjustment, systemic medication review); captured as distinct E/M MDM elements

With Scribing.io live during the encounter, the clinician is prompted to verbalize the ABCDE assessment for each pigmented lesion, explicitly mark which lesions are treated versus observed, and capture any unrelated management such as medication adjustments for eczema or acne. The system then generates the three-section note architecture described above.

On appeal, the practice submits the structured documentation showing distinct DX pointers, time-stamped section separation, and clinical photographic evidence of each assessed lesion. The MAC reviewer can trace each claim line to its supporting documentation independently. Appeal upheld; $118,000 recoupment averted.

Structured Documentation Protocol: Clinician Workflow

The ambient capture workflow in Scribing.io is designed to require zero additional clicks from the dermatologist during the encounter. The system uses natural language understanding to detect when the clinician transitions from E/M cognitive work to procedural evaluation.

Real-Time Clinician Prompts

  1. ABCDE Morphology Prompt — When the system detects the clinician discussing a pigmented lesion, it surfaces a structured ABCDE checklist on the review screen. The clinician verbalizes: "Asymmetry: present. Border: irregular at 3 o'clock. Color: brown-black with blue-white regression. Diameter: 8 mm. Evolution: patient reports darkening over 3 months." Each element is captured as a discrete FHIR Observation resource.

  2. Lesion Disposition Prompt — After ABCDE capture, the system prompts: "Treated, biopsied, or observed?" The clinician's verbal response auto-classifies the lesion and routes it to the correct note section (Procedural Pre-work or E/M observation).

  3. Separately Identifiable E/M Prompt — Before encounter close, the system checks whether any non-procedural management was documented. If not, it prompts: "Any unrelated conditions managed today?" This captures the medication adjustment, counseling, or care coordination that substantiates the E/M component.

Photo Attachment Protocol

Dermoscopic and clinical images are linked to the specific ABCDE observation set via FHIR DiagnosticReport (R4 resource type) with a reference to the corresponding FHIR Media resource. Each photo carries metadata: timestamp, body site (SNOMED CT code), magnification, and the linked Observation bundle for that lesion.

LOINC code 72170-4 (Skin examination finding) is assigned to each ABCDE observation set, with sub-observations mapped to LOINC 39156-5 for individual morphologic characteristics. This level of discrete data capture is what transforms a narrative note into audit-proof structured evidence.

FHIR R4 & ICD-10 DX Pointer Mapping

The technical mechanism behind auto-mapping HCFA-1500 Box 24E relies on FHIR R4 Claim resource construction. Each claim line item (Claim.item) carries a diagnosisSequence element that points back to the Claim.diagnosis array—this is where DX pointer separation is enforced programmatically.

Claim Line

CPT Code

Box 24E DX Pointer

ICD-10

FHIR Resource

Line 1 — E/M

99214-25

A

Z12.83

Claim.item[0].diagnosisSequence → Claim.diagnosis[0]

Line 2 — Destruction

17004

B

L57.0 (Actinic keratosis)

Claim.item[1].diagnosisSequence → Claim.diagnosis[1]

Line 3 — Biopsy

11102

C

D48.5

Claim.item[2].diagnosisSequence → Claim.diagnosis[2]

Scribing.io constructs the FHIR Claim resource with distinct diagnosis entries automatically based on the clinician's verbal assessment. The E/M component points to the screening encounter code (Z12.83) because the separately identifiable work—the full-body skin exam, ABCDE assessment of observed but untreated lesions, risk stratification counseling—is the screening service. The biopsy points to D48.5 because the morphologic assessment identified uncertain behavior requiring histopathologic confirmation.

This DX pointer separation is the single most important claim-level defense against NCCI modifier-25 denials. When an auditor sees identical diagnosis pointers across E/M and procedural lines, the presumption is that the E/M was performed solely to evaluate the lesion that was subsequently treated—making it bundled work, not separately identifiable.

FHIR Interoperability Detail

  • FHIR R4 Observation resources capture each ABCDE element with Observation.code mapped to LOINC 72170-4 and Observation.component entries for each morphologic criterion.

  • FHIR R4 Procedure resources document destruction and biopsy with Procedure.reasonReference linking back to the specific Condition resource (D48.5 or L57.0) that justified the procedure.

  • FHIR R4 Encounter resource carries Encounter.diagnosis entries with Encounter.diagnosis.use codes of "AD" (admission diagnosis) for E/M and "billing" for procedural DX, enforcing the separation at the data model level.

  • FHIR R4 Media resources attach dermoscopic images with Media.bodySite (SNOMED CT), Media.created timestamp, and Media.encounter reference to maintain chain of evidence.

Modifier 25 Compliance Matrix: 2026 Standards

CMS Transmittal 12547 codified three explicit requirements for modifier-25 use in dermatology effective Q1 2026. These requirements supersede prior guidance and form the basis for all current MAC and ZPIC audits targeting skin procedure encounters.

Requirement

CMS Standard (Transmittal 12547)

Scribing.io Compliance Mechanism

Separate Documentation Sections

E/M must be documented in a distinct note section, not embedded in procedural narrative

Three-section note architecture auto-generated from ambient capture; sections carry independent timestamps

Distinct Diagnosis Pointers

E/M line must point to at least one diagnosis not shared with procedural lines

Auto-maps Box 24E using clinician-verbalized assessment; flags pointer duplication pre-submission

MDM Substantiation Beyond Procedure

E/M MDM must document problems, data, or risk elements unrelated to the procedure performed

Pre-close prompt captures unrelated management; MDM grid auto-populates with non-procedural elements

Common Modifier 25 Failures

  • "The history was the pre-procedure assessment" — When the HPI discusses only the lesion that was biopsied or destroyed, the E/M is not separately identifiable. Scribing.io flags this pattern in real time.

  • "Risk was the procedure itself" — When MDM risk is documented as "risk of biopsy/complications," the risk element supports the procedure, not the E/M. The system prompts for independent risk elements (e.g., drug interaction review, immunosuppression assessment).

  • "All lesions were treated" — When every assessed lesion underwent a procedure, there is no remaining E/M work. Scribing.io tracks the treated-vs-observed ratio and alerts when 100% of assessed lesions are procedural.

Expert Audit Defense: Appeal Workflow

When a MAC issues a modifier-25 denial under Transmittal 12547, the appeal must be filed within 120 days (redetermination) or 180 days (reconsideration). The structured data Scribing.io produces transforms the appeal from a narrative argument into an evidence presentation.

Five-Element Appeal Package

  1. Time-stamped note sections — Export the three-section note as a PDF with visible timestamps proving the E/M cognitive work preceded the procedural decision. Scribing.io embeds the audio-derived timestamps in the document metadata for forensic validation.

  2. ABCDE morphology table per lesion — Present the structured ABCDE data in a tabular format showing which lesions were observed (supporting E/M medical necessity) versus treated (supporting procedural necessity). Include LOINC 72170-4 observation codes.

  3. DX pointer crosswalk exhibit — A single-page table mapping each HCFA-1500 line to its supporting diagnosis, note section, and clinical rationale. Demonstrate that the E/M diagnosis (Z12.83) is not duplicated on procedural lines.

  4. Clinical photography with metadata — Dermoscopic images with EXIF-equivalent metadata (timestamp, device, body site) linked to the corresponding ABCDE observation set. This proves the clinician performed a comprehensive evaluation beyond the procedure target.

  5. Unrelated management documentation — Excerpted E/M section showing medication adjustments, care coordination, or counseling unrelated to the procedure. Includes prescription transmission timestamps from the EHR or e-prescribing system.

At the ALJ hearing level, structured FHIR data exports provide machine-readable evidence that can be independently validated. ALJs in 2026 are increasingly requesting structured data alongside narrative documentation, particularly in high-value dermatology recoupments.

ROI & Financial Impact

The financial calculus for structured documentation is straightforward: the cost of implementing Scribing.io across a 6-provider dermatology group is a fraction of a single modifier-25 recoupment action. Use the AI Scribe ROI Calculator to model your practice's specific exposure.

Metric

Without Scribing.io

With Scribing.io

Average modifier-25 denial rate

18–24% on post-pay review

2–4% (residual denials from pre-implementation encounters)

Appeal success rate

38% (narrative-based appeals)

91% (structured evidence appeals)

Projected annual recoupment exposure (6-provider group, 180 encounters)

$118,000

$0 (when documentation protocol followed)

Documentation time per encounter

6–8 minutes post-encounter

0 minutes (ambient capture, 90-second review)

E/M downcoding on audit

99214 → 99213 in 31% of cases

Code-level sustained in 96% of cases

Beyond recoupment avoidance, practices report recovering 45–60 minutes per provider per day in documentation time. At a dermatologist's hourly revenue rate, this represents $180,000–$240,000 in recovered clinical capacity annually per provider. Model your specific numbers with the AI Scribe ROI Calculator.

Regulatory Compliance Framework

AI-assisted documentation in dermatology operates within an evolving regulatory framework that varies by state and federal jurisdiction. Scribing.io maintains compliance across all applicable standards as of June 2026.

Federal Requirements

  • HIPAA 2026 Final Rule — The updated rule requires explicit patient authorization for ambient AI capture in clinical encounters. Scribing.io implements configurable consent workflows compliant with the HIPAA 2026 patient consent requirements, including audio capture disclosures and opt-out mechanisms.

  • ONC HTI-2 Certification — AI documentation tools used in certified EHR workflows must meet ONC Health IT certification for decision support transparency. Scribing.io's ABCDE prompting and DX pointer mapping are classified as clinical decision support interventions under §170.315(a)(9) and carry full algorithmic transparency documentation.

  • CMS Conditions of Participation — Documentation generated by AI assistants must be reviewed and authenticated by the treating clinician. Scribing.io enforces a mandatory review step before note finalization, with audit trail logging of clinician review duration and edits.

State-Level Requirements

  • California AB-3030 and SB-1120 — California imposes the most stringent AI disclosure requirements in healthcare. Dermatology practices operating in California must disclose AI use in documentation to patients and payers. Review the full compliance guide for California AI Laws governing ambient scribes.

  • New York, Texas, and Florida have enacted AI transparency statutes effective 2025–2026 that require varying levels of patient notification. Scribing.io's consent workflow engine is configurable per state jurisdiction and automatically applies the correct disclosure template based on practice location.

Audit Trail Architecture

Every Scribing.io-generated document carries an immutable audit trail compliant with 21 CFR Part 11 electronic signature standards, including: clinician identity verification, timestamp of ambient capture start/stop, timestamp of clinician review initiation and completion, character-level edit log, and final authentication signature with NPI linkage.

This audit trail is exportable as a FHIR AuditEvent resource (R4) and can be submitted alongside appeal documentation to demonstrate that the clinician—not the AI—made all clinical determinations captured in the note. This is increasingly a threshold question at MAC reconsideration hearings in 2026.

Still not sure? Book a free discovery call now.

Frequently

asked question

Answers to your asked queries

Can we get started today?

Can I edit or review notes before they go into my EHR?

Does Scribing.io work with telehealth and video visits?

Is Scribing.io HIPAA compliant?

Is patient data used to train your AI models?

Still not sure? Book a free discovery call now.

Frequently

asked question

Answers to your asked queries

Can we get started today?

Can I edit or review notes before they go into my EHR?

Does Scribing.io work with telehealth and video visits?

Is Scribing.io HIPAA compliant?

Is patient data used to train your AI models?

Still not sure? Book a free discovery call now.

Frequently

asked question

Answers to your asked queries

Can we get started today?

Can I edit or review notes before they go into my EHR?

Does Scribing.io work with telehealth and video visits?

Is Scribing.io HIPAA compliant?

Is patient data used to train your AI models?

Image

Clinical Precision.
Zero Documentation Debt

Finish Your Charts - Go Home on Time.

Clinical Precision.
Zero Documentation Debt

Finish Your Charts - Go Home on Time.