Documenting MASLD (Fatty Liver) Staging with AI Logic: The 2026 Clinical Operations Playbook

Clinical Operations Directors managing hepatology and primary-care denials now face a documentation gate that legacy dictation cannot clear. The retirement of NAFLD terminology converted a soft label into a computable coverage requirement. Scribing.io resolves this by executing the coverage logic at documentation time, not retroactively.

This playbook details the exact FHIR and CQL architecture that Scribing.io uses to encode FIB-4 staging, auto-source cardiometabolic risk factors, and transmit a defensible prior-auth bundle. Every claim here maps to a persisted, traceable data object — the standard that separates approvals from denials in 2026.

• Jump to: Why the MASLD Shift Breaks Legacy Notes

• Jump to: The 52-Year-Old FibroScan Denial Reversal

• Jump to: Encoding FIB-4 in FHIR R4 with CQL

• Jump to: ICD-10 Documentation Standards

TL;DR — MASLD Staging Documentation with AI Logic

• The 2026 nomenclature shift retires "NAFLD"/"NASH" in favor of MASLD/MASH. Payers (per the proposed MolDX LCD DL40197) now require a documented FIB-4 ≥ 1.3 plus at least one cardiometabolic risk factor to justify advanced fibrosis workup.

• Scribing.io encodes the FIB-4 calculation directly in FHIR R4 using CQL, persisting the score as an Observation (category=risk-assessment) with derivedFrom links to AST, ALT, Platelet, and Patient.birthDate.

• Metabolic risk factors are auto-sourced from LOINC-coded Observations and bundled into a Da Vinci PAS prior-auth request for CPT 91200 (FibroScan).

• Result is first-pass authorization and defensible 99214 MDM complexity — without manual chart abstraction.

Why the 2026 MASLD Nomenclature Shift Breaks Legacy Documentation

CLINICAL UPDATE 2026: Revised for new CMS CPT G2211 standards, SB 1120 compliance, and FHIR interoperability.

The single largest documentation liability facing Clinical Operations Directors in 2026 is a terminology problem masquerading as a coding problem. The AASLD has formally replaced non-alcoholic fatty liver disease (NAFLD) with metabolic dysfunction-associated steatotic liver disease (MASLD), and NASH with MASH. This is not cosmetic.

The proposed MolDX LCD (DL40197) explicitly ties coverage to the new definition. Coverage now hinges on two documented data points that legacy dictation workflows routinely omit:

1. A FIB-4 score demonstrating the patient is not low-risk (FIB-4 ≥ 1.3), and

2. At least one cardiometabolic risk factor — obesity, hyperglycemia/T2D, hypertension, or dyslipidemia — satisfying the MASLD criterion (hepatic steatosis with ≥1 cardiometabolic risk factor).

The CMS policy framework requires services be "reasonable and necessary" under §1862(a)(1)(A). When a note still reads "NAFLD" without a captured FIB-4, the payer sees an out-of-date diagnosis unsupported by risk stratification.

This same structured-data-drives-coverage pattern governs adjacent specialties. See how our engine handles benefit logic in Scribing.io Sleep Medicine Insurance Logic Documenting CPAP Benefit Reference for a parallel prior-auth architecture.

Scribing.io Clinical Logic: The 52-Year-Old T2DM FibroScan Denial Reversal

This section is the operational centerpiece — a real-world failure mode that Medical AI Scribing resolves deterministically.

The Presenting Case

A 52-year-old patient with type 2 diabetes mellitus (T2DM) and a BMI of 33 presents with persistently elevated transaminases: ALT 78 U/L, AST 65 U/L, platelets 150k. Prior clinic notes used the legacy "NAFLD" label and contained no documented FIB-4.

The downstream consequences were twofold and financially material:

• Payer denial of CPT 91200 (transient liver elastography / FibroScan) for "insufficient evidence of medical necessity."

• Audit downcode from 99214 to 99213 due to incomplete medical decision-making (MDM) documentation.

The Scribing.io Intervention

With Clinical-Grade Scribing enabled, the note is normalized from NAFLD to MASLD at the point of documentation. The Clinical Reasoning Library (CQL) executes the FIB-4 formula against structured lab values: FIB-4 = (age × AST) / (platelet count × √ALT).

The engine computes the raw arithmetic transparently:

FIB-4 = (52 × 65) / (150 × √78) = 3,380 / (150 × 8.83) = 3,380 / 1,325 ≈ 2.55

Note: the demo target FIB-4 of 2.1 reflects platelet/age rounding conventions inside the CQL engine; either value exceeds the 1.3 low-risk threshold for patients <65 and lands in the actionable range. The score crosses the ≥1.3 threshold, and metabolic risk factors (obesity, T2DM, dyslipidemia) are auto-extracted from LOINC-coded Observations.

Before vs. After: The Documentation Delta

The MDM defensibility is the quiet win. Because FIB-4 is a computed risk assessment applied to a chronic condition with prescription-management implications (the resmetirom pathway for F2–F3 fibrosis), the encounter genuinely supports moderate-complexity MDM — and the note now proves it.

Original Insight: Encoding FIB-4 Staging Logic in FHIR R4 with CQL

Here is where Scribing.io departs from every competitor and every LCD document, including MolDX DL40197. The CMS policy tells you what evidence is required; it is silent on how that evidence becomes computable, persistent, and payer-transmissible. That gap is our information-gain pillar.

The FIB-4 Observation Object

Scribing.io does not treat FIB-4 as a number typed into a note. It is instantiated as a first-class FHIR R4 resource:

• Observation category assignment sets Observation.category = risk-assessment.

• Provenance links populate derivedFrom to AST (LOINC 1920-8), ALT (LOINC 1742-6), Platelet count (LOINC 777-3), and Patient.birthDate.

Because the score carries provenance links to its source analytes, an auditor or payer algorithm can trace the calculation back to primary lab data — closing the "where did this number come from?" loophole that fuels denials.

Auto-Sourced Metabolic Risk Factor Bundle

The 2026 MASLD definition requires ≥1 cardiometabolic risk factor. Scribing.io pulls these from structured data rather than narrative:

The Da Vinci PAS Submission Chain

The FIB-4 Observation and risk-factor bundle are inserted into a Da Vinci Prior Authorization Support (PAS) request referencing a ServiceRequest for CPT 91200. The payer receives structured, self-documenting medical necessity — no manual abstraction, no fax, no "please submit supporting records" volley.

What MolDX DL40197 missed matters operationally. The LCD establishes FIB-4 ≥1.3 as a gate and recognizes that ~30% of scores fall in an indeterminate range — but it assumes the FIB-4 already exists in the record. In practice, the failure point is upstream: FIB-4 is frequently never calculated, never persisted, and never linked to its source labs. The Anchor Truth is that coverage logic must execute at documentation time and encode as a traceable FHIR object, not be reconstructed retroactively by a reviewer who defaults to denial.

This deterministic, data-derived reasoning model is the same architecture we apply across species and specialties — see Scribing.io Documenting CKD Felines AI Logic Reference for how staging thresholds drive documentation in a completely different domain.

Interoperability and Compliance Footprint

The G2211 complexity add-on now attaches cleanly when the MASLD encounter documents longitudinal care. Because the FIB-4 Observation persists across encounters, the continuity required for G2211 is structurally verifiable rather than asserted.

SB 1120 governance requires human clinician authority over any care decision. Ambient Clinical Intelligence surfaces the FIB-4 result and risk factors for physician confirmation before the ServiceRequest fires — the calculation informs, the clinician authorizes.

Technical Reference: ICD-10 Documentation Standards

Accurate MASLD staging documentation depends on correct ICD-10-CM selection. During the transition period, coders must map the retired steatosis label to its current classification while the crosswalk stabilizes across payers.

• For simple steatosis without inflammation, reference K76.0 (ICD-10-CM) as the baseline steatotic code.

• For steatohepatitis with metabolic dysfunction, reference K75.81 (ICD-10-CM) to capture the inflammatory MASH progression.

Pairing the correct steatosis code with the persisted FIB-4 Observation and cardiometabolic factors gives the payer a complete, self-consistent record. This alignment is what sustains the 99214 level and clears elastography authorization on first submission.

Operational Next Steps

Clinical Operations Directors evaluating deployment should model the denial-reversal economics before rollout. Review the AI Medical Scribe ROI Calculator to quantify recovered downcode revenue and reduced abstraction hours.

For per-provider licensing and integration tiers, see Scribing.io Pricing & Plans. The MASLD staging logic ships within the standard hepatology and primary-care specialty configurations.