ICD Coding for Non-Cancer Terminal Diagnoses: Hospice Documentation and LCD Alignment Requirements

Hospice agencies face intensifying Medicare Administrative Contractor (MAC) scrutiny on non-cancer terminal diagnoses. The HHS Office of Inspector General (OIG) has identified terminal prognosis documentation failures as the leading hospice claim denial driver. Centers for Medicare and Medicaid Services (CMS) Comprehensive Error Rate Testing (CERT) data confirm a 6.8% improper payment rate for non-hospital hospice claims in 2024, representing approximately $1.6 billion.

Non-cancer diagnoses represent approximately 77% of Medicare hospice beneficiaries, according to the CMS Hospice Monitoring Report (April 2025): dementia and Alzheimer’s disease 21.6%, cardiac conditions 19.2%, stroke 10.7%, and respiratory conditions 9.9%. These diagnoses carry a structurally higher audit burden than cancer because terminal prognosis requires disease-specific staging assessments, functional decline evidence, and clinical documentation that cancer diagnoses typically do not.

Key Takeaways

• Non-cancer diagnoses account for approximately 77% of Medicare hospice beneficiaries (CMS Hospice Monitoring Report, April 2025), but carry a structurally higher documentation burden than cancer diagnoses because terminal prognosis must be supported by disease-specific staging assessments and quantifiable functional decline evidence under MAC-published Local Coverage Determinations (LCDs).
• LCD requirements differ by diagnosis category and by MAC jurisdiction. The four most audited non-cancer diagnoses — dementia, CHF, COPD, and renal failure — each require specific functional staging tools, clinical thresholds, and supporting complication evidence that must appear in the clinical record before and at certification.
• ICD-10-CM specificity is a direct component of LCD alignment. Accurate terminal diagnosis coding requires the most specific code the documentation supports. Unspecified diagnosis codes for dementia, heart failure, or renal disease fail to establish the clinical severity that the LCD requires, and are a leading driver of documentation insufficiency findings in MAC medical review.
• The etiology/manifestation coding convention applies to dementia diagnoses. The underlying etiology code (e.g., G30.x for Alzheimer's disease) must be sequenced first, with the dementia manifestation code (F02.80 or F02.81) sequenced second. Reversing this sequence or omitting the etiology code violates ICD-10-CM Official Guidelines and produces a claim that misrepresents the clinical picture.
• Recertification documentation must demonstrate continued decline against LCD criteria, not restate the initial certification findings. MAC reviewers identify documentation recycling as a specific audit finding across all non-cancer diagnosis categories.

Why Non-Cancer Terminal Diagnoses Carry Higher Hospice Denial Risk

Cancer diagnoses follow an observable trajectory that MAC reviewers can assess from standard oncology documentation. Non-cancer diagnoses progress non-linearly, with periods of stabilization that complicate documentation without negating terminal prognosis. CMS addresses this through disease-specific LCDs that establish the clinical evidence required to support a six-month prognosis for each category.

How MAC Reviewers Approach Non-Cancer LCD Claims

LCDs are published by each MAC jurisdiction. CMS Medicare Benefit Policy Manual Chapter 9 sets the overarching hospice benefit standards, but disease-specific clinical thresholds are defined at the MAC level and vary by contractor. Agencies should verify the active LCD for each non-cancer diagnosis from their MAC’s coverage database. The CMS Medicare Coverage Database at cms.gov/medicare-coverage-database provides direct access to all active and retired LCDs by contractor.

MAC medical review evaluates non-cancer claims against a two-part framework: disease-specific criteria using staging tools (Functional Assessment Staging (FAST) scale for dementia, New York Heart Association (NYHA) classification for heart failure) and non-disease-specific baseline criteria requiring Karnofsky Performance Status (KPS) or Palliative Performance Score (PPS) below 70% and dependence on two or more activities of daily living (ADLs).

Both parts must be met. An ICD code that does not reflect the clinical severity these criteria require creates a documentation inconsistency before the reviewer reaches the physician narrative.

LCD Requirements by Non-Cancer Diagnosis Category

The four non-cancer diagnosis categories that generate the highest hospice audit volume are Alzheimer's disease and related dementias, congestive heart failure (CHF), chronic obstructive pulmonary disease (COPD), and renal failure. The table below maps the specific LCD clinical thresholds for each, sourced from the CMS Medicare Coverage Database (LCD L34538 and L33393, representative of Palmetto GBA and CGS jurisdictions).

Dementia: FAST Scale and ICD Specificity

Alzheimer’s disease hospice claims must demonstrate FAST Stage 7A or beyond. The LCD (L34538) references Stage 7A as the threshold: speech limited to one to five intelligible words per day. MAC reviewers expect the specific FAST stage to be documented in the clinical record at certification and recertification.

A dementia claim coded to G30.9 (Alzheimer’s disease, unspecified) without FAST staging documentation, or coded to F03.90 (unspecified dementia, unspecified severity) rather than the specific Alzheimer’s etiology code, presents both a specificity failure and an LCD alignment gap.

The etiology/manifestation convention requires G30.x sequenced first, with F02.80 (without behavioral disturbance) or F02.81 (with) sequenced second. CMS MM13882 confirms that claims using the manifestation code as primary, or coded to debility or adult failure to thrive, will be returned. For non-Alzheimer’s dementias (F01.5x vascular, G31.83 Lewy body, G31.09 frontotemporal), the FAST scale does not apply; these claims rely on KPS/PPS and ADL dependence.

CHF: NYHA Classification and Documentation Requirements

CHF terminal prognosis documentation requires NYHA Class IV classification, defined as inability to carry on any physical activity without discomfort and presence of symptoms at rest. The LCD also requires that the patient be optimally treated with diuretics and vasodilators, or that a documented medical contraindication to those treatments exists. An ejection fraction of 20% or less is supportive but not required.

The ICD-10-CM code set for CHF has a specificity structure that directly affects LCD alignment. The four primary types are: I50.21 (acute systolic CHF), I50.22 (chronic systolic CHF), I50.31 (acute diastolic CHF), I50.32 (chronic diastolic CHF), and I50.41/I50.42 (combined systolic and diastolic).

A claim coded to I50.9 (heart failure, unspecified) or I50.1 (left heart failure) when the clinical record documents systolic or diastolic type fails the specificity standard and signals to the MAC reviewer that the coding does not reflect the documented clinical picture. Where hypertension is the underlying cause, the combination code I11.0 (hypertensive heart disease with heart failure) must be used with an additional code for the specific heart failure type.

COPD: Spirometry Values and Functional Evidence

COPD terminal prognosis requires disabling dyspnea at rest, FEV1 below 30% predicted post-bronchodilator, and SpO2 at or below 88% at rest on room air. Lab testing is not required for hospice eligibility, but where spirometry and oxygen saturation values exist in the clinical record, they must be consistent with the coded diagnosis.

The ICD-10-CM COPD code structure requires selecting from J44.0 (COPD with acute lower respiratory infection), J44.1 (COPD with acute exacerbation), or J44.9 (COPD, unspecified), depending on current clinical status. J44.9 is frequently over-used in hospice coding when J44.1 is clinically supported by documented exacerbation.

This matters for LCD alignment because recurrent exacerbations are themselves a supporting criterion for terminal prognosis, and coding J44.9 when J44.1 is supported underrepresents the clinical trajectory. Red Road’s guide to hospice coding outsourcing covers how disease-specific ICD expertise affects denial rates across COPD and other non-cancer categories.

Renal Failure, Stroke, and Liver Disease

End-stage renal disease (ESRD) hospice claims require specific creatinine clearance values (below 10 cc/min, or below 15 cc/min for diabetics) and serum creatinine thresholds (above 8.0 mg/dl, or above 6.0 for diabetics), combined with a clinical decision to forgo dialysis or transplant.

The ICD-10-CM distinction between N18.5 (CKD Stage 5) and N18.6 (ESRD) matters: N18.6 applies only when the patient is on renal replacement therapy. Hospice patients forgoing dialysis are typically N18.5. Coding N18.9 (unspecified) when Stage 5 is documented is a specificity failure.

Stroke claims require attention to the acute versus late effect coding distinction. Acute codes (I63.xx) apply only during the acute phase. Hospice patients are typically post-acute and should be coded using late effect codes (I69.xx) reflecting the neurological sequelae driving the terminal prognosis.

Liver disease hospice claims typically require documentation of hepatic encephalopathy, ascites, or hepatorenal syndrome to support terminal prognosis, with ICD-10-CM codes from the K70.xx (alcoholic liver disease) or K74.xx (cirrhosis) categories coded with appropriate complication codes.

Clinical Documentation Elements That Support Terminal Prognosis 

Functional Decline Indicators, Nutritional Status, and Physician Narrative

Every non-cancer LCD requires the following documentation at certification and recertification. Each element must appear in the clinical record, not only in the physician narrative.

• Functional staging current to the certification period: FAST Stage 7A+ for dementia, NYHA Class IV for CHF, FEV1 and SpO2 values for COPD, creatinine clearance for renal failure
• KPS or PPS below 70% with dependence on two or more ADLs documented in the clinical record
• Nutritional evidence where applicable: unintentional weight loss of 10% or more in six months or serum albumin below 2.5 gm/dl
• Physician narrative that applies the LCD criteria to the specific patient’s clinical trajectory, not a generic restatement

MAC reviewers identify generic or recycled physician narratives as a specific audit finding. The narrative must connect the coded diagnoses to the patient’s specific functional and clinical evidence and explain why life expectancy is six months or less.

Recertification Documentation

Recertification carries the same LCD standards as initial certification. MAC reviewers reject recertifications that reaffirm without demonstrating continued decline: for dementia, documented complications during the period; for CHF, continued symptomatic status at rest or functional decline; for COPD, documented exacerbations or declining oxygen saturation. Red Road’s guide to hospice IDG documentation and certification period requirements covers the interdisciplinary team documentation standards that support recertification defensibility.

Acceptable vs. Denial-Risk ICD Coding Patterns for Non-Cancer Hospice Claims

The table below maps the ICD coding patterns that support LCD alignment against the patterns that generate MAC review findings, for the four primary non-cancer diagnosis categories. Code selection must be validated against current ICD-10-CM Official Guidelines and the applicable MAC LCD for each certification period.

Both patterns are systematic and preventable through coder-level review before claim submission.  Red Road’s guide to hospice RCM denial reasons covers the documentation-level causes behind these coding patterns and the operational controls that prevent them from reaching the payer.

How Specialized Hospice Coding Support Addresses LCD Alignment Risk

Non-cancer hospice coding requires coders trained in both ICD-10-CM Official Guidelines and the LCD-specific requirements for each diagnosis category. Applying general coding logic without hospice-specific LCD training produces the specificity failures and sequencing errors described above as pattern-level errors, not isolated mistakes.

Red Road’s Coding and OASIS Review service applies LCD alignment review at the coding stage, before claim submission, covering: primary diagnosis specificity against the applicable MAC LCD; etiology/manifestation sequencing for dementia; complication code capture; physician narrative consistency; and recertification documentation currency.

For hospice agencies that want to understand where their current non-cancer coding is generating denial risk, Red Road’s guide to working with LCDs and terminal diagnoses covers the operational framework for integrating LCD compliance into the coding and documentation workflow.

Bottom Line

Non-cancer hospice claims carry a higher documentation burden than cancer claims because terminal prognosis requires disease-specific staging assessments, functional decline evidence, and physician narrative connecting those findings to a six-month prognosis. An unspecified ICD code signals to a MAC reviewer that the clinical severity the LCD requires may not be established before they review a single progress note.

Each category has distinct requirements: FAST Stage 7+ with etiology/manifestation sequencing for dementia; NYHA Class IV with systolic/diastolic specification for CHF; FEV1 and SpO2 thresholds with exacerbation status for COPD; creatinine clearance with CKD stage for renal failure. None are met by an unspecified code.

Review Your Non-Cancer Hospice Coding Protocols

For hospice clinical directors, billing leads, and coders: pull your last 90-day denial report and identify which non-cancer diagnosis categories are generating the highest denial concentration. If dementia, CHF, COPD, or renal failure claims show a disproportionate share of insufficient documentation findings, the root cause is typically an ICD specificity gap, a missing staging assessment, or a physician narrative that does not apply the LCD requirements to the specific patient.

Learn more about Red Road’s Coding and OASIS Review services for hospice agencies.

Regulatory Sources Referenced

• CMS Hospice Monitoring Report, April 2025 — Non-Cancer Diagnosis Share of Hospice Beneficiaries FY2024 (cms.gov)
• CMS 2024 CERT FFS Supplemental Improper Payment Data — 6.8% Non-Hospital Hospice Improper Payment Rate / $1.6 Billion (cms.gov)
• OIG Hospice Care Featured Topic — $27.5 Billion Annual Medicare Hospice Spend / 1.8 Million Beneficiaries; Terminal Prognosis Documentation as Leading Denial Driver (oig.hhs.gov)
• CMS Medicare Coverage Database: LCD L34538 (Palmetto GBA) — Hospice: Determining Terminal Status; FAST Stage 7 Threshold for Alzheimer’s and Related Disorders (cms.gov/medicare-coverage-database)
• CMS Medicare Coverage Database: LCD L33393 (CGS) — Hospice: Determining Terminal Status; NYHA Class IV for Heart Disease; COPD and Renal Failure Criteria (cms.gov/medicare-coverage-database)
• CMS Medicare Coverage Database: LCD L34567 — Hospice Alzheimer’s Disease and Related Disorders; Reisberg FAST Scale Stage 7 Substages (cms.gov/medicare-coverage-database)
• ICD-10-CM Official Guidelines for Coding and Reporting FY 2025 — Etiology/Manifestation Convention; Combination Codes; Unacceptable Principal Diagnoses (cms.gov)
• CMS MM13882 — Principal Diagnosis Code Reporting Update for Hospice: Debility and Adult Failure to Thrive as Unacceptable Principal Diagnoses (cms.gov)
• CMS Medicare Benefit Policy Manual, Publication 100-02, Chapter 9 — Coverage of Hospice Services; Certification of Terminal Illness; Physician Narrative Requirements (cms.gov)
• 42 CFR § 418.22 — Certification of Terminal Illness: Physician Narrative and Clinical Findings Requirements