Under the Prior Resource Utilization Groups (PPS) model, primary diagnosis selection in home health carried limited reimbursement consequences. The Centers for Medicare and Medicaid Services (CMS) restructured that relationship with the Patient-Driven Groupings Model (PDGM), effective January 1, 2020. Specifically, the primary ICD-10 diagnosis code on a 30-day claim determines which of 12 clinical groupings the period falls into, which directly sets the base payment rate. A sequencing error that places a patient in the wrong clinical group does not produce a minor coding discrepancy. It produces a payment variance that compounds across every episode until the error is identified and corrected.

The audit consequences follow the same logic. The OIG’s 2024 nationwide home health compliance audit series identified unsupported diagnosis codes and invalid Face-to-Face (F2F) encounter documentation among the most common deficiencies in home health claims. CMS’s Comprehensive Error Rate Testing (CERT) program reported a 7.7% improper payment rate for home health in 2023, representing approximately $1.2 billion. The ICD coding decisions that drive PDGM grouping placement are among the most audited compliance points in the home health benefit.

The PDGM clinical grouping structure, ICD-10-CM sequencing rules specific to home health, documentation requirements for defensible code assignment, and the coding errors that generate reimbursement variance and audit exposure are each examined below.

Key Takeaways

• Under PDGM, primary diagnosis selection is a reimbursement-critical decision. The principal diagnosis on a 30-day claim determines clinical group assignment and base payment rate. A sequencing error that moves a patient into the wrong clinical group creates a payment variance on every claim it affects until the error is corrected.
• Approximately 40% of ICD-10-CM codes acceptable under PPS are not acceptable as primary diagnoses under PDGM. Symptom codes, manifestation codes, and most therapy treatment diagnosis codes cannot serve as primary diagnosis under PDGM grouping rules.
• Home health ICD coding follows Section I of the ICD-10-CM Official Guidelines, not the outpatient coding guidelines that govern physician and clinic settings. The primary diagnosis must reflect the focus of care for the home health episode, which is not necessarily the admitting diagnosis or the most acute condition.
• Comorbidity coding directly affects payment. Low comorbidity adjustment adds 6.01% to the case-mix weight; High comorbidity adjustment adds 12.95%. Incomplete secondary diagnosis coding leaves a measurable reimbursement adjustment uncaptured on every affected period.
• OASIS item responses and coded diagnoses must be clinically consistent. Misalignment between OASIS functional and clinical data and the coded primary diagnosis is one of the most consistent findings in MAC medical review and OIG compliance audits.

How PDGM Made Primary Diagnosis Selection a Reimbursement-Critical Decision

PDGM calculates payment for each 30-day period through four sequential variables: admission source, timing, clinical grouping, and functional impairment level, with a comorbidity adjustment applied on top. The primary diagnosis code on the claim determines clinical grouping entirely. Of the four variables, clinical grouping and comorbidity adjustment are the two where ICD coding decisions carry direct financial consequences.

The 12 PDGM Clinical Groupings

CMS designed the 12 clinical groupings to represent the most common reasons for home health services. The groupings divide into four categories: three rehabilitation groups, seven Medication Management Teaching and Assessment (MMTA) subgroups, Complex Nursing Interventions, and Behavioral Health. MMTA subgroups account for approximately 55% of all projected 30-day periods under PDGM, according to the CMS Federal Register. Each clinical group is mapped to a defined set of acceptable ICD-10-CM primary diagnosis codes.

If the primary diagnosis code on a claim does not fall within an accepted grouping code set, CMS returns the claim to the agency as ungroupable rather than denying it outright. The reference table below maps each clinical group to its primary care focus and representative ICD-10-CM code categories.

How a Sequencing Error Compounds Across Episodes

Before PDGM, therapy treatment diagnosis codes such as M62.81 (muscle weakness, generalized), R26.89 (other abnormalities of gait), and R13.12 (dysphagia, oropharyngeal phase) were routinely used as primary diagnoses. Under PDGM, these codes are ungroupable as primary diagnoses for most claims. An agency that has not updated its coding protocols since PPS may be producing a pattern of returned claims without identifying the root cause. Red Road’s complete guide to home health coding and OASIS review covers the PDGM grouping logic in operational detail.

Because PDGM recalculates grouping independently for each 30-day period, a recurring sequencing error on a patient population affects every claim until the coding protocol is corrected. Claims still process and pay — at the wrong rate. As a result, the variance often remains invisible until agencies compare paid claims against the PDGM groupings the documentation actually supported.

ICD-10-CM Guidelines Specific to Home Health Coding

The ICD-10-CM Official Guidelines for Coding and Reporting contain setting-specific rules. Home health follows Section I of the Official Guidelines, not the outpatient encounter guidelines (Section IV) that govern physician office and clinic coding. The most consequential difference is the sequencing standard.

How Home Health Sequencing Rules Differ from Other Settings

In inpatient hospital coding, the principal diagnosis is the condition established after study to be chiefly responsible for the admission. In outpatient coding, the primary diagnosis reflects the reason for the encounter. In home health, the primary diagnosis must reflect the focus of care for the home health episode as defined by the plan of care. The same patient record will produce different primary diagnoses depending on which guideline set the coder applies.

A patient admitted to home health following hip replacement may have Z96.641 (presence of right artificial hip joint) documented on the hospital discharge record. That code is not appropriate as the primary home health diagnosis under PDGM. The focus of the home health episode is the skilled service being provided, which would typically be coded from surgical aftercare or musculoskeletal groupings. Coders applying inpatient sequencing logic to home health claims produce this grouping error as a repeatable pattern across every post-surgical patient.

Symptom Codes, Unspecified Codes, and PDGM Acceptability

Symptom codes may be used as a primary diagnosis in home health when a definitive diagnosis has not been established. However, not all symptom codes are acceptable or groupable as primary diagnoses under PDGM. Before using a symptom code as primary, the coder must verify that it appears on the CMS accepted primary diagnosis list. An ungroupable symptom code used as primary produces a returned claim.

Unspecified codes carry additional compliance scrutiny. In particular, when a MAC reviewer examines a claim with an unspecified primary diagnosis code, the finding is direct: either the physician documentation did not support specificity, or the coder did not query when it would have. ICD-10-CM Official Guidelines require code assignment at the highest level of specificity supported by documentation. In home health, where referral documentation quality varies substantially across sources, specificity failures are predictable and preventable.

Combination Codes vs. Multiple Codes in Home Health

Where a combination code exists in ICD-10-CM, it must be used rather than separate codes for etiology and manifestation. This rule has direct PDGM consequences because the combination code and a separately coded etiology may be assigned to different clinical groups.

Diabetic peripheral neuropathy illustrates the point: the correct code is E11.42 (type 2 diabetes with diabetic neuropathy, unspecified) or a more specific subcategory, not E11.9 plus G62.9 separately. Using multiple codes where a combination code exists violates ICD-10-CM Official Guidelines and may produce a different PDGM grouping than the correctly coded combination would generate.

Common ICD Coding Errors That Trigger Case-Mix Audits

The OIG’s nationwide home health audit series found unsupported diagnosis codes and invalid F2F encounter documentation as two of the three most common deficiency categories. The PEPPER report shows which clinical groupings generate billing patterns that differ from peer agencies in the same jurisdiction. In other words, this comparative billing analysis surfaces patterned sequencing errors at the population level, before an individual claim is reviewed. Groupings where an agency is a consistent outlier, high or low, are the first place to examine coding protocols.

Primary Diagnosis Sequencing Errors

The most consequential error under PDGM is placing the wrong condition in the primary diagnosis position. Common patterns include: coding the referring diagnosis rather than the home health focus of care, sequencing the most acute condition rather than the one driving skilled services, and applying the etiology code where a combination or manifestation code is clinically correct.

A patient with chronic systolic congestive heart failure (I50.22), being seen primarily for venous stasis wound care, should have the wound condition as primary, not the heart failure. Sequencing cardiac as the primary place, the patient in MMTA-Cardiac rather than Wounds, a different payment rate, and different medical necessity criteria for that episode. The error is process-level: it recurs on every similarly situated patient as long as the coder’s protocol defaults to the most clinically significant condition rather than the focus of care.

Specificity Failures and Unspecified Code Use

Specificity failures are audit-readable. When a MAC reviewer sees an unspecified primary diagnosis where clinical documentation would support a more specific code, the finding is direct. Under the ICD-10-CM Official Guidelines, code assignment must reflect the highest level of specificity the documentation supports. Unspecified codes may also produce a lower-case-mix weight than the correct specific code would generate, creating simultaneous audit exposure and underpayment.

The appropriate response to insufficient documentation is a physician query, not an unspecified code. Agencies that accept inadequate referral documentation and code to unspecified levels without querying are producing a predictable pattern of specificity failures that is visible in CERT audit data and PEPPER outlier statistics.

Misalignment Between Physician Orders, OASIS, and ICD Codes

PDGM requires that physician documentation, OASIS item responses, and coded diagnoses present a consistent clinical picture. Misalignment between these three data sources is a specific MAC audit trigger.

A patient coded in the Neuro/Stroke Rehabilitation grouping should have OASIS responses reflecting neurological functional deficits. OASIS responses reflecting primarily wound or cardiac limitations create an internal inconsistency that auditors identify directly. Red Road’s guide to OASIS-E documentation requirements covers the M-item responses that must align with the coded clinical picture.

Surgical Aftercare Codes Under PDGM

Z48.xx surgical aftercare codes are appropriate as primary diagnoses when the focus of care is post-operative management. They are not appropriate when the patient has developed a complication or residual condition that is now the primary focus of care. A patient recovering from knee replacement who develops a postoperative wound infection should carry the infection code as primary, not Z48.1, because the infection — not aftercare — is the focus of the home health episode.

Using aftercare codes past their clinically appropriate window, or when a complication has superseded the post-surgical management phase, is a consistent MAC audit finding. The coder must assess whether the episode is still focused on post-operative management or whether a new or complicating condition has become the clinical focus.

Documentation Requirements That Support ICD Code Assignment

What Physician Documentation Must Establish

Physician certification and plan of care documentation must establish the primary diagnosis, the skilled care rationale, the patient’s functional and clinical status, and medical necessity for home health services. For ICD coding, the documentation must support the specificity level of the assigned code. Where documentation does not support a specific code, the coder must query the physician before assigning. Coding to unspecified levels without querying when the clinical picture warrants greater specificity is a compliance failure.

Referral documentation from hospitals and physician offices frequently lacks the specificity that PDGM-accurate coding requires. Consequently, agencies that accept insufficient referral documentation without querying produce a pattern of under-specified codes, generating a reimbursement gap and a documentation risk profile that accumulates across the patient population. Clinical documentation integrity requires agencies to embed the query process in the coding workflow — not treat it as optional. 

How OASIS Responses Must Align with Coded Diagnoses

OASIS items M1021 (primary diagnosis) and M1023 (other diagnoses) must be completed in alignment with the coded diagnoses and the plan of care. Achieving that alignment requires interdisciplinary review — coders, clinicians, and the reviewing physician must work from the same clinical picture. Where OASIS responses and coded diagnoses diverge, the inconsistency indicates either a documentation gap or a coding error that needs correction before claim submission. OASIS-E, effective January 1, 2023, expanded the standardized patient assessment data elements. Coders and clinicians must apply current OASIS-E guidance to ensure their responses accurately reflect the coded clinical picture. Red Road’s OASIS-E documentation guide covers the item-level requirements that support defensible primary diagnosis coding.

Face-to-Face Encounter Documentation as a Coding Control Point

The F2F encounter requirement mandates that a physician or qualified non-physician practitioner certify that the patient had an in-person encounter related to the primary reason for home health services within 90 days before or 30 days after the start of care. The F2F documentation is both a certification requirement and a coding control point.

When F2F documentation is inconsistent with the coded primary diagnosis, the inconsistency is audit-visible. If the F2F encounter note describes a cardiac assessment and the coded primary diagnosis is in the Musculoskeletal Rehab grouping, the documentation does not support the coded reason for home health services. The OIG’s nationwide audit series identified invalid F2F encounters as one of the three most common deficiency types in home health claims.

• Timing: The encounter must occur within 90 days before the start of care or 30 days after. Claims with F2F encounters outside this window are non-compliant regardless of clinical documentation quality.
• Practitioner qualification: The certifying physician, nurse practitioner, clinical nurse specialist, or physician assistant must meet Medicare home health benefit qualification requirements.
• Clinical content: The encounter note must document that the encounter was related to the primary reason for home health services. A routine office visit note that does not reference the home health-related condition does not meet the F2F content standard.

Sequencing and Comorbidity Coding Under PDGM

Primary Diagnosis Must Reflect the Focus of Care

ICD-10-CM Official Guidelines for home health state that the primary diagnosis is the condition most responsible for the provision of home health services during the episode. This is not the most acute condition, the hospital discharge diagnosis, or the condition the patient has had the longest. It is the condition that is the clinical reason for the skilled services being provided under the plan of care.

Coders applying this rule must understand what skilled services are planned and why. A nurse visit to manage a wound following vascular surgery requires a wound-related primary diagnosis, not a vascular code, even if the underlying vascular condition is clinically more significant. A physical therapy visit for gait training after stroke requires a neurological primary diagnosis, not the hypertension that contributed to the stroke. Applying the focus-of-care principle requires coders to understand the skilled services being delivered — not just the diagnosis list.

Z-Codes as Primary Diagnosis in Home Health

Z-codes are appropriate as primary diagnoses in home health in specific circumstances. Post-surgical aftercare (Z48.xx) is the most common. Z-codes for rehabilitation (Z50.xx), monitoring of therapeutic drug use, and aftercare following joint replacement are all used in home health. The key constraint is that Z-codes used as primary must appear on the CMS accepted primary diagnosis list for PDGM. Not all Z-codes are groupable as primary diagnoses, and the accepted list is updated with each annual ICD-10-CM revision.

Comorbidity Coding and the PDGM Payment Adjustment

Secondary diagnoses affect payment through comorbidity adjustment. CMS defines 12 low-comorbidity subgroups and 34 high-comorbidity subgroups. A 30-day period can receive only one comorbidity adjustment, but that adjustment adds 6.01% (Low) or 12.95% (High) to the case-mix weight.

For an agency with an average episode reimbursement of $2,500 per period, failing to code documented comorbidities that qualify for a High adjustment represents a $324 reimbursement gap per period. Across a patient population where those comorbidities are documented but not coded, the cumulative underpayment adds up to a material gap.  Claims pay at the lower rate without generating a denial, making the gap invisible in denial reports and visible only through case-mix analysis.

Secondary diagnoses must reflect active conditions being treated or affecting care during the episode. Resolved conditions, historical diagnoses not affecting current care, and conditions documented only in referral records without evidence of current management should not be coded. All coded secondary diagnoses should match across the claim, OASIS M1023, and the clinical record.

How Specialized ICD Coding Support Reduces PDGM Audit Exposure

PDGM-accurate ICD coding requires home health-specific training in both the ICD-10-CM Official Guidelines and the PDGM grouping rules that govern which codes are acceptable as primary diagnoses. Generalist coders applying outpatient or inpatient sequencing logic produce the errors described above as protocol-driven failures, not isolated mistakes. Because PDGM grouping logic differs substantially from traditional ICD coding environments, some agencies apply a secondary coding review process focused specifically on home health reimbursement accuracy and coding governance. Red Road’s Coding and OASIS Review service uses coders trained specifically in home health ICD-10 coding and PDGM clinical grouping logic — not generalist coding protocols.

The review process covers primary diagnosis sequencing against PDGM grouping criteria, specificity verification against physician documentation, OASIS M1021 and M1023 alignment, comorbidity capture against PDGM adjustment subgroup definitions, and F2F documentation consistency with the coded primary diagnosis. Each of these review points corresponds directly to a deficiency category in OIG audit findings and MAC medical review outcomes.

For agencies evaluating whether their current coding is generating patterned grouping errors or reimbursement variance, Red Road’s coding quality metrics guide covers the performance indicators that surface PDGM grouping accuracy problems before they accumulate into medical review risk.

Bottom Line

PDGM changed the financial stakes of primary diagnosis selection. Under PPS, a sequencing error produced a modest HHRG adjustment. Under PDGM, the same error may shift a patient across clinical groups, creating cumulative reimbursement distortion that persists until the coding protocol is corrected. Because affected claims still process and pay, the variance accumulates as an operational blind spot — visible through PDGM grouping analysis but absent from denial reports.

The audit exposure follows the same pattern. OIG compliance audits and MAC medical review target documentation-coding consistency, specificity standards, and F2F documentation validity. Agencies using generalist coders without PDGM-specific training are producing predictable patterns of grouping errors, specificity failures, and OASIS misalignments that surface in PEPPER outlier data before they surface in internal QA review.

Review Your ICD Coding Protocols for PDGM Accuracy

A practical starting point: pull your PEPPER report and identify which clinical groupings show outlier billing patterns relative to peers. Consistent deviation in a specific grouping may reflect a systematic sequencing pattern rather than a clinical census difference. Compare that pattern against your current primary diagnosis protocols and the CMS accepted primary diagnosis list for the current fiscal year. For agencies that want a more structured review, a grouping variance analysis — comparing PDGM clinical group distribution against coded diagnosis data by coder and by patient population — identifies where coding protocol adjustments or coder education workflows are needed before audit selection surfaces them. The ADR prevention checklist covers the documentation standards that support defensible primary diagnosis coding across the most common PDGM groupings.

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

Regulatory Sources Referenced

• CMS PDGM Overview — 12 Clinical Groupings, Comorbidity Adjustments, and 432 Payment Groups (cms.gov/medicare/payment/prospective-payment-systems/home-health)
• OIG Medicare Home Health Compliance Audit: Bridge Home Health — Nationwide Audit Series Findings (oig.hhs.gov, December 2024)
• CMS CERT 2023 — 7.7% Improper Payment Rate / $1.2 Billion in Home Health Improper Payments (cms.gov)
• ICD-10-CM Official Guidelines for Coding and Reporting FY 2024 — Section I General Guidelines; Home Health Sequencing Standards (cms.gov)
• CMS PDGM Grouper Software and Accepted Primary Diagnosis Code List — Updated Annually with ICD-10-CM Revisions (cms.gov)
• CMS Federal Register: PDGM Final Rule — MMTA Grouping Share (~55%), Comorbidity Adjustment Rates (6.01% Low / 12.95% High) (federalregister.gov)
• CMS OASIS-E Guidance Manual — M1021/M1023 Diagnosis Coding Requirements, Effective January 1, 2023 (cms.gov)
• CMS Medicare Benefit Policy Manual, Publication 100-02, Chapter 7 — Face-to-Face Encounter Requirements for Home Health Services (cms.gov)