Identifying Undercoding Opportunities: A Strategic Guide for Healthcare Leaders

Introduction: The Hidden Cost of Undercoding in Healthcare

Undercoding occurs when documented clinical conditions—particularly clinically relevant chronic diseases and complications—are not fully captured in coded diagnoses. This is rarely a result of a single point failure. More often, it reflects small, systematic gaps across documentation, coding workflows, and feedback loops between clinicians and coding teams. Over time, these gaps translate into measurable financial and operational consequences.

From a revenue optimization perspective, undercoding can drive avoidable revenue leakage across multiple payment models, including Medicare Advantage risk adjustment, MSSP/ACO arrangements, and other value-based contracts. When Hierarchical Condition Category (HCC) coding is incomplete, the organization’s risk adjustment factor (RAF) may be understated, resulting in payments that do not match the true acuity and complexity of the patient population. Undercoding also distorts internal analytics: leaders may underestimate disease burden, under-resource care management programs, and misalign staffing, outreach, and clinical interventions.

The impact extends beyond finance. Inaccurate risk scores can compromise population health stratification, quality reporting, and longitudinal care planning. If chronic conditions are not consistently assessed, documented, and coded each year (where required by the model), patients who would benefit from proactive management may not be appropriately flagged for care coordination. In this way, undercoding becomes both a financial and clinical operations problem—one that should be addressed through disciplined governance, compliant workflows, and modern tooling that supports accurate clinical documentation and coding.

For healthcare leaders, prioritizing undercoding detection is not about “coding more.” It is about coding accurately and defensibly, supported by clear documentation, aligned clinician-coder workflows, and a compliance-first risk adjustment strategy.

Understanding the Root Causes of Undercoding

Undercoding is typically multifactorial. The most effective remediation plans start with an honest assessment of why diagnoses are missed, omitted, or not supported strongly enough to code. Common root causes include the following.

Documentation gaps and missing clinical specificity

Even when clinicians are providing high-quality care, the medical record may not consistently contain the elements required to code certain conditions—especially within HCC coding and risk adjustment contexts, where documentation must demonstrate that conditions were assessed and addressed during the encounter.

Examples of documentation issues that commonly drive undercoding include:

Chronic conditions listed on a problem list but not evaluated, monitored, assessed, or treated during the visit (i.e., not “actively managed” in documentation).
Failure to document disease status (e.g., “CKD stage 3a” vs. “CKD”).
Missing linkages between conditions and manifestations (e.g., diabetes with chronic kidney disease vs. diabetes without complications, when clinically true and documented).
Conditions treated by specialists not reconciled into the primary care note, leading to “silent” comorbidities across settings.
Annual recapture failures, where conditions documented in prior years are not re-assessed and documented in the current period.

Training limitations in complex HCC coding rules

HCC models and risk adjustment rules are complex and evolve over time. Even experienced coders and CDI professionals can struggle with:

Condition hierarchy logic (e.g., only the most severe HCC in a related group counting).
Model-specific nuances across Medicare Advantage, ACA, and other programs.
Documentation requirements for conditions frequently audited (e.g., depression, vascular disease, COPD, malnutrition).
The difference between clinical plausibility and codable support.

When training does not keep pace with regulatory updates and model changes, the default behavior can become conservative coding that inadvertently results in undercoding—especially in high-velocity outpatient environments.

Time pressure and workflow inefficiencies

In both ambulatory and inpatient contexts, time constraints can push teams toward a “minimum viable coding” approach:

Clinicians may prioritize immediate clinical tasks and omit status updates for stable conditions.
Coders may focus on primary diagnoses and high-visibility problems while missing secondary HCC-relevant conditions embedded in the narrative.
High volumes and short turnaround times can reduce the opportunity for clarification queries or follow-up documentation.

In risk adjustment programs, where annual recapture is critical, time pressure is amplified during peak periods (e.g., year-end), increasing the likelihood of missed opportunities.

Lack of real-time feedback between coders and clinicians

Undercoding persists when clinicians and coders operate in parallel rather than as a coordinated system. Without tight feedback loops:

Clinicians may not understand why a condition could not be coded (e.g., insufficient assessment language).
Coders may avoid querying due to time constraints or uncertainty about clinician responsiveness.
Teams may not have shared dashboards showing missed HCC opportunities, query response rates, and recapture performance.

Fear of overcoding and an overly conservative posture

A compliance-aware culture is essential, but fear of audits can sometimes swing organizations too far toward undercoding. This often appears as:

Avoidance of coding conditions perceived as “audit-prone,” even when supported.
Reluctance to code complications and manifestations unless explicitly stated in a specific format.
Coding only what is “obvious,” rather than what is clinically true and documented.

A mature compliance posture does not equate to conservative coding; it equates to accurate coding with strong documentation support and defensible processes.

Key Strategies for Identifying Undercoding Opportunities

Identifying undercoding opportunities requires a structured program that combines audit discipline, operational redesign, and technology enablement. The strongest approaches are systematic and repeatable—not one-time remediation events.

1) Conduct retrospective chart audits with a clear methodology

Retrospective audits remain a foundational tool for identifying undercoding. However, they are only as useful as their design. Effective audits typically:

Use statistically valid sampling where appropriate (and targeted sampling for high-impact populations such as high RAF variance panels).
Compare documented conditions (problem list, assessment/plan, historical diagnoses, specialist notes) versus coded diagnoses submitted on claims.
Identify both “missed codes” and “missing documentation” scenarios.
Classify root cause categories (documentation vs. coding vs. workflow gaps) to guide remediation.

To increase reliability and reduce bias, organizations often incorporate independent reviewers and standardized abstraction templates. Audit results should translate directly into targeted education and workflow changes, not just retrospective corrections.

2) Leverage AI-powered documentation analysis to surface missed HCC opportunities

Natural language processing (NLP) and machine learning can analyze large volumes of clinical text to identify conditions likely present but not coded. This is particularly valuable when undercoding is driven by:

Diagnoses documented in narrative sections but not captured in structured fields.
Comorbidities mentioned in specialist reports, hospital summaries, or scanned documents.
Subtle documentation patterns (e.g., repeated medication regimens implying chronic disease, when clinically appropriate and documented).

AI tools can prioritize charts for review by highlighting high-probability gaps and providing explainable evidence trails (e.g., where in the note the condition is documented). This can reduce reviewer burden and support scalable risk adjustment programs.

Used appropriately, AI should function as decision support—not as an autonomous coding engine. Organizations increasingly deploy AI to triage charts, recommend queries, and standardize review quality. Solutions such as Arkangel AI are often positioned in this “assistive” layer to augment chart review and reduce missed opportunities, while maintaining clinician and coder accountability for final decisions.

3) Implement concurrent reviews to catch gaps before claim submission

Concurrent (or near-real-time) coding and documentation review can prevent undercoding from becoming a “post-claim” problem. Benefits include:

The ability to query clinicians while the encounter is still fresh.
Reduced rework and fewer retroactive corrections.
Improved clinician learning through immediate feedback.

Concurrent review models can be deployed in multiple ways:

Point-of-care prompts or clinician-facing documentation cues for chronic disease status updates.
Daily or weekly coder/CDI review of high-priority visits (e.g., annual wellness visits, transitional care management, high-risk clinic days).
Shared worklists for recapture candidates based on prior-year diagnoses and current visit schedules.

The operational goal is to make correct coding the default outcome, rather than a later clean-up effort.

4) Use benchmarking to detect RAF and coding pattern anomalies

Benchmarking can help leaders understand whether undercoding is likely occurring at scale. Useful approaches include:

Comparing RAF scores and HCC prevalence rates against peer organizations with similar demographics.
Monitoring year-over-year RAF drift, especially unexpected declines.
Evaluating provider-level variation in HCC capture (adjusted for panel risk and visit volume).
Tracking recapture rates by condition category (e.g., CHF, COPD, CKD, diabetes complications).

Benchmarking should be interpreted carefully. A low RAF score does not automatically prove undercoding (it may reflect healthier populations), and a high RAF does not prove accuracy. The value lies in identifying outliers that warrant deeper audit and workflow review.

5) Establish cross-functional governance involving clinicians, coders, and compliance

Undercoding remediation often fails when it is assigned solely to coding or revenue cycle teams. A durable program requires cross-functional ownership, typically including:

Clinical leadership (to align documentation practices with clinical reality and care workflows).
Coding and CDI leadership (to ensure technical accuracy and consistent query standards).
Compliance and legal stakeholders (to design defensible processes and audit readiness).
Analytics and population health leaders (to connect risk adjustment accuracy with care management priorities).

Effective governance sets clear policies around documentation standards, query practices, escalation pathways, and quality assurance. It also creates shared accountability: clinicians are responsible for accurate documentation; coders are responsible for applying coding rules correctly; compliance ensures guardrails; leadership ensures resourcing and measurement.

Balancing Revenue Optimization with Compliance

Organizations can—and should—pursue revenue optimization through accurate coding. The key is to ensure that every coded diagnosis is clinically supported, properly documented, and compliant with payer and regulatory rules. In risk adjustment environments, this balance is particularly critical because scrutiny is high and payment impact is significant.

Coding accuracy must be the goal—not “more codes”

Revenue optimization efforts should be framed as accuracy initiatives:

Capture the true disease burden that is already present and managed.
Ensure documentation supports the level of specificity required.
Avoid adding diagnoses without clear evidence of assessment and management.

Leaders should reinforce that the objective is correct representation of patient complexity to support appropriate reimbursement and resource allocation—not aggressive revenue capture.

Align risk adjustment workflows with CMS and OIG expectations

Compliance frameworks should be built around current guidance and audit realities. Key considerations include:

Diagnoses must be supported by documentation in the medical record and meet reporting requirements for the relevant program year.
Documentation should demonstrate evaluation and/or management of the condition during the encounter, consistent with widely adopted standards (often operationalized using MEAT—Monitor, Evaluate, Assess/Address, Treat—though organizations may vary in terminology).
Risk adjustment programs are subject to ongoing oversight, including Risk Adjustment Data Validation (RADV) in Medicare Advantage and other audit mechanisms.

Leaders should ensure policies reflect CMS guidance and anticipate OIG priorities, particularly around high-risk diagnosis categories and unsupported coding patterns.

Build strong audit trails and defensible documentation practices

A compliance-forward undercoding initiative should include:

Clear documentation standards for chronic conditions (status, assessment, plan).
Standardized query templates and criteria for when queries are appropriate.
Retention of evidence supporting why a diagnosis was coded (e.g., note excerpts, lab trends, imaging summaries, specialist documentation) where allowable and appropriate.
Routine internal audits to validate that coding practices are consistent and supported.

Audit readiness is not a one-time event; it is a continuous capability.

Train staff on the difference between correcting undercoding and avoiding upcoding

Education should address common gray areas:

When a condition is historical vs. active.
When a diagnosis is suspected vs. confirmed (and applicable outpatient coding rules).
The documentation needed to support complications and manifestations.
Appropriate use of problem lists and past medical history sections.

Clinicians and coders should share a common understanding of what constitutes sufficient support for a coded diagnosis—especially for HCC coding.

Add compliance checkpoints to the undercoding detection workflow

Practical safeguards include:

Pre-submission QA reviews for high-impact HCC categories.
Randomized secondary review of AI-suggested opportunities to validate precision and reduce bias.
Threshold triggers for provider-level outliers (both high and low) requiring review.
Regular compliance committee reporting with metrics on queries, recapture rates, denial trends, and audit outcomes.

These checkpoints help ensure that revenue optimization initiatives remain defensible and aligned with regulatory expectations.

Practical Takeaways

Establish an undercoding program goal focused on accuracy and documentation integrity, not volume of codes.
Use targeted retrospective audits to identify the most common missed conditions and isolate whether the root cause is documentation, coding, or workflow.
Prioritize high-impact encounter types (e.g., annual wellness visits, chronic care follow-ups, transitional care management) for concurrent review.
Implement closed-loop clinician-coder feedback, including standardized queries, response SLAs, and periodic education based on real examples.
Track recapture rates by condition category, not just overall RAF, to identify where undercoding is concentrated (e.g., CKD staging, diabetes complications, CHF specificity).
Adopt AI as a triage and decision-support layer—with clear governance, explainability expectations, and human validation—to scale chart review without sacrificing compliance.
Embed compliance checkpoints (QA sampling, audit trails, outlier detection) in the workflow to mitigate upcoding risk while correcting undercoding.
Align performance measurement across teams: combine coding accuracy, query quality, clinician documentation quality, and audit outcomes into leadership dashboards.

Future Outlook: AI and Predictive Analytics in Undercoding Detection

The next phase of undercoding detection will be shaped by automation, interoperability, and evolving regulatory scrutiny. Leaders should anticipate several trends.

AI and NLP will become standard for documentation review—if governance matures

NLP-driven chart review can scale far beyond manual auditing, enabling organizations to:

Identify missed HCC codes across large patient populations.
Surface documentation elements that support coding (or highlight what is missing).
Standardize review quality and reduce variability between reviewers.

However, AI performance depends on data quality and workflow integration. Leaders should expect to invest in:

Model governance (validation, monitoring, drift detection).
Explainability requirements (why a suggestion was made, where evidence appears).
Human-in-the-loop processes to ensure coding decisions remain accountable and compliant.

Predictive analytics will help prioritize patients likely to be undercoded

Rather than reviewing every chart equally, predictive models can help flag:

Patients with prior-year HCCs who have not had recapture documentation this year.
Patients with utilization patterns (ED visits, hospitalizations, specialist notes) suggesting rising acuity.
Medication and lab patterns consistent with chronic disease severity that may not be reflected in coded data.

This prioritization can improve ROI and reduce clinician burden by focusing effort where undercoding risk is highest.

Real-time coding assistance will shift undercoding prevention upstream

Undercoding is easiest to correct at the point of documentation. Emerging workflows include:

EHR-integrated prompts reminding clinicians to document status and assessment of active chronic conditions during relevant visits.
Automated pre-visit planning summaries that highlight conditions needing annual reassessment.
Smart queries generated during or immediately after the encounter, routed to the appropriate clinician.

The most effective tools will be those that minimize disruption and respect clinical workflow, while improving the completeness of documentation.

Regulatory evolution will keep compliance pressure high

Risk adjustment models, audit approaches, and documentation expectations continue to evolve. Leaders should plan for:

Continued scrutiny of unsupported diagnoses and documentation insufficiency.
Greater emphasis on data provenance and traceability (especially as AI enters workflows).
Model updates that may change which conditions map to payment categories and how hierarchies are applied.

Organizations that build strong compliance infrastructure now—policy, training, auditing, and governance—will be better positioned to adapt to changes without reactive overhauls.

Conclusion: Taking Action to Close the Undercoding Gap

Undercoding is a persistent, enterprise-wide challenge that undermines revenue optimization, distorts risk adjustment accuracy, and can weaken population health decision-making. The downstream effects—lost reimbursement, underestimated RAF scores, and misallocated resources—are often invisible until leaders examine coding patterns, documentation quality, and benchmark variance in a structured way.

Closing the undercoding gap requires more than isolated education or periodic audits. It calls for a coordinated strategy that combines:

disciplined chart audit programs,
concurrent review workflows,
clinician-coder feedback mechanisms,
compliance-first governance, and
thoughtfully implemented AI decision support.

When executed well, accurate HCC coding becomes a strategic capability: it supports appropriate reimbursement, improves the fidelity of risk stratification, and strengthens the organization’s ability to plan and deliver care for complex populations. Healthcare leaders should assess current coding accuracy, identify high-yield quick wins (such as recapture workflows for common chronic conditions), and invest in sustainable processes and tools that make accurate documentation and coding the default outcome.