Clinical Documentation Improvement: A Complete Guide for Outpatient Practices

What You Need to Know

• Clinical Documentation Improvement is the process of making clinical notes accurate, specific, and complete so the record reflects the true complexity of care delivered.
• Outpatient practices face the same documentation risks as large health systems: unspecified diagnoses, missing chronic condition documentation, underdocumented MDM, and late note completion.
• Practices of any size can close those gaps by auditing denial patterns, tightening EHR templates, enforcing same-day note closure, and using ambient AI documentation at the point of care.

What Is Clinical Documentation Improvement?

Clinical Documentation Improvement (CDI) is a set of practices and tools that make clinical notes accurate, specific, and complete. Every diagnosis, comorbidity, and procedure performed should appear in the record, supported by language that justifies the level of care provided.¹

CDI originated in inpatient settings, where hospitals employed dedicated specialists to review records and query physicians before claims submission. The same principles have moved into outpatient care. Value-based reimbursement models risk-adjust payment using diagnosis data. Quality reporting programs score practices on documented outcomes. Both forces have pushed CDI downstream into ambulatory settings.²³

Clinical Documentation Improvement comes down to one question: does the note say what actually happened? When it does not, the consequences reach further than most clinicians expect.

Why Does Clinical Documentation Improvement Matter in Outpatient Settings?

Most Clinical Documentation Improvement content targets large health systems with dedicated CDI teams, coders, and compliance staff. Outpatient practices across all sizes face the same documentation risks, typically without that infrastructure in place.⁴

Reimbursement accuracy

In Medicare Advantage and value-based contracts, payment is tied to hierarchical condition category (HCC) risk scores. Those scores depend on the diagnoses in the record. A diagnosis present at the visit but missing from the note does not affect the score. Underdocumented diagnoses suppress risk scores and reduce reimbursement for the actual complexity of care delivered.⁵

Claim denial rates

Payers deny claims when documentation does not support the level of service billed or when a diagnosis lacks the specificity ICD-10 requires. A note recording "hypertension" without specifying whether the condition is controlled, or whether chronic kidney disease is present, creates a coding gap. That gap sits between the E/M level billed and the documentation on file.⁶

Quality reporting and MIPS

Medicare's Merit-based Incentive Payment System (MIPS) scores practices on documented quality measures. A measure met clinically but not recorded does not count. Documentation gaps translate into quality scores that affect future payment rates.⁷

Continuity of care

Complete notes support the next clinician who sees the patient. Vague documentation forces the next provider to reconstruct context from patient recall. Note specificity is a patient safety issue as much as a billing one.

What Are the Key Elements of Clinical Documentation Improvement?

A Clinical Documentation Improvement program, whether staffed by specialists or built into workflow tools, works on the same documentation elements. The principles apply equally to a solo practice and a multi-specialty group.

Diagnostic specificity

ICD-10 requires a level of specificity that earlier coding systems did not. "Diabetes" is not a billable code. The note must support a specific type, the presence or absence of complications, and whether those complications are controlled. CDI means capturing this specificity during the encounter, not after.²

Medical necessity documentation

Every billed service must be medically necessary, and the note must say why. The connection between the complaint, the exam findings, and the clinical decision-making must be explicit enough for a reviewer to follow without inference. Notes that record what was done but not why are a major source of payer denials.⁶

Timeliness and completeness

Notes completed days after an encounter rely on memory. Errors of omission increase with every hour between the visit and the note. Clinical Documentation Improvement frameworks treat same-day or next-day closure as a clinical standard, not an administrative preference. Across a 1.7 billion-note US outpatient dataset, average note completion time dropped 11.1% between 2020 and 2023, yet average note length grew 8.1%. Speed gains from tools have not solved the underlying problem of incomplete capture.⁸

Chronic disease management documentation

Chronic conditions monitored but not actively treated still need to appear in the note to maintain HCC risk accuracy. In many outpatient CDI programs, clinicians are encouraged to document every active chronic condition at each encounter. Current status must be noted even when that condition is not the reason for the visit.⁵

Procedure and service documentation

E/M level selection depends on the documented complexity of medical decision-making or total time. CDI for E/M focuses on capturing the full scope of problems addressed, the data reviewed, and the risk of the treatment plan. All three are required to support higher-level codes.¹

What Are the Most Common Documentation Gaps in Outpatient Practices?

The gaps that create the most downstream problems cluster around the same documentation habits.⁹

• Unspecified diagnoses: a category code used when a more specific code is available and supported by the note.
• Missing chronic condition documentation: active chronic problems not noted at visits where a different condition is the primary reason for the encounter.
• Incomplete HPI: history of present illness sections that describe symptoms without onset, duration, severity, and associated factors required to support the E/M level billed.
• Underdocumented MDM: medical decision-making sections that list a plan without documenting the complexity of problems addressed, data reviewed, or risk of the management options.
• Omitted secondary diagnoses: comorbidities that affect management of the presenting condition but do not appear in the assessment and plan.
• Timing gaps: notes signed days after an encounter that cannot accurately reflect the clinical picture at the time of the visit.

How Can Outpatient Practices Implement Clinical Documentation Improvement Without a Dedicated CDI Program?

Large health systems hire Clinical Documentation Improvement specialists to review records and query physicians. Most outpatient practices do not run a dedicated CDI program. The practical alternative: move CDI upstream into the encounter itself, capturing the information needed before the visit ends.⁴

Step 1: Audit current documentation patterns

Start with a denial report from your billing team or a short audit of recent notes against coding guidelines. This shows whether the primary problem is diagnostic specificity, E/M level support, chronic condition capture, or timing. The fix for each differs.⁶

Step 2: Build specificity into visit templates

EHR note templates that prompt for diagnosis specificity at the field level are a low-cost CDI intervention. A template requiring clinicians to select a specific ICD-10 code reduces unspecified diagnoses at the point of entry. This requires a review of existing templates against common gap patterns, not new software.¹

Step 3: Establish same-day note closure as a team standard

Note closure timing is a CDI variable practices can change without new technology. Same-day or same-shift closure, with a clear protocol for late-running visits, directly reduces memory-dependent documentation. Same-day completion changes what gets captured, not just when.¹⁰

Step 4: Use ambient AI documentation to capture at the point of care

Ambient AI documentation is a major structural Clinical Documentation Improvement intervention available to outpatient practices right now. A clinician who can stay present in the room, paying attention to the patient rather than the keyboard, produces a conversation the AI captures as it happens. The note reflects what was actually said and assessed, not a reconstructed summary written later.¹¹

AI scribes that generate structured SOAP notes with suggested ICD-10 and CPT codes give clinicians a draft that prompts for specificity and completeness. This provides a CDI-like checkpoint at the end of every encounter.¹¹

Step 5: Create a feedback loop between coding and documentation

Denial patterns are documentation data. A billing team or RCM partner that tracks the diagnosis codes and E/M levels most tied to denials gives clinicians concrete information about where documentation is falling short. Reviewed monthly, this feedback loop delivers the same function as the Clinical Documentation Improvement specialist query process, without dedicated staff.⁶

How Does Technology Support Clinical Documentation Improvement?

Technology does not replace the clinical judgment CDI requires. It removes the friction that prevents clinicians from documenting that judgment completely.¹

EHR templates and structured data

Modern EHRs support structured data fields, problem lists, and diagnosis-specific documentation prompts. Using these features, rather than defaulting to free-text narrative, improves the searchability and auditability of the record. It also reduces the work required to find billable diagnoses at coding time.¹²

Ambient AI documentation

Ambient AI scribes capture the clinical encounter as it happens and generate a structured note draft quickly after visit completion. The clinician reviews and signs a complete draft rather than building a note from memory.¹¹

A 2026 UCSF study of 1,565 physicians across 1.2 million ambulatory encounters found that AI scribe adopters generated 1.81 more relative value units per week compared to non-adopters. They also handled 0.80 more patient encounters per week, with no increase in claim denial rates.¹³

A 2025 UCLA study found AI scribe technology reduced documentation time and improved note quality across the practices studied.¹²

The AMA's 2024 national survey of about 18,000 physicians found after-hours EHR time accounted for 22.5% of total EHR use. That figure remained persistent even as overall burnout declined year over year. Documentation that follows clinicians home is a structural problem. Ambient capture can help address it.¹⁴

Coding assistance

AI tools that suggest ICD-10 and CPT codes based on the documented encounter create a CDI checkpoint at the end of every visit. The suggestions prompt the clinician to confirm documented diagnoses are complete and specific enough to support the codes. They also flag conditions discussed during the visit that were omitted from the assessment and plan.¹¹

What Should Practices Look for in a Clinical Documentation Improvement-Enabling AI Scribe?

For outpatient practices evaluating AI scribes for CDI, these questions narrow the field quickly.

• Structured notes, not transcripts: a verbatim transcript requires as much clinical work to turn into a note as typing from scratch. The value is a structured SOAP draft the clinician reviews and edits.
• ICD-10 and CPT code suggestions in the draft: code suggestions at note review prompt the clinician to confirm every documented diagnosis is coded and every coded service is documented. This is the Clinical Documentation Improvement function that replaces the specialist query process.
• EHR integration: workflow value depends on how easily the draft gets into the record. Copy/paste from a separate app adds steps. Direct integration removes them.
• HIPAA compliance: any tool capturing audio of patient encounters requires a Business Associate Agreement (BAA). Verify the vendor provides one and review the audio storage and retention policy before deployment.
• Free trial without credit card requirements: the only way to evaluate accuracy and workflow fit is in actual clinical use.

What Are the HIPAA Considerations for AI Scribe Tools?

Any AI scribe that captures, processes, or stores protected health information (PHI) in audio or text form is subject to HIPAA. Verify the following before deployment.¹⁵

• Business Associate Agreement: the vendor must provide a BAA covering audio capture, note generation, and any data retention.
• Audio storage policy: vendors differ on whether audio is deleted after transcription, retained for a defined period, or stored long-term. Know what your vendor stores and for how long.
• Data residency: confirm whether data is processed and stored in US data centers and whether storage meets your state's security requirements.
• Patient consent: most practices inform patients at the start of the visit that an AI tool will assist with documentation. Whether written consent is required varies by state. Review with your compliance team before deployment.

Do not rely on this guide or any content source for patient disclosure language. Practice-specific legal review is required.

How Commure Scribe Supports Clinical Documentation Improvement in Outpatient Practices

Commure Scribe is an ambient AI medical scribe used by 20,000+ clinicians across outpatient practices. The features most relevant to CDI address the same gaps covered in this article: diagnostic specificity, E/M level support, denial reduction, and same-day note closure.

ICD-10 and CPT code suggestions

After each encounter, suggested ICD-10 and CPT codes appear in a separate tab. The clinician reviews them against the note before finalizing. This functions as a CDI checkpoint at the end of every visit: it prompts review of whether documented diagnoses are specific enough to support the codes, and whether any conditions discussed were omitted from the assessment and plan. Codes are suggested, not automated. The clinician confirms before the note is finalized.

Note quality and capture

Commure Scribe captures the clinical conversation and generates a structured SOAP note after the clinician clicks End Recording. Transcription accuracy is 99.4%. Clinicians report that the plan section is often more detailed and better structured than their manual notes, reflecting the AI catching clinical nuances documented in the conversation. 90%+ of providers reduce clinical documentation time and digital fatigue. Chart close time averages 43 seconds.

Clinicians report feeling present in the room rather than managing a keyboard during the encounter. The conversation that drives medical decision-making gets captured as it happens, not reconstructed afterward.

Denial reduction

Practices using Commure Scribe see a 25% reduction in denials on average. More complete, specific notes reduce the gap between the level of service billed and the documentation on file.

Data practices

Commure Scribe is HIPAA-compliant and SOC 2 certified. Audio is stored and encrypted. It is not used for AI training or any purpose other than generating the clinical note. Default audio retention is one year active, then archived for a minimum of six years per HIPAA requirements. A Business Associate Agreement is available. All data is stored onshore.

EHR compatibility

Copy/paste into any web-based EHR is available on all tiers. Write-back integration with 69 EHRs is available for larger practices.