ICD-10 Coding Accuracy : Common Errors Across Specialties

Welcome! If you are here, it is because you already know one of the hardest truths in modern medicine: the clinical work may happen in the exam room, the OR, or at the bedside, but the financial consequences often happen quietly in the diagnosis field. Physicians are being asked to carry an impossible load : deliver excellent care, document defensibly, supervise teams, manage inboxes, respond to payer rules, and somehow protect revenue integrity at the same time. That tension is real. It is not laziness, and it is not lack of intelligence. It is administrative overload colliding with a coding system that rewards microscopic specificity.

We often spend our days worrying about high-profile denials or complex NCCI bundling rules, because those problems are visible. They generate edits, rejections, and appeals. But there is a quieter and often more dangerous revenue threat living inside your claims data every single day: ICD-10 coding errors. These mistakes do not always announce themselves. They do not always trigger an immediate rejection. Instead, they chip away at payment accuracy, reduce risk-adjustment capture, distort quality reporting, and create a false picture of the complexity you manage.

Unlike a modifier usage rules mistake that results in an immediate "hard denial" you can appeal, ICD-10 errors frequently slip through the front door. They are "soft errors" that result in undercoded claims, mismatched DRG groupings, missed HCC risk-adjustment opportunities, and preventable downstream audit exposure. As a physician who codes, I see this daily. I have walked in your shoes, balancing patient care with the administrative burden of clinical documentation while trying to make sure the medical record says enough : but not too much, clearly enough : but still fast enough, specific enough : but without turning every note into a compliance dissertation.

That physician perspective matters. When a coder asks for specificity, what they are really asking for is a clinically accurate story that can survive reimbursement logic. When a payer asks for specificity, what they are really doing is deciding whether your patient is "sick enough" on paper to justify the resources you used. If your documentation says "diabetes" but the patient has diabetic CKD, neuropathy, and vascular disease, the medical reality and the reimbursement reality immediately separate. That gap is where revenue leaks begin.

It is why I founded Dr. Sloan-Kelly Consulting: to ensure your expertise is reflected accurately in your revenue. My focus has never been simply getting claims out the door. It is protecting the integrity of physician work, translating clinical complexity into compliant code selection, and helping practices reduce the hidden drag caused by undercoding, overuse of unspecified codes, poor sequencing, and missed 2026 code updates. Today, we are diving deep into the FY 2026 ICD-10-CM standards to identify exactly what is going wrong and how we can fix it together before your next audit cycle begins.

The Five Most Common ICD-10 Errors Across Specialties

Navigating the 70,000+ codes in the ICD-10-CM manual requires more than just a software search; it requires clinical intuition. Here are the five "usual suspects" that I see triggering audits and draining revenue across every specialty I consult for.

1. Unspecified Codes When Specificity Exists

The "Unspecified" code (often ending in .9) is the ultimate red flag for payers in 2026. While sometimes necessary, overusing them suggests weak documentation discipline, poor coder-query workflows, or both. Unspecified coding patterns are easy for payer analytics to identify, and once your ratios drift higher than your peers, your claims can be routed for manual review even before a formal audit begins.

• The Error: Using R10.9 (Abdominal pain, unspecified) when the notes clearly state the patient has "right lower quadrant pain."
• The Fix: Always code to the highest level of specificity documented. In this case, R10.31. Payers now flag practices with high unspecified-to-specified ratios for manual review.
• Clinical Case Study: A 27-year-old patient presents to urgent care with 18 hours of worsening right lower quadrant pain, nausea, guarding, and rebound tenderness. The physician documents "acute RLQ abdominal pain, concern for appendicitis" and sends the patient to the ED. The superbill is submitted with R10.9 instead of R10.31. That seems minor, but it weakens the clinical picture, makes medical necessity review less defensible, and creates unnecessary ambiguity if imaging, transfer, or surgical consultation follows. The coder should capture the exact documented site of pain and, when supported later in the encounter sequence, update diagnosis coding to reflect confirmed findings.

2. Missing Laterality

Laterality is not a suggestion; it is a requirement for thousands of codes. In orthopedics, wound care, dermatology, podiatry, ophthalmology, and vascular medicine, missing side-specific detail is one of the fastest ways to produce avoidable denials or rework.

• The Error: Submitting a claim for cellulitis of the leg using L03.11 (Cellulitis of lower limb) without specifying which side.
• The Fix: You must use the sixth character. L03.115 for the right leg or L03.116 for the left. Missing this simple character can lead to an automated line-item denial.
• Clinical Case Study: A 72-year-old patient with venous insufficiency and edema presents with erythema, warmth, and swelling of the left lower leg. The physician documents "cellulitis, left lower leg; no abscess." The claim is submitted without laterality specificity. If the code does not match the documentation, the practice now has a preventable discrepancy between note content and billed diagnosis. That creates appeal risk, slows payment, and can confuse follow-up care when subsequent encounters are for the opposite limb or bilateral disease. The coding team should build a hard-stop rule: if the note states right, left, bilateral, upper, or lower, the final ICD-10 code must mirror that specificity.

3. Missing Acuity and Seventh-Character Extensions

ICD-10-CM requires us to track the "story" of the encounter. Is this the first time we’re seeing them for this injury, or are they in follow-up? Is the condition acute, chronic, healing routinely, delayed, or resulting in sequela? Those details materially change the code.

• The Error: Using a fracture code without the seventh character (A, D, or S) to denote the encounter type.
• The Fix: Ensure every injury code includes the correct extension. S82.001A (Initial encounter for closed fracture of right patella) vs. S82.001D (Subsequent encounter).
• Clinical Case Study: A 46-year-old patient is seen by orthopedics after slipping on wet pavement and sustaining a closed nondisplaced fracture of the right patella. At the first active treatment visit, the diagnosis should reflect an initial encounter with the proper seventh character. Six weeks later, the same patient returns for routine healing follow-up, but the claim still carries the initial-encounter extension. That is incorrect. The seventh character is not about whether your practice has seen the patient before; it is about the phase of treatment. Reusing the initial-encounter code across follow-up visits distorts the clinical timeline and creates unnecessary payer scrutiny.

4. Wrong Principal vs. Secondary Diagnosis Sequencing

On inpatient claims, sequencing dictates the DRG (Diagnosis Related Group). If the principal diagnosis: the condition established after study to be chiefly responsible for the admission: is wrong, your entire payment is wrong. Even in outpatient environments, sequencing errors can alter the clinical logic that supports testing, infusion, observation, or specialist involvement.

• The Error: Sequencing a secondary complication like dehydration before the primary pneumonia.
• The Fix: Review the "Code First" and "Use Additional Code" instructions in the FY 2026 Tabular List.
• Clinical Case Study: A frail 81-year-old patient is admitted with fever, productive cough, hypoxia, and confusion. Workup confirms community-acquired pneumonia with dehydration secondary to poor oral intake. If dehydration is sequenced as principal while pneumonia is pushed to secondary position, the coded story no longer reflects the true reason for admission. That can shift DRG assignment, weaken utilization review support, and create discrepancies between physician narrative, treatment intensity, and billed data. Sequencing is not a clerical detail; it is the architecture of the claim.

5. Missing Combination Codes

One of the most frequent errors I find in my coding accuracy consulting work is the "double-coding" of conditions that should be one single code. Combination codes matter because they connect cause and manifestation, and that connection often carries reimbursement, HCC, and audit significance.

• The Error: Reporting Type 2 Diabetes (E11.9) and Chronic Kidney Disease (N18.31) separately.
• The Fix: Use the combination code E11.22 (Type 2 diabetes mellitus with diabetic chronic kidney disease). This captures the manifestation and the disease in one high-weighted code.
• Clinical Case Study: A 68-year-old established patient presents for diabetes follow-up. The physician documents Type 2 diabetes, diabetic chronic kidney disease, CKD stage 3a, hypertension, and long-term insulin use. The claim goes out with E11.9 and N18.31 only. That misses the causal diabetic relationship. The correct reporting should include E11.22 plus N18.31, with any additional applicable status or treatment codes as supported. This is exactly the kind of silent undercoding that lowers condition capture, weakens RAF reporting, and makes a medically complex patient look routine on paper.

Specialty-Specific ICD-10 Traps: What You Need to Know for 2026

Every specialty has its own unique hurdles. In my 21 years of physician-led consulting, I’ve identified these "traps" that consistently trip up even the most experienced coders.

• Endocrinology & Primary Care: A massive change in the FY 2026 codebook is the introduction of E11.A (Type 2 diabetes mellitus without complications in remission). You must have the word "remission" in the documentation to use this. Do not use it for Type 1 or secondary diabetes, and do not use it as a shortcut for "controlled" diabetes. "Controlled," "at goal," or "A1c improved" are not the same as remission. If the patient has active diabetic complications, then E11.A is not the right code family. This is a high-risk education point for family medicine, internal medicine, endocrinology, and obesity medicine because remission language is increasing in GLP-1 era charting. Also, watch for the expanded SDoH codes, especially the Z59.86- financial insecurity series. In 2026, details like utility insecurity, housing instability, transportation barriers, and other documented social risk factors are not just community-health talking points; they influence care planning, support medical necessity narratives, and increasingly affect risk stratification and value-based reporting. For example, Z59.861 for utility-related financial insecurity should be assigned only when clearly supported in the note, not inferred from a missed payment or generalized hardship.
• Cardiology: The common trap is failing to specify the type and acuity of heart failure. Reporting I50.9 (Heart failure, unspecified) is no longer acceptable as a routine habit. You need to distinguish between systolic (I50.2x), diastolic (I50.3x), or combined (I50.4x) and whether it is acute, chronic, or acute-on-chronic. Also watch for documentation that links hypertensive heart disease, CKD, or pulmonary edema, because the presence or absence of those relationships changes code assignment and can materially affect severity capture.
• OB/GYN: Trimester specificity is paramount. Many codes require a final character to denote the trimester (1, 2, or 3). Additionally, for multiple gestations, you must now use the seventh character to identify which fetus is affected by a specific complication. A common trap is copying forward a diagnosis from a prior trimester or failing to update the fetus designation after ultrasound confirmation. Those are small documentation misses with big coding consequences.
• Mental Health/Neurology: The FY 2026 updates have refined the phenotype structure for Multiple Sclerosis, replacing the old one-size-fits-all G35 approach with precise alpha-subcode options effective October 1, 2025: G35.A for relapsing-remitting MS, G35.B for primary progressive MS (with further specificity at G35.B0 unspecified, G35.B1 active, G35.B2 non-active), G35.C for secondary progressive MS (G35.C0 unspecified, G35.C1 active, G35.C2 non-active), and G35.D for multiple sclerosis unspecified. The codes use alpha characters, not numeric — a common documentation error in 2026 is providers and coders defaulting to legacy G35 or attempting numeric subcodes that do not exist in the FY 2026 code set. This matters because neurology, behavioral health, rehab, and infusion-based practices are no longer coding a generic disease label when the neurologist has already documented a distinct course pattern. If the physician documents relapsing-remitting, secondary progressive, primary progressive, progressive-relapsing, tumefactive, or other defined phenotype language supported by the 2026 manual, coders should move beyond unspecified MS reporting. The trap is failing to update long-standing problem lists and charge tickets that still default to legacy G35 logic. Phenotype accuracy supports treatment rationale, infusion authorization, disability documentation, and care-complexity reporting.
• Orthopedics: Laterality and the 7th character (A, D, S) remain the biggest traps. Ensure your providers are documenting whether the fracture is "displaced" or "nondisplaced," open or closed, and whether healing is routine, delayed, nonunion, or malunion when applicable. Those details completely change the code selection. In 2026, orthopedic coding accuracy still depends less on software and more on whether the operative note and follow-up assessment tell the full healing story.

Expert Tip: If you're struggling with these transitions, you can find my full documentation standardization framework on Amazon Kindle for $9.99, or grab the paperback for $29.99 to keep at your desk.

Why ICD-10 Accuracy Kills Revenue Silently

You might think, "If I’m getting paid, why does it matter?" Here is the reality: ICD-10 accuracy is about more than just the current check.

1. HCC Risk Adjustment: In Medicare Advantage plans, your payment is tied to the patient’s RAF (Risk Adjustment Factor) score. RAF is built from demographic factors plus documented, submitted, and accepted diagnosis data that maps to HCC categories for the applicable risk year. In plain language, the plan and CMS are estimating the expected cost of caring for that patient next year based on the disease burden you captured this year. If you code "Diabetes" instead of "Diabetes with Peripheral Angiopathy," or fail to connect diabetes to CKD, neuropathy, or vascular manifestations when documented, you may lose a meaningful HCC relationship. That missing specificity can reduce the patient’s captured illness burden enough to drop the RAF score by 0.3 or more in some scenarios, especially when multiple chronic conditions are involved. That is not a rounding error. Across a panel of Medicare Advantage patients, it becomes a major revenue compression event.
2. Granular Specificity Changes Risk Math: A single missing character can matter. If the provider documents a condition with a causal relationship, severity level, manifestation, or phenotype and the final ICD-10 code fails to include it, the patient may map to a lower-weighted condition profile or fail to map to the expected HCC category at all. Think of RAF capture as cumulative. Each accurately documented chronic condition adds to the clinical picture, but only if the final code set preserves the detail. When one specific element is dropped, the patient can appear less complex than they truly are. In value-based contracts, that mismatch follows the practice long after the original encounter is paid.
3. DRG Mismatches: For hospital-based providers, an incorrect ICD-10 code can shift a case from a high-paying DRG to a lower one, costing the facility thousands per encounter. A sequencing error, a missing complication/comorbidity relationship, or failure to capture acute-on-chronic status can all change grouping logic. The physician may have delivered high-acuity care, but if the final diagnosis coding does not tell that story correctly, the DRG will not either.
4. MIPS and Quality Reporting: Many quality measures are triggered by specific diagnosis codes. If your ICD-10 coding is sloppy, your quality scores will suffer, leading to negative payment adjustments. This is especially problematic when chronic diseases are documented generally but not specifically enough to trigger the correct denominator or risk stratification framework.
5. Audit Exposure: Payers use algorithms to look for "unspecified" outliers, unsupported complication coding, and abnormal chronic-condition prevalence patterns. If your practice uses I10 (Essential hypertension) exclusively without ever documenting secondary causes or manifestations when present, or if your diabetes coding never reflects documented complications, you become a target for a pre-payment or post-payment audit.

If you're noticing that your unspecified codes are piling up or your revenue doesn't seem to match the complexity of the patients you see, it's time for an expert eye. I invite you to schedule a confidential ICD-10 coding accuracy audit consultation so we can look at your data together and stop the leaks.

The 4-Step ICD-10 Hygiene Audit You Can Run This Month

You don't have to wait for a disaster to improve your coding. I recommend every practice administrator and lead physician perform this "Hygiene Audit" once a quarter.

• Step 1: Run an Unspecified-Code Utilization Report. Use your billing software to pull a report of all codes ending in .9. If more than 20% of your top 50 codes are unspecified, you have a documentation problem.
  How-To Tip: Sort the report by provider, specialty, and diagnosis frequency. Do not review unspecified codes as one giant list. Break them into your top revenue-producing service lines first. Then ask a simple question for each code: was specificity unavailable, or was it documented and simply not captured? That distinction tells you whether the fix belongs with the physician, coder, template, or EHR workflow.
• Step 2: Audit Your Top 20 Codes for Specificity. Look at your highest-volume codes. Are you using laterality? Are you capturing the most recent FY 2026 changes, like the new MS phenotypes or the E11.A remission code?
  How-To Tip: Pull the top 20 diagnosis codes by claim volume for the last 60 to 90 days. Next to each code, place the corresponding documentation requirement: laterality, acuity, linkage language, complication detail, stage, trimester, encounter status, or remission status. This creates a quick comparison grid your coding lead can review in under an hour. If a high-volume code repeatedly lacks the same element, that is a teachable systems issue, not a one-off mistake.
• Step 3: Perform a Clinical-to-Code Cross-Check. Randomly select 10 encounters. Does the physician's note mention details (like "left side," "acute," or "due to financial insecurity") that didn't make it into the final ICD-10 code? This is where the most revenue is "lost."
  How-To Tip: Choose encounters from different providers and different visit types: new patient, established patient, procedure follow-up, and chronic care. Highlight every diagnostic specificity word in the note before you look at the claim. Then compare that list to the final code set. This "highlight first, code second" method quickly exposes dropped severity terms, missed cause-and-effect relationships, and overlooked SDoH documentation such as the Z59.86 financial insecurity series.
• Step 4: Target Provider Education. Don't try to teach the whole book. Focus on the top 3 traps relevant to your specialty. Show the doctors the difference in medical decision making weight when specificity is captured.
  How-To Tip: Keep the education specialty-specific and brief. A 10-minute huddle with three real examples from the provider's own documentation is more effective than a 40-slide lecture. Build mini cheat sheets with phrases physicians can drop directly into the assessment, such as "diabetic CKD stage 3a," "acute on chronic diastolic heart failure," or "utility-related financial insecurity affecting medication storage." If you want behavior change, make the documentation language easy to use in real time.

Moving Forward with Confidence

Coding shouldn't be a source of anxiety. It is a tool for clinical and fiscal integrity. Whether you are navigating the complexities of telehealth billing in 2026 or refining your surgical coding, accuracy is your best defense.

If you are ready to take your practice to the next level, I highly recommend our Premium Bundle on our physician training platform. For just $47, you get the full "Medical Train Right" experience: my ebook PDF, a 90-day action plan workbook, regulatory quick-reference cards for 2026, and audit-ready checklists. It’s a complete system designed to keep you compliant and profitable.

Thank you for your dedication to excellence in medicine and administration. Let's make sure your revenue reflects the incredible work you do every day.

Dr. Dreama Sloan-Kelly, MD, CCS, CPC CEO/President, Dr. Sloan-Kelly Consulting LLC

---

Don't miss the next update. Each month I send The DocsDoc Brief : clinical documentation, coding, compliance, and AI insights : straight to your inbox. 👉 Subscribe to The DocsDoc Brief here